UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Alternative approaches to the analysis of consumer spatial behavior Taylor, Stuart Martin 1974

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1974_A1 T39.pdf [ 11.48MB ]
Metadata
JSON: 831-1.0100000.json
JSON-LD: 831-1.0100000-ld.json
RDF/XML (Pretty): 831-1.0100000-rdf.xml
RDF/JSON: 831-1.0100000-rdf.json
Turtle: 831-1.0100000-turtle.txt
N-Triples: 831-1.0100000-rdf-ntriples.txt
Original Record: 831-1.0100000-source.json
Full Text
831-1.0100000-fulltext.txt
Citation
831-1.0100000.ris

Full Text

ALTERNATIVE APPROACHES TO THE ANALYSIS OF CONSUMER SPATIAL BEHAVIOUR b y STUART MARTIN TAYLOR 6.A. Un i v e r s i t y of B r i s t o l , 196,9 -M.A. U n i v e r s i t y of B r i t i s h Columbia, 1972 A THESIS SUBMITTED IN PARTIAL.FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY i n the Department of Geography We accept t h i s t h e s i s as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA In p r e s e n t i n g t h i s t h e s i s in p a r t i a l f u l f i l m e n t o f the r equ i r emen t s f o r an advanced degree at the U n i v e r s i t y o f B r i t i s h Co lumb i a , I ag ree that the L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r r e f e r e n c e and s tudy . I f u r t h e r agree t h a t p e r m i s s i o n f o r e x t e n s i v e c o p y i n g o f t h i s t h e s i s f o r s c h o l a r l y purposes may be g r a n t e d by the Head o f my Department o r by h i s r e p r e s e n t a t i v e s . It i s u n d e r s t o o d that c o p y i n g o r p u b l i c a t i o n o f t h i s t h e s i s f o r f i n a n c i a l g a i n s h a l l not be a l l owed w i thout my w r i t t e n p e r m i s s i o n . Department o f The U n i v e r s i t y o f B r i t i s h Co lumbia Vancouver 8, Canada Date /6<-ABSTRACT Recent s t u d i e s have sought to e x p l a i n consumer s p a t i a l behaviour i n terms of the p s y c h o l o g i c a l c h a r a c t e r i s t i c s of the consumer. A v a r i e t y of p s y c h o l o g i c a l a t t r i b u t e s have been suggested as p o t e n t i a l independent v a r i a b l e s i n c l u d i n g p e r c e p t i o n s , b e l i e f s , a t t i t u d e s and v a l u e s . T h i s study seeks to s y n t h e s i s e these d i s p a r a t e approaches through the i d e n t i -f i c a t i o n and measurement of consumer d i s p o s i t i o n s which are seen to f u n c t i o n as c e n t r a l components i n the d e c i s i o n -making process l e a d i n g to behaviour. P a r t i c u l a r emphasis i s placed upon the development of a r e l i a b l e and v a l i d set of d i s p o s i t i o n a l s c a l e s which i t i s hoped might have ap-p l i c a b i l i t y beyond t h i s s i n g l e study. This r e s e a r c h makes a f u r t h e r c o n t r i b u t i o n to the ex-i s t i n g l i t e r a t u r e by comparing the a b i l i t y of a d i s p o s i t i o n a l model to p r e d i c t r e t a i l choices with that of t r a d i t i o n a l shopping behaviour models based on measures of the l o c a t i o n -a l and b i o g r a p h i c a l c h a r a c t e r i s t i c s of the consumer. Interview responses and a f a c t o r a n a l y s i s of q u e s t i o n -n a i r e data i d e n t i f i e d f i v e major d i s p o s i t i o n s u n d e r l y i n g r e t a i l choices with respect to c l o t h i n g purchases. These were l a b e l l e d status-orientation, convenience-orientation. fashion-orientation, price o r i e n t a t i o n and q u a l i t y - o r i e n t a t i o n . A set of L i k e r t s c a l e s was developed to measure these dispo-s i t i o n s and a p r e t e s t was conducted to ensure that e s t a b l i s h e d psychometric standards were met. The major data c o l l e c t i o n phase comprised a q u e s t i o n n a i r e survey of a s t r a t i f i e d random sample of households drawn from the C i t y of Vancouver and adjacent m u n i c i p a l i t i e s . The data provided i n f o r m a t i o n on the respondent's d i s p o s i t i o n s , shop-ping behaviour and b i o g r a p h i c a l c h a r a c t e r i s t i c s . L o c a t i o n r e l a t i v e to r e t a i l f a c i l i t i e s was subsequently determined from the respondent's address. A s e r i e s of d i s c r i m i n a n t analyses were performed on the q u e s t i o n n a i r e data to compare the e f f i c a c y of l o c a t i o n a l , b i o g r a p h i c a l and d i s p o s i t i o n a l v a r i a b l e s as d i s c r i m i n a t o r s between r e t a i l patronage groups. The groups were d e f i n e d on the b a s i s of the type of s t o r e and type of shopping centr e p a t r o n i z e d on each of four shopping t r i p s f o r c l o t h i n g . The r e s u l t s suggest that r e t a i l choices may be the outcome of a two-stage d e c i s i o n process: the f i r s t stage at a macro s c a l e i n v o l v i n g the i s o l a t i o n of one or more f e a s i b l e a l t e r n a t i v e shopping areas p r i m a r i l y on the b a s i s of l o c a t i o n a l c o n s i d e r -a t i o n s ; the second stage, at a micro s c a l e i n v o l v i n g the se-l e c t i o n of a s t o r e w i t h i n one of the p r e s e l e c t e d shopping areas with d i s p o s i t i o n a l f a c t o r s being fundamental i n d i r e c t -ing the e v a l u a t i o n of a l t e r n a t i v e s and the choice of a p r e f e r r e d s t o r e . i i ACKNOWLEDGEMENTS Es p e c i a l thanks are due to my supervisor, Dr. Walter Hardwick, who has done so much to expedite my graduate programme. I am a l s o g r a t e f u l to the other members of my committee - Drs. C o l l i n s , Forbes, Harris and Ley - for t h e i r w i l l i n g n e s s to devote t h e i r time and energy to reviewing t h i s study. I am a l s o g r a t e f u l to the group of geography undergraduate students who were w i l l i n g to assume the unenviable r o l e of interviewers. My thanks also go to the anonymous respondents whose w i l l i n g n e s s to complete questionnaires never ceases to amaze me. I thank Miss Raymonde Thibeault f o r typing the manuscript i n such an exemplary manner. F i n a l l y , I g r a t e f u l l y acknowledge the f i n a n c i a l support I have received as a doctoral fellow from The Canada Council over the past two years. i i i TABLE OF CONTENTS Page ABSTRACT i ACKNOWLEDGEMENTS i i i TABLE OF CONTENTS i v LIST OF FIGURES v i i i LIST OF TABLES ix CHAPTER Page INTRODUCTION ' 1 Major Objectives of the Study 2 The Broader Context of the Study 4 Outline of Chapters 6 1 APPROACHES TO THE ANALYSIS OF CONSUMER 8 SPATIAL BEHAVIOUR A. INTRODUCTION 8 B. THE ARRAY OF CONSUMER BEHAVIOUR MODELS 9 (i) Comprehensive Models 1Q (ii) U t i l i t y Models 13 ( i i i ) Social Class and Life-Cycle Models 17 (iv) Image Models 20, (v) Perceptual Preference Models 27 (vi) Learning Models 31 (vii) Attitude and Value Models 34 C. SYNTHESIS AND THE CONTEXT OF THE 45 PRESENT STUDY D. SUMMARY 48 2 RESEARCH DESIGN 49 A. INTRODUCTION 49 B. RESEARCH PURPOSES 49 iv CHAPTER Page 2 C. RESEARCH DESIGN 51 (i) L i k e r t Scales 5 1 ( i i ) The Assessment of R e l i a b i l i t y 54 ( i i i ) The Assessment of V a l i d i t y 56 (iv) Research Design Components 63 D. SUMMARY 65 3 THE IDENTIFICATION AND MEASUREMENT OF CONSUMER 67 DISPOSITIONS A. INTRODUCTION 67 B. CONSTRUCT IDENTIFICATION 67 C. THE DEVELOPMENT OF A SET OF LIKERT SCALES 72 (i) Generation of an Item Pool 72 ( i i ) P r e t e s t i n g and Item Analysis 80 ( i i i ) Construct V a l i d i t i e s 8 5 (iv) P r e d i c t i v e V a l i d i t i e s 87 D. REVISED SCALES 90 4 QUESTIONNAIRE DESIGN, SAMPLE DESIGN AND DATA COLLECTION PROCEDURES A. INTRODUCTION B. QUESTIONNAIRE DESIGN 91 (i) The Measurement of Behaviour 92 1. Name of store patronized 94 2. Type of store patronized 94 3. Store l o c a t i o n 95 4. T r i p o r i g i n 95 5. Location of t r i p o r i g i n 96 6. Distance estimate 96 7. Travel mode 96 8. T r i p purpose 96 9. Past purchases 96 ( i i ) B i ographical Measures 97 ( i i i ) General Design Considerations 97 V CHAPTER Page 4 C. SAMPLE DESIGN 9 8 (i) Population D e f i n i t i o n ^8 ( i i ) S t r a t i f i c a t i o n by Area 99 ( i i i ) Sample Se l e c t i o n ^9 D. INTERVIEWER TRAINING 1 0 3 E. DATA RETURNS AND CODING PROCEDURES 1 0 3 F. SUMMARY 105 5 DATA ANALYSIS 106 A. INTRODUCTION 106 B. SAMPLE CHARACTERISTICS, INTERVIEWER BIAS 107 AND SCALE STATISTICS (i) S p a t i a l D i s t r i b u t i o n of Respondents 107 ( i i ) Biographical C h a r a c t e r i s t i c s 110 ( i i i ) Interviewer C h a r a c t e r i s t i c s and 110 Response Bias C. LIKERT SCALE RELIABILITIES AND VALIDITIES 112 D. SHOPPING BEHAVIOUR: DESCRIPTIVE STATISTICS H 4 (i) Store and Shopping Centre Patronage 114 ( i i ) Trip O r i g i n H 8 ( i i i ) Travel Mode I I 9 (iv) T r i p Purpose 125 (v) Past Purchases 126 (vi) Conclusion 129 E. ANALYSIS OF SHOPPING BEHAVIOUR 129 (i) The A n a l y t i c a l Framework 129 ( i i ) S t a t i s t i c a l Procedures 134 ( i i i ) Analyses of Store and Shopping Centre l 3 ^ Patronage (iv) Analysis of Department Store Patronage 177 (v) . Analysis of S p e c i f i c Shopping Centre Patronage 182 (vi) Regression Analysis of Distance 185 Tra v e l l e d to Shop ( v i i ) Male-Female Differences i n Shopping 190 Behaviour and Di s p o s i t i o n s F. CONCLUSION 194 v i CHAPTER Page 6 SUMMARY AND CONCLUSIONS 196 Summary 196 Consumer Di s p o s i t i o n s and S p a t i a l Behaviour 199 Implications f o r Market Research 200 Limitations of the Present Study and Proposals 201 for Future Research BIBLIOGRAPHY 204 APPENDIX A: CONSUMER SHOPPING SURVEY 213 APPENDIX B: LETTER INTRODUCING THE SURVEY 224 APPENDIX C: INTERVIEWER'S MANUAL 226 APPENDIX D: INTERVIEWER'S ASSIGNMENT SHEET 230 v i i LIST OF FIGURES FIGURE 1.1 A Comprehensive Model of Consumer Behaviour 1.2 A D i s p o s i t i o n a l Model of Consumer S p a t i a l Behaviour 2.1 Flow Diagram of the Research Design 4.1 . The S p a t i a l D i s t r i b u t i o n of Commercial Floorspace i n Greater Vancouver 4.2 Travel Time to Downtown Vancouver 5.1 Sample and Population D i s t r i b u t i o n s by E l e c t o r a l D i s t r i c t 5.2 S p a t i a l D i s t r i b u t i o n of Respondents and Selected R e t a i l Locations 5.3 Biographical C h a r a c t e r i s t i c s of the Sample 5.4 T r i p D i s t r i b u t i o n by Store Type 5.5 T r i p D i s t r i b u t i o n by Shopping Centre 5.6 T r i p D i s t r i b u t i o n by O r i g i n 5.7 T r i p D i s t r i b u t i o n by Travel Mode 5.8 T r i p D i s t r i b u t i o n by T r i p Purpose 5.9 T r i p D i s t r i b u t i o n by Past Purchases 5.10 A Diagramatic Representation of the A n a l y t i c a l Framework v i i i LIST OF TABLES TABLE Page 1.1 A Suggested C l a s s i f i c a t i o n of Store Image Components 22 1.2 The Hypothesised Structure of Shopping Centre 24 Cognitions 3.1 Item-Scale Co r r e l a t i o n s 75 3.2 Pretest Measures of Scale R e l i a b i l i t i e s 84 3.3 Pretest Measures of Scale V a l i d i t i e s 86 3.4 Group Means and F S t a t i s t i c s 88 3.5 Group C l a s s i f i c a t i o n Matrix 89 5.1 F i n a l Scale R e l i a b i l i t y and V a l i d i t y C o e f f i c i e n t s 113 5.2 D i s t r i b u t i o n of Shopping Trips by Store Type and 117 Shopping Centre 5.3 D i s t r i b u t i o n of Shopping Trips by T r i p O r i g i n and 120 Store Type 5.4 D i s t r i b u t i o n of Shopping T r i p s by T r i p O r i g i n and 121 Shopping Centre 5.5 D i s t r i b u t i o n of Shopping Trips by Travel Mode and 123 Store Type 5.6 D i s t r i b u t i o n of Shopping Trips by Travel Mode and 124 Shopping Centre 5.7 D i s t r i b u t i o n of Shopping Trips by T r i p Purpose 127 and Store Type 5.8 D i s t r i b u t i o n of Shopping T r i p s by T r i p Purpose 128 and Shopping Centre 5.9 D i s t r i b u t i o n of Shopping Trips by Past Purchases and 130 Store Types 5.10 D i s t r i b u t i o n of Shopping Trips by Past Purchases and 131 Shopping Centre 5.11 An a l y s i s 1: Summary of Results 140 i x TABLE Page 5.12 Analysis 2: Summary of Results 143 5.13 Analysis 3: Summary of Results 145 5.14 Analysis 4: Summary of Results 147 5.15 Analysis' 5: Summary of Results 149 5.16 Analysis 6: Summary of Results 151 5.17 Analysis 7: Summary of Results 153 5.18 Analysis 8: Summary of Results 155 5.19 Analysis 9: Summary of Results 157 5.20 Analysis 10: Summary of Results 159 5.21 Analysis 11: Summary of Results 162 5.22 Analysis 12: Summary of Results 164 5.23 Analysis 13: Summary of Results 166 5.24 Analysis 14: Summary of Results 168 5.25 Analysis 15: Summary of Results 170 5.26 A n a l y s i s 16: Summary of Results 172 5.27 Average Percentage of Cases C o r r e c t l y Assigned to 175 Store Patronage Groups 5.28 Average Percentage of Cases C o r r e c t l y Assigned to 176 Shopping Centre Patronage Groups 5.29 Analysis of Department Store Patronage: Summary of 179 Results 5.30 Analysis of Downtown Department Store Patronage 181 5.31 Analysis of Shopping Centre Patronage 183 5.32 Regression Equations on Distance T r a v e l l e d to Shop 187 x TABLE Page 5.33 D i s t r i b u t i o n of Shopping T r i p s by Sex and 192 Store Type 5.34 D i s t r i b u t i o n of Shopping T r i p s by Sex and 192 Shopping Centre 5.35 Male-Female Group Means on D i s p o s i t i o n a l Scales 193 and F S t a t i s t i c s x i INTRODUCTION The analysis of consumer s p a t i a l behaviour has a t t r a c t e d consider-able research i n t e r e s t w i t h i n both geography and marketing, and over time various approaches have evolved. In geography, i n i t i a l emphasis was placed upon s p a t i a l e q u i l i b r i u m models of r e t a i l a c t i v i t y within which consumer behaviour was assumed rather than examined i n d e t a i l . There has been a subsequent re a c t i o n to the broad and overly s i m p l i s t i c assumptions inherent i n such models with the r e s u l t that e x p l i c i t atten-t i o n has been paid to the analysis of shopping patterns at a disaggre-gate l e v e l . In t h i s context, two sets of independent v a r i a b l e s have been p r i n c i p a l l y employed: namely, measures of the demographic and socio-economic c h a r a c t e r i s t i c s of the consumer; and, more recent l y , measures of the consumer's cognition of the r e t a i l environment i n terms of a range of p s y c h o l o g i c a l v a r i a b l e s i n c l u d i n g images, b e l i e f s and a t t i t u d e s . The study of consumer s p a t i a l behaviour in.marketing has t r a d i t i o n -a l l y emphasized the development of h e u r i s t i c devices to meet the imme-diate demands of the business world f o r r e l a t i v e l y simple methods f o r p r e d i c t i n g p o t e n t i a l patronage at given r e t a i l l o c a t i o n . To a large extent, the gravity model and i t s d e r i v a t i v e s have been the b a s i c operational procedures. In part as a reaction to the pragmatism charac-t e r i s t i c of t h i s t r a d i t i o n , increased attention has been paid i n recent years to various aspects of consumer psychology i n seeking to understand the process which antecedes behaviour. This development has r e s u l t e d i n the construction of a number of models of consumer behaviour ranging - 1 -2. i n complexity from those which are purpotedly comprehensive to others which are p a r t i a l focussing upon s e l e c t e d components of the more general models. Previous research therefore demonstrates various approaches to the analysis of consumer s p a t i a l behaviour and these are described and categorized i n the following chapter. Three major approaches emerge based upon: a. the l o c a t i o n of the consumer r e l a t i v e to r e t a i l f a c i l i -t i e s of varying s i z e (or the locational approach). b. the demographic and socio-economic c h a r a c t e r i s t i c s of the consumer (or the biographical approach). c. the consumer's cognition of the r e t a i l environment (sub-sequently termed the dispositional approach). Since the f i r s t two have been frequently discussed i n previous studies, greater emphasis i s placed i n the l i t e r a t u r e review upon examining various cognitive models of consumer behaviour. Major Objectives of the Study In l i g h t of the e x i s t i n g l i t e r a t u r e , the research described i n t h i s thesis has two major ob j e c t i v e s . The f i r s t i s to develop and t e s t a cognitive model of consumer s p a t i a l behaviour. The model presented at the conclusion of Chapter 1 draws upon previous studies of personal a t t i t u d e s and values. I t i s argued that a consumer's behaviour r e f l e c t s h i s d i s p o s i t i o n s towards the r e t a i l environment. Such d i s p o s i t i o n s are seen to summarize the values the consumer seeks to s a t i s f y by h i s beha-viour being l i n k e d to s p e c i f i c r e t a i l a l t e r n a t i v e s through the formation 3. of a t t i t u d e s . In order to o p e r a t i o n a l i z e the model i t was necessary to i d e n t i f y s a l i e n t consumer d i s p o s i t i o n s and to develop a research instrument to measure them. The i d e n t i f i c a t i o n of consumer d i s p o s i t i o n s was based upon a s e r i e s of unstructured interviews. With regard to the development of a measuring instrument, r e l a t e d work i n environmental psychology i n d i c a -tes the p o t e n t i a l usefulness of L i k e r t s c a l i n g . Following t h i s precedent, a preliminary set of scales was developed and a p r e t e s t conducted to assess the r e l i a b i l i t y and v a l i d i t y of the proposed instrument. On the b a s i s of the p r e t e s t r e s u l t s a f i n a l set of scales was determined. These scales formed one component of the questionnaire used i n the major data c o l l e c t i o n phase. The data obtained served as the b a s i s f o r assessing the r e l a t i o n s h i p between consumer d i s p o s i t i o n s and s p a t i a l behaviour. The second major obje c t i v e was to compare the e f f i c a c y of l o c a t i o n -a l , b i o g r a p h i c a l and d i s p o s i t i o n a l v a r i a b l e s as p r e d i c t o r s of shopping behaviour. To t h i s end, questions p e r t a i n i n g to the demographic and socio-economic c h a r a c t e r i s t i c s of the consumer were included i n the f i n a l questionnaire i n a d d i t i o n to the L i k e r t s c a l e s . The shopping t r i p o r i g i n of each respondent was also determined to permit the measurement of l o c a t i o n a l v a r i a b l e s . In the a n a l y s i s phase, m u l t i v a r i a t e s t a t i s t i -c a l procedures were employed to compare the a b i l i t y of the d i f f e r e n t models to discriminate between r e t a i l patronage groups. 4. The Broader Context of the Study In a broader context, the present study r e f l e c t s the current o r i e n -t a t i o n to the analysis of human s p a t i a l behaviour. This o r i e n t a t i o n , sometimes termed 'cognitive behaviouralism', emphasizes the exploration and understanding of psychological processes as a necessary and perhaps s u f f i c i e n t condition f o r the explanation of man's movements between d i f f e r e n t points i n space. To t h i s end, there i s a burgeoning volume of conceptual and empirical contributions from geographers, psychologists, planners and others. To date however the e f f o r t s of cognitive b e h a v i o u r a l i s t s have tended to be long on the cogni t i v e and short on the behavioural component. There are probably two r e l a t e d reasons f o r t h i s . F i r s t l y , there i s a l o g i c a l sequence whereby the i d e n t i f i c a t i o n and measurement o f process v a r i a b l e s precedes the explanation of c a u s a l l y r e l a t e d behaviour. I t i s probable that t h i s i n i t i a l phase w i l l continue f o r some time before the complexities of environmental cognition and the nature of the cognition-behaviour l i n k are reasonably w e l l understood. Secondly, ques-tions concerning a person's cognition of h i s environment are i n t r i g u i n g i n themselves and i t can be argued that t h e i r treatment i s an important facet of man-environment research i r r e s p e c t i v e of whether the behavioural impl i c a t i o n s are s p e c i f i e d . I t seems however that the p o s s i b i l i t i e s f o r c l a r i f y i n g the c o g n i t i o n -behaviour l i n k are perhaps greater with respect to c e r t a i n aspects of human s p a t i a l behaviour than others. For various reasons i t seems that shopping behaviour i s one such aspect. 5. One reason f o r t h i s i s that the study of r e t a i l structure and associated consumer movements has been a t o p i c to which geographers have paid considerable atte n t i o n . Although the e x p l i c i t recognition of the p o t e n t i a l s i g n i f i c a n c e of cognitive v a r i a b l e s i n t h i s research area i s of recent o r i g i n , the legacy of past studies i s an understanding of some important c h a r a c t e r i s t i c s of consumer s p a t i a l behaviour. -Secondly, considerable research e f f o r t has been expended by both marketers and consumer psychologists to i d e n t i f y the factors which lead shoppers to s e l e c t one store or shopping centre rather than another. The importance of cognitive v a r i a b l e s i n t h i s context has been recognised f o r some time as the numerous 'store image' studies t e s t i f y . Nonetheless, the immediacy of the demands of the business world have tended to over-shadow the goal of theory development with the r e s u l t that a w e l l - a r t i -culated and e m p i r i c a l l y tested theory of consumer s p a t i a l behaviour has not been forthcoming although valuable groundwork has been l a i d . The nature of the behaviour i t s e l f i s a t h i r d reason f o r optimism about achieving a c l a r i f i c a t i o n of the r e l a t i o n s h i p between consumer cognition and shopping behaviour. Most people undertake shopping t r i p s quite frequently and are therefore r e l a t i v e l y f a m i l i a r with the process of choosing between a l t e r n a t i v e r e t a i l opportunities. These circum-stances increase the l i k e l i h o o d of a consumer being able to a r t i c u l a t e the factors i n f l u e n c i n g h i s d e c i s i o n of where to shop. This i s important since i t i s l i k e l y to expedite the development of r e l i a b l e and v a l i d research instruments designed to measure consumer cognition. Furthermore, a consumer's s e l f and environmental knowledge p e r t a i n i n g to shopping decisions w i l l probably enhance the r e l i a b i l i t y and v a l i d i t y of the 6. cognitive measures subsequently obtained. As i n d i c a t e d by the research o b j e c t i v e s , the p o s s i b i l i t y of demon-s t r a t i n g the r e l a t i o n s h i p between consumer cognition and shopping beha-viour has been a major motivating force behind the research d e t a i l e d i n t h i s t h e s i s . Furthermore, an attempt has been made to r e l a t e t h i s study to previous research i n the f i e l d by comparing the e f f i c a c y of a cogni-t i v e model with that of l o c a t i o n a l and b i o g r a p h i c a l models as p r e d i c t o r s of consumer s p a t i a l behaviour. Outline of Chapters The structure of the d i s s e r t a t i o n i s perhaps most e a s i l y compre-hended from the following b r i e f o u t l i n e of chapters. Chapter 1 establishes the context of the study i n presenting a review of the various approaches adopted i n the ana l y s i s of consumer s p a t i a l behaviour, three of which were subsequently incorporated i n the research design. Various consumer behaviour models are discussed and i l l u s t r a t i v e empirical studies are described. P a r t i c u l a r consider-ation i s given to a t t i t u d e and value models and, following from t h i s , the chapter concludes with the presentation of a d i s p o s i t i o n a l model of consumer s p a t i a l behaviour. Chapter 2 o u t l i n e s the research design. The research purposes are itemized and the major components of the design are summarized with e s p e c i a l emphasis being placed upon the development of a research i n s t r u -ment to measure consumer d i s p o s i t i o n s , and upon the contingent consider-ations of r e l i a b i l i t y and v a l i d i t y . 7. Chapter 3 focusses on the preliminary stages of the research design and d e t a i l s the i d e n t i f i c a t i o n of s a l i e n t d i s p o s i t i o n s and the subsequent development of a L i k e r t s c a l i n g instrument. The p r e t e s t i n g of a set of preliminary scales i s described together with the c a l c u l a t i o n o f r e l i a -b i l i t y and v a l i d i t y indicees. On the basis of the p r e t e s t r e s u l t s a f i n a l set of d i s p o s i t i o n a l scales i s determined. Chapter 4 describes the major data c o l l e c t i o n phase, and, more s p e c i f i c a l l y , the design o f the f i n a l questionnaire, the sample design, the t r a i n i n g of the interviewers, and the data c o l l e c t i o n procedures. Chapter 5 d e t a i l s the data a n a l y s i s . Various d e s c r i p t i v e s t a t i s -t i c s are presented to document the c h a r a c t e r i s t i c s of the sample and sele c t e d aspects o f the respondents' shopping behaviour. Primary emphasis i s placed upon the r e s u l t s of a s e r i e s of mu l t i v a r i a t e analyses performed to compare the e f f i c a c y of l o c a t i o n a l , b i o g r a p h i c a l and dispo-s i t i o n a l models as pr e d i c t o r s of r e t a i l patronage and distance t r a v e l l e d to shop. F i n a l l y , male-female d i f f e r e n c e s i n shopping behaviour and di s p o s i t i o n s are described. Chapter 6 summarizes the study. The r e l a t i o n s h i p between consu-mer d i s p o s i t i o n s and s p a t i a l behaviour i s reconsidered i n l i g h t of the r e s u l t s of the a n a l y s i s . The market research implications are discussed and areas f o r furth e r research are proposed. CHAPTER 1 APPROACHES TO THE ANALYSIS OF CONSUMER SPATIAL BEHAVIOUR A. INTRODUCTION The increasing a t t e n t i o n paid by geographers i n recent years to the analysis of human s p a t i a l behaviour has been nowhere more c l e a r than i n r e l a t i o n to the study of consumer behaviour. In part t h i s r e -f l e c t s a t r a d i t i o n a l geographical i n t e r e s t i n the analysis o f the commercial structure of c i t i e s . Indeed, the perceived inadequacy of the assumptions about consumer behaviour inherent i n models of t e r t i a r y a c t i v i t y has been a major force motivating the development of d i f f e r e n t approaches. In large measure, the d i r e c t i o n of these developments has been guided by research i n the areas of marketing and consumer psychology. The purpose of t h i s chapter i s to present a b r i e f review of t h i s research i n e s t a b l i s h i n g the context of the present study. The chapter begins with a d e s c r i p t i o n of a comprehensive model of con-sumer behaviour. At t e n t i o n i s then focussed on various p a r t i a l models which to date have been applied i n the analysis of consumer s p a t i a l • behaviour. The review culminates i n a d e t a i l e d d e s c r i p t i o n of a t t i t u d e and value models, and, following from t h i s , a d i s p o s i t i o n a l model of consumer s p a t i a l behaviour i s proposed. -8 -B. THE ARRAY OF CONSUMER BEHAVIOUR MODELS Recent reviews of consumer behaviour models by Sheth (1967) and Hansen (1972) have sought to develop model typologies. Hansen i d e n t i -f i e s no less than twenty eight d i f f e r e n t model types each d i s t i n g u i s h e d by a unique combination of dependent v a r i a b l e s , independent v a r i a b l e s , u n i t of a n a l y s i s , temporal component, underlying process, l e v e l of aggregation and context of a p p l i c a t i o n . Models range i n scope from those which purport to be f u l l y integrated and comprehensive (e.g., the d e c i s i o n process models of Nic o s i a (1966), Howard and Sheth (1969) and Engel et al. (1968) to others which are p a r t i a l and s e l e c t i v e (e.g., p e r s o n a l i t y models (Brody and Cunningham, 1968). To a large extent the l a t t e r are subsumed by the former. Seven of the model types i d e n t i f i e d by Hansen appear to have par-t i c u l a r relevance to the ana l y s i s of consumer s p a t i a l behaviour and these are discussed i n the following sections of t h i s chapter. They are: (i) Comprehensive models. ( i i ) U t i l i t y Models. ( i i i ) S o c i a l c l a s s and l i f e - c y c l e models. (iv) Image models. (v) Perceptual preference models. (vi) Learning models. ( v i i ) A t t i t u d e and value models. 10. (i) Comprehensive Models Comprehensive models based l a r g e l y on s a l i e n t economic and psy-c h o l o g i c a l theory commonly assume that behaviour i s the outcome of a sequential decision-making process, the conduct of which r e f l e c t s the i n t e r a c t i o n of complex sets of s i t u a t i o n a l and d i s p o s i t i o n a l v a r i a b l e s . A b r i e f d e s c r i p t i o n of the model developed by Engel et al. w i l l serve to i l l u s t r a t e the t y p i c a l components of such decision-making frame-works . As Figure 1.1 shows, the operation of the model i s determined by. the 'central c o n t r o l u n i t ' which functions at the i n t e r f a c e between the input s t i m u l i and output responses of the system. The 'central c o n t r o l u n i t ' consists of memory and thought processes which are con-d i t i o n e d by the p e r s o n a l i t y c h a r a c t e r i s t i c s of the i n d i v i d u a l , i n terms of h i s behavioural t r a i t s and motives, and by h i s stored informa-t i o n l e v e l based on past experience. These conditioning components f i n d combined expression i n the i n d i v i d u a l ' s a t t i t u d e s and values. The decision-making process proposed i n the model i s t r i g g e r e d when a need i s aroused. The need may be i n t e r n a l l y generated or may be a response to an external stimulus emanating from the p h y s i c a l or s o c i a l environment. The nature of the need controls the perception of incoming s t i m u l i which are compared to determine what action, i f any, i s to be taken to solve the problem created by the need. A search f o r a l t e r n a t i v e s o l u t i o n s to the problem i s begun i f i t appears that a course of action can lead to a s a t i s f a c t o r y s o l u t i o n . Once the a l t e r n a t i v e s have been i s o l a t e d they are subject to evaluation FIGURE 1.1 11. A COMPREHENSIVE MODEL OF CONSUMER BEHAVIOUR [_SEARCHj SOURCE; Engel, K o l l a t and Blackwell C 1 9 6 8 I , p. 5Q. 12. to determine the preferred s o l u t i o n . I t i s only at t h i s stage, having made t h i s sequence of p r i o r d e c i s i o n s , that the i n d i v i d u a l enters i n t o a purchase s i t u a t i o n , and even then i t i s f a r from c e r t a i n that an actual purchase w i l l be made. But whatever the ultimate outcome, the experience derived w i l l be added to the stored information to influence future behaviour. Hence the model can be regarded as dynamic. The basic components of the model w i l l not vary with the nature of the product under consideration, but c l e a r l y the r e l a t i v e importance of the various stages within the d e c i s i o n process w i l l be appreciably modified by the p o t e n t i a l purchase. For example, the search of a l t e r -native solutions w i l l obviously be more extensive i f the purchase of a car i s proposed, as opposed to a carton of milk; i n the l a t t e r case, where the purchase i s probably based on habit, t h i s step may be com-p l e t e l y bypassed. Termination of the process can occur at v i r t u a l l y any stage; i n many instances t h i s i s l i k e l y to take place p r i o r to the i n d i v i d u a l entering i n t o an actual purchasing s i t u a t i o n . In addition, the out-comes at each stage may be modified by the e f f e c t s of p s y c h o l o g i cal response sets such as the i n d i v i d u a l p e r c e i v i n g r i s k and doubt i n buy-ing s i t u a t i o n s . The model has not been tested e m p i r i c a l l y . This i s hardly s u r p r i -s ing, given the complexity of the i n t e r r e l a t i o n s h i p s between model components and the inexactness with which c e r t a i n of the components are defined. Likewise, the other two s i g n i f i c a n t comprehensive models of consumer behaviour - those of N i c o s i a and Howard and Sheth - remain e s s e n t i a l l y untested with the exception of one empirical study which 13. was designed to t e s t the Howard-Sheth model (Farley and Ring, 1970) with somewhat inconclusive r e s u l t s . I t i s true however, as Zaltman et al. (1973) poin t out i n evaluating the three models, that each, but p a r t i c u l a r l y the N i c o s i a and Howard-Sheth models, derive some support as well as t h e i r i n s p i r a t i o n from empirical studies conducted p r i o r to t h e i r construction. Regardless of empirical v e r i f i c a t i o n , these com-prehensive models serve an important function i n proposing b a s i c con-s t r u c t s and t h e i r i n t e r r e l a t i o n s h i p s and thereby a s s i s t i n g the develop-ment of p a r t i a l models o f consumer behaviour which, being l e s s complex, are more amenable to empirical t e s t i n g . ( i i ) U t i l i t y Models U t i l i t y models draw l a r g e l y on microeconomic theory i n t r e a t i n g consumer behaviour as the outcome of s e l e c t i n g an a l t e r n a t i v e (whether i t be a brand, product or store) so as to maximize a u t i l i t y f u n c t i o n . This approach n e c e s s a r i l y assumes that a consumer has knowledge of a l l p o s s i b l e choice a l t e r n a t i v e s and i s hence able to evaluate each a l t e r n a t i v e and assign a f i x e d u t i l i t y to each. Rational economic behaviour i s assumed whereby each i n d i v i d u a l seeks to maximize h i s o v e r a l l s a t i s f a c t i o n within the l i m i t s of h i s income i n r e l a t i o n to p r e v a i l i n g p r i c e s . Furthermore, as Hurst (1972, p. 213) points out, i n minimizing the costs of acquiring goods, the economically r a t i o n a l consumer w i l l seek to minimize tran s p o r t a t i o n costs and hence patron-i z e s the nearest f a c i l i t y o f f e r i n g the required commodity. The concept of u t i l i t y maximization as the basis f o r consumer behaviour i s i m p l i c i t i n the c l a s s i c a l c e n t r a l place model ( C h r i s t a l l e r , 14. 1966; Berry and Garrison, 1958; Berry, 1967) and i s e x p l i c a t e d i n the subsequent r e v i s e d form of the model (Rushton, 1969, 1970, 1971; Clark and Rushton, 1970) where s p a t i a l behaviour i s seen to r e s u l t from the a p p l i c a t i o n of subjective preference functions to a set of s p a t i a l a l t e r n a t i v e s . These preference functions are i n turn regarded as the outcome of a t r a d e - o f f between distance separation and the r e l a t i v e a t t r a c t i o n of the opportunity ( i . e . , i t s place u t i l i t y ) . The u t i l i t y p r i n c i p l e i s also e x p l i c i t i n trading area models which have evolved from the e a r l y mechanistic formulation of R e i l l y (1931), through the work of Baumol and Ide (1956), to the p r o b a b i l i s -t i c model developed by Huff (1962). In each case, the u t i l i t y assigned Huff's s t o c h a s t i c model of consumer s p a t i a l behaviour assumes that consumers c a l c u l a t e a p o s i t i v e measure of u t i l i t y f o r each shop-ping centre within a defined s e t , and that they d i s t r i b u t e t h e i r r e t a i l patronage of a l t e r n a t i v e s s p a t i a l l y i n p r o b a b i l i s t i c fashion. Huff states that the p r o b a b i l i t y of a person's t r a v e l l i n g to a given shop-ping centre i s p r o p o r t i o n a l to the u t i l i t y of that centre i n r e l a t i o n to the u t i l i t y of other centres, where the u t i l i t y of a center i s determined by i t s s i z e (in terms of square footage of s e l l i n g space) and the time necessary to reach i t expressed a l g e b r a i c a l l y i n the following terms: S . J P. i i n S. j=l (T. , ) X i : where P^j i s the p r o b a b i l i t y of a consumer at i t r a v e l l i n g to shopping centre j , Sj i s the s i z e of the shopping centre j , T^j i s the t r a v e l time separating i and j , and X i s a parameter that varies with the type of merchandise under consideration. This r e f l e c t s that d i f f e r e n t types of merchandise support d i f f e r e n t amounts of consumer search. The value of the exponent was determined e m p i r i c a l l y on the basis of l i n e a r c o r r e l a t i o n analysis to give a ' b e s t - f i t ' s o l u t i o n , i n a s i m i l a r manner that R e i l l y determined the value of the exponent of the distance term i n h i s model of r e t a i l g r a v i t a t i o n . 15. to each r e t a i l a l t e r n a t i v e i s seen to be a d i r e c t function of i t s a t t r a c -tiveness, defined in terms of a measure of mass (e.g., the number of i n - s t o r e items; or the commercial floorspace of a shopping centre) and an inverse function of the cost i n c u r r e d i n reaching i t , u s u a l l y equa-ted with p h y s i c a l distance or t r a v e l time. U t i l i t y models, such as Huff's, have been frequently employed i n trade area studies and despite t h e i r s i m p l i c i t y and the questionable nature of the underlying assumptions, they have proved to be p a r t i c u -l a r l y accurate i n p r e d i c t i n g consumer t r i p s f or convenience goods (Bruch and Gauthier, 1968). Nevertheless, they have been the subject of repeated c r i t i c i s m . Bucklin (1967), f o r example, i d e n t i f i e s three major problems associated with t r a d i t i o n a l t r a d i n g area models: problems i n the measurement of v a r i a b l e s (both measures of mass and c o s t ) , d i f f i c u l t i e s i n determining the value of the exponent f o r the cost (distance) f a c t o r , and issues regarding the number and type of v a r i a b l e s included. The l a s t c r i t i c i s m i s the most far-reaching and l e d Bucklin to develop a model of shopping behaviour which added shopping plan, demo-graphic and motivational v a r i a b l e s to the measures of distance and s i z e of f a c i l i t i e s t r a d i t i o n a l l y employed. S i m i l a r l y , within geography, d i s s a t i s f a c t i o n with the lack o f realism c h a r a c t e r i s t i c of u t i l i t y model assumptions has motivated the adoption of more complex approaches i n v o l v i n g the measurement of p s y c h o l o g i c a l v a r i a b l e s (Downs, 1970, 1970a; Harvey, 1969; Golledge, 1970; Winkel, 1970). In Huff's defence, i t should be recognized that he acknowledged that the u t i l i t y of a shopping centre to a consumer was based upon a 16. host of d i f f e r e n t f a c t o r s , but he maintained that for the purpose of making reasonably accurate p r e d i c t i o n s of consumer s p a t i a l behaviour i t was s u f f i c i e n t to "discover and s p e c i f y only a few relevant v a r i a -b l e s " (Huff, 1962, p. 17). I t i s s i g n i f i c a n t that Bucklin's r e s u l t s do not c o n t r a d i c t t h i s contention since i n h i s study distance measures proved to be the most e f f e c t i v e p r e d i c t o r s of shopping centre patronage (Bucklin, 1967, 1967a). Nonetheless, the problem remains that a sim-p l i s t i c p r e d i c t i v e model may deviate so f a r from r e a l i t y as to lack t h e o r e t i c a l value. Success i n the development of accurate p r e d i c t i v e models i s not a j u s t i f i c a t i o n f o r a neglect of research which i s aimed at i n c r e a s i n g our knowledge of the complex process whereby consumer r e t a i l preferences are formed. This argument has been e x p l i c i t l y stated by Dalrymple and Thompson (1969) i n c r i t i c i z i n g trade area research; they contend that future advances i n the f i e l d are dependent on " g e n e r a l i z i n g and develop-ing a t h e o r e t i c a l structure from e x i s t i n g research" i n an attempt to " r e l a t e such studies to the large body of research on consumer behaviour, which l a r g e l y ignores the l o c a t i o n a l aspects of the purchasing d e c i s i o n " (p. 105 f f . ) . Elsewhere, Thompson has stated the view that such a r e -o r i e n t a t i o n of research e f f o r t w i l l n ecessitate "less emphasis on census and other published data as the p r i n c i p a l inputs to explanatory models, more emphasis on the sampling of consumer a t t i t u d e s , and atten-t i o n to survey research" (Thompson, 1966, p. 17). The extent to which t h i s view has been shared by other consumer researchers i s obvious from the content of the relevant marketing and geographic l i t e r a t u r e appearing i n the l a s t ten years, and to a large 17. extent the consumer behaviour models described i n the remainder of t h i s chapter i l l u s t r a t e the product of the approach advocated by Thompson. ( i i i ) S o c i a l Class and L i f e - C y c l e Models Frequent attempts have been made to account f o r d i f f e r e n c e s i n consumer behaviour in terms of v a r i a t i o n s i n s o c i a l c l a s s and stage i n the family l i f e - c y c l e . Martineau (1958), i n c o l l a b o r a t i o n with s o c i o -l o g i s t Lloyd Warner, was amongst the f i r s t to demonstrate the r e l a t i o n -ship between s o c i a l c l a s s membership and consumption patterns, and h i s work has been of p a r t i c u l a r i n t e r e s t to geographers because he p a i d s p e c i f i c attention to s o c i a l c l a s s v a r i a t i o n s i n s p a t i a l behaviour i n terms of store patronage. In one of a s e r i e s of Chicago studies, f o r example, s i g n i f i c a n t d i f f e r e n c e s were shown i n the s o c i a l c l a s s membership of patrons of two f u r n i t u r e s t o r e s . Results demonstrated that while the one store appealed p r i m a r i l y to the middle and upper cl a s s e s , the other appealed to the lower cla s s e s . This segmentation of the market was found to occur even though both stores had f u r n i t u r e i n a l l p r i c e ranges and professed to s e l l to everyone. Brown and F i s k (1965) r e l a t e d the store preferences of housewives drawn from four s o c i a l groups and a q u a l i t y ranking of s i x P h i l a d e l p h i a department stores. I t was found that the store hierarchy and s o c i a l c l a s s hierarchy were d i r e c t l y r e l a t e d i n that the highest q u a l i t y store was p r e f e r r e d by housewives i n the upper s o c i a l c l a s s group, while the store preferences of successively lower s o c i a l classes corresponded with decreasing q u a l i t y ranking of stores. 18. Levy (1966) has made some i n t e r e s t i n g observations about the r e l a -t i o n s h i p between s o c i a l status and shopping behaviour, suggesting that s o c i a l status appears to a f f e c t how people f e e l about where they should shop. The r e s u l t i s that the same products may be purchased i n d i f f e r -ent channels of d i s t r i b u t i o n by members of d i f f e r e n t s o c i a l c l a s s e s . He notes that i n the purchase of cosmetics, upper middle c l a s s women are more apt to shop i n department stores than are lower status women who are more prone to shop i n v a r i e t y stores. Drug stores seem equally a t t r a c t i v e or s u i t a b l e to a l l . In addition he makes the general obser-vation that there are sharp d i f f e r e n c e s i n the status reputation of department stores and that consumers tend to s o r t themselves out i n terms of where i t i s appropriate f o r themselves to shop. Most e s t a b l i s h -ments w i l l have customers of more than one s o c i a l c l a s s , but the pro-p o r t i o n a l representation of d i f f e r e n t classes w i l l vary, and t h e i r purchasing patterns may also d i f f e r . Rich and J a i n (1968) report s i g n i f i c a n t d i f f e r e n c e s in the shopping behaviour of consumers varying i n t h e i r s o c i a l c l a s s and stage i n the family l i f e - c y c l e . The data were extensive, c o n s i s t i n g of 4000 personal and telephone interviews i n Cleveland and New York. S o c i a l c l a s s was s t r a t i f i e d using Warner's Index of Status C h a r a c t e r i s t i c s (Warner, 1949), and stage i n the family l i f e - c y c l e was defined i n terms of four cate-gories using a composite measure of age, m a r i t a l status, and number of c h i l d r e n i n the household. The r e s u l t s revealed s i g n i f i c a n t between group d i f f e r e n c e s on both v a r i a b l e s i n r e l a t i o n to such behavioural c r i t e r i a as shopping frequency, downtown shopping and store preference. Despite these d i f f e r e n c e s , Rich and J a i n conclude that t h e i r f i n d i n g s question the usefulness of l i f e - c y c l e and s o c i a l c l a s s concepts i n understanding consumer behaviour i n view of recent changes i n income d i s t r i b u t i o n , education, l e i s u r e time and the movement to suburbia. In a recently completed study of consumer s p a t i a l behaviour i n Metropolitan Vancouver, Gayler (1974) reports s i g n i f i c a n t d i f f e r e n c e s i n the shopping patterns of consumers i n d i f f e r e n t s o c i a l c l a s s and family status groups. In t h i s case, taxonomic v a r i a b l e s and consumer groups were defined using the r e s u l t s of a f a c t o r analysis of various demographic and socio-economic measures. I t was found that the higher the s o c i a l c l a s s group, the more frequently shopping goods were pur-chased and the greater the distance t r a v e l l e d f o r groceries and dress. In terms of family status, o l d e r f a m i l i e s without c h i l d r e n often t r a -v e l l e d s i g n i f i c a n t l y shorter distances than other groups. Furthermore, a s i g n i f i c a n t p o l a r i z a t i o n of preference by s o c i a l c l a s s f o r downtown department stores was shown, although s i g n i f i c a n t d i f f e r e n c e s i n de-partment store patronage between family status groups were not e v i -dent. Consistent with the r e s u l t s of these various studies, Engel et at. (1968, pp. 305-6) place strong emphasis on the usefulness of s o c i a l c l a s s as a basis for d e l i m i t i n g homogeneous consumer groups. They recommend s o c i a l c l a s s as the most s i g n i f i c a n t v a r i a b l e f o r purposes of market segmentation on the basis that s o c i a l c l a s s groups are r e l a t i v e l y homogeneous with respect to income, p s y c h o l o g i c a l v a r i a b l e s , r e s i d e n t i a l l o c a t i o n (within urban areas) and patterns of a c t i v i t y , and are furthermore amenable to q u a n t i f i c a t i o n on the b a s i s of straightforward measures of demographic and socio-economic a t t r i b u t e s . 20. Although such claims are somewhat exaggerated, i t i s probably true that s o c i a l c l a s s and l i f e - c y c l e v a r i a b l e s have been the most f r e -quently employed i n market segmentation studies to date. In t h i s con-text, Carman (1965) presents a d e t a i l e d examination of the a p p l i c a t i o n of s o c i a l class i n market segmentation. An inherent l i m i t a t i o n of s o c i a l c l a s s and l i f e - c y c l e models however i s t h e i r f a i l u r e to provide explanations of consumer behaviour i n terms of the p s y c h o l o g i c a l v a r i a b l e s and processes which c o n s t i t u t e the core components of the comprehensive models r e f e r r e d to i n Section (i) above. In so f a r as the postulated structure of these comprehensive models i s v a l i d , i t follows that s o c i a l c l a s s and l i f e -c y cle measures are only u s e f u l i n accounting f o r behavioural d i f f e r -ences to the extent that they are congruent with the more ba s i c psy-c h o l o g i c a l v a r i a b l e s . This l i m i t a t i o n i s an important f a c t o r i n ex-p l a i n i n g the recent move away from models of consumer behaviour based on demographic and socio-economic v a r i a b l e s i n favour of models i n which the component v a r i a b l e s d i r e c t l y r e f l e c t the p s y c h o l o g i c a l processes assumed to unde r l i e behaviour. Attention now turns to a consideration of four such models. (iv) Image Models The basic p r o p o s i t i o n of image models i s that consumers evaluate a set of a l t e r n a t i v e s (e.g., products, stores or shopping centres) on the basis of t h e i r perceived a t t r i b u t e s or 'image'. I t i s f u r t h e r assumed that the a l t e r n a t i v e with the most p o s i t i v e image w i l l be s e l e c -ted. A v a r i a t i o n of t h i s model i s the image-congruence model which i s founded on the pr o p o s i t i o n that the a l t e r n a t i v e chosen i s the one most congruent with the consumer's s e l f image. In both cases the image i s regarded as a multidimensional concept. Of the various psychological models of consumer behaviour, image models have probably the longest h i s t o r y and are the type which have been most e x p l i c i t l y applied i n the study of shopping behaviour. Here again Martineau's work was i n i t i a t o r y . He was among the f i r s t to argue that a shopper's choice of store was r e l a t e d to store image which he i n t e r p r e t e d as "the way i n which the store i s defined i n the shopper's mind, p a r t l y by i t s f u n c t i o n a l q u a l i t i e s and p a r t l y by an aura of psychological a t t r i b u t e s " (Martineau, 1958a). Various categorizations of the components of store image have subsequently been devised (Fisk, 1961; Kunkel and Berry, 1968; Stephenson, 1969). Hansen (1972, p. 362) argues that three image dimensions are fundamental, namely: p r i c e aspects, f u n c t i o n a l aspects and s o c i a l aspects. He recognizes that p r i c e image may deviate from the actual p r i c e l e v e l , r e f l e c t i n g v a r i a t i o n s i n the consumer's income and other factors such as c r e d i t a v a i l a b i l i t y and sales o f f e r s . The fu n c t i o n a l image i s seen to depend upon merchandise and appearance a t t r i b u t e s , while the s o c i a l image i s based on the consumer's percep-t i o n of who patronizes a p a r t i c u l a r store and i n turn what kind of statement he would be making about himself by p a t r o n i z i n g i t . The most d e t a i l e d image c l a s s i f i c a t i o n i s probably that of Kunkel an'd Berry (1968) who l i s t twelve major components and f o r t y s i x image determinants (see Table 1.1). This scheme was the r e s u l t of a content analysis of unstructured interviews with 744 department store customers 22. TABLE 1.1 A SUGGESTED CLASSIFICATION OF STORE IMAGE COMPONENTS 01 P r i c e of Merchandise a. low p r i c e s b. f a i r or competitive p r i c e s c. high or non-competitive p r i c e s d. values, except with s p e c i f i c regard to premiums, such as stamps, or q u a l i t y of merchan-dise 08 Services a. c r e d i t b. d e l i v e r y c. restaurant f a c i l i t i e s d. other services ( g i f t consultants, layaway plans, baby s t r o l l e r s , e s c a l a t o r s , etc.) 02 Quality of Merchandise a. good or poor q u a l i t y of merchan-dise b. good or poor department(s), ex-cept with respect to assortment, fashion, etc. c. stock brand names 03 Assortment of Merchandise a. breadth of merchandise b. depth of merchandise c. c a r r i e s a brand I l i k e 04 Fashion of Merchandise 09 Sales Promotions a. s p e c i a l s a l e s , i n c l u d i n g q u a l i t y or assortment of sales merchandise b. stamps and other promo-tions c. fashion shows and other s p e c i a l events 10 A d v e r t i s i n g a. s t y l e and q u a l i t y of a d v e r t i s i n g b. media and ve h i c l e s used c. r e l i a b i l i t y of a d v e r t i s i n g 05 Sales Personnel a. a t t i t u d e of sales personnel b. knowledgeability of sales per-sonnel c. number of sales personnel d. good or poor s e r v i c e 06 07 Locational Convenience a. l o c a t i o n from home b. l o c a t i o n from work c. access d. good or poor l o c a t i o n Other Convenience Factors a. parking b. hours store i s open c. convenience with regard to other stores d. store layout with respect to convenience e. convenience (in general) 11 Store Atmosphere a. layout of store without respect to convenience b. external and i n t e r n a l decor of store c. merchandise d i s p l a y d. customer type e. congestion f. good f o r g i f t s , except with respect to q u a l i t y , assortment or fashion of merchandise g. "prestige" store 12 Reputation on Adjustments a. returns b. exchange c. reputation f o r f a i r n e s s (SOURCE: Kunkel and Berry (1968), p. 26.) 23. which y i e l d e d nearly 4000 statements de s c r i b i n g store perceptions. Of those, approximately 99 per cent could be c l a s s i f i e d i n t o the twelve categories shown i n Table 1.1. Si m i l a r but less d e t a i l e d image schemes are reported by F i s k (1961) and Stephenson (1969) f o r food stores. I t i s to be expected that image dimensions w i l l vary dependent upon the type of store under considera-t i o n which suggests that i t i s probably necessary to define the s a l i e n t image determinants independently f o r d i f f e r e n t purchase s i t u a t i o n s . The adoption of image models by geographers f o r the purpose of analysing human s p a t i a l behaviour, i n c l u d i n g shopping behaviour, i s evidenced by the work of Downs (Downs, 1970, 1970a; Downs and Stea, 1973). In one s p e c i f i c study (Downs, 1970), he approached the problem of explaining consumer s p a t i a l behaviour by attempting to measure the images shoppers held of a l t e r n a t i v e shopping centres i n B r i s t o l , England. An 'image' he defined as being "the product of the process of c o l l e c t i n g , coding and evaluating information about the s p a t i a l environment" (p. 15), a view consistent with a decision process model of behaviour. Downs hypothesised that shoppers hold images of shopping centres based on t h e i r evaluations of them i n terms of nine cognitive categories. Since the nature of the images i s seen to determine the consumer's preferences, the centre with the most favourable image was assumed to be the most preferred. This l e d to the t e s t i n g of the hypothesis that the image of the area r e g u l a r l y used would be weighted to the more favourable end of a set of semantic d i f f e r e n t i a l scales (Table 1.2). Unfortunately, Downs encountered an i n s o l u b l e data c o l l e c t i o n problem 24. TABLE 1.2 THE HYPOTHESISED STRUCTURE OF SHOPPING CENTRE COGNITIONS (SOURCE: Downs (1970), p. 22.) Nine Hypothesized Cognitive Categories and T h i r t y - S i x Semantic D i f f e r e n t i a l Scales (1) P r i c e 1 Competitive 2 Many bargains 3 Good value f o r money 4 Many p r i c e cuts (2) Structure and Design 5 Well designed 6 Simple layout 7 Designed with shoppers i n mind 8 Wide pavements (3) Ease of I n t e r n a l Movement and Parking 9 Easy to cross roads 10 Easy to part 11 Not congested 12 Easy to walk around i n (4) V i s u a l Appearance 13 Well kept stores 14 Tidy 15 Clean 16 A t t r a c t i v e (5) Reputation 17 Good reputation 18 Generally w e l l known 19 Generally popular 20 Recommend to friends (6) Range of Goods 21 Good choice 22 Wide range 23 Well stocked 24 Can get i t (7) Service 25 H e l p f u l 26 F r i e n d l y s e r v i c e 27 Good service 28 p o l i t e (8.) Shopping Hours 29 Late c l o s i n g 30 Convenient opening times 31 Good f o r evening shopping 32 Always somewhere open (9) Atmosphere 33 Busy 34 Relaxed atmosphere 35 Personal uncompetitive few bargains poor value f o r money few p r i c e cuts badly designed complicated layout not designed with shoppers i n mind narrow pavements d i f f i c u l t to cross roads d i f f i c u l t to park congested d i f f i c u l t to walk around i n badly kept stores untidy d i r t y u n a t t r a c t i v e bad reputation generally l i t t l e known generally unpopular wouldn't recommend to friends poor choice narrow range badly stocked can't get i t unhelpful un f r i e n d l y s e r v i c e poor s e r v i c e rude e a r l y c l o s i n g inconvenient opening times bad f o r evening shopping never anywhere open not busy tense atmosphere impersonal 2 5 . since respondents were unable to describe t h e i r images o f shopping centres other than i n terms of i n d i v i d u a l stores, and i t was impossible to assume that respondents considered the same stores when d e s c r i b i n g shopping centre images. He was able however to t e s t the hypothesis that the consumer's image of a major shopping centre and her associated preferences were organized i n terms of nine cognitive dimensions. By means of f a c t o r a n a l y s i s , he demonstrated the existence of e i g h t c o g n i t i v e categories: s e r v i c e q u a l i t y , p r i c e , structure and design, shopping hours, i n t e r n a l pedestrian movement, shop range and q u a l i t y , v i s u a l appearance and t r a f f i c conditions. This r e s u l t was seen to accord w e l l with the hypo-t h e s i s . However, Downs recognised that t h i s was an ideographic study, r e l a t e d to a p a r t i c u l a r urban shopping centre, and hence there was no means of assessing the generality of the cognitive structure produced by the a n a l y s i s . A further a p p l i c a t i o n of an image model i n a s i m i l a r context i s reported i n a recent doctoral d i s s e r t a t i o n (Brown, 1974). In t h i s case, image and behavioural data were obtained from a s o c i a l l y homo-geneous sample group r e s i d i n g approximately midway between two shop-ping centres of s i m i l a r s i z e and composition i n Edmonton. I t was thereby intended to hold constant three of the. major v a r i a b l e s which are known to influence shopping behaviour, namely: distance, centre s i z e , and the socio-economic c h a r a c t e r i s t i c s of the consumer. In so doing, i t was hoped to i s o l a t e the r e l a t i o n s h i p between the consumer's image and her patronage behaviour. As i n Down's study, images were measured using a semantic d i f f e r e n t i a l s c a l i n g instrument. Image dimensions and scale scores were derived using f a c t o r a n a l y s i s . When these measures were r e l a t e d to the behavioural data, the research hypo-theses were e s s e n t i a l l y confirmed, demonstrating that consumers d i s -t r i b u t e d t h e i r patronage between the two centres i n accordance with how favourably they perceived them. I t should be added that the usefulness of t h i s and s i m i l a r studies l i e s not i n a r r i v i n g at such s e l f - e v i d e n t conclusions, but i n seeking to c l a r i f y the process whereby consumers evaluate a set of r e t a i l a l t e r n a t i v e s . The studies of both Downs and Brown d i f f e r from many previous image studies i n focussing on shopping centres rather than stores. Comparing the r e s u l t s with those described i n store image studies suggests that several of the dimensions have relevance at both s c a l e s . The v a l i d i t y of shopping centre image measures must remain i n doubt however i f , as Downs found, consumers are unable to di s a s s o c i a t e the image of a shop-ping centre from that of p a r t i c u l a r stores within i t . Dornoff and Tatham (1972) report an i n t e r e s t i n g a p p l i c a t i o n of an image congruence model i n the analysis of shopping behaviour. Extend-ing the work on s e l f theory by Rogers (1965), Dornoff and Tatham hypothesised that i n d i v i d u a l images of pr e f e r r e d stores would vary systematically with self-images, i d e a l self-images and images of best f r i e n d . Data on personal and store images were obtained from 84 r e s -pondents i n C i n c i n n a t i using a pretested semantic d i f f e r e n t i a l s c a l i n g instrument comprising 28 b i p o l a r s c a l e s . The r e l a t i o n s h i p s between the three types of personal image - r e a l s e l f , i d e a l s e l f and image of best f r i e n d - and the images of three classes of preferred store - a supermarket, a department store and a s p e c i a l t y store - were tested. In general, the r e s u l t s supported the hypothesis that an i n d i v i d u a l ' s personal image i s congruent with h i s store image with the greatest congruity being between i d e a l s e l f and stores most p r e f e r r e d and the l e a s t being for image of best f r i e n d and most p r e f e r r e d stores. More-over, v a r i a t i o n s i n the importance of the three types of personal image were observed f o r d i f f e r e n t types of store. In the s e l e c t i o n of a department store, the i d e a l self-image had greater s a l i e n c e than best f r i e n d image; whereas, i n s p e c i a l t y store s e l e c t i o n , the image of best f r i e n d had greater influence than the r e a l s e l f , r e f l e c t i n g the s o c i a l v i s i b i l i t y of s p e c i a l t y store shopping. For the super-market, the d i r e c t i o n of influence was reversed with greater influence being exerted by r e a l self-image than by image of best f r i e n d . These r e s u l t s , though l i m i t e d , are i n d i c a t i v e of the p o t e n t i a l usefulness of image congruence models i n the analysis of consumer dec i s i o n processes. Furthermore, they serve to c l a r i f y the t h e o r e t i -c a l underpinnings of t r a d i t i o n a l image models. (v) Perceptual Preference Models Perceptual preference models are s i m i l a r to image models i n t r e a t i n g behaviour as the outcome of the consumer evaluating a l t e r -natives along a set of perceptual dimensions. They d i f f e r from image models i n that the dimensions are not predefined. Models of t h i s type represent one of the f a s t e s t growing areas i n consumer research associated with the development and a p p l i c a t i o n of non-metric m u l t i -dimensional s c a l i n g techniques (Green and Carmone, 1971; Green and Rao, 1972; Romney, Nerlove and Shepard, 1972). Their p o t e n t i a l a p p l i c a t i o n i n geographic research has been reviewed by Golledge and Rushton (1970). Non-metric multidimensional s c a l i n g has the advantage over other multidimensional s c a l i n g procedures (e.g., f a c t o r analysis) that i t does not require i n t e r v a l data. Nor i s i t necessary to determine i n advance s a l i e n t a t t r i b u t e s of objects to function as scale constructs as i s the case with image models. I t i s termed a 'non-metric' t e c h n i -que since the input data are o r d i n a l although they are seen to possess l a t e n t i n t e r v a l p r o p e r t i e s . This follows from the pioneer research i n t h i s area by Coombs (1950) who noted that, where the existence of a u n i t of measurement i s not assumed, i t i s s t i l l p o s s i b l e to order the magnitude of the i n t e r v a l s between objects, and hence to define an "ordered metric" f o r a psychological s c a l e . The data are provided by respondents making judgements of the r e l a t i v e s i m i l a r i t y of the objects under study. This may involve respondents s p e c i f y i n g which two objects of a set of three they per-ceive to be the most or l e a s t s i m i l a r ; a l t e r n a t i v e l y , they may be asked to i n d i c a t e a most s i m i l a r p a i r among a s e t of p a i r e d objects; or t h i r d l y , data may be derived from a rank-ordering of k - 1 objects i n terms of t h e i r s i m i l a r i t y to the kth. From t h i s data i t i s p o s s i b l e to rank order each object p a i r i n terms of 'subjective' s i m i l a r i t y . The rank order i s then submitted to a computer algorithm which develops a s p a t i a l c o n f i guration repre-senting the data. The dimensionality of the space produced i s minimi-zed on the basi s of a predetermined 'goodness of f i t 1 c o n s t r a i n t . The dimensions represent the perceived c h a r a c t e r i s t i c s of the objects; each object has a point l o c a t i o n i n the space. The p o s i t i o n s of the respondents on these dimensions are represented by ' i d e a l ' points derived from each respondent's rank ordered preferences f o r the objects. The l a t e n t i n t e r v a l p r o p e r t i e s of the data are such that i t i s p o s s i b l e to i n f e r a measure of s i m i l a r i t y between two objects from the distance separating the two points representing them. I t follows t h a t the c l o s e r two points are to each other the more s i m i l a r they are perceived to be by the respondents, and the c l o s e r two ' i d e a l ' points the more s i m i l a r are the preferences of the two respondents thus represented. To date non-metric multidimensional s c a l i n g has been most f r e -quently applied i n the design of new products f o r e x i s t i n g markets, but an a p p l i c a t i o n i n the analysis of consumer s p a t i a l behaviour i s reported i n a recent paper by Doddridge (1972). In t h i s study data were obtained from two small sample groups r e s i d i n g w i t h i n a d i s t r i c t of Sydney, A u s t r a l i a . Respondents i n the two groups d i f f e r e d only i n t h e i r length of residence i n the area, group means being two years and nine years r e s p e c t i v e l y . The research task involved each respondent ranking nineteen centres i n order of preference and i n d i c a t i n g the frequency of use of, and money spent at, each centre i n the preceding s i x month period. I t was proposed that respondents i n the two groups would d i f f e r i n the perceptual dimensions employed to evaluate a l t e r -native shopping centres f o r the purchase of c l o t h i n g , hence demonstra-t i n g d i f f e r e n t states of l e a r n i n g and perception. This p r o p o s i t i o n was supported by the data: whereas l e s s well-informed respondents evaluated a l t e r n a t i v e s on two dimensions defined as 'likeness to the 30. c i t y ' and 'expected value f o r money and e f f o r t ' , longer term residents were found to employ perceptual dimensions l a b e l l e d as 'number of unoccupied parking spaces close to shops' and 'learnt value f o r money and e f f o r t ' . The dimensions as defined c l e a r l y suggest consistent between group differences i n lear n i n g s t a t e s , although i t should be noted that the d e f i n i t i o n of perceptual dimensions derived using non-metric multidimensional s c a l i n g procedures i s a p a r t i c u l a r l y d i f f i c u l t problem and i s open to a good deal of subjective i n t e r p r e t a t i o n . In addi t i o n , Doddridge found support f o r propositions l i n k i n g the percep-t u a l evaluations of shopping centres and measures of shopping t r a v e l and purchasing behaviour. Work i n progress at the Un i v e r s i t y of B r i t i s h Columbia i n d i c a t e s a further a p p l i c a t i o n of a perceptual preference approach to the analy-s i s of r e t a i l patronage (Forbes and Wiley, i n progress). In t h i s case, respondents were required to provide p a i r e d comparison ratings of s i x Vancouver department stores plus an i d e a l store using a ten poin t s i m i l a r i t y - d i s s i m i l a r i t y scale. I t i s hoped using t h i s data to demon-st r a t e s i g n i f i c a n t r e l a t i o n s h i p s between these measures and other perceptual, demographic and behavioural v a r i a b l e s . I t would be premature at t h i s point i n time to evaluate the con-t r i b u t i o n of perceptual preference models to consumer behaviour research since t h e i r p o t e n t i a l i s only now being explored. Nonetheless, the e f f o r t s and progress i n t h i s area to date suggest that t h i s type of model may prove to be amongst the most productive yet employed. At the same time, the use o f non-metric multidimensional s c a l i n g techniques has attendant problems (Day, 1972) and the l i m i t a t i o n s which these impose have yet to be f u l l y i d e n t i f i e d , (vi) Learning Models The models described thus f a r are e s s e n t i a l l y s t a t i c , although i t i s widely recognised that the behaviour of consumers i s a dynamic process r e f l e c t i n g the operation of le a r n i n g processes. Consumer researchers have therefore drawn upon stimulus-response and cognitive lea r n i n g theories i n developing models of consumer behaviour (Haines, 1969; Howard and Sheth, 1969). The p o t e n t i a l a p p l i c a t i o n of l e a r n i n g models i n the study of consumer s p a t i a l behaviour has also received s p e c i f i c treatment by geographers (Golledge, 1967, 1969; Golledge and Brown, 1967; Hudson, 1970, 1971). Golledge (1967) has i d e n t i f i e d e s s e n t i a l l y two phases i n the temporal development of consumer s p a t i a l behaviour; the search phase and the h a b i t u a l response phase, which he summarizes as follows: "As search proceeds and l e a r n i n g about the system takes p l a c e , the range of f e a s i b l e opportunities becomes more sharply defined and the range of s p a t i a l behaviour becomes more l i m i t e d u n t i l asymptotic pat-terns evolve". This p r o p o s i t i o n of course assumes s t a b i l i t y i n the supply environment and the non-occurrence o f ' i r r a t i o n a l ' purchases i n the form of impulse buying f o r example. The i d e n t i f i c a t i o n of a temporal trend i n s p a t i a l behaviour i s i n i t s e l f i n s u f f i c i e n t to account f o r the s p e c i f i c l o c a t i o n s which are patronized. Certain expectations are reasonable however given that s p a t i a l search involves costs f o r the i n d i v i d u a l which he balances against the b e n e f i t s gained from the search phase. In t h i s regard, 32 . Cyert and March (1963) have s p e c i f i e d a set of rules that provide an i n i t i a l b asis f o r a n t i c i p a t i n g the conduct of search a c t i v i t y . These are: (1) search i n the neighbourhood of the problem s i t u a t i o n (presu-mably proximate to the t r i p o r i g i n ) ; (2) search i n the neighbourhood of the current a l t e r n a t i v e ; and (3) i f neighbourhood search procedures do not provide a s a t i s f a c t o r y s o l u t i o n , then successively use "more d i s t a n t " search procedures. The product of the search phase i s regarded by Golledge and Brown (1967) as the accumulation of information about the s p a t i a l system i n which the i n d i v i d u a l finds himself. I n i t i a l t r i a l and e r r o r behaviour i s seen to give way with i n c r e a s i n g knowledge and the reinforcement of s a t i s f y i n g solutions to an e q u i l i b r i u m state comprising the h a b i t u a l response phase. Golledge and Brown r e l a t e t h i s process to the le a r n i n g process described by H u l l (1962), Spence (1951) and Bush and M o s t e l l e r (1955), and following the lead of the l a t t e r proceed to develop a s t o c h a s t i c model of the market decision process. Golledge (1970) reports the r e s u l t s of a study based on a sample o f new migrants to Madison, Wisconsin, which e s s e n t i a l l y confirm model expectations. Results i n d i c a t e d a decrease i n the number of places v i s i t e d f o r groceries over time; a greater d i v e r s i t y i n t r i p behaviour among recent in-migrants than among long term residents; and a marked decrease i n new store t r i a l s over time. Hudson (1970) on the other hand, following a comprehensive exami-nation of the relevance of various l e a r n i n g theories to consumer spa-t i a l choice, concluded that stimulus-response theories s i m i l a r to that of H u l l were inappropriate and that a more promising approach to under-3 3 . standing such behaviour was v i a cogni t i v e l e a r n i n g t h e o r i e s . In t h i s regard, he argued that a l e a r n i n g model of s p a t i a l behaviour derived e s s e n t i a l l y from the work of Tolman (1932, 1952) and Lewin (1936, 1936a, 1951) could be subsumed under K e l l y ' s Theory of Personal Con-st r u c t s (Kelly, 1955). In a subsequent paper, Hudson (1971) proposed a research design f o r t e s t i n g the r e l a t i o n s h i p between consumer personal constructs measured at three d i f f e r e n t points i n time and s p a t i a l behaviour. In order to maximize the e f f e c t s of learning, he decided to obtain data from u n i v e r s i t y students i n B r i s t o l who had made recent i n t e r or i n t r a -urban r e s i d e n t i a l moves. Constructs r e l a t i n g to each of eleven fami-l i a r food stores and an i d e a l store were to be measured on an eleven poin t scale. I t was hoped using t h i s data to demonstrate increases i n knowledge of the shopping environment over time; greater accuracy i n store images with increased patronage frequency; greater conformity over time between asp i r a t i o n s and perceived shopping opportunities; and consistency between construct and behaviour changes. Furthermore, he suggested the p o s s i b i l i t y of developing a p r o b a b i l i s t i c model of store patronage based on the deviation of store image scores from the scores f o r the i d e a l store. In a s i m i l a r context, Doddridge (1972), i n the study previously described under the heading of perceptual preference models, has demonstrated changes in. consumer perceptions of s p a t i a l a l t e r n a t i v e s over time which she argued were i n d i c a t i v e of a cognitive l e a r n i n g process. There c l e a r l y remains much more scope f o r the f u r t h e r i n v e s t i g a t i o n 34. of the temporal dynamics of consumer s p a t i a l behaviour, and, i n t h i s regard, the a p p l i c a t i o n of cognitive learning theories i n the manner suggested by Hudson and i n part demonstrated by Doddridge appears to be a p o t e n t i a l l y f r u i t f u l avenue of i n q u i r y , and one which i s l i k e l y to y i e l d more i n s i g h t than can be provided by stimulus-response models. ( v i i ) A t t i t u d e and Value Models The a p p l i c a t i o n of a t t i t u d e and value models has been of major s i g n i f i c a n c e i n consumer research. This work has b u i l t upon s o c i a l psychological research i n qhich c e n t r a l importance has frequently been placed on a t t i t u d e s and values as a basis f o r i n t e r p r e t i n g behaviour. An a t t i t u d i n a l component i s included and indeed assumes a key p o s i t i o n i n each of the three major comprehensive models r e f e r r e d to i n an e a r l i e r s e c t i o n . One s u b f i e l d of Nicosia's d e c i s i o n model focusses on consumer a t t i t u d e formation (Nicosia, 1966); Howard and Sheth (1969) i n t h e i r theory of buyer behaviour recognise a t t i t u d e as a major l e a r n -ing construct and as the p r i n c i p a l intermediary between brand compre-hension and behavioural i n t e n t i o n ; while Engel et al. (1968) ascribe to a t t i t u d e s and values the function of summarizing the i n f l u e n c e of s a l i e n t consumer a t t r i b u t e s within the 'central c o n t r o l u n i t ' of t h e i r model of consumer motivation and behaviour (see Figure 2.1). Indeed Engel et al. go so f a r as to suggest that a t t i t u d e s comprise the most important component i n an i n d i v i d u a l ' s "map of h i s world" (p. 165). Empirical studies of consumer behaviour i n v o l v i n g a t t i t u d e measure-ment have been numerous, while varying markedly i n rigour and i n the methodological techniques employed. Each of the accepted measurement techniques have been used i n c l u d i n g Thurstone, L i k e r t , Guttman and semantic d i f f e r e n t i a l s c a l i n g . The published proceedings of the A t t i t u d e 35. Research Committee of the American Marketing A s s o c i a t i o n provide a compendium of the types of study conducted i n t h i s area (Adler and Crespi, 1966, 1968; King and T i g e r t , 1971; Haley, 1972). Consistent with the work i n s o c i a l psychology, consumer behaviour researchers have recognised that the structure of an a t t i t u d e i s d i v i -s i b l e i n t o three components: cognition (or b e l i e f ) , a f f e c t (or eva-luation) and conation (or i n t e n t i o n ) . Consumer studies vary i n t h e i r treatment of these components but most demonstrate an attempt to account f o r behavioural d i f f e r e n c e s i n terms of b e l i e f and/or evalua-t i o n measures. A t t i t u d e models range i n complexity from those based upon unidimensional measures of a f f e c t to those, i n v o l v i n g multiple a t t r i b u t e measures of both cognition and a f f e c t . The l a t t e r are becoming i n c r e a s i n g l y favoured over simple measues of preference because of t h e i r greater 'diagnostic' value. As Wilkie and Pessemier (1973) i n d i c a t e i n a recent review, the major purpose of m u l t i - a t t r i b u t e a t t i t u d e models i s to increase understanding of consumer a t t i t u d e structures and hence to provide information which has d i r e c t a p p l i c a -t i o n i n the formulation of marketing strategy. Day (1972) furt h e r d i s t i n g u i s h e s between unidimensional measures of a f f e c t and more complex models i n regarding the former as i l l u s t r a t i v e of a purely p r e d i c t i v e approach while the l a t t e r demonstrate an explanatory or s t r u c t u r a l approach. To the extent that consumer researchers, have attended to a t t i t u d e theory, the work of Rosenberg (1956, 1960, 1965) and Fishbein (1963, 1965, 1967, 1967a, 1972) has been most frequently c i t e d (Day, 1972; Cohen et al. 1972; Wilkie and Pessemier, 1973). Rosenberg sought to delineate v a r i a b l e s that were assumed to covary with a t t i t u d e which he defines as "a r e l a t i v e l y stable a f f e c -t i v e response to an object" (1956, p. 367). He regarded one of these variables to be the i n t e n s i t y o f a person's values and a second to be the perceived importance of the a t t i t u d e object i n leading to or b l o c k i n g the attainment of h i s values. He hypothesised therefore that a person's a t t i t u d e toward a given object would be "accompanied by a c o g n i t i v e structure made up of b e l i e f s about the p o t e n t i a l i t i e s of that object f o r a t t a i n i n g or blocking the r e a l i z a t i o n of valued s t a t e s " (1956, p. 367). I t follows that the more the object leads to attainment of p o s i t i v e l y valued states or blocks the attainment of negatively valued states the more the person would have a p o s i t i v e a t t i t u d e toward the object. Rosenberg's hypothesis has been expressed a l g e b r a i c a l l y as follows: n A_ = / I.V. where A„ = a t t i t u d e toward object 0. 0 , L . l l 0 J i = l I. = perceived instrumentality, the extent to which the person believes that the object 0 w i l l lead to or block the attainment of value i . V. = value importance, value i ' s importance to the respondent as a source of s a t i s f a c t i o n . n = number of values. Fishbein, w r i t i n g within a behaviour theory framework, developed an" a t t i t u d e model based upon processes of mediated g e n e r a l i z a t i o n and conditioning, which he has summarized i n the following way: 37. An i n d i v i d u a l holds many b e l i e f s about any given object; that i s many d i f f e r e n t c h a r a c t e r i s t i c s , a t t r i b u t e s , values, goals and concepts are p o s i t i v e l y or negatively associated with any given object; associated with each of these 'related objects' i s a mediating evaluative response - an a t t i t u d e ; these evaluative responses summate; through the mediation process, the summated evaluative response i s associated with the a t t i t u d e object; and thus on future occasions the a t t i -tude object 'w i l l e l i c i t t h i s summated evaluative response -t h i s a t t i t u d e ... According to the theory, then, an i n d i v i -dual's a t t i t u d e toward any object i s a function of (1) the strength of h i s b e l i e f s about the objects and (2) the eva-l u a t i v e aspect of those b e l i e f s . (1965, p. 117). Expressed a l g e b r a i c a l l y the model takes the following form: n A„ = > B.a. where: A„ = a t t i t u d e toward object 0. 0 .L. l l 0 1 J 1=1 B. = the strength of b e l i e f i about the a t t i t u d e object 0, that i s , the p r o b a b i l i t y that 0 i s r e -l a t e d to some other object x.. i a. = the evaluative aspect of B^, that i s the evaluation of x.. l n = number of b e l i e f s . Thus despite d i f f e r e n c e s i n t h e i r t h e o r e t i c a l premises, Rosenberg and Fishbein have developed remarkably s i m i l a r models, both of which regard an i n d i v i d u a l ' s a t t i t u d e toward any object as a function of h i s b e l i e f s about the object and the evaluative aspects of those b e l i e f s . As applied to consumer a t t i t u d e s , Cohen et al. (1972) note that, given a s et of s a l i e n t or relevant a t t r i b u t e s used by consumers to choose among a l t e r n a t i v e s , both models require information d e t a i l i n g : (1) the extent to which the consumer believes the a l t e r n a t i v e i s r e l a t e d to or possesses each a t t r i b u t e , and (2) h i s evaluation of or s a t i s f a c -t i o n he would derive from each a t t r i b u t e . 38. To date the a p p l i c a t i o n of these models i n consumer research has been r e s t r i c t e d to questions of product and brand choice, although there i s no reason why they should not be extended to the analysis of s p a t i a l choices. Wilkie and Pessemier (1973) provide an e x c e l l e n t review of work i n t h i s area and i n d i c a t e the s i g n i f i c a n t issues which have arisen and i n p a r t remain to be solved. These issues include the determination of the number and type of s a l i e n t a t t r i b u t e s f o r i n c l u s i o n i n the model; the u t i l i t y and measurement of importance weights; the conceptualisation and measurement of b e l i e f s ; the mathematical manipulation of model com-ponents; and the comparison of t h i s type of model with a l t e r n a t i v e formulations. I t i s encouraging to witness the rigour t y p i c a l of recent work employing m u l t i - a t t r i b u t e a t t i t u d e measures, which i s i n marked contrast to the haphazard and ad hoa approaches i n e a r l i e r consumer at t i t u d e studies where too often p s y c h o l o g i c a l t o o l s were borrowed with i n s u f f i c i e n t consideration as to t h e i r appropriateness. Wilkie and Pessemier (1972, p. 438) also report that i n a l l cases where the p r e d i c t i v e v a l i d i t y of m u l t i - a t t r i b u t e a t t i t u d e measures has been tested against behavioural c r i t e r i a p o s i t i v e r e s u l t s have been achieved. They therefore conclude that there i s l i t t l e question that brand a t t i t u d e s w i l l p r e d i c t brand preferences or c o n t r o l l e d choice behaviour s i g n i f i c a n t l y b e t t e r than chance assignments of preference or choice. I t remains an open question whether a s i m i l a r conclusion can be reached i n r e l a t i o n to consumers 1 s p a t i a l choices, but there seems l i t t l e doubt that the p o s s i b i l i t y i s worthy of exploration. Given the emphasis placed upon a t t i t u d e measurement i n marketing research, i t i s s u r p r i s i n g that a t t i t u d e models have received such 3 9 . scant attention i n the analysis of consumer s p a t i a l behaviour by geo-graphers. This i s perhaps i n d i c a t i v e of an over-emphasis having been placed on Harvey's p r e c i p i t o u s judgement regarding the p o t e n t i a l u t i l i t y of a t t i t u d i n a l measures i n the development of cognitive-beha-v i o u r a l theory (Harvey, 1969). At the same time, the uncertainty of the attitude-behaviour l i n k demonstrated by s o c i a l p s y c h o l o g i c a l studies cannot be ignored (Thomas, 1972). There i s reason to suspect however that the r e l a t i v e n e u t r a l i t y and f a m i l i a r i t y of the behaviour commonly inv e s t i g a t e d by geographers (including consumer s p a t i a l behaviour) w i l l suppress the major confounding in f l u e n c e t y p i c a l of t r a d i t i o n a l a t t i t u d e research i n s o c i a l psychology, where, frequently, controver-s i a l opinions have been sought i n r e l a t i o n to r e l a t i v e l y u n f a m i l i a r behavioural s i t u a t i o n s . The p o t e n t i a l u t i l i t y of a t t i t u d e models i n the study of s p a t i a l behaviour has been the subject of general d i s c u s s i o n by Murphy and Golledge (1972) and Golledge (1970). Furthermore, Murphy's d i s s e r t a t i o n i l l u s t r a t e s a preliminary exploration of the r e l a t i o n s h i p between a t t i -t u d i n a l measures and behavioural c r i t e r i a i n c l u d i n g shopping centre patronage (Murphy, 1970). The r e s u l t s confirmed that consumers were more l i k e l y to patronize f a c i l i t i e s towards which, they held a p o s i t i v e a t t i t u d e than those towards which they were n e u t r a l or negatively predisposed. However, the questionable adequacy of the research i n s t r u -ment ( a f i v e construct semantic d i f f e r e n t i a l s c a l e ) , the sample and the data c o l l e c t i o n procedures p r o h i b i t d e f i n i t i v e conclusions. Espe-c i a l l y r e s t r i c t i v e i s that, by choosing to employ b i p o l a r a d j e c t i v a l constructs as h i s a t t i t u d i n a l measures, Murphy f a i l s to i d e n t i f y the 40. a t t r i b u t e s of the locations which generate the consumer's evaluation of a l t e r n a t i v e s . A more i n s t r u c t i v e study, r e l a t i n g measures of the cognitive component of consumer atti t u d e s ( i . e . , b e l i e f s ) to s p a t i a l behaviour and membership i n demographic groups, i s that reported by Margulis (1972). Data were c o l l e c t e d from 300 respondents i n Newark, New Jersey, d e t a i l i n g responses to twenty two b e l i e f statements measured on a ten-p o i n t scale, demographic c h a r a c t e r i s t i c s , and the distances t r a v e l l e d to purchase f u r n i t u r e . The b e l i e f statements r e l a t e d to various c r i t e -r i a , i n c l u d i n g merchandise q u a l i t y , c r e d i t a v a i l a b i l i t y and shopping convenience, which were assumed to underlie consumer choices. Using Friedman's two-way analysis of variance by ranks, Margulis demonstrated s i g n i f i c a n t r e l a t i o n s h i p s between scores on c e r t a i n b e l i e f statements and the various demographic measures, which were: e t h n i c i t y , income, family s i z e , age and education. S i g n i f i c a n t r e l a t i o n s h i p s were also found between b e l i e f s and the l o c a t i o n of the purchase. When b e l i e f measures were c o r r e l a t e d with distances t r a v e l l e d to shop, twenty of the twenty two c o e f f i c i e n t s were s i g n i f i c a n t , a s u r p r i s i n g l y high pro-portion e s p e c i a l l y as many of the statements were not obviously distance r e l a t e d . I t i s also i n t e r e s t i n g that each of the s i g n i f i c a n t r e l a t i o n -ships was negative i n d i c a t i n g that the more p o s i t i v e the shopper's b e l i e f about a p a r t i c u l a r store the shorter the distance t r a v e l l e d . This r e s u l t perhaps suggests that consumers b o l s t e r t h e i r b e l i e f s to j u s t i f y to themselves and others t h e i r unwillingness to search out more d i s t a n t opportunities. C e r t a i n l y the r e s u l t s are i n t r i g u i n g and i n d i c a t e the p o t e n t i a l u t i l i t y of follow-up studies. One l i m i t a t i o n 41. should be noted however and that i s the treatment of each b e l i e f s t a t e -ment as an independent v a r i a b l e rather than the more common procedure i n psychometrics of combining scores on r e l a t e d statements. Given the v u l n e r a b i l i t y of s i n g l e statement responses to measurement e r r o r , the r e l i a b i l i t y and v a l i d i t y of the data i s enhanced by obtaining summative scale scores. This of course requires p r i o r d e f i n i t i o n of the s a l i e n t s c a l e constructs, which i s another issue unfortunately not mentioned by Margulis i n r e l a t i o n to h i s s e l e c t i o n of b e l i e f statements. I t i s commonly recognized that a consumer's a t t i t u d e s towards spe-c i f i c stimulus objects are symptomatic of underlying and more general value o r i e n t a t i o n s . However, i n many instances a t t i t u d e s and values are not o p e r a t i o n a l l y d i s t i n g u i s h e d but are rather assumed to be essen-t i a l l y synonymous as i s the case i n Engel et al. 's model of consumer behaviour (1968, p. 166). While t h i s assumption may be j u s t i f i e d on pragmatic grounds, i t does ignore the d i s t i n c t c h a r a c t e r i s t i c s of con-sumer values which warrant separate a t t e n t i o n . Seminal work on the study of value systems i s that of Rokeach (1968, 1969). He has defined a 'value' as an "enduring b e l i e f that a p a r t i c u l a r mode of conduct or a p a r t i c u l a r end-state of existence i s personally and s o c i a l l y p referable to a l t e r n a t i v e modes of conduct or end-states of existence" (1969, p. 550). Values are then ends i n themselves i n contrast to at t i t u d e s towards s p e c i f i c objects which are means towards the attainment of valued end-states. I t i s therefore reasonable to regard p a r t i c u l a r a t t i t u d e s and b e l i e f s as being derived from values (Bern, 1970, p. 17). Rosenberg's a t t i t u d e model, c i t e d p r e v iously, which incorporates value r e l a t e d measures as determinants 42. of a f f e c t , lends further support f o r t h i s view. Consistent with h i s d e f i n i t i o n , Rokeach dis t i n g u i s h e s between 'terminal values' r e l a t e d to end-states of existence (e.g., e q u a l i t y , s a l v a t i o n , s e l f - f u l f i l l m e n t , and freedom) and 'instrumental values' r e l a t e d to broad modes of conduct (e.g., courage, honesty, f r i e n d s h i p , and c h a s t i t y ) . Rokeach (1969) presents t h i r t y s i x such terminal and instrumental values which have been found to d i f f e r s i g n i f i c a n t l y among d i f f e r e n t c r i t e r i o n groups. An a l t e r n a t i v e c l a s s i f i c a t i o n i s suggested by Janis (1959) who recognises the following four types of values: (1) a n t i c i p a t e d u t i l i -t a r i a n gains f o r s e l f ; (2) a n t i c i p a t e d u t i l i t a r i a n gains f o r others; (3) a n t i c i p a t e d approval and disapproval from s e l f ; (4) a n t i c i p a t e d approval and disapproval from others. Hansen (1972) regards t h i s l a t t e r c l a s s i f i c a t i o n scheme as p a r t i c u l a r l y appropriate to the kinds o f values s a l i e n t i n consumer choice processes although there i s as yet no empirical evidence to support t h i s contention. To date studies r e l a t i n g consumer values to behaviour have r e l i e d upon the use of motivational and p e r s o n a l i t y research techniques deve-loped i n psychology. Edwards' Personal Preference Schedule (Edwards, 1954) i s a p e r s o n a l i t y inventory which has been frequently applied i n studies of consumer behaviour. I t was used by Evans (1958, 1968) i n attempting to discriminate between Ford and Chevrolet owners, and by Koponen (1960) and Massy et al. (1968) i n seeking to account f o r d i f f e r -ences on a number of consumer purchase v a r i a b l e s . Various other studies i l l u s t r a t e the use of other p e r s o n a l i t y i n v e n t o r i e s . O v e r a l l , the r e s u l t s achieved have not been impressive; as Wells (1966, p. 187) 4 3 . s u c c i n c t l y s t a t e s : "The findings of these studies have been very con-s i s t e n t . Almost always they have r e s u l t e d i n s t a t i s t i c a l l y s i g n i f i -cant c o r r e l a t i o n s that have been too small to be of much p r a c t i c a l value ". I t would be premature however to conclude on the b a s i s of t h i s evidence that there i s l i t t l e to be gained i n understanding consumer behaviour from the measurement of consumer values. I t may w e l l be that the apparent lack of success i n previous studies i s l a r g e l y a t t r i b u t a b l e to the research instruments that have been used. I t i s questionable whether te s t s developed i n psychology f o r c l i n i c a l and other uses can be expected to account for d i f f e r e n c e s i n a seemingly unrelated area such as consumer choice behaviour. In t h i s regard, Engel al. (1968, p. 150) advocate the development of t e s t instruments s p e c i f i c a l l y designed for consumer research. In the few cases where t h i s approach has been adopted the r e s u l t s have been encouraging. White (1966), f o r example, has demonstrated c l e a r d i f f e r e n c e s i n product preferences between groups d i f f e r i n g i n t h e i r scores on s p e c i f i c a l l y designed value dimensions. An e a r l y study, i n d i c a t i v e of the types of value o r i e n t a t i o n s s a l i e n t to consumer s p a t i a l choices, i s that of Stone (1954) i n which Chicago respondents were asked whether they would p r e f e r to patronize l o c a l stores or large chain stores and why. On the b a s i s of t h e i r d i f f e r e n t o r i e n t a t i o n s Stone recognized four major classes o f shopper which were: the 'economic' shopper, who was extremely s e n s i t i v e to p r i c e , q u a l i t y and assortment of merchandise; the 'personalising' shopper who formed strong r e l a t i o n s with store personnel which were c r u c i a l i n determining store choice; the ' e t h i c a l ' shopper who was w i l l i n g to s a c r i f i c e lower p r i c e s and wider s e l e c t i o n i n order to support the small r e t a i l e r i n the face of competition from the chain stores; and the 'apathetic' shopper who found shopping an onerous task, and hence l o c a t i o n a l convenience proved c r u c i a l to her s e l e c t i o n of a store rather than p r i c e , q u a l i t y of goods, r e l a t i o n s with store personnel or e t h i c a l considerations. In a more recent study Kenny-Levick (1969) attempted to demonstrate the extent to which Stone's typology could be used to segment the gro-cery trade market i n L i v e r p o o l , England. A sample of 554 housewives were asked why they p r e f e r r e d to shop at t h e i r f i r s t choice grocery store. Responses were c l a s s i f i e d i n terms of twelve categories which were then collapsed to form f i v e broader based d i v i s i o n s . Kenny-Levick i d e n t i f i e d these f i v e categories with Stone's four shopper types plus a miscellaneous category. He then compared the percentage representation of consumers i n the four corresponding categories. He found that a s i g n i f i c a n t d i f f e r e n c e e x i s t e d between the two d i s t r i b u -t i o n s . L i v e r p o o l housewives seemed to be more 'apathetic' and l e s s ' e t h i c a l ' than t h e i r Chicago counterparts. The 'personalizing' f a c t o r was stronger i n Chicago, while the 'economic' f a c t o r was of equal and the greatest importance to both sample groups. Kenny-Levick con-cluded that L i v e r p o o l shoppers could best be described i n terms of a sevenfold typology, the f i r s t four being coincident with Stone's, while the remaining three were defined as 'time-saving', 'enhancement of self-image' and 'pleasure seeking'. Although both of these studies are open to methodological c r i t i c i s m , the r e s u l t s are suggestive as 45. to the types of value orientations which may be salient in the analysis of consumer spatial behaviour. There i s clearly much scope for further exploration of the relation-ships between consumer values and spatial behaviour. The use of values rather than specific attitudes as independent variables in consumer research is particularly appealing because of their more pervasive character and hence their potentially wider explantory power. As Levy (1966) has noted, a l l too often consumer variables most highly related to behaviour have been so close to the behaviour as to be redundant in explaining i t . There is reason to hope that models based upon appro-priate measures of consumer values would avoid this frustrating c i r c u l a r i t y while, at the same time, proving effective in accounting for behavioural differences. C. SYNTHESIS AND THE CONTEXT OF THE PRESENT STUDY The purpose of this review of consumer behaviour models has been to i l l u s t r a t e the various approaches open to the student of consumer spatial behaviour, some of which have been extensively explored already, others p a r t i a l l y so, and others hardly at a l l . Although each model has been discussed separately, clear interrelationships exist between them, and particularly between those which are based upon psychological variables as their integration within comprehensive models demonstrates. There seems, for example, to be a close similarity between image and attitude models, especially where the latter involve the use of seman-t i c d i f f e r e n t i a l measures based on the work of Osgood et al. (1957). This similarity has been noted in a geographic context by Downs (1970a). The present study r e l a t e s and contributes to the e x i s t i n g l i t e r a -ture i n two ways. F i r s t l y , a d i s p o s i t i o n a l model of consumer s p a t i a l behaviour i s developed and tested. This representes an extension of the research described i n the previous section under the heading o f a t t i t u d e and value models. The term dispositions i s used following the lead of Craik (1969, 1970) and McKechnie (1972) who have employed the term i n environmental psychology to describe those p s y c h o l o g i c a l dimensions, underlying s p e c i f i c a t t i t u d e s and behaviour, which are used by the i n d i v i d u a l to describe, comprehend and evaluate a set of environmental objects. This d e f i n i t i o n was seen to coincide c l o s e l y with the c h a r a c t e r i s t i c s of the p s y c h o l o g i c a l constructs included i n the present study. Following from the previous d i s c u s s i o n of a t t i t u d e and value models, the r o l e of d i s p o s i t i o n a l v a r i a b l e s within a model of consumer s p a t i a l behaviour i s hypothesised to be as shown i n Figure 1.2. This flow diagram i s e s s e n t i a l l y a s i m p l i f i e d and modified version of Engel et al. ' s model shown i n Figure 1.1. Purchasing behaviour i s regarded as the outcome of a sequential process whereby the consumer i n t e r a c t s with and c o l l e c t s information concerning the r e t a i l environ-ment leading to the formation of d i s p o s i t i o n s , which summarize the values the consumer seeks to s a t i s f y through h i s behaviour. These d i s p o s i t i o n s are l i n k e d to s p e c i f i c r e t a i l a l t e r n a t i v e s ( i . e . , stores or shopping centres) through the formation of a t t i t u d e s . I t i s on the b a s i s of these a t t i t u d e s that the a l t e r n a t i v e s are evaluated and a p r e f e r r e d a l t e r n a t i v e decided upon. The conduct of s p a t i a l behaviour 47. i s contingent upon t h i s d e c i s i o n . The degree of s a t i s f a c t i o n attained gives r i s e to r e i n f o r c i n g or modifying feedback e f f e c t s , but, since a s t a t i c form of the model was tested i n the present study, these dynamic components are omitted. FIGURE 1.2 A DISPOSITIONAL MODEL OF CONSUMER SPATIAL BEHAVIOUR DISPOSITIONS XL SPECIFIC ATTITUDES EVALUATION OF ALTERNATIVES DECISION SPATIAL BEHAVIOUR RETAIL ENVIRONMENT (Alternative Stores/Centres) 48. The research presented i n t h i s d i s s e r t a t i o n i n f a c t i s l i m i t e d to a p a r t i a l t e s t of t h i s model since the data set was r e s t r i c t e d to d i s -p o s i t i o n a l and behavioural measures. The intervening components and i n t e r r e l a t i o n s h i p s were therefore i n f e r r e d rather than demonstrated, t h e i r v e r i f i c a t i o n would c l e a r l y require that a t t i t u d i n a l measures also be obtained. The second contribution of t h i s study to the e x i s t i n g l i t e r a t u r e i s to compare the e f f i c a c y of a d i s p o s i t i o n a l model with t h a t of two other types of model as p r e d i c t o r s of consumer s p a t i a l behaviour. The two other models are based on l o c a t i o n a l and b i o g r a p h i c a l v a r i a b l e s and e s s e n t i a l l y represent the u t i l i t y and s o c i a l c l a s s categories as pre-v i o u s l y o u t l i n e d . Since the e x i s t i n g l i t e r a t u r e i s replete with ex-amples of the l a t t e r two model types, emphasis i n the f o l l o w i n g two chapters i s placed on the development of a d i s p o s i t i o n a l model. D. SUMMARY The l i t e r a t u r e review presented i n t h i s chapter has established the context of the present study. Various consumer behaviour models have been described with p a r t i c u l a r emphasis being placed on t h e i r a p p l i c a t i o n to the analysis of consumer s p a t i a l behaviour. The research presented i n subsequent chapters b u i l d s on previous studies i n develop-ing and t e s t i n g a d i s p o s i t i o n a l model of consumer s p a t i a l behaviour and i n comparing the p r e d i c t i v e power of such a model with that of models ba'sed on l o c a t i o n a l and b i o g r a p h i c a l v a r i a b l e s . In the f o l l o w i n g chapter, attention switches from a discussion of the general context of the study to,a d e t a i l e d consideration of the research objectives and the research design. CHAPTER 2 RESEARCH DESIGN A. INTRODUCTION This chapter begins with a statement of the research purposes of t h i s study. The research design formulated to achieve these purposes i s then presented and i t s major components are summarized. Since the construction and development of a psychometric s c a l i n g instrument i s c e n t r a l to the design, the assessment of the r e l i a b i l i t y and v a l i d i t y of such measurement to o l s i s considered. B. RESEARCH PURPOSES The objectives of t h i s study, which are i n p a r t substantive and i n part methodological, can be itemized as follows: 1. To provide empirical evidence of the r e l a t i o n s h i p between consumer d i s p o s i t i o n s and s p a t i a l behaviour. 2. To compare models based upon l o c a t i o n a l , b i o g r a p h i c a l and d i s p o s i t i o n a l measures as p r e d i c t o r s of shopping behaviour using data c o l l e c t e d from a s t r a t i f i e d random sample of households within the c i t y of Vancouver and immediate environs. 3. To assess* the incremental v a r i a t i o n i n behaviour accounted f o r by the i n c l u s i o n of d i s p o s i t i o n a l v a r i a -bles i n the set of p r e d i c t o r s . - 49 -50. 4. To demonstrate the construction, development and t e s t -ing of a set of L i k e r t scales i n a geographical context. Two major considerations governed the establishment of the f i r s t three o b j e c t i v e s . In the f i r s t p lace, a strong argument has been made by various researchers that the i d e n t i f i c a t i o n and measurement of s a l i e n t cognitive v a r i a b l e s would s i g n i f i c a n t l y a i d i n the development of models of s p a t i a l behaviour by combining the q u a l i t i e s of p r e d i c t i v e accuracy and t h e o r e t i c a l adequacy (Downs, 1971; Golledge, 1970; Harvey, 1969; Winkel, 1971). I t seemed necessary therefore, i n the context of e x i s t -ing cognitive-behavioural research i n geography, to supplement the paucity of empirical studies documenting the r e l a t i o n s h i p between cog-n i t i v e measures and v a r i a t i o n s i n behaviour. Hence the establishment of the f i r s t research o b j e c t i v e . Secondly, since a considerable amount of research has already been completed i n the s p e c i f i c area of consumer s p a t i a l behaviour, i t was des i r a b l e to formulate a research design which would allow comparison of a model based upon cognitive v a r i a b l e s with models based upon measures commonly employed i n e x i s t i n g studies. This r a t i o n a l e accounts f o r the i n c l u s i o n of the second and t h i r d o bjectives l i s t e d above. Of somewhat l e s s e r importance i s the fourth o b j e c t i v e which i s large l y methodological. E x i s t i n g studies have demonstrated the a p p l i c a -t i o n of various psychometric s c a l i n g techniques i n geographic contexts, i n c l u d i n g the semantic d i f f e r e n t i a l (Brown, 1974; Down, 1970), and non-metric multidimensional s c a l i n g (Doddridge, 1972). I t i s somewhat sur-p r i s i n g that to date there appears to have been no reported use of L i k e r t 51. s c a l i n g procedures. This i s p a r t i c u l a r l y so since there i s evidence to suggest that measurements based upon L i k e r t scales are more r e l i a b l e and more accurately p r e d i c t behaviour than those obtained by other methods ( T i t t l e and H i l l , 1967). In a ddition, the evidence from geographical studies employing psy-chometric s c a l i n g procedures i n d i c a t e s an unfortunate f a i l u r e to assess the r e l i a b i l i t y and v a l i d i t y of the research instruments. This i s not to suggest that the data obtained i s n e c e s s a r i l y suspect, but simply that routine t e s t i n g procedures have not been followed. An a d d i t i o n a l metho-d o l o g i c a l objective of the present study therefore was to construct, develop and t e s t a s c a l i n g instrument i n accordance with es t a b l i s h e d psychometric standards. C. RESEARCH DESIGN The research design formulated to achieve the stated objectives appears i n the form of a flow diagram i n Figure 2.1. Components 1 to 9 e x c l u s i v e l y concern the development of a r e l i a b l e and v a l i d L i k e r t sca-l i n g instrument. I t i s important, therefore, that b r i e f consideration be given at the outset to L i k e r t scales and to the contingent notions of r e l i a b i l i t y and v a l i d i t y i n psychometric measurement. (i) L i k e r t Scales L i k e r t scales take t h e i r name from the researcher responsible f o r o r i g i n a t i n g t h i s psychometric technique ( L i k e r t , 1932). They are a p a r t i c u l a r type of summative s c a l e s p e c i f i c a l l y designed f o r the measu-rement of a t t i t u d e s . In essence, a L i k e r t scale consists of a set of FIGURE 2.1 FLOW DIAGRAM OF THE RESEARCH DESIGN 53. a set of statements which subjects respond to on a five point scale ran-ging from 'strongly agree' to 'strongly disagree". Each of the statements is designed to represent a facet of the attitude in question, and the attitude score for any respondent i s simply a summation of the scores on each of the items (statements) contributing to the scale. As with other scaling procedures based upon a summative model, Likert scales assume that contributing items are monotonically related to under-lying traits and that a summation of item scores is approximately linear-ly related to the t r a i t (Nunnally, 1967, p. 70 f f . ) . Previous psycholo-gical research indicates that summative scales have a number of advantages over other scaling methods: "they (1) follow from an appealing model; (2) are rather easy to construct; (3) usually are highly reliable; (4) can be adapted to the measurement of many different kinds of attitudes; and (5) have produced meaningful results in many different studies to date" (Nunnally, 1967, p. 531). Further support for the selection of Likert scaling as a technique for measuring individual dispositions in this study i s derived from i t s use in a comparable context by environmental psychologists. In this regard, the work of McKechnie (1970, 1972) is instructive. Faced with the task of assessing individual orientations to both natural and b u i l t environments, McKechnie developed nine Likert scales describing d i f f e r -ent attitudes, beliefs and preferences with respect to a range of envi-ronmental settings and experiences. In a subsequent study, the relation-ships between environmental dispositions,specific environmental attitudes and leisure a c t i v i t i e s were examined as a basis for identifying environ-mental l i f e - s t y l e groups. 5 4 . ( i i ) The Assessment of R e l i a b i l i t y In the administration of any tes t of a t t i t u d e s , a b i l i t i e s , aptitudes or whatever, i t i s assumed that any s i n g l e respondent has a 'true score' on the a t t r i b u t e i n question, and that the purpose of the t e s t items i s to provide a r e l i a b l e estimate of that 'true score'. I t i s recognized that departures from hypothetical 'true scores' w i l l occur due to the combined e f f e c t of various errors of measurement. The extent to which these errors are minimized increases r e l i a b i l i t y , and hence the r e s u l t s obtained i n any sing l e administration of a t e s t should be repeatable on subsequent occasions. The sources of measurement e r r o r are e s s e n t i a l l y of two kinds: those which are inherent within a p a r t i c u l a r set of t e s t items; and those which occur between d i f f e r e n t forms of a t e s t given at d i f f e r e n t times. E r r o r s of the f i r s t type l a r g e l y r e f l e c t errors due to content sampling. C l e a r l y , the set of items selected to make up a p a r t i c u l a r t e s t form i s but a sample of the universe of items which e x i s t s to mea-sure the a t t r i b u t e i n question. There i s therefore the p o s s i b i l i t y of a non-representative sample of items being selected with the r e s u l t that the estimated scores would depart appreciably from 'true scores'. I t i s important therefore to have a means o f assessing the amount of erro r a r i s i n g from t h i s source. The theory of measurement e r r o r (Nunnally, 1967, Ch. 6) provides a basis f o r such an assessment. I t emerges that the e r r o r a r i s i n g from the sampling of items i s p r e d i c t a b l e from the average c o r r e l a t i o n between items on a p a r t i c u l a r t e s t form, and the sub-sequent c a l c u l a t i o n of c o e f f i c i e n t alpha as the index of r e l i a b i l i t y . 55. There are other sources of e r r o r within a- set of t e s t items which a r i s e f o r example from respondents guessing answers or from subjective scoring procedures, but, since these are not s i g n i f i c a n t f o r the type of t e s t developed i n the present study, they w i l l receive no fu r t h e r consideration. Three major sources of e r r o r between tes t s have been i d e n t i f i e d (Nunnally, 1967, p. 209 f f . ) . The f i r s t a r i s e s from systematic d i f f e r -ences i n the content of d i f f e r e n t forms of a t e s t such that the two sets of items place emphasis on d i s t i n c t aspects of the a t t r i b u t e being measured. This problem tends to be more c r i t i c a l with respect to t e s t s of a b i l i t y (e.g., s p e l l i n g tests) than to a t t i t u d e s c a l i n g . S u b j e c t i v i t y of scoring can r e s u l t i n erro r between tes t s but again t h i s i s usually not c r i t i c a l i n the measurement of at t i t u d e s where objective scoring procedures are normally employed. The t h i r d major source of e r r o r , and perhaps the most d i f f i c u l t to counter, i s that due to the f a c t that people change with respect to the a t t r i b u t e . Measures of a t t i t u d e are c l e a r l y not immune to such v a r i a t i o n s , since experiences and the unfolding of events between t e s t s may s i g n i f i c a n t l y i n f l u e n c e response patterns. In many instances, the non-conformity of a t t i t u d e s and behaviour may be a t t r i b u t a b l e to such changes intervening between the measurement of a t t i -tude and the monitoring of contingent behaviour, p a r t i c u l a r l y when the at t i t u d e r e l a t e s to issues of current controversy. The standard procedure f o r assessing v a r i a t i o n s between t e s t s i s to administer a l t e r n a t i v e forms at various time i n t e r v a l s . The c o r r e l a t i o n between the scores obtained using the a l t e r n a t i v e forms i s then the basis f o r assessing r e l i a b i l i t y . I d e a l l y , a l t e r n a t i v e forms should be adminis-56. tered on several occasions to i n v e s t i g a t e short, mid and long term v a r i a -tions (for example r e t e s t i n g a f t e r two weeks, s i x weeks and s i x months i s d e s i r a b l e ) . There are some obvious p r a c t i c a l obstacles to r e t e s t i n g and p r i n c i p a l l y that of access to the same group of respondents. I t i s not s u r p r i s i n g therefore that the administration of a l t e r n a t i v e forms has normally employed a 'captive' student group. Unfortunately, i n the present study, which drew upon a city-wide sample from the general popu-l a t i o n , i t was not p o s s i b l e to perform r e t e s t s . ( i i i ) The Assessement of V a l i d i t y Cronbach and Meehl (1955) d i s t i n g u i s h four types of v a l i d i t y : pre-d i c t i v e , concurrent, content and construct v a l i d i t y . The f i r s t two d i f -f e r only to the extent of the temporal separation between the administra-t i o n of the t e s t instrument and the measurement of the relevant c r i t e r i o n . For the purposes of t h i s o u t l i n e therefore, i t i s necessary only to con-s i d e r p r e d i c t i v e , content and construct v a l i d i t y . Although i t i s important to appreciate the relevance of each of these, i t i s also necessary to recognise the interdependence that e x i s t s between them. Predictive validity concerns the use of a research instrument to estimate some important form of behaviour where the l a t t e r i s r e f e r r e d to as the ' c r i t e r i o n ' . A normal procedure i s to administer a t e s t , and then to obtain an independent c r i t e r i o n measure on the same subjects, and subsequently to compute the c o r r e l a t i o n of the two sets of r e s u l t s . The measure of c o r r e l a t i o n i s then taken to be a d i r e c t i n d i c a t i o n of the p r e d i c t i v e v a l i d i t y of the research instrument. C l e a r l y , both sound theory and common sense need to be applied i n the s e l e c t i o n of p r e d i c t o r 57. instruments f o r i n v e s t i g a t i o n i f the danger of spurious p r e d i c t i o n i s to be avoided. To t e s t the p r e d i c t i v e v a l i d i t y of a s p e c i f i c research instrument would seem to be an immediate and obvious concern whenever p s y c h o l o g i c a l s c a l i n g instruments are used, and yet there are few cases i n the consumer s p a t i a l behaviour l i t e r a t u r e which document such t e s t i n g . Psychological t e s t s have been devised and administered which purport to measure v a r i a -b l e s which d i r e c t l y impinge upon shopping behaviour, and yet there has been a f a i l u r e to demonstrate that the scores obtained c o n s i s t e n t l y p r e d i c t r e t a i l choices. In some cases, t h i s f a i l u r e can be explained i n terms of the l i m i t e d aims of the research i n v e s t i g a t i o n (Kunkel and Berry, 1968), where the concern was rather with construct i d e n t i f i c a t i o n than with hypothesis t e s t i n g . In another case (Downs, 1970), unforeseen data problems prevented the t e s t i n g of the p r e d i c t i v e v a l i d i t y of the research i n s t r u -ment. With respect to t h i s most simple form of v a l i d i t y t e s t i n g , imme-diate studies are obviously needed i f the p r e d i c t i v e accuracy of the t e s t instruments i s to be assessed. I t i s necessary to demonstrate convincingly that i n d i v i d u a l s , who d i f f e r i n terms of t h e i r cognition of the r e t a i l environment as measured by the t e s t instrument, also e x h i b i t d i f f e r e n t patterns of behaviour. V a l i d a t i o n i n t h i s context b a s i c a l l y involves the t e s t i n g of two r e l a t e d hypotheses: that consum-ers with markedly d i f f e r e n t t e s t scores e x h i b i t d i f f e r e n t r e t a i l choices; and that i n d i v i d u a l s with s i m i l a r t e s t score p r o f i l e s e x h i b i t s i m i l a r r e t a i l choices. 58. Content Validity has two d i s t i n c t aspects. F i r s t l y , confirmation of content v a l i d i t y demands that the t e s t items c o n s t i t u t e a representa-t i v e sample of the universe of items which could p o s s i b l y be included i n that t e s t . I t i s i n t h i s context that a p r i o r notion of the composi-t i o n of the universe of items assumes great importance. This i s not a problem i f previous research has e f f e c t i v e l y i s o l a t e d the var i a b l e s relevant to the area of i n v e s t i g a t i o n . An a priori approach can be more problematic, however, where the nature of the va r i a b l e s i s not w e l l understood. The p o s s i b i l i t y i s then increased of the t e s t items c o n s t i -t u t i n g a non-representative sample of the universe. Studies of store s e l e c t i o n undertaken under the heading of 'store image' research can be c r i t i c i s e d from the poin t of view that i n s u f f i -c i e n t consideration has been given to content v a l i d i t y . Such studies have l a r g e l y assumed that the var i a b l e s s a l i e n t to store choice coincide with the objec t i v e a t t r i b u t e s of the r e t a i l environment, such as p r i c e l e v e l , q u a l i t y of goods, personnel, i n t e r i o r design and so on. While i t i s obvious that such a t t r i b u t e s are of fundamental importance, there i s evidence to suggest that more s p e c i f i c aspects of these generalised notions c o n s t i t u t e the c r i t i c a l v a r i a b l e s (Taylor, 1972). I f t h i s i s true, an attempt must be made i n the course of t e s t construction to i d e n t i f y these v a r i a b l e s i f the requirements of content v a l i d i t y are to be s a t i s f i e d . I t can be argued that t e s t items suggested by a subsample represen-t a t i v e of p o t e n t i a l respondents b e t t e r meet the requirements of content v a l i d i t y than items imposed a priori by the researcher. The procedure f o r e l i c i t i n g t e s t items can be based upon structured or unstructured 5 9 . interview techniques. The repertory g r i d test. (Kelly, 1955; Bannister and Mair, 1968) i l l u s t r a t e s a method of psychometric s c a l i n g which begins with e l i c i t i n g scale constructs from the respondent. This method has been suggested and applied i n consumer research (Harrison and Sarre, 1971; Hudson, 1970 and 1971, Sampson 1971). An a l t e r n a t i v e approach i s to content analyse unstructured i n t e r -views i n which respondents are requested to a r t i c u l a t e s a l i e n t aspects of t h e i r a t t i t u d e s and behaviour. Again i n the context of consumer spa-t i a l behaviour research there are precedents f o r employing t h i s approach (Kunkel and Berry, 1968). C l e a r l y , the use of interview procedures has attendant problems and i s c e r t a i n l y not a foolproof method of ensuring content v a l i d i t y ; nonetheless such a procedure f o r t e s t item i d e n t i f i c a -t i o n can perhaps be more e a s i l y defended than a wholly a priori, approach. A second aspect of content v a l i d i t y r e l a t e s to the manner i n which the t e s t items are presented. Of p a r t i c u l a r importance i s the choice o f response format. I t i s necessary to decide whether responses should be of a ' t r u e - f a l s e 1 , 1always-never', 'strongly agree-strongly disagree' or whatever other kind. The way i n which the items are expressed w i l l normally i n d i c a t e which format c o n s t i t u t e s the most sensible choice; but, t h i s does not n e c e s s a r i l y r u l e out the need to administer the t e s t using d i f f e r e n t response formats as a check on the consistency of responses. These comments are perhaps s u f f i c i e n t to i n d i c a t e that, although steps can be taken to increase confidence i n content v a l i d i t y , i t s assessment must almost i n e v i t a b l y proceed on an i n t u i t i v e b a s i s , which i s substantiated by the observation that: "content v a l i d i t y mainly rests upon an appeal to the p r o p r i e t y of content and the way that i t i s 60. presented" (Nunnally, 1967, p. 83). Construct validity i s the t h i r d aspect of v a l i d a t i o n i n psychometric research. Cronbach and Meehl (1955, p. 287) state that: "construct v a l i d a t i o n i s involved whenever a t e s t i s to be i n t e r p r e t e d as a measure of some a t t r i b u t e ( s ) or q u a l i t y ( i e s ) which i s not 'operationally defined'." Constructs, by d e f i n i t i o n , cannot be measured d i r e c t l y , and yet i t i s frequently the case that the measurement o f a p a r t i c u l a r construct i s b a s i c to an understanding of behaviour. I t i s common, for example, i n analysing s p a t i a l behaviour to invoke abstract v a r i a b l e s to explain patterns of behaviour. S o c i a l c l a s s i s an example of a construct common-l y employed to which can be added the various c o g n i t i v e v a r i a b l e s sug-gested as behavioural antecedents (Gayler, 1974; Downs, 1970). Construct measurement i s dependent upon the d e f i n i t i o n of a domain of observable v a r i a b l e s which r e l a t e to t h a t construct. Test scores on the observable v a r i a b l e s then serve as a measure of the construct. In the case of s o c i a l c l a s s , income, education and occupational status e i t h e r separately or i n composite have commonly functioned as appropriate observable v a r i a b l e s . In the measurement of p s y c h o l o g i c a l constructs such as a t t i t u d e s then the observable v a r i a b l e s often take the form of a set of statements which v a r i o u s l y express the construct domain. The assessment of construct v a l i d i t y r e s t s upon the extent to which the observable v a r i a b l e s do i n f a c t measure the construct they are pur-ported to measure. This assessment i s l a r g e l y based upon an examination of the structure of the c o r r e l a t i o n s among the observable v a r i a b l e s , p r e f e r a b l y based upon several sets o f sample respondents. Where the measures vary systematically i n r e l a t i o n to each other over a number of 61. t e s t s i t u a t i o n s , i t i s poss i b l e to assert that they are measuring a s i n -g l e construct. On the other hand, i f the v a r i a b l e s tend to s p l i t up in t o c l u s t e r s such that the members of a c l u s t e r c o r r e l a t e h i g h l y with one another and c o r r e l a t e much l e s s with the members of other c l u s t e r s , i t can be concluded that a number of d i f f e r e n t constructs are being measured. In the case of there being a complete absence of c l u s t e r i n g , then t h i s would seem to i n d i c a t e that each v a r i a b l e i s measuring a d i f f e r -ent construct. The assessment of construct v a l i d i t y should be a strong force g u i -ding the course of research i n v e s t i g a t i o n , and i t s neglect may lead to the acceptance of a construct which i n f a c t should be replaced by two or more new constructs, or perhaps, should even be abandoned altogether. Factor a n a l y s i s i s c e n t r a l to the i d e n t i f i c a t i o n , measurement and v a l i d a t i o n of ps y c h o l o g i c a l constructs. The e s s e n t i a l purpose of t h i s technique i s to f i n d c l u s t e r s o f r e l a t e d v a r i a b l e s (or factors) such that the members of each c l u s t e r c o r r e l a t e more hig h l y among themselves than they do with v a r i a b l e s not included i n the c l u s t e r . This charac-t e r i s t i c i s obviously coincident with the conditions of construct v a l i -dation, and immediately i n d i c a t e s the usefulness of f a c t o r analysis as a t o o l f o r the assessment of construct v a l i d i t y . At the same time, f a c t o r analysis provides a means of obtaining a measure f o r each respondent on each of a set of constructs i n the form of f a c t o r scores. These are derived by obtaining l i n e a r expressions f o r the f a c t o r s i n terms o f the observed v a r i a b l e s , and subsequently s u b s t i t u t i n g scores on those v a r i a b l e s f o r each i n d i v i d u a l . 62. A further use of f a c t o r analysis i s as a means of exploring sets of v a r i a b l e s i n order to i d e n t i f y the constructs presumed to underlie those v a r i a b l e s (Kerlinger, 1964, p. 680). In a research context where the relevant constructs are e s s e n t i a l l y i l l - d e f i n e d , f a c t o r analysis ( i f employed with due caution) can therefore provide a u s e f u l means of construct i d e n t i f i c a t i o n and c l a r i f i c a t i o n , and i t i s to t h i s end that i t has l a r g e l y been applied i n the analysis of consumer s p a t i a l beha-viour. However, i t i s necessary to add a word of caution with respect to the i n t e r p r e t a t i o n of the r e s u l t s of f a c t o r a n a l y s i s . This w i l l also demonstrate the need to recognise the e s s e n t i a l interdependence of a l l three types o f v a l i d i t y when assessing a p a r t i c u l a r t e s t instrument. As has often been pointed out, the output of a fa c t o r analysis i s no more and no l e s s than an expression of the data contained i n the input v a r i a b l e s (Armstrong, 1967). With t h i s i n mind, l e t i t be assumed that f a c t o r analysis i n d i c a t e s the existence of a p a r t i c u l a r set of con-s t r u c t s , and, moreover, that subsequent studies demonstrate that these constructs s a t i s f y the standards o f construct v a l i d i t y i n that each con-s t r u c t i s c o n s i s t e n t l y defined by a set of i n t e r c o r r e l a t e d observable v a r i a b l e s . Is i t reasonable to conclude therefore that a set of abstract v a r i a b l e s which w i l l f u r t h e r the explanation and p r e d i c t i o n of behaviour has been s u c c e s s f u l l y i s o l a t e d ? Obviously not, u n t i l s u f f i c i e n t assur-ance can be gained regarding the content v a l i d i t y of the input v a r i a b l e s and the p r e d i c t i v e v a l i d i t y of the t e s t scores. In other words, construct v a l i d i t y , as confirmed by f a c t o r a n a l y s i s , i s a necessary but not s u f f i -c i e n t condition for assuming the o v e r a l l v a l i d i t y of a p a r t i c u l a r t e s t . 6 3 . The extent to which r e l i a b i l i t y and v a l i d i t y assessment c o n s t i t u t e s an i n t e g r a l p art of the research design employed i n t h i s study should be c l e a r from the following section and the subsequent chapters. (iv) Research Design Components The sequential s t r u c t u r e of the research design i s i l l u s t r a t e d i n Figure 2.1 and the various components can be summarized as follows: Components 1 to 3 together comprise the construct i d e n t i f i c a t i o n phase of the research. The purpose o f t h i s phase was to i d e n t i f y the c r i t e r i a important to consumers when choosing where to shop f o r major c l o t h i n g goods, and subsequently to reduce these c r i t e r i a to a set of underlying constructs as the basis f o r developing a L i k e r t s c a l i n g instrument. Consistent with the previous d i s c u s s i o n of content v a l i d i t y , the c r i t e r i a were not s e l e c t e d on an a priori basis but rather were i d e n t i f i e d using a content analysis of unstructured interviews with shoppers. In order to i s o l a t e the constructs presumed to underlie these c r i t e r i a , a convenience sample rated each of the c r i t e r i a on a ten-point scale of importance and the r e s u l t s obtained were f a c t o r analysed. Thus i l l u s t r a t i n g the use of f a c t o r a n a l y s i s as a means of construct i d e n t i f i c a t i o n . Component 4 concerns the generation of an item pool to provide statements for a preliminary set of L i k e r t s c a l e s . Items were wr i t t e n to represent the various facets of the content domains summarized by each of the constructs i d e n t i f i e d i n the previous phase of the research. Components 5 and 6 involve the p r e t e s t i n g of the research i n s t r u -ment as a basis f o r assessing the r e l i a b i l i t y and v a l i d i t y of the 64. preliminary L i k e r t s c a l e s . To t h i s end, data were c o l l e c t e d from two sample groups: the f i r s t comprising a c l a s s of introductory psychology students and the second comprising shoppers interviewed i n two contrast-i n g i n - s t o r e s i t u a t i o n s . Components 7 and 8 deal with the ana l y s i s of the p r e t e s t data. Standard procedures f o r assessing the r e l i a b i l i t y and v a l i d i t y of the L i k e r t scales were followed i n c l u d i n g the c a l c u l a t i o n of item-scale c o r r e l a t i o n s , c o e f f i c i e n t alphas, and construct v a l i d i t y c o e f f i c i e n t s . In addition, i t was p o s s i b l e to make an i n i t i a l assessment of the pre-d i c t i v e v a l i d i t i e s of the scales by determining how accurately they d i s -criminated between shoppers p a t r o n i z i n g the two d i f f e r e n t stores i n which interviews were conducted. Component 9 concerns the development of the f i n a l s c a l e s . The r e s u l t s obtained from the p r e t e s t served to i n d i c a t e what modifications were necessary t o increase the r e l i a b i l i t y and v a l i d i t y o f the s c a l e s , and the appropriate r e v i s i o n s were made. Component 10 involves the design of the questionnaire to be adminis-tered i n the major data c o l l e c t i o n phase. The research objectives d i c -tated that three sets of data be c o l l e c t e d from each respondent i n order to document cognitions, behaviour and b i o g r a p h i c a l c h a r a c t e r i s t i c s . P r i o r phases of the research defined the cognitive component, but d e c i -sions as to the design of the other sections of the questionnaire had to be made at t h i s stage. Component 11 deals with the sample design. Various considerations l e d to the choice of a s t r a t i f i e d random sample of households within the m u n i c i p a l i t i e s of Vancouver, North Vancouver and West Vancouver. 65. Component 12 concerns the t r a i n i n g of the interviewers responsible for conducting the major data c o l l e c t i o n phase. The interviewers, drawn from students e n r o l l e d i n a second year geography c l a s s , received i n -c l a s s t r a i n i n g plus a written i n s t r u c t i o n manual. Component IS involves the major data c o l l e c t i o n phase. The i n t e r -viewers obtained data from 351 respondents d i s t r i b u t e d i n accordance with the sample design. Component 14 deals with the coding and analysis of the data. The information was coded on the questionnaires and was subsequently key-punched. Various analyses were performed but major emphasis was placed upon comparing the a b i l i t y of co g n i t i v e , l o c a t i o n a l and bi o g r a p h i c a l v a r i a b l e s to p o s t d i c t membership i n predefined c r i t e r i o n groups; the c r i t e r i a being "type of store' and 'type of shopping area' patronized on each of four shopping t r i p s . The ba s i c s t a t i s t i c a l technique employed was stepwise discriminant a n a l y s i s . R e l i a b i l i t y and v a l i d i t y s t a t i s t i c s were computed for each of the L i k e r t scales as a recheck on the p r e t e s t r e s u l t s . Component 15 concerns the reporting of the r e s u l t s and needs no further elaboration. D. SUMMARY This chapter began with an o u t l i n e of the research purposes. Atten-t i o n then focussed on the research design and p a r t i c u l a r l y the develop-ment of a s c a l i n g instrument to measure consumer d i s p o s i t i o n s . The s u i t a -b i l i t y of L i k e r t s c a l i n g as a measurement procedure was discussed and the importance of r e l i a b i l i t y and v a l i d i t y considerations i n scale 66. construction was stressed. F i n a l l y , the major components of the research design were b r i e f l y summarized. In the following chapter the preliminary phases of the design are described i n d e t a i l . CHAPTER 3 THE IDENTIFICATION AND MEASUREMENT OF CONSUMER DISPOSITIONS A. INTRODUCTION The development of a psychometric s c a l i n g instrument to measure consumer d i s p o s i t i o n s i s c e n t r a l to the achievement of the research o b j e c t i v e s itemized i n the previous chapter. This chapter d e t a i l s the i d e n t i f i c a t i o n of the s a l i e n t d i s p o s i t i o n a l v a r i a b l e s and the subsequent construction of a L i k e r t s c a l i n g instrument as an appropriate measure-ment device. A content analysis of unstructured interviews served to i d e n t i f y the c r i t e r i a employed by shoppers to evaluate a l t e r n a t i v e r e t a i l oppor-t u n i t i e s . Each of the c r i t e r i a was subsequently rated by a convenience sample and the data obtained were f a c t o r analysed to i s o l a t e underlying constructs. Preliminary L i k e r t scales were constructed to provide a more r e f i n e d measure of the major constructs, and these were pretested using data obtained from two sample groups. Standard psychometric pro-cedures f o r assessing the r e l i a b i l i t y and v a l i d i t y of the scales were followed and the necessary r e v i s i o n s were made to produce a set of f i n a l s c a l e s . B. CONSTRUCT IDENTIFICATION The model of consumer s p a t i a l behaviour o u t l i n e d at the conclusion of Chapter 1 argues that a dec i s i o n as to where to shop i s the outcome of a shopper evaluating a l t e r n a t i v e r e t a i l opportunities i n accordance - 67 -6 8 . with a set of predefined d i s p o s i t i o n s . In order to o p e r a t i o n a l i s e t h i s model therefore i t was necessary to i d e n t i f y s a l i e n t consumer d i s p o s i -t i o n s and to e s t a b l i s h a means of measuring them. The e x i s t i n g l i t e r a t u r e i s suggestive as to the nature of these d i s p o s i t i o n a l constructs, which have been v a r i o u s l y termed "cognitive categories", "evaluative c r i t e r i a " , "cognitive dimensions" and "image determinants" (Fisk, 1961; Becker, 1967; Kunkel and Berry, 1968; Engel et ail., 1968; Stephenson, 1969; Downs, 1970). S a l i e n t constructs have been defined on both an a p r i o r i and empi-r i c a l basis l a r g e l y i n the context of store image research and have been c l a s s i f i e d under such general headings as 'location/distance', 'assort-ment of merchandise', 'price', 'advertising', 'personnel' and 'services' (Engel et al. , 1968, pp. 452-3). However, the s p e c i f i c constructs generated by d i f f e r e n t researchers have v a r i e d dependent upon the type of purchase, and there i s no evidence to suggest the g e n e r a l i t y of constructs across r e t a i l functions. Furthermore, the u t i l i t y of the constructs remains i n doubt because there has been a general f a i l u r e to v a l i d a t e them against behavioural c r i t e r i a . In the context of the present study therefore i t was not c l e a r how much confidence could be placed i n the constructs i d e n t i f i e d by resear-chers h i t h e r t o . As a r e s u l t , content v a l i d i t y standards were more l i k e l y to be met by i d e n t i f y i n g s a l i e n t d i s p o s i t i o n a l constructs as a f i r s t stage of the research design. As a f i r s t step to t h i s end a s e r i e s of unstructured interviews were conducted with consumers. Though the respondents comprised a small convenience sample (N = 27) some e f f o r t was made to maximize demographic 69. and socio-economic heterogeneity. In each interview the respondent was asked to describe the most recent shopping t r i p f o r c l o t h i n g goods, and to s p e c i f y , where p o s s i b l e , those factors which had influenced store choice. I n t e r j e c t i o n s by the interviewer were minimized to avoid i n t r o -ducing a needless'source of response b i a s . The i n f l u e n t i a l c r i t e r i a itemized by each respondent were recorded verbatim. A content analysis of interview responses revealed a t o t a l of 172 d i f f e r e n t c r i t e r i a . Of these 111 had been mentioned as p o s i t i v e i n f l u -ences (e.g., "good q u a l i t y merchandise") and 61 as negative influences (e.g., "aggressive sales s t a f f " ) . Many of the c r i t e r i a were mentioned only once. While i t was c l e a r that many of the c r i t e r i a could be cate-gorized under such general headings as ' p r i c e 1 , ' q u a l i t y ' , ' l o c a t i o n ' , etc., a subjective procedure of construct i d e n t i f i c a t i o n was rejected i n favour of an examination of the str u c t u r e of c o r r e l a t i o n s amongst the c r i t e r i a using f a c t o r a n a l y s i s . To achieve t h i s , a questionnaire was designed which required respondents to i n d i c a t e on a 10 po i n t scale the importance they ascribed to each of the c r i t e r i a when choosing where to shop f o r c l o t h i n g goods. Questionnaires were completed by students e n r o l l e d i n an introductory geography c l a s s and t h e i r parents (where p o s s i b l e ) . Information was obtained from 120 respondents. A fa c t o r analysis was performed on 70 of the c r i t e r i a , which remai-ned a f t e r the elimination of l a r g e l y synonomous and high l y i n t e r c o r r e l a -ted c r i t e r i a . The analysis y i e l d e d ten factors accounting f o r 56 per cent of the o r i g i n a l variance. Of these, the f i r s t f i v e , accounting f o r 44 per cent of the t o t a l variance and 72 per cent of the common fa c t o r variance a f t e r r o t a t i o n , were the most c l e a r l y i n t e r p r e t a b l e . 70. They were described as follows: Factor 1: 'Fashion O r i e n t a t i o n ' - c r i t e r i a r e l a t i n g to contemporary fashion loaded highly on t h i s f a c t o r , suggesting that fashion o r i e n t a t i o n i s a major consumer d i s p o s i t i o n i n f l u -encing the evaluation of a l t e r n a t i v e c l o t h i n g stores. Factor 2: 'Price Orientation" - i n t h i s case, f a c t o r d e f i n i n g c r i t e r i a were r e l a t e d to p r i c e considerations, confirming the r e s u l t s of previous studies that p r i c e o r i e n t a t i o n i s a b a s i c d i s -p o s i t i o n a l construct. Factor 3: 'Quality O r i e n t a t i o n 1 - c r i t e r i a loading h i g h l y on the t h i r d f a c t o r i n d i c a t e d that q u a l i t y considerations c o n s t i t u t e an-other b a s i c consumer d i s p o s i t i o n when choosing where to shop. Notions of q u a l i t y appeared to extend beyond the character-i s t i c s of the merchandise to include general concerns about the r e l i a b i l i t y and trustworthiness of a store. Factor 4: 'Convenience O r i e n t a t i o n ' - the importance of l o c a t i o n a l c r i -t e r i a i s confirmed by the emergence of t h i s f a c t o r . The term 'convenience' was chosen as the appropriate d e s c r i p t o r since the f a c t o r d e f i n i n g c r i t e r i a were not r e s t r i c t e d to considerations of the r e l a t i v e l o c a t i o n s of home/workplace and store, but extended to time-related concerns. Factor 5: 'Status O r i e n t a t i o n ' - the c r i t e r i a d e f i n i n g t h i s f a c t o r suggested that status concerns, as expressed i n c h a r a c t e r i s -71. t i c s of the merchandise, c l i e n t e l e and store l o c a t i o n , form an important d i s p o s i t i o n a l construct f o r shoppers when s e l e c t i n g a c l o t h i n g s t o r e . In general, these major constructs coincide with the r e s u l t s of previous studies. The 'convenience' construct, f o r example, was i d e n t i -f i e d by Berry (1968), Downs (1970), F i s k (1961) and Stephenson (1969); while the 'qu a l i t y ' construct i s i n p a r t coincident with what Stephenson termed 'dependability' and Downs and Berry r e f e r r e d to as 'reputation'. 'Price o r i e n t a t i o n ' i n e v i t a b l y emerges i n t h i s and r e l a t e d studies as a b a s i c consumer d i s p o s i t i o n r e f l e c t i n g v a r i a t i o n s i n buying power and d i f f e r i n g views on personal expenditure. The emergence of 'fashion o r i e n t a t i o n ' as a major d i s p o s i t i o n probably r e f l e c t s the f a c t that attention i n t h i s study was focussed on c l o t h i n g purchases. The 'status' construct may well represent a p s y c h o l o g i c a l formulation of notions previously embedded i n s o c i a l c l a s s measures. The f a c t o r structure further suggests that s a l i e n t consumer d i s -p o s i t i o n s transcend a simple c l a s s i f i c a t i o n i n terms of general and tangible c r i t e r i a . Downs (1970) has made a s i m i l a r observation i n suggesting that the process by which consumers c o l l e c t , code and eva-luate information concerning the r e t a i l environment may w e l l r e f l e c t the importance o f i n t a n g i b l e influences which may lack p h y s i c a l r e f e r -ents. To recognize the coincidence of the constructs i d e n t i f i e d i n t h i s study with those documented by previous research i s not to admit to the redundancy of the preliminary phases of the research design. C l e a r l y , 7 2 . the development of an e f f e c t i v e research instrument demands f a r more than a l i s t of s a l i e n t constructs; namely, an understanding of the con-tent domain for which the construct i t s e l f i s but a u s e f u l d e s c r i p t o r . I t i s i n t h i s regard that the i n i t i a l phase of the research design was p a r t i c u l a r l y p r o f i t a b l e since i t y i e l d e d a set of c r i t e r i a associated with each construct, which, as subsequent sections of t h i s chapter show, could function as the basis for constructing a r e f i n e d s c a l i n g i n s t r u -ment. C. THE DEVELOPMENT OF A SET OF LIKERT SCALES Having i d e n t i f i e d the s a l i e n t d i s p o s i t i o n a l constructs, the next phases of the research design (components 4 to 9) concerned the construc-t i o n of a set of L i k e r t scales to function as d i s p o s i t i o n a l measures. This task involved the generation of a s u i t a b l e item pool, the p r e t e s t -in g of preliminary s c a l e s , and the subsequent formulation of a set of f i n a l s c a l e s . (i) Generation of an Item Pool Item w r i t i n g i s almost c e r t a i n l y as much an a r t as i t i s a science. Various approaches have been suggested but the ' r a t i o n a l ' method i s the one most frequently employed and most strongly supported (Jackson, 1971). The ' r a t i o n a l ' approach involves w r i t i n g items r e l a t e d to the con-s t r u c t s i n question on the basis of an i n t u i t i v e grasp of the content domain of each construct, and subsequently deciding the d i r e c t i o n i n which a p a r t i c u l a r item should be scored. This s u b j e c t i v e procedure i s followed by an empirical phase to ensure, by means of item a n a l y s i s , 73. that a given item contributes to the o v e r a l l variance of the scale of which i t i s supposedly a measure. Item w r i t i n g i s f a r from being a capricious exercise since the nature of the construct i s usually s u f f i c i e n t l y w e l l understood to pro-vide c l e a r guidance i n the s e l e c t i o n of s e n s i b l e items. This was espe-c a i l l y true i n the present study where the construct i d e n t i f i c a t i o n phase had served to e s t a b l i s h the c r i t e r i a associated with each of the f i v e major constructs. The generation of a s u i t a b l e item pool i n t h i s study was f u r t h e r circumscribed by the established c r i t e r i a f o r evaluating L i k e r t scale statements ( L i k e r t , 1932, p. 44 f f . ) , which are l i s t e d below i n verbatim form. 1. "Each statement should be of such a nature that persons with d i f f e r e n t points of view, so f a r as the p a r t i c u l a r a t t i t u d e i s concerned, w i l l respond to i t d i f f e r e n t i a l l y . " 2. " I t i s e s s e n t i a l that a l l statements be expressions of desired behaviour and not statements of fact. Two persons with decided-l y d i f f e r e n t a t t i t u d e s may, nevertheless, agree on questions of f a c t . " 3. Each p r o p o s i t i o n should be stated i n " c l e a r , concise, s t r a i g h t -forward statements." P o t e n t i a l ambiguity a r i s i n g from confus-i n g grammatical construction, double-barrelled statements, et c . should be eliminated. 74. 4. "In general i t would seem de s i r a b l e to have the questions so worded that the modal reaction to some i s more toward one end of the a t t i t u d e continuum and to others more i n the middle or toward the other end." 5. "To avoid any space e r r o r or any tendency to a stereotyped response i t seems desi r a b l e to have the d i f f e r e n t statements so worded that about one-half of them have one end of the a t t i t u d e continuum corresponding to the l e f t or upper part of the r e a c t i o n a l t e r n a t i v e s ( i . e . , 'strongly agree') and the other h a l f have the same end of the a t t i t u d e continuum corres-ponding to the r i g h t or lower p a r t of the reaction a l t e r n a -t i v e s ( i . e . , 'strongly disagree')." On the b a s i s of these c r i t e r i a and the f a c t o r structure obtained i n the construct i d e n t i f i c a t i o n phase, an item pool of 75 statements was generated. The statements were equally d i s t r i b u t e d between the f i v e d i s -p o s i t i o n a l constructs ( i . e . , 15 statements per s c a l e ) . In accordance with c r i t e r i o n 5 above, one h a l f of the statements f o r each scale (either 7 or 8) expressed a p o s i t i v e d i s p o s i t i o n and the remainder were negatively worded. Neutral and extreme statements were avoided since they tend to create less variance than statements that are l e s s extreme. The statements are l i s t e d i n Table 3.1. TABLE 3.1 ITEM-SCALE CORRELATIONS STATUS SCALE I enjoy buying expensive clothes. I l i k e to shop i n elegant surroundings. I r e g u l a r l y read high fashion magazines. Shopping at exclusive shores i s beyond my means, I f e e l uneasy i n an exclusive store. I'm not i n t e r e s t e d i n what happens i n the world of high fashion. I avoid stores which look or sound expensive. I appreciate highly personalised s e r v i c e . I patronise s p e c i a l t y stores even though I could p o s s i b l y buy s i m i l a r goods more cheaply elsewhere. Few people can a f f o r d to shop where I do. I l i k e to shop where the sales s t a f f know me and appreciate my needs. I l i k e to know that what I buy can't be bought fo r l e s s elsewhere. * I usually have ray clothes custom made. I don't appreciate the opinions and advice of store personnel.* I never consider shopping i n a high income neighbourhood.* Pretest .5252 .5152 .4848 .4520 .4454 .4269 .4219 .3806 .3792 .3338 .3230 .3050 .2744 .2141 .1606 F i n a l .6192 .6138 .5774 .5749 .6474 .5873 .5940 .4241 .4712 .5020 .2999 .1105 N = 197 N = 351 Statements excluded from the f i n a l scale. TABLE 3.1 (cont'd.) CONVENIENCE SCALE Pretest F i n a l I shop as infrequ e n t l y as p o s s i b l e . I regard shopping as a nec e s s i t y rather than a pleasure. I'm w i l l i n g to spend a l o t of time and miles to get what I want. I'm quite w i l l i n g to t r a v e l to the other side of the c i t y to shop. I don't regard distance as very important when choosing where to shop. Shopping forms one of my main l e i s u r e a c t i v i t i e s . I only patronise stores which are e a s i l y a c c e s s i b l e to my home. The n i c e t i e s of s t y l e and fashion are not important to me. I r a r e l y shop outside my home neighbourhood. I enjoy looking around stores even though I may have no i n t e n t i o n of buying anything. I spend considerable time shopping around before making a purchase. I would be quite content to do a l l my shopping by phone. * I commonly buy the f i r s t t h i n g I t r y on. I'm not prepared to l i m i t my choice of store f o r the sake of convenience. I l i k e stores where i t ' s easy to f i n d what I want.* .6284 .6256 .6230 .5992 .5770 .5338 .4880 .4845 .4786 .4777 .4113 .3962 .3274 .3264 .0032 .4043 .5612 .6245 .6359 .5570 .5362 .5338 .4450 .5150 .5152 .4830 .4477 Statements excluded from the f i n a l score. 77. TABLE 3.1 (cont'd.) FASHION SCALE L i g h t i n g e f f e c t s and rock music create a store atmosphere which appeals to me. I l i k e to dress i n the l a t e s t fashions. I would not consider shopping i n a trendy boutique. I don't see myself as one of the younger set. I shop at stores which appeal mainly to the younger age group. I avoid stores which make use of l i g h t i n g and sound e f f e c t s . I p r e f e r stores which are quiet and peaceful. Stores which cater f o r the avant garde appeal to me. My taste i n clothes i s conservative I l i k e to shop i n b r i g h t and c o l o u r f u l surroundings. I would enjoy shopping on Carnaby S t r e e t . I'm a t t r a c t e d to stores which make shopping e x c i t i n g . What I buy^  i s n ' t i nfluenced by the l a t e s t fashions. I don't mind m^  clothes being considered old-fashioned. I appreciate the use of imaginative d i s p l a y s and decor.* Pre t e s t .6490 .6484 .5879 .5849 .5642 .5438 .5405 .5045 .4880 .4791 .4657 .4298 .3916 .3346 .2184 F i n a l .7283 .6307 .5590 .5649 .5959 .6185 .5591 .5646 .6361 .4407 .5615 .5448 Statements excluded from the f i n a l scale. 78. TABLE 3.1 (Cont'd.) PRICE SCALE I don't patronise stores which have a reputation f o r low p r i c e s . I enjoy, bargain hunting. I buy at stores which u n d e r s e l l t h e i r competitors. I t doesn't worry me to know that other people own clothes i d e n t i c a l to mine. I'm not a t t r a c t e d to stores i n low income areas. I don't mind shopping i n the poorer parts of the c i t y . I p r e f e r to buy several cheaper items than one expensive one. a I l i k e to f e e l that my patronage i s r e a l l y appreciated. I consider i t e s s e n t i a l that the merchandise be neatly arranged. I f i n d i t hard to r e s i s t sales o f f e r s . I t ' s e s s e n t i a l to me that the sales s t a f f be w e l l - t r a i n e d . I f garments i n the store window aren't p r i c e d I won't go i n . The stores I patronise appeal mainly to the higher income groups.* if I avoid crowded stores. I l i k e to see the p r i c e s of garments c l e a r l y marked. I p r e f e r stores which base t h e i r appeal on low prices.* 3 Pretest .4835 .4659 .4649 .4606 .4253 .4228 .3875 .3504 .3440 .3421 .3348 .2608 .2591 .1807 .1726 F i n a l .6121 .5361 .6277 .4045 .5396 .5244 .2351 .4200 .3262 .3780 .2401 .6345 Statements excluded from the f i n a l scale. Statement t r a n s f e r r e d to the q u a l i t y scale. Statement t r a n s f e r r e d from the q u a l i t y scale. TABLE 3.1 (cont'd.) QUALITY SCALE Pretest F i n a l I'm more i n c l i n e d to buy expensive items infrequently than to make frequent cheap purchases. I take care to ensure that the garments I buy are well-made. I'm w i l l i n g to s a c r i f i c e q u a l i t y f o r low p r i c e s . I'm put o f f by l o t s of signs and adver-tisements i n a store. I'm always suspicious of low p r i c e s . I would describe myself as a d i s c r i m i n a t -ing shopper. I'm a t t r a c t e d by t a s t e f u l d i s p l a y s and decor. My f i r s t concern i s with the q u a l i t y of what I buy. The stores I patronise don't base t h e i r reputation on q u a l i t y . I t doesn't worry me to buy mass-produced clothes. I'm not deterred by impolite personnel. I don't c a r e f u l l y examine the t a i l o r i n g of the goods I buy. My choice of store i s n ' t influenced by a t t r a c t i v e displays and decor.* I p r e f e r stores which base t h e i r appeal on p r i c e . 3 The stores I patronise tend to be located i n middle-income neighbourhoods. I p r e f e r to buy several cheaper items than one expensive one.'3 .5662 .4597 .4537 .4345 .4144 .4094 .4059 .3837 .3607 .3457 .3158 .3104 .2596 .2213 .1123 .5843 .5280 .5530 .3322 .4146 .5246 .4602 .5341 .5993 .4311 .3118 .5644 Statements excluded from the f i n a l scale. Statement t r a n s f e r r e d to the p r i c e s c a l e . Statement t r a n s f e r r e d from the p r i c e scale. 8 0 . ( i i ) P r e t e s t i n g and Item Analysis As previously i n d i c a t e d , i t i s desi r a b l e to p r e t e s t an i n i t i a l set of scales to provide an empirical base f o r assessing r e l i a b i l i t y and v a l i d i t y . The c h a r a c t e r i s t i c s of the p r e t e s t respondents should corres-pond as f a r as p o s s i b l e with the c h a r a c t e r i s t i c s of the sample group to whom the f i n a l instrument i s to be administered. Furthermore, Nunnally suggests that the s i z e of the p r e t e s t sample should be at l e a s t f i v e times the number of items on each scale (Nunnally, 1967, p. 532). The p r e t e s t o b j e c t i v e i n the present study was to obtain data from a s u f f i c i e n t l y large and diverse group of subjects to permit an ade-quate assessment of the r e l i a b i l i t i e s and v a l i d i t i e s of the f i v e p r e l i -minary sc a l e s . The p r e t e s t sample (N = 197) comrpised a cla s s of introductory psychology students (N = 144) and two groups o f shoppers (N's = 51 and 32) who completed questionnaires i n contrasting i n - s t o r e s i t u a t i o n s . The i n - s t o r e sampling was conducted f o r two reasons: f i r s t l y , to increase the heterogeneity of the pr e t e s t sample; and, secondly, to allow an i n i t i a l assessement of the p r e d i c t i v e v a l i d i t i e s of the s c a l e s . The respondents were requested to provide s e l f - r a t i n g s on each of the 75 statements (presented i n a questionnaire booklet) using a f i v e p o i n t scale ranging from 'strongly agree' to 'strongly disagree'. The ordering of the statements was randomized to minimize problems of re s -ponse b i a s . The data obtained from the p r e t e s t sample formed the basis f o r an item a n a l y s i s , i n v o l v i n g : the c a l c u l a t i o n of item-scale c o r r e l a t i o n s to determine how e f f e c t i v e l y i n d i v i d u a l statements functioned as measures of the scale construct to which they were i n i t i a l l y assigned; and the computation of c o e f f i c i e n t alphas to e s t a b l i s h scale r e l i a b i l i t i e s . The c a l c u l a t i o n of item-scale c o r r e l a t i o n s involves c o r r e l a t i n g scores on an i n d i v i d u a l item with the t o t a l , score on the scale to which that item contributes. The t o t a l score i s obtained by simply summing the scores on the co n t r i b u t i n g items a f t e r f i r s t reversing scores on negatively worded statements. The item-scale c o r r e l a t i o n s computed fo r the 75 statements are l i s t e d i n Table 3.1. The i n t e r p r e t a t i o n of item-scale c o r r e l a t i o n s i s based on the r a t i o n a l e that items c o r r e l a t i n g highly with t o t a l scores have more variance r e l a t i n g to the common fa c t o r among the items, and hence make a greater c o n t r i b u t i o n to scale r e l i a b i l i t y (Nunnally, 1967, p. 261). I t follows that, i n the i n t e r e s t s of r e l i a b i l i t y , items c o r r e l a t i n g h i g h l y with t o t a l scores should be retained, while those e x h i b i t i n g i n s i g n i f i c a n t c o r r e l a t i o n s might be u s e f u l l y eliminated. The cut-o f f p o i n t between us e f u l and dispensable items depends upon the magnitude of the c o r r e l a t i o n s and the number of items required to comprise the f i n a l s c a l e . In the case of item-scale c o r r e l a t i o n s obtained f o r the p r e t e s t items (Table 3.1), c o e f f i c i e n t s above .30 approximately were found to be s i g n i f i c a n t at the .05 l e v e l , a f t e r allowance had been made f o r the i n f l a t i o n of the c o r r e l a t i o n s due to the co n t r i b u t i o n of the score on * an " i n d i v i d u a l statement to the t o t a l s c a l e score. The use of the .05 Procedures f or c o r r e c t i n g t h i s i n f l a t i o n of c o r r e l a t i o n s are d i s -cussed by Nunnally (1967, p. 262 f f . ) and G u i l f o r d (1965, pp. 502-504). An equation suggested f o r c a l c u l a t i n g corrected scores and the one 82. s i g n i f i c a n c e l e v e l as the c r i t e r i o n f o r d i s t i n g u i s h i n g between us e f u l and dispensable items would appear from the psychometric l i t e r a t u r e to be a conservative procedure. The c o r r e l a t i o n s of an item with t o t a l scores on the four scales to which i t d i d not contribute were also c a l c u l a t e d . In a few cases items were found to c o r r e l a t e more hig h l y with another s c a l e . Where t h i s c o r r e l a t i o n exceeded the .05 l e v e l s i g n i f i c a n c e l e v e l , the item was regarded as an acceptable measure of that s c a l e . The statements f o r which t h i s was true are indexed i n Table 3.1. That a l i m i t e d number of statements were i n c o r r e c t l y assigned i s not too s u r p r i s i n g i n the present context where a l l 5 scales concern d i s p o s i t i o n s toward a common c r i t e r i o n . The main purpose i n computing item-scale c o r r e l a t i o n s i s to iden-t i f y statements which enhance the o v e r a l l r e l i a b i l i t y of a given s c a l e . As o u t l i n e d i n the previous chapter, a standard procedure for assessing scale r e l i a b i l i t y i s to c a l c u l a t e c o e f f i c i e n t alpha, which i n d i c a t e s the er r o r due to content sampling - the major source of v a r i a t i o n w i t h i n a t e s t instrument. The c a l c u l a t i o n of c o e f f i c i e n t alpha i s based l a r g e l y on the average c o r r e l a t i o n among items using the following formula: employed i n the present study i s : r r a - a. y i y 1 where r , = c o r r e l a t i o n of item 1 with t o t a l scores y. a •- standard deviation of t o t a l scores. y a, = standard d e v i a t i o n of item 1. r = c o r r e l a t i o n of item 1 with sum of scores on a l l items exclusive of item 1. 83. kr. . hi. rkk 1 - (k - 1) r. . ID where r = c o e f f i c i e n t alpha JCK. r\_. = average c o r r e l a t i o n among items k = number of items C o e f f i c i e n t alpha i s a p a r t i c u l a r l y meaningful measure since i t represents the expected c o r r e l a t i o n of one k-item with other hypothetical te s t s drawn from the same domain; and furthermore, the square root of c o e f f i c i e n t alpha i n d i c a t e s the c o r r e l a t i o n between scores on a k-item t e s t and hypothetical 'true scores' on the construct measured by the t e s t (Nunnally, 1967, pp. 193-4). Table 3.2 shows f o r each scale the average c o r r e l a t i o n among items, c o e f f i c i e n t alpha and the square root of c o e f f i c i e n t alpha c a l c u l a t e d from the p r e t e s t data. In terms of a s a t i s f a c t o r y l e v e l of r e l i a b i l i t y , (Nunnally (1967, p. 226) suggests that i n the e a r l y stages of research on p r e d i c t o r t e s t s or hypothetical measures of a construct, r e l i a b i l i t i e s of .50 or above are adequate. The r e l i a b i l i t i e s i n the p r e t e s t mea-sured up to t h i s standard, although there was c l e a r l y room f o r improve-ment. Two l o g i c a l procedures can be followed to increase r e l i a b i l i t y : eliminate statements having low item-scale c o r r e l a t i o n s ; and introduce new statements s i m i l a r i n content to o r i g i n a l items h i g h l y c o r r e l a t e d with t o t a l scores. 84. TABLE 3.2 PRETEST MEASURES OF SCALE RELIABILITIES SCALE AVG. INTER-ITEM CORRELATION COEFFICIENT ALPHA SQUARE ROOT OF COEFF. ALPHA FASHION .1924 .7812 .8838 CONVENIENCE .1616 .7430 .8620 STATUS .0810 .5693 .7545 QUALITY .0708 .5333 .7303 PRICE .0661 .5150 .7176 In r e v i s i n g the f i v e d i s p o s i t i o n a l scales the former procedure was followed but the l a t t e r was not. E s s e n t i a l l y the three statements f o r each scale e x h i b i t i n g the lowest item-scale c o r r e l a t i o n s were discarded. Those items which were eliminated are indexed i n Table 3.1. The r a -t i o n a l e f o r not introducing new statements was to avoid c r e a t i n g an overly cumbersome research instrument. There i s a c l e a r c o n f l i c t between the demands of psychometric theory and the p r a c t i c a l i t i e s of survey research. While i t i s de s i r a b l e from the point of view of r e l i a b i l i t y to administer t e s t s comprising a large number of items, consideration has t o be given to the imposition of such a research instrument upon a volunteer respondent. Whereas a captive student group may be w i l l i n g to endure lengthy sessions completing p s y c h o l o g i c a l t e s t s , the same w i l l i n g -ness can hardly be expected from a sample of the general population. I t follows that the desire i n the present study to obtain complete information from the f i n a l sample, while nujiimizing negative r e a c t i o n , d i c t a t e d that the length of the research instrument be c a r e f u l l y l i m i t e d . 85. ( i i i ) Construct V a l i d i t i e s In addition to the c a l c u l a t i o n of c o e f f i c i e n t alphas, which G u i l f o r d (1965, p. 399) argues can be int e r p r e t e d as a measure of ' i n t r i n -s i c v a l i d i t y ' , construct v a l i d i t y c o e f f i c i e n t s were obtained f o r each of the f i v e s c a l e s . As noted i n the previous chapter, f a c t o r analysis i s c e n t r a l to the assessment of construct v a l i d i t y . A reasonable procedure to follow i s to perform a f a c t o r analysis using raw scores on scale items as input v a r i a b l e s , to extract as many factors as there are o r i g i n a l s c a l e s , and to c a l c u l a t e f a c t o r scores. The construct v a l i d i t y of the scales can then be estimated by i n t e r c o r r e l a t i n g scale scores and f a c t o r scores. A f a c t o r analysis was performed using scores on the 75 p r e t e s t statements, and f i v e factors were extracted. Various rotations were performed and a b e s t - f i t s o l u t i o n was obtained using an oblique (biquar-timin) r o t a t i o n , confirming a c e r t a i n degree of i n t e r r e l a t i o n s h i p between the scale constructs. The f i v e f a c t ors together accounted f o r 60 per cent of the common f a c t o r variance. C o r r e l a t i n g summative scale scores and f a c t o r scores y i e l d e d the set of c o e f f i c i e n t s shown i n Table 3.3 V a l i d i t y c o e f f i c i e n t s appear on the p r i n c i p a l diagonal. Scanning across the rows of t h i s table shows that three of the scales - 'fashion', 'convenience' and 'qu a l i t y ' -were c l e a r l y v a l i d a t e d by corresponding f a c t o r scales 1, 2 and 3. The 'status' scale was l e s s strongly v a l i d a t e d by f a c t o r 4, and indeed corre-l a t e d almost as highly with f a c t o r 2. The 'price' scale f a i l e d to correspond with a unique v a l i d a t i n g f a c t o r showing stronger r e l a t i o n s h i p s 86. with factors 3 and 4 than with f a c t o r 5 with which i d e a l l y i t should have been most highly c o r r e l a t e d . TABLE 3.3 PRETEST MEASURES OF SCALE VALIDITIES FACTOR 1 FACTOR 2 FACTOR 3 FACTOR 4 FACTOR 5 FASHION .8944 -.4359 -.0107 -.1171 .2137 CONVENIENCE -.4330 .8738 -.2116 .0926 .1196 QUALITY .0867 -.1182 .7674 .0243 -.3390 STATUS .2868 -.5118 .2127 .5530 -.3165 PRICE .0205 -.0544 -.6448 -.4113 .2210 The interdependence of r e l i a b i l i t y and v a l i d i t y considerations, as noted by G u i l f o r d (1965, p. 399), i s confirmed by these p r e t e s t r e s u l t s since scales having highest c o e f f i c i e n t alphas also had highest v a l i d i t y c o e f f i c i e n t s . O v e r a l l , the f a c t o r structure s a t i s f a c t o r i l y v a l i d a t e d four of the f i v e s c a l e s . The lack of c l e a r v a l i d a t i o n i n the case of the 'price' scale was not regarded as a s u f f i c i e n t basis f o r i t s exclusion from the f i n a l research instrument. I t s retention was based upon i t s s a t i s f a c t o r y , i f moderate, l e v e l of r e l i a b i l i t y and the suspicion that a higher v a l i d i t y c o e f f i c i e n t would have emerged had the p r e t e s t sample been l e s s r e s t r i c t e d i n range. This suspicion was confirmed when v a l i d i t y c o e f f i c i e n t s were recomputed using data obtained from the f i n a l s t r a t i f i e d random sample (see below p.113). 87. (iv) P r e d i c t i v e V a l i d i t i e s As i n d i c a t e d above, the p r e t e s t sample included 83 respondents who completed questionnaires i n i n - s t o r e s i t u a t i o n s - 51 of these were i n t e r -viewed i n what can perhaps best be described as a "fashionable boutique" while the remaining 32 were patrons of a major department store. The information obtained from these two groups of customers provided a useable, though r e s t r i c t e d , data base f o r making a preliminary assess-ment of the p r e d i c t i v e v a l i d i t i e s of the f i v e d i s p o s i t i o n a l s c a l e s . In order to make t h i s assessment i t was necessary to t e s t f o r s i g n i f i c a n t d ifferences between the mean scale, scores of the two groups and further to demonstrate how accurately the scale scores d i s c r i m i n a -ted between the groups. Stepwise discriminant analysis provided a means of achieving these two r e l a t e d purposes (Cooley and Lohnes, 1971). This technique derives l i n e a r combinations of o r i g i n a l v a r i a -bles which maximally discriminate between predefined subject groups. I t s use i s most appropriate i n s i t u a t i o n s where the dependent v a r i a b l e (or c r i t e r i o n ) i s nominal and the independent variables (or predictors) are i n t e r v a l or r a t i o scaled. Linear combinations of o r i g i n a l v a r i a b l e s (or discriminant functions) are derived i n an i t e r a t i v e fashion - the va r i a b l e included i n the discriminant function at each stage being the one which maximizes between to within group variance as determined by a s e r i e s of uni v a r i a t e analyses of variance on each of the va r i a b l e s not previously included. The i t e r a t i v e procedure i s terminated when the F p r o b a b i l i t y of va r i a b l e s not included i n the discriminant func-t i o n i s i n a l l cases above a predetermined s i g n i f i c a n c e l e v e l (usually .05). The data obtained from the p r e t e s t subsample was c l e a r l y appro-p r i a t e f o r discriminant analysis purposes; the nominal c r i t e r i o n varia-b l e being 'store type 1 and the i n t e r v a l p r e d i c t o r s being the f i v e d i s -p o s i t i o n a l scales. Table 3.4 shows the group means, i n i t i a l F values and F probabi-l i t i e s f o r each of the f i v e s c a l e s . TABLE 3.4 GROUP MEANS AND F STATISTICS MEAN B MEAN D F VALUE F PROB. FASHION 53.16 43.12 53.93 .0000 PRICE 45.33 42.25 6.75 .0108 QUALITY 51.37 54.25 5.71 .0183 CONVENIENCE 35.88 37.19 0.89 .3496 STATUS 45.31 46.44 0.77 .7698 Boutique shoppers ** Department store shoppers I t emerged that the means for the two groups were s i g n i f i c a n t l y d i f f e r e n t f o r three of the f i v e s c a l e s , namely, 'fashion', 'price' and ' q u a l i t y ' . These s i g n i f i c a n t d i f f e r e n c e s can be inter p r e t e d by reference to the group means. The most s t r i k i n g d i f f e r e n c e was i n r e l a -t i o n to 'fashion o r i e n t a t i o n ' where the mean score f o r the 'boutique' group was more than ten points higher than that f o r the department store patrons - an i n t u i t i v e l y reasonable r e s u l t . Less s t r i k i n g , but s t i l l s i g n i f i c a n t , were the dif f e r e n c e s with respect to 'price', the department store shoppers placed the greater emphasis on 'quality' con-siderations . Other r e s u l t s obtained from the discriminant analysis were also r e v e a l i n g . Only one of the f i v e scales - 'fashion' - was included i n the discriminant function. The degree of covariance between t h i s scale and the others were s u f f i c i e n t to r a i s e the F p r o b a b i l i t i e s above the c r i t i c a l .05 l e v e l and hence prevented the i n c l u s i o n of any of the four remaining s c a l e s . Following the termination of the stepwise procedure, scores on the discriminant functions are c a l c u l a t e d f o r each of the o r i g i n a l subjects. These scores are the basis f o r p r e d i c t i n g membership i n the c r i t e r i o n groups. Each subject and the group centroids are represented by a p o i n t l o c a t i o n i n the geometric space defined by the discriminant functions. Subjects are assigned to groups on the basis of proximity to the group centroids. The coincidence between predicted and a c t u a l group membership then serves as a b a s i s f o r assessing how accurately the v a r i a bles introduced i n t o the discriminant function d i s t i n g u i s h between the members of the d i f f e r e n t groups. The discriminant analysis of the p r e t e s t data y i e l d e d the c l a s s i -f i c a t i o n matrix shown i n Table 3.5. TABLE 3.5 GROUP CLASSIFICATION MATRIX DEPARTMENT BOUTIQUE TOTAL ACTUAL MEMBERSHIP DEPARTMENT 26 (81%) 6 (19%) 32 BOUTIQUE 9 (18%) 42 (82%) 51 TOTAL PREDICTED MEMBERSHIP 35 48 90. The figures i n d i c a t e that a high degree of p r e d i c t i v e accuracy was achieved. O v e r a l l , approximately 82% of the subjects were assigned to the correct group, although only one of the f i v e scales - 'fashion' -was included i n the discriminant function. This c l e a r l y demonstrates the p r e d i c t i v e v a l i d i t y of the 'fashion' scale a l b e i t i n a r e s t r i c t e d s i t u a t i o n . In general, the r e s u l t s of -the discriminant analysis confirmed, i n varying degrees, the p r e d i c t i v e v a l i d i t y of at l e a s t three of the f i v e s c a l e s . C l e a r l y conclusions were te n t a t i v e given the r e s t r i c t e d range of shopping behaviour represented within the p r e t e s t subsample. Never-the l e s s , the a b i l i t y to demonstrate s i g n i f i c a n t d i f f e r e n c e s i n the d i s -p o s i t i o n s of the two c r i t e r i o n groups and to p r e d i c t group membership with such a high degree of accuracy served as a reassuring conclusion to the p r e t e s t a n a l y s i s . D. REVISED SCALES On the basis of the p r e t e s t r e s u l t s , i t was decided to r e t a i n a l l f i v e s c a l e s . Revisions were made l a r g e l y i n accordance with the item-scale c o r r e l a t i o n s shown i n Table 3.1. In the i n t e r e s t s of i n c r e a s i n g scale r e l i a b i l i t i e s and decreasing the length of the research instrument (and hence the time required for i t s administration), the three s t a t e -ments on each scale having lowest item-scale c o r r e l a t i o n s were discarded. Each of the rev i s e d scales therefore comprised twelve statements, making a t o t a l of 60 statements f o r i n c l u s i o n i n the f i n a l research instrument. An item a n a l y s i s of the rev i s e d scales using the data obtained i n the major data c o l l e c t i o n phase (see below p.112) subsequently confirmed the e f f e c t i v e n e s s of these changes. CHAPTER 4 QUESTIONNAIRE DESIGN, SAMPLE DESIGN AND DATA COLLECTION PROCEDURES A. INTRODUCTION The purpose of t h i s chapter i s to d e t a i l the components of the research design r e l a t e d d i r e c t l y to the major data c o l l e c t i o n phase; namely, questionnaire and sample design and data c o l l e c t i o n procedures. One component of the questionnaire comprised the f i v e d i s p o s i t i o n a l scales described i n the previous chapter. Two other components were designed to obtain information about the shopping behaviour and b i o -graphical c h a r a c t e r i s t i c s of consumers. The u t i l i t y of the information obtained by questionnaire survey methods stands or f a l l s to a large extent on the c h a r a c t e r i s t i c s of the sample from which i t i s derived. The r a t i o n a l e underlying the sample design employed i n t h i s study i s presented together with the mechanics of sample s e l e c t i o n . Questionnaires were administered by a team of t r a i n e d interviewers and the procedures followed f o r ensuring systematic and consistent data c o l l e c t i o n methods are o u t l i n e d . B. QUESTIONNAIRE DESIGN The questionnaire was designed to obtain three categories of i n f o r -mation: measures of consumer d i s p o s i t i o n s , consumer shopping behaviour and consumer b i o g r a p h i c a l c h a r a c t e r i s t i c s . The construction of a measurement instrument f o r the f i r s t of these was described i n the previous chapter. In t h i s s e c t i o n , the measurement of behavioural and - 91 -92. bio g r a p h i c a l v a r i a b l e s i s considered, (i) The Measurement of Behaviour Previous questionnaire surveys of shopping behabiour demonstrate d i f f e r e n t approaches to the measurement of behaviour. In some cases, respondents have been asked about t h e i r most recent shopping t r i p (Bucklin, 1967); i n others, about a s e r i e s of recent t r i p s (Brown, 1974); and i n others, about t h e i r 'normal' shopping behaviour (Rich, 1963). Each approach has c e r t a i n advantages and disadvantages. Asking consumers to report information about the most recent shop-ping t r i p has the advantage that accuracy of r e c a l l i s l i k e l y to be maximized since the time lapse between behaviour and report i s usually short. I t has the disadvantage that, f o r varying reasons, a s i n g l e shopping t r i p may be an a t y p i c a l example of a p a r t i c u l a r shopper's behaviour. The main advantage of obtaining information about a s e r i e s of past shopping t r i p s i s that i t permits a check on the consistency of r e l a -t ionships between dependent and independent v a r i a b l e s . The main d i s -advantage i s that accuracy of r e c a l l may be suspect i n cases where respondents have made r e l a t i v e l y infrequent shopping t r i p s - t h i s problem i s accentuated f o r middle and high order goods when the time lapse between purchases may be quite extensive. The a l t e r n a t i v e strategy of monitoring behaviour at regular time i n t e r v a l s using a panel group introduces the problem of data contamination. I t i s very d i f f i c u l t to assess the extent to which behaviour i s influenced by the consumer's awareness that shopping decisions have subsequently to be reported. 9 3 . Asking respondents to i n d i c a t e where they- most frequently shop f o r a c e r t a i n good has the advantage that i t serves to define groups of shoppers i n terms of ' t y p i c a l ' behaviour - f o r example, the regular downtown shopper as opposed to the person who normally patronizes a suburban shopping centre. I t i s reasonable to speculate that member-ship i n ' t y p i c a l ' c r i t e r i o n groups of t h i s kind i s more c l e a r l y r e l a t e d to independent p r e d i c t o r v a r i a b l e s than i s group membership defined on the basis of a s i n g l e shopping t r i p . (The r e s u l t s reported i n the following chapter i n f a c t confirm t h i s speculation.) The main d i s -advantage o f requesting information concerning ' t y p i c a l ' shopping t r i p s i s that for some respondents such a t r i p may not be definable. A consumer may f o r example patronize a number of d i f f e r e n t stores or shopping areas with approximately equal frequency. In which case, to request that a s i n g l e f a c i l i t y be s p e c i f i e d must r e s u l t e i t h e r i n miss-ing data or the reporting of deceptive information. Given that each of these three approaches to the documentation of behaviour has attendant advantages and disadvantages, i t was decided i n the present study not to r e l y on any one of them, but rather to obtain information r e l a t e d to a l l three. Respondents were therefore asked to describe t h e i r three most recent shopping t r i p s , and also to report, i f p o s s i b l e , where they most frequently shopped. I t was hoped that t h i s body of information would provide a basis for assessing the s t a b i l i t y of the r e l a t i o n s h i p s between consumer c h a r a c t e r i s t i c s and behaviour over a sequence of purchases. Questions were devised to e l i c i t the following information about each of the respondent's three most recent shopping t r i p s : 94. 1 . Name of store patronized. Studies of consumer s p a t i a l behaviour have tended to concentrate on shopping centre rather than store patronage (Brown, 1974; Buck l i n , 1967; Doddridge, 1972; Down, 1970). I m p l i c i t i n t h i s emphasis i s the assump-t i o n that shopping decisions involve two interdependent l o c a t i o n a l com-ponents of d i f f e r i n g s p e c i f i c i t y ; namely, choice of store, and choice of shopping centre. The r e l a t i v e importance of these components with respect to a s p e c i f i c shopping decision i s l i k e l y to vary between consu-mers. Ultimately, of course, a l l purchases have to be made at s p e c i f i c r e t a i l establishments and hence store choice can hardly be regarded as an i n c i d e n t a l consideration. There i s not the same need however f o r a consumer to make a conscious choice between shopping centres, and studies of consumer s p a t i a l behaviour which assume that decisions are formulated at that aggregate l e v e l may i n part be proceeding on the basis of a misconception. In t h i s context, Downs provides evidence of the i n a b i l i t y of shoppers to d i s a s s o c i a t e evaluations of shopping centres from evalua-t i o n s of stores within centres (Downs, 1970). S u f f i c e i t to say that the ambiguous r e l a t i o n s h i p between store choice and shopping centre patronage i s a p o t e n t i a l source of confusion which cannot be ignored i n the analysis of consumer s p a t i a l behaviour. For t h i s reason, an attempt was made i n t h i s study to account for behavioural d i f f e r e n c e s with respect to both store and shopping centre choices. 2. Type of store patronized. A procedure f o r c l a s s i f y i n g the stores s p e c i f i e d by the respondents was required since i t was i n e v i t a b l e , given the large number of c l o t h i n g o u t l e t s , that many stores would be mentioned very i n f r e q u e n t l y . Therefore, 95. i n order to increase the a n a l y t i c a l u t i l i t y of-the data, the respondent was asked to assign the store p r e v i o u s l y s p e c i f i e d to one of the follow-ing s i x categories: high-priced specialty, medium-priced specialty, boutique, department store, department store-budget floor, budget priced store. This c l a s s i f i c a t i o n scheme i s s i m i l a r to that used i n an e a r l i e r study of shopping behaviour (Rich, 1963) and has the desired advantage of being based e s s e n t i a l l y on denotative c r i t e r i a . 3. Store l o c a t i o n . Respondents were asked to sp e c i f y the loc a t i o n s of the stores they had patronized by i n d i c a t i n g , i f a p p l i c a b l e , the name of a shopping centre or otherwise the i n t e r s e c t i o n nearest to the store. For purposes of an a l y s i s , i t was necessary to c l a s s i f y these l o c a t i o n s . The r e t a i l l i t e r a t u r e suggests a number of approximately equivalent c l a s s i f i c a t i o n schemes, employing s i z e of f a c i l i t y and character of c o n t r o l ( i . e . , ' c o n t r o l l e d ' vs 'autonomous' centres) as the ba s i c taxonomic c r i t e r i a (Berry, 1963; K e l l y , 1956). Following Bucklin (1967), three l o c a t i o n a l categories were adopted: primary centres - r e s t r i c t e d to Downtown Vancouver and Downtown New Westminster; secondary centres - r e s t r i c t e d to planned shopping centres containing a major department store; tertiary centres - a l l remaining commercial d i s t r i c t s which were e s s e n t i a l l y e i t h e r ribbon developments or smaller planned shopping centres. 4. T r i p o r i g i n . Although most shopping t r i p s o r i g i n a t e from the home, previous r e -search has demonstrated that other o r i g i n s - p a r t i c u l a r l y workplace -are also important (Leigh, 1965). Respondents were required to s t i p u -l a t e whether the t r i p o r i g i n a t e d from the home, workplace, or another 6 ' 9 6 . l o c a t i o n which was to be s p e c i f i e d . 5. Location of t r i p o r i g i n . In order to permit the accurate measurement of distances separating the consumer's l o c a t i o n and a l t e r n a t i v e r e t a i l o pportunities, the re s -pondent was asked to sp e c i f y the hundred block address of the t r i p o r i g i n . 6. Distance estimate. There i s an increasing body of research evidence (Briggs, 1973) to suggest the importance of cognitive distance as a behavioural determinant. A very l i m i t e d attempt to obtain such a measure was made by asking r e s -pondents to estimate the p h y s i c a l distance separating t h e i r t r i p o r i g i n and d e s t i n a t i o n . 7. Travel mode. Respondents were asked to sp e c i f y which of four modal a l t e r n a t i v e s they had employed to reach t h e i r shopping d e s t i n a t i o n s . 8. T r i p purpose. The s e l e c t i o n of a p a r t i c u l a r r e t a i l f a c i l i t y may be influenced by the consumer combining a p a r t i c u l a r purchase with other purposes which may or may not involve shopping. In order to d i s t i n g u i s h s i n g l e pur-chase and multiple purpose t r i p s , respondents were asked to s p e c i f y t h e i r main purpose from a set of f i v e a l t e r n a t i v e s . 9. Past purchases. As a basi s f o r assessing the degree of commitment to a s p e c i f i e d store, respondents were asked to record the number of purchases made there i n the previous two years. These nine questions were repeated f o r each of the three most recent c l o t h i n g purchases. Questions 1, 2 and 3 were repeated again to deter-mine the store and shopping area patronized most often. ( i i ) Biographical Measures The b i o g r a p h i c a l v a r i a b l e s commonly employed i n past research were the ones included i n the present study; namely: age, sex, marital status, number of children at home, before taxes annual household income, * education and occupation of head of household. Since measures of t h i s kind are so commonly employed i n survey research, there i s l i t t l e need f o r f u r t h e r elaboration here. ( i i i ) General Design Considerations The f i n a l questionnaire comprised three main sections: Section 1 ** c o n s i s t i n g of the 60 L i k e r t statements , Section 2 comprising questions r e l a t e d to past c l o t h i n g purchases; and Section 3 dealing with b i o g r a -p h i c a l information. The ordering of these three sections was by no means a r b i t r a r y . There i s always a danger of data contamination i n circumstances such as t h i s where r e l a t e d measures are juxtaposed within a s i n g l e sur-vey instrument. I t i s necessary therefore to design the questionnaire This was recorded i n the form of a job d e s c r i p t i o n , which f o r a n a l y t i c a l purposes was assigned to one of the following o r d i n a l cate-gories: professional, managerial, white collar, skilled, semi-skilled, unskilled. ** The statements were s e r i a l l y ordered such that statement 1 was a measure of scale 1, statement 2 of scale 2, and so on to statement 6 which was again a measure of scale 1 beginning the second f i v e statement sequence. This ordering serves to minimize p o s s i b l e response bias 98. so as to minimize contamination e f f e c t s . In the present case the greatest danger of contamination appeared to be between the d i s p o s i -t i o n a l measures and the behavioural data. I t was decided that t h i s p o t e n t i a l problem could be c o n t r o l l e d most e f f e c t i v e l y by r e q u i r i n g respondents to complete the d i s p o s i t i o n a l scales before recording t h e i r past behaviour. The r a t i o n a l e f o r t h i s ordering was that response to statements can be d i s t o r t e d e i t h e r d e l i b e r a t e l y or subconsciously to match previously recorded behaviour with r e l a t i v e ease; whereas to d i s t o r t information about behaviour so as to correspond with p r e v i o u s l y stated d i s p o s i t i o n s requires that a respondent d e l i b e r a t e l y l i e - a contingency which most survey researchers would p r e f e r not to contem-p l a t e . The b i o g r a p h i c a l questions were po s i t i o n e d on the l a s t page of the questionnaire since they request information property regarded as c o n f i d e n t i a l and are therefore the most l i k e l y to arouse non-coopera-t i o n . In the i n t e r e s t s of obtaining complete information i t seemed prudent therefore to leave these questions u n t i l l a s t . The questionnaire was presented to the respondent i n the form of a ten page booklet, a copy of which appears as Appendix A. C. SAMPLE DESIGN (i) Population D e f i n i t i o n The population was defined as a l l households within the c i t y of Vancouver and the fe d e r a l e l e c t o r a l d i s t r i c t of Capilano, which encom-passes West Vancouver D i s t r i c t , North Vancouver C i t y and the western h a l f r e s u l t i n g from a succession of statements r e l a t e d to the same scale. At the same time, the ordering i s s u f f i c i e n t l y systematic to expedite the coding and analysis of the data. 99. of North Vancouver D i s t r i c t . These geographical l i m i t s were based on the desire that sample respondents be aware of and have access to a range of a l t e r n a t i v e r e t a i l o p p o r t u n i t i e s . On the basis of the s p a t i a l d i s t r i b u t i o n of commercial floorspace i n Greater Vancouver (see Figure 4.1), i t seemed f a i r to assume that households within the defined geo-graphical area had reasonable access to the major commercial centre i n Downtown Vancouver and to planned shopping centres and other shopping d i s t r i c t s d i s t r i b u t e d throughout the area. A d d i t i o n a l support f o r t h i s assumption derived from the f a c t that a l l households within the sample area were within approximately 30 minutes d r i v i n g time of Downtown Vancouver (see Figure 4.2) and were considerably shorter time distances from other r e t a i l f a c i l i t i e s . ( i i ) S t r a t i f i c a t i o n by Area In the i n t e r e s t s of drawing a sample of the population which was p r o p o r t i o n a l l y representative of a l l parts of the urban region defined above, i t was decided to s t r a t i f y on a census enumeration area base. I t was f u r t h e r hoped that s t r a t i f y i n g the sample i n t h i s way would help to maximize behavioural and b i o g r a p h i c a l variance given the known d i s -t r i b u t i o n s of commercial f a c i l i t i e s (see Figure 4.1) and socio-economic groups within the sample area (Patterson, forthcoming). ( i i i ) Sample Selection The number of households i n the sample area as determined by the 1971 census was 189,124 representing 967 census enumeration areas (E.A.s) of approximately equal population size-. I t was decided that a FIGURE 4.1 100. THE SPATIAL DISTRIBUTION OF COMMERCIAL FLOORSPACE IN GREATER VANCOUVER i i i . -i # •/  1 • £ek \ " i a • C O M M E R C I A L F L O O R S P A C E FLOOR AREA SQUARE FEET A CENTRE WITH LESS THAN 15,000 SO. FT. " FLOOR AREA. CENTRE WITH UNKNOWN FLOOR AREA O (LESS THAN 15,000 SO. FT. BUT NOT INCLUDED IN TABULATIONS). NOTES : BASED ON MUNICIPAL ASSESSMENT DATA AND L.MR.PB. ESTIMATES. DATA VARIES IN YEAR AND DEFINITION ANO THUS ONLY APPROXIMATES ACTUAL FLOOR SPACE. SEE TABLES FOR COMMENTS. NUMBERS AND LETTERS REFER TO TABLES. GREATER VANCOUVER REGIONAL DISTRICT PLANNING DEPARTMENT FEBRUARY, 1978 SCALE IN MILES 0 I 2 FIGURE 4.2 TRAVEL TIME TO DOWNTOWN VANCOUVER* 101. * Map shows 5 minute isochrones from i n t e r s e c t i o n of G r a n v i l l e and Georgia i n Downtown Vancouver at 4:30 p.m. by f a s t e s t route (data c o l -l e c t e d i n summer of 1968 by G.V.R.D. Planning Department). s u f f i c i e n t l y large sample pool (approximately 483) would be obtained by a random s e l e c t i o n of households from alternate enumeration areas as l i s t e d i n the Canada Census Street Directory f o r Greater Vancouver. This d i r e c t o r y l i s t s the block-face addresses occurring within each enumeration area. Two block faces were randomly s e l e c t e d from a l t e r n a t e E.A.s. This procedure defined a range of household addresses from which a s i n g l e household was subsequently select e d . L e t t e r s were d e l i v e r e d to the f i r s t ten households within the s p e c i f i e d range of addresses. This l e t t e r served to introduce the study and to state i t s general object-ives - a copy of the l e t t e r appears as Appendix B. I t was hoped that t h i s p r i o r contact would enhance the response rate. A f t e r about two days a t r a i n e d interviewer (see below} returned to a randomly selected household within the ten which had received a l e t t e r i n the hope of f i n d i n g a s u i t a b l e and w i l l i n g respondent. The c r i t e r i a f o r determining s u i t a b i l i t y were those of age and sex. Inter-viewers were i n s t r u c t e d that respondents must be seventeen or above, and the sex of the respondent required from each E.A. was also s t i p u l a t e d . (In t h i s way, the sample was e f f e c t i v e l y s t r a t i f i e d by sex as w e l l as by area). I f a s u i t a b l e and w i l l i n g respondent was not found at the randomly selected household, the interviewer was i n s t r u c t e d to proceed to the next household (in order of ascending address) where a l e t t e r had been deli v e r e d . This procedure was repeated u n t i l a s u i t a b l e and w i l l i n g respondent was found. D. INTERVIEWER TRAINING Data c o l l e c t i o n was undertaken by a team of t r a i n e d interviewers. The interviewers were drawn from a second year geography cl a s s and the research task formed an i n t e g r a l p a r t of t h e i r course requirements. Each interviewer received two hour long periods of i n s t r u c t i o n . The f i r s t took the form of a c l a s s l e c t u r e i n which the purposes and procedures of the research design were o u t l i n e d i n the context of the pert i n e n t research l i t e r a t u r e . In the second session, d e t a i l e d i n s t r u c -t i o n s concerning the methods of data c o l l e c t i o n were given. In addi-t i o n , each interviewer received an i n s t r u c t i o n manual (see Appendix C) which summarized the procedures to be followed. The r e s p o n s i b i l i t y of each interviewer was to d e l i v e r l e t t e r s to ten designated addresses i n each of four E.A.s l i s t e d on an assignment sheet (see Appendix D) and subsequently to complete four interviews - one i n each E.A. E. DATA RETURNS AND CODING PROCEDURES Data c o l l e c t i o n was completed i n the second h a l f o f November 1973. A t o t a l of 351 completed questionnaires were returned. This f i g u r e represents a 73% response rate - the remaining 27% were l a r g e l y accoun-ted f o r by a p a r t i a l or t o t a l f a i l u r e of c e r t a i n interviewers to f u l f i l l t h e i r assigned r e s p o n s i b i l i t i e s . The geographical d i s t r i b u t i o n and bio g r a p h i c a l c h a r a c t e r i s t i c s of the 351 respondents - described i n the following chapter - were such that the sample was riot s e r i o u s l y compro-mised, and, as a r e s u l t , no e f f o r t was made to supplement the returns. 104. The questionnaire (see Appendix A) was designed such that the coding could be completed on the booklet i t s e l f . In large measure, t h i s task simply involved the t r a n s c r i p t i o n of data codes; the coding of distance measures was somewhat more involved however. Preparatory to data a n a l y s i s , i t was necessary to obtain measures of the distances separating respondents from various r e t a i l f a c i l i t i e s ; namely: the store patronized, Downtown Vancouver, the nearest planned shopping centre; and the nearest other shopping d i s t r i c t containing a * c l o t h i n g s t o r e . These distances were determined by p l o t t i n g the respondent's t r i p o r i g i n on a large scale map of the Greater Vancouver area and subsequently measuring the main a r t e r i a l distances to each of the four f a c i l i t i e s using a plenimeter. Time distance to Downtown Vancouver was also coded on the basis of a r e l a t i v e l y recent survey (G.V.R.D., 1970). Also coded were the s i z e s (in square feet) of the shopping area patronized, the nearest planned shopping centre and the nearest other shopping d i s t r i c t . These f i g u r e s were obtained from a survey of com-mercial floorspace (G.V.R.D., 1970a) which was updated to include recent changes i n the r e t a i l s t r u c t u r e . Following the completion of the coding phase the data were tr a n s -fered to cards i n preparation f o r computer a n a l y s i s . * F i e l d survey served to define these commercial f a c i l i t i e s . F. SUMMARY This chapter has described questionnaire design, sample design and data c o l l e c t i o n procedures. In each case, an e f f o r t was made to mini-mize sources of bias and hence strengthen the data. A f t e r coding and key-punching, the data were analysed and the r e s u l t s are presented i n the following chapter. CHAPTER 5 DATA ANALYSIS A. INTRODUCTION ' The data from the questionnaire completed by the 351 respondents were analysed and the r e s u l t s of the analyses are presented i n t h i s chapter. The chapter i s divided i n t o three main sections. The f i r s t includes a d e s c r i p t i o n of the s p a t i a l d i s t r i b u t i o n and bi o g r a p h i c a l c h a r a c t e r i s t i c s of the respondents as a check on the representative-ness of the sample. The p o s s i b l e b i a s introduced by the interviewer i s also considered. Measures of the r e l i a b i l i t y and v a l i d i t y of the f i n a l L i k e r t scales are then presented and compared with the corresponding r e s u l t s i n the p r e t e s t . The second section i s devoted to a d e t a i l e d d e s c r i p t i o n of the shopping behaviour of the respondents. A s e r i e s of tables i s presented to show the d i s t r i b u t i o n of shopping t r i p s i n terms of both store and shopping centre patronage. The r e l a t i o n s h i p s between selected behaviour-a l v a r i a b l e s and r e t a i l patronage are demonstrated. In the t h i r d s ection attention i s focussed on the major objective of the analysis which was to demonstrate how e f f e c t i v e l y measures of the locational, biographical and dispositional c h a r a c t e r i s t i c s of the re s -pondents, both separately and i n composite, discriminated between pre-defined c r i t e r i o n groups; where the c r i t e r i a were type of store and type of shopping centre patronized on each of four shopping t r i p s . The - 106 -107. sample s i z e was also s u f f i c i e n t to permit a r e s t r i c t e d analysis of the e f f i c a c y of these same measures as di s c r i m i n a t o r s between patrons of s p e c i f i c department stores and i n d i v i d u a l shopping centres. The p r e d i c -t i v e power of the three sets of measures i s furt h e r compared by means of a regression analysis i n which distance travelled to shop was the dependent v a r i a b l e and the independent v a r i a b l e s were again various l o c a -t i o n a l , b i o g r a p h i c a l and d i s p o s i t i o n a l measures. S p e c i f i c consideration i s a lso given to male-female d i f f e r e n c e s i n shopping behaviour and d i s -p o s i t i o n s . B. SAMPLE CHARACTERISTICS, INTERVIEWER BIAS AND SCALE STATISTICS (i) S p a t i a l D i s t r i b u t i o n of Respondents. The purpose of s t r a t i f y i n g on an enumeration base was to ensure that the sample be p r o p o r t i o n a l l y representative of subareas within the t o t a l sampling area. In order to determine how w e l l t h i s aim was * achieved the sample and population d i s t r i b u t i o n s were compared f o r each of the s i x e l e c t o r a l d i s t r i c t s which together make up the sampling area. The r e s u l t i n g frequency d i s t r i b u t i o n s are shown i n histogram form i n Figure 5 . 1 . For each e l e c t o r a l d i s t r i c t , bar A represents the percentage of the sample, bar B the percentage of census enumeration areas and bar C the percentage of households. Since the sample was drawn on an enumeration area base, bars A and B should be approximately coincident. In no case does the d i s p a r i t y exceed two percentage p o i n t s , e f f e c t i v e l y demonstrat-* Population d i s t r i b u t i o n f i g u r e s based on 1971 Canada Census r e s u l t s . 108. FIGURE 5.1 SAMPLE AND POPULATION DISTRIBUTIONS BY ELECTORAL DISTRICT CAPILANO VAN. CENTRE VAN. EAST VAN. KNGSWY VAN. QUADRA VAN. SOUTH A B C A B C A B C A B C A B C A B C 16.8 18.2 17.4 19.9 21.3 24.4 18.2 16.3 15.4 14.0 13.4 13.5 16.0 14.5 13.6 14.8 15.9 15.5 A = % sample B = % population (E.A.s) C = % population (hhlds) ing at t h i s r e l a t i v e l y coarse l e v e l the achievement of the desired a r e a l representation. Mapping respondent l o c a t i o n s (see Figure 5.2) f u r t h e r served to confirm the approximate coincidence between the sample d i s t r i -bution and the population density surface. FIGURE 5.2 SPATIAL DISTRIBUTION OF RESPONDENTS AND SELECTED RETAIL LOCATIONS Department Stores B The Bay E Eatons S Sears W Woodwards Shopping Centres * Park Royal + Oakridge 110. (ii) Biographical Characteristics. A further concern was that the sample be representative of the popu-lation in terms of salient demographic and socio-economic variables. Histograms, showing the frequency and percentage distribution of res-pondents on each of the biographical measures obtained, appear in Figure 5.3. At the time of writing, the 1971 Canada census s t a t i s t i c s for the population on each of these variables were not available. Nevertheless, in light of past s t a t i s t i c s , i t would seem that the sample characteris-t i c s , described by the histograms, show no major departures from popu-lation distributions. There i s perhaps a disproportionate representa-tion of younger respondents without children and also of highly educated people engaged in professional occupations. But these biases are merely speculative and are unlikely to invalidate the results of the analyses which follow. ( i i i ) Interviewer Characteristics and Response Bias. An attested source of response bias in survey research stems from such characteristics of the interviewer as sex, age and education (Kahn and Cannel, 1957). Since the interviewers used in the data collection phase in this study were a l l drawn from the same undergraduate class, homogeneity with respect to age and educational background was assumed. As a result attention was restricted to assessing the effects of the inferviewer 1s sex. Of the 351 questionnaires, 234 were administered by male interviewers and 117 by females. FIGURE 5.3 BIOGRAPHICAL CHARACTERISTICS OF THE SAMPLE o • O CT\ CN C\ ON ON O fi N W ^ ^ LT\ MD t- CD N C I^V) ACE 4 0 p E R C E N T 2 0 SEX CD J O R O s ™ a m S j n MARITAL STATUS o H cy <o v ir»eo # CHILDREN vo y • t~ PC A u O fe O CK W • £ t/i o to K E ^ M n r U W K X B. H a CM . Fran T CO (\l W C\J O ro r cc CM vo ^ r l H CO W vfi 6 -H H W N W INCOME (I'OOO) i t n I — eg EDUCATION j j j j j x j j ij a t,1 Sri M o o K W iri O £ D n » a £ OCCUPATION p E E s a ti, w E J E< Z E< W | j 3 to fa « EMPLOYMENT STATUS 112. A s e r i e s of u n i v a r i a t e analyses of variance were performed to t e s t f o r s i g n i f i c a n t d i f f e r e n c e s i n the responses e l i c i t e d by male and female interviewers. Only i n the case of the respondent's occupation was the F s t a t i s t i c s i g n i f i c a n t at the .05 l e v e l r e f l e c t i n g the o v e r a l l tendency for subjects interviewed by males to record higher status occupations. The extent to which t h i s d i s p a r i t y r e f l e c t s a response bias i s a matter for conjecture. The f a c t that no other s i g n i f i c a n t d i f f e r e n c e s emerged i s reassuring. C. LIKERT SCALE RELIABILITIES AND VALIDITIES Relevant s t a t i s t i c s were c a l c u l a t e d to check the r e l i a b i l i t i e s and v a l i d i t i e s of the f i n a l L i k e r t scales. The same procedures as were described i n the p r e t e s t analysis (in Chapter 3) were employed. The reader i s r e f e r r e d to that chapter for a d e s c r i p t i o n of these . s t a t i s t i c s and the basis f o r t h e i r use. The f i n a l item-scale c o r r e l a t i o n s are l i s t e d i n Chapter 3 (Table 3.1, pp. 75-79 ) to ease comparison with the p r e t e s t r e s u l t s . In almost a l l cases the c o r r e l a t i o n s increased. In part t h i s increase merely r e f l e c t s the r e s u l t of reducing the number of c o n t r i b u t i n g statements f o r each scale from f i f t e e n to twelve. This reduction has the i n e v i t a b l e e f f e c t of increasing the proportion of scale variance accounted f o r by i n d i v i d u a l item variance. The c a l c u l a t i o n of c o e f f i c i e n t alphas served to demonstrate that t h i s apparent improvement i n the item and scale r e l i a b i l i t i e s was not merely an a r t i f a c t . Table 5.1 l i s t s f o r each scale the mean i n t e r - i t e m c o r r e l a t i o n , c o e f f i c i e n t alpha, the square root of c o e f f i c i e n t alpha and 113. TABLE 5.1 FINAL SCALE RELIABILITY AND VALIDITY COEFFICIENTS AVG. INTER-ITEM COEFFICIENT SQ. RT. OF CONSTRUCT SCALE CORRELATION ALPHA COEFF. ALPHA VALIDITY FASHION .2820 .8249 .9182 .9699 (.1924) (.7812) (.8838) (.8944) CONVENIENCE .2013 .7573 .8703 .8325 (.1616) (.7430) (.8620 (.8738) STATUS .1845 .7310 .8550 .8181 (.0810) (.5693) (.7545) (.5530) QUALITY .1681 .7079 .8413 .6376 (.0708) (.5333) (.7303) (.7674) PRICE .1381 .6406 .8004 .7652 (.0661) (.5150) (.7176) (.2210) * construct v a l i d i t y c o e f f i c i e n t based on the scores obtained from the 351 respondents. To ease comparison, the corresponding s t a t i s t i c s from the p r e t e s t (see Tables 3.2 and 3.3) are shown i n parentheses. In general there i s a marked improvement i n the f i n a l s c a le s t a t i s -t i c s over the p r e t e s t l e v e l s . This improvement i s p a r t l y a t t r i b u t a b l e to the modification of the scales i n l i g h t of the p r e t e s t r e s u l t s , and p a r t l y to the use of a les s constrained sample group i n the major data c o l l e c t i o n phase, which increases item and scale variances and tends to sharpen construct d i f f e r e n c e s . This r e s u l t s from the f a c t that the responses of a heterogeneous set of subjects are more l i k e l y to represent a complete range of d i s p o s i t i o n s . * The p r i n c i p a l diagonal values of the matrix obtained from c o r r e l a -t i n g summative scale and f a c t o r scale scores. D. SHOPPING BEHAVIOUR: DESCRIPTIVE STATISTICS (i) Store and Shopping Centre Patronage. The 351 questionnaires provided p o t e n t i a l information on 1404 (351 x 4) shopping t r i p s . Figures 5.4 and 5.5 i l l u s t r a t e the d i s t r i b u -t i o n of these t r i p s by store type and shopping centre. In both cases, there .was some missing data which l a r g e l y r e f l e c t s the i n a b i l i t y of ce r t a i n respondents to r e c a l l l e s s recent shopping t r i p s or to s t i p u l a t e a "most frequent" t r i p . FIGURE 5.4 TRIP DISTRIBUTION BY STORE TYPE MISSING DATA HIGH PRICE SPECIALTY MEDIUM PRICE SPECIALTY BOUTIQUE DEPARTMENT DEPT./BUDGET BUDGET 0 10 20 30 40 50 60 j i I t 1 I 1 Percentage of Tr i p s * In the case of t r i p d i s t r i b u t i o n by store type (Figure 5.4), the most s t r i k i n g feature i s the predominance of department store shopping, The s i x categories represent approximately 200 d i f f e r e n t stores mentioned by the respondents. The r e l i a b i l i t y of the c l a s s i f i c a t i o n scheme was attested by the minimal discrepancy between respondents i n c l a s s i f y i n g the same store. 78 80 21Q 99 245 118 115. which accounted for 57 per cent of the observations. The other categories are small by comparison, the smallest being department store - budget f l o o r which for purposes of further analysis was added to the budget store category. FIGURE 5.5 TRIP DISTRIBUTION BY SHOPPING CENTRE MISSING DATA 122 DOWNTOWN SECONDARY CENTRE 433 TERTIARY CENTRE 150 699 0 10 I 20 30 i 40 50 __i 60 — i Percentage of Trips The d i s t r i b u t i o n of t r i p s by shopping centre (Figure 5.5) i n d i c a t e s the dominance of downtown shopping. This r e f l e c t s two f a c t o r s : the con-centration of c l o t h i n g stores i n Downtown Vancouver; and the f a c t that the sample was drawn l a r g e l y from within the trade area of the c e n t r a l business d i s t r i c t . Nevertheless, the other two categories, and p a r t i c u -l a r l y secondary centres, accounted f o r s i g n i f i c a n t percentages of the t r i p d e s t i n a t i o n s . In a r e l a t e d study Bucklin (1967) found the following percentage d i s t r i b u t i o n of t r i p s to centres of varying s i z e : primary -50 per cent, secondary - 17 per cent, t e r t i a r y - 33 per cent. The two sets of r e s u l t s are not s t r i c t l y comparable however since i n Bucklin's study respondents (in Oakland, C a l i f o r n i a ) were asked to s p e c i f y where they had l a s t made any s i n g l e item purchase over $5.00 i n value. 116. The u n i v a r i a t e d i s t r i b u t i o n s shown i n Figures 5.4 and 5.5 are com-* bined i n the following b i v a r i a t e table (Table 5.2) which provides addi-t i o n a l information about the shopping patterns of the sample group. The table shows both frequency and h o r i z o n t a l percentage figures with the l a t t e r appearing i n parentheses. In percentage terms, department stores located downtown and i n secon-dary centres accounted f o r approximately 35 per cent and 26 per cent r e s p e c t i v e l y of a l l shopping t r i p s . The next l a r g e s t b i v a r i a t e category - downtown budget stores - accounted f o r a considerably smaller percentage of t r i p s (6.5%). For each store type, other than department sto r e s , the l a r g e s t per-centage of t r i p s was to downtown locatio n s followed by t e r t i a r y centres and t h i r d l y secondary centres. There are however v a r i a t i o n s within t h i s o v e r a l l trend. Whereas, for example, over 60 per cent of t r i p s to boutique and budget stores involved downtown l o c a t i o n s , the same was true f o r only 40 per cent approximately of t r i p s to high and medium p r i c e d s p e c i a l t y o u t l e t s . In these l a t t e r cases an almost equal percentage of t r i p s as to stores located i n t e r t i a r y centres. This f i n d i n g i s somewhat contrary to t r a d i t i o n a l c e n t r a l place notions whereby s p e c i a l t y stores are expected to seek downtown l o c a t i o n s i n order to maximize t h e i r market area. I t i s supportive however of Leigh's conclusion that s p e c i a l t y stores i n Vancouver show a tendency to locate i n suburban centres proximate to the r e s i d e n t i a l areas i n which t h e i r c l i e n t e l e are concentrated (Leigh, 1965). * Tabulations exclude missing data f o r t h i s and subsequent b i v a r i a t e t a b l e s . TABLE 5.2 DISTRIBUTION OF SHOPPING TRIPS BY STORE TYPE AND SHOPPING CENTRE r . ™ ™ , ^ ™ SECONDARY TERTIARY „ « m 1 1 T DOWNTOWN ^ ^ „ m r > r , TOTAL CENTRE CENTRE High P r i c e S p e c i a l t y 31 (41%) 19 (25%) 26 (34%) 76 Medium P r i c e S p e c i a l t y 77 (39%) 47 (24%) .72 (37%) 196 Boutique 55 (61%) 17 (19%) 18 (20%) 90 Department 452 (58%) 332 (42%) 0 ( 0%) 784 Budget 84 (63%) 17 (13%) 33 (25%) 134 To t a l 699 (55%) 432 (34%) 149 (12%) 1280 Chi-Square = 318.02 Degrees of Freedom = 8 S i g n i f i c a n t at .001 l e v e l 118. The bivariate distribution of shopping trips (.Table 5.2) shows a strong relationship between store and shopping centre choices. The chi-square value confirms a significant relationship. This value i s however a r t i f i c a l l y inflated because of the absence, by definition of the shop-ping centre classification scheme (see p. 95 ), of department stores in tertiary centres. (ii) Trip Origin. Figure 5.6 illustrates the distribution of shopping trips by origin. The breakdown by category i s as expected with the largest percentage of trips originating from the home (75%), followed by the workplace (14%) and thirdly other locations (5%) which in many cases were the homes of relatives or friends. FIGURE 5.6 TRIP DISTRIBUTION BY ORIGIN MISSING DATA 66 HOME WORK 145 OTHER 53 0" 10 20 30 40 50 60 70. 80 i i i i i i i i i Percentage of Trips Smaller N results from the fact that 'trip origin' was not recor-ded for "most frequent" t r i p ; the same applies to 'travel mode', 'trip purpose' and 'past purchases'. 119. T r i p o r i g i n was also cross-tabulated with 'store type and shopping centre and the r e s u l t s are shown i n Tables 5.3 and 5.4. The percentage figures i n Table 5.3 reveal c e r t a i n aggregate d i f f e r -ences i n the types of store patronized by shoppers having d i f f e r e n t t r i p o r i g i n s . Most marked i s the higher percentage of t r i p s to medium p r i c e d s p e c i a l t y stores amongst those whose t r i p s o r i g i n a t e d from a workplace or other locations and the correspondingly smaller percentage of t r i p s to departments sto r e s . The chi-square s t a t i s t i c i s again s i g n i f i c a n t i n d i c a t i n g a systematic r e l a t i o n s h i p between t r i p o r i g i n and store type. Table 5.4 shows that a s t i l l stronger r e l a t i o n s h i p e x i s t s within the data between t r i p o r i g i n and shopping centre patronage. In t h i s case, there i s a contrast between t r i p s o r i g i n a t i n g from a workplace, of which 71 per cent were to a downtown d e s t i n a t i o n , and those o r i g i n a t i n g from other l o c a t i o n s which were more evenly d i s t r i b u t e d between d i f f e r e n t centres with t e r t i a r y centres being the modal category. Where the home was the o r i g i n , downtown locatio n s were again the most frequent followed by secondary and t e r t i a r y centres i n that order. The marked tendency f o r t r i p s beginning at a workplace to end at a downtown store i s l a r g e l y explained by the f a c t that such t r i p s were i n most cases undertaken by people working i n the c e n t r a l business d i s t r i c t . ( i i i ) Travel Mode. The d i s t r i b u t i o n of shopping t r i p s by t r a v e l mode i s shown i n Figure 5.7. These figures confirm the importance of the automobile as the major means of transportation to shopping destinations (see Claus and Hardwick (1972) for a d e t a i l e d consideration of automobile o r i e n t e d r e t a i l i n g ) . 1 TABLE 5.3 DISTRIBUTION OF SHOPPING TRIPS BY TRIP ORIGIN AND STORE TYPE HIGH PRICE MEDIUM PRICE BOUTIQUE DEPARTMENT BUDGET TOTAL SPECIALTY SPECIALTY * Home 51 ( 6%) 118 (15%) 61 ( 8%) 465 (59%) 93 (12%) 788 Work 9 ( 6%) 37 (25%) 14 (10%) 72 (50%) 13 ( 9%) 145 Other 3 ( 5%) 20 (38%) 5 ( 9%) 18 (34%) 7 (13%) 53 TOTAL 63 ( 6%) 175 (18%) 80 ( 8%) 555 (56%) 113 (11%) 986 Chi Square = 28.88 Degrees of Freedom = 8 S i g n i f i c a n t at .001 l e v e l ro o TABLE 5.4 DISTRIBUTION OF SHOPPING TRIPS BY TRIP ORIGIN AND SHOPPING CENTRE DOWNTOWN ~ TOTAL CENTRE CENTRE Home 397 (51%) 286 (37%) 88 (11%) 771 Work 100 (71%) 22 (16%) 19 (13%) 141 Other 12 (32%) 11 (29%) 15 (39%) 38 Tota l 509 (54%) 319 (34%) 122 (13%) 950 Chi-Square = 50.88 Degrees of Freedom = 4 S i g n i f i c a n t at .001 l e v e l 122. The frequency of pedestrian t r i p s (13 per cent) perhaps exceeds expecta-t i o n and i n large measure r e f l e c t s the tendency of people who work down-town to walk to stores i n the immediate v i c i n i t y of t h e i r workplace. FIGURE 5.7 TRIP DISTRIBUTION BY TRAVEL MODE MISSING DATA 49 CAR BUS 232 WALK 134 OTHER 36 0 10 20 30 40 50 60 I 1 1 I I i l Percentage of Tri p s S i g n i f i c a n t d i f f e r e n c e s emerge i n the store and shopping centre choices of consumers using d i f f e r e n t modes of t r a v e l (Tables 5.5 and 5.6). Most s t r i k i n g i s the ass o c i a t i o n of pedestrian t r i p s and s p e c i a l t y store patronage (Table 5.5). Compared with the other t r a v e l modes, a greater proportion (39 per cent) of such t r i p s were to one or other of the spe-c i a l t y store categories. Whereas the proportion of these same t r i p s to department stores was comparatively l e s s (38 per cent). This f i n d i n g seems to underline the tendency noted above for s p e c i a l t y stores to locate i n commercial d i s t r i c t s proximate to the homes or workplaces of t h e i r c l i e n t e l e . Clear d i f f e r e n c e s can be seen i n the t r a v e l modes employed to reach d i f f e r e n t types of shopping centres (Table 5.6) . There i s an approximately TABLE 5.5 ' DISTRIBUTION OF SHOPPING TRIPS BY TRAVEL MODE AND STORE TYPE HIGH PRICE MEDIUM PRICE n^n r i M ^ " o u i u " ^ BOUTIQUE DEPARTMENT BUDGET TOTAL SPECIALTY SPECIALTY y Car 48 ( 8%) 102 (17%) 49 ( 8%) 348 (58%) 54 (9%) 601 Bus 4 ( 2 % ) 30 (13%) 20 ( 9%) 143 (62%) 34 (15%) 231 Walk 9 ( 7%) 43 (32%) 10 ( 7%) 51 (38%) 21 (16%) 134 Other 3 ( 8 % ) 4 (11%) 1 ( 3 % ) 23 (64%) 5 (14%) 36 TOTAL 64 ( 6%) 179 (18%) 80 ( 8%) 565 (56%) 114 (11%) 1002 Chi Square = 48.84 Degrees of Freedom = 12 S i g n i f i c a n t at .001 l e v e l to OJ TABLE 5.6 DISTRIBUTION OF SHOPPING TRIPS BY TRAVEL MODE AND SHOPPING CENTRE DOWNTOWN „„„^„ „ ™ , m r , r , TOTAL CENTRE CENTRE Car 235 (40%) 268 (46%) 82 (14%) 585 Bus 195 (85%) 26 (11%) 9 ( 4%) 230 Walk 84 (65%) 16 (12%) 29 (22%) 129 Other 10 (40%) 11 (44%) 4 (16%) 25 Tot a l 524 (54%) 321 (33%) 124 (13%) 969 Chi-Square = 168.46 Degrees of Freedom = 6 S i g n i f i c a n t at .001 l e v e l 125. even s p l i t of automobile t r i p s to downtown and secondary centres. Bus t r a n s i t i s predominantly used to reach downtown l o c a t i o n s , r e f l e c t i n g the downtown focus of most p u b l i c t r a n s i t routes i n Greater Vancouver, while access to secondary centres i s l a r g e l y r e s t r i c t e d to those with cars (84 per cent of t r i p s to secondary centres were by c a r ) . The per-centage d i s t r i b u t i o n of pedestrian t r i p s confirms the as s o c i a t i o n with downtown locatio n s noted above. I t also emerges that t r i p s to t e r t i a r y centres are frequently made on foot, which i s some confirmation of cen-t r a l place expectations that these smaller centres have comparatively l o c a l i s e d trade areas. (iv) T r i p Purpose. Approximately 62 per cent of t r i p s were motivated by the s p e c i f i c need to purchase c l o t h i n g (Figure 5.8). In a furt h e r 23 percent of cases, c l o t h i n g purchases were made i n the course of a t r i p where the stated purpose was to f u l f i l l other shopping needs. The other t r i p purpose categories accounted f o r comparatively small percentages. Of these the l a r g e s t was the r e c r e a t i o n a l category suggesting that shopping f o r clothes i s regarded by some as a l e i s u r e a c t i v i t y . FIGURE 5.8 TRIP DISTRIBUTION BY TRIP PURPOSE MISSING DATA 46 CLOTHES OTHER SHOPPING 245 BUSINESS 35 SOCIAL 26 RECREATIONAL 52 0 10 20 30 40 50 60 70 • i i i i i • • 126. Interp r e t a t i o n of the b i v a r i a t e table r e l a t i n g t r i p purpose and store type (Table 5.7) i s r e s t r i c t e d by the number of low frequency en-t r i e s (which accounts for the i n f l a t e d chi-square value). One obvious observation can be made however i n terms of the coincidence of t r i p s to s p e c i a l t y and boutique sto r e s , which by d e f i n i t i o n market c l o t h i n g goods e x c l u s i v e l y , and the stated t r i p purpose of purchasing c l o t h i n g . No s i g n i f i c a n t d i f f e r e n c e s i n the t r i p purposes of shoppers s e l e c t -i n g d i f f e r e n t types of shopping centre were found (Table 5.8). Some support for the contention that multi-purpose shopping t r i p s are t y p i c -a l l y to suburban shopping centres (Claus and Hardwick, 1972, p. 76) derives from the f a c t that t r i p s to secondary centres were proportion-a l l y higher (38 per cent) where a multi-purpose shopping t r i p was in d i c a t e d . (v) Past Purchases. Most purchases were made at stores from which the respondents had obtained other c l o t h i n g goods i n the previous two year period (Figure 5.9). This r e f l e c t s the widely recognised importance of past purhcase FIGURE 5.9 TRIP DISTRIBUTION BY PAST PURCHASES ONE 198 2 TO 5 455 6 TO 10 156 OVER 10 153 0 10 20 30 40 50 « 1 1 1 i i Percentage of T r i p s TABLE 5.7 DISTRIBUTION OF SHOPPING TRIPS BY TRIP PURPOSE AND STORE TYPE Clothing purchase Other shopping Business S o c i a l Recreational TOTAL 65 ( 6%) 179 (18%) 80 ( 8%) 566 (56%) 115 (11%) 1005 Chi Square = 51.45 Degrees of Freedom = 16 S i g n i f i c a n t at .001 l e v e l HIGH PRICE MEDIUM PRICE „ m m T „ m r ^ r , . ™ , . . ™ ™ r , r r r ^ m ^ m , T „„„_,,.„,.„„, r,„ T 1^ T, T m„ BOUTIQUE DEPARTMENT BUDGET TOTAL SPECIALTY SPECIALTY * 51 ( 8%) 135 (21%) 62 (10%) 344 (53%) 56 ( 9%) 648 8 ( 3%) 27 (11%) 7 ( 3%) 160 (65%) 43 (18%) 245 2 ( 6 % ) 6 (18%) 4 (12%) 18 (53%) 4 (12%) 34 2 ( 8 % ) 0 ( 0 % ) 2 ( 8%) 18 (69%) 4 (15%) 26 2 ( 4%) 11 (21%) 5 (10%) 26 (50%) 8 (15%) 52 to TABLE 5.8 DISTRIBUTION OF SHOPPING TRIPS BY TRIP PURPOSE AND SHOPPING CENTRE Clothing Purchase 338 (53%) 206 (33%) 88 (14%) 632 Other Shopping 125 (51%) 92 (38%) 26 (11%) 243 Business 22 (67%) 5 (15%) 6 (18%) 33 S o c i a l 13 (65%) 5 (25%) 2 (10%) 20 Recreational 27 (64%) 13 (31%) 2 ( 5%) 42 To t a l 525 (54%) 321 (33%) 124 (13%) 970 Chi-Square = 12.41 Degrees of Freedom = Not s i g n i f i c a n t 8 129. experience and store l o y a l t y as factors i n f l u e n c i n g shopping decisions. Repeat purchasing patterns vary s i g n i f i c a n t l y between type of store, however, being most pronounced amongst department store patrons and l e a s t common amongst s p e c i a l t y store shoppers (Table 5.9). This pattern sug-gests that for most consumers shopping at s p e c i a l t y stores i s a r e l a t i v e l y infrequent p r a c t i c e , probably l i m i t e d to the rare s p e c i a l occasion pur-chase. S i g n i f i c a n t d i f f e r e n c e s again emerged i n the case of the r e l a t i o n -ship between repeat purchasing and shopping centre patronage (Table 5.10). Downtown and secondary centre shoppers were the more i n c l i n e d to make frequent repeat purchases, which l a r g e l y r e f l e c t s the influence of the major department store l o c a t i o n s . (vi) Conclusion. The preceding r e s u l t s i n d i c a t e s i g n i f i c a n t r e l a t i o n s h i p s between shopping t r i p v a r i a b l e s and r e t a i l patronage and l a r g e l y confirm the findings of previous studies (e.g., Bucklin, 1967). These r e s u l t s are i n s t r u c t i v e i n providing a comprehensive d e s c r i p t i o n of the shopping behaviour of the respondents. This serves as a basis for proceeding to the next section i n which attention turns from data d e s c r i p t i o n to a consideration of the main a n a l y t i c a l objectives of t h i s study. E. ANALYSIS OF SHOPPING BEHAVIOUR (i)- The A n a l y t i c a l Framework. The c e n t r a l objective of t h i s research i s to assess the e f f i c a c y of three d i f f e r e n t sets of measures, both s i n g l y and i n composite, as TABLE 5.9 DISTRIBUTION OF SHOPPING TRIPS BY PAST PURCHASES AND STORE TYPES One 2 to 5 6 to 10 Over 10 HIGH PRICE SPECIALTY 16 ( 8%) 37 ( 7%) 10 ( 6%) 2 ( 1%) MEDIUM PRICE SPECIALTY 58 (29%) 99 (20%) 19 (12%) 3 ( 2%) BOUTIQUE 21 (11%) 43 ( 9%) 4 ( 3%) 11 ( 7%) DEPARTMENT 71 (36%) 259 (52%) 110 (70%) 123 (80%) BUDGET 31 (16%) 57 (12%) 13 ( 8%) 14 ( 9%) TOTAL 197 495 156 153 TOTAL 65 ( 6%) 179 (18%) 79 ( 8%) 563 (56%) • 115 (11%) 1001 Chi Square = 98.71 Degrees of Freedom = 12 S i g n i f i c a n t at .001 l e v e l CO o TABLE 5.10 DISTRIBUTION OF SHOPPING TRIPS BY PAST PURCHASES AND SHOPPING CENTRE DOWNTOWN TOTAL CENTRE CENTRE One 95 (53%) 51 (29%) 32 (18%) 178 2 to 5 262 (54%) 147 (30%) 73 (15%) 482 6 to 10 78 (51%) 59 (38%) 17 (11%) 154 Over 10 88 (58%) 62 (41%) 1 ( 1%) 151 T o t a l 523 (54%) 319 (33%) 123 (13%) 965 Chi-Square = 30.71 Degrees of Freedom = 6 S i g n i f i c a n t at .001 l e v e l 132. discriminators between r e t a i l patronage groups. To r e c a p i t u l a t e , the p r e d i c t i v e measures comprised the l o c a t i o n a l , b i o g r a p h i c a l and d i s p o s i -t i o n a l c h a r a c t e r i s t i c s of the consumer. The c r i t e r i o n v a r i a b l e was shopping behaviour categorized e s s e n t i a l l y i n terms of the type of store and type of shopping centre patronized on a given shopping t r i p . To a l i m i t e d extent i t was p o s s i b l e to categorize the c r i t e r i o n v a r i a b l e at a disaggregate l e v e l r e f l e c t i n g the s p e c i f i c department stores and shopping centres selected by the consumer. The questionnaire data provided a basis f o r e s t a b l i s h i n g the s t a b i l i t y of c r i t e r i o n - p r e d i c t o r r e l a t i o n s h i p s over a sequence of purchases, which i t was f e l t c o n s t i t u -ted a stronger b a s i s f o r inference than i f the analysis was r e s t r i c t e d to a s i n g l e shopping d e c i s i o n . I t i s u s e f u l to conceive of the analysis phase i n cubic form as i l l u s t r a t e d i n Figure 5.10. The dimensions of the cube are the c r i t e r i a , p r e d i c t o r s and purchasing t r i p s f o r which data were a v a i l a b l e . Each of the dimensions has four l a b e l l e d segments corresponding to the categories of behavioural c r i t e r i a , the sets of p r e d i c t i v e measures and the sequence of past purchases. The cube i s therefore d i v i s i b l e i n t o 64 ( 4 x 4 x 4 ) separate c e l l s corresponding to 64 p o s s i b l e c r i t e r i o n - p r e d i c t o r analyses of which 16 are l a b e l l e d i n the diagram. No attempt i s made i n t h i s chapter to provide an exhaustive d e s c r i p -t i o n of the r e s u l t s obtained from each of these 64 a n a l y t i c a l combinations. Rather, examples of the analysis of each c r i t e r i o n - p r e d i c t o r combination w i l l be presented followed by a summary of the o v e r a l l r e s u l t s which w i l l r e f l e c t analyses not o u t l i n e d i n d e t a i l . I t i s hoped i n t h i s way to produce a report which i s both comprehensive i n i t s coverage and 133. FIGURE 5.10 A DIAGRAMATIC REPRESENTATION OF THE ANALYTICAL FRAMEWORK C R I T E R I A C r i t e r i a C . = store type s t C = centre type C = s p e c i f i c stores s C^ = s p e c i f i c centres Purchases T^ = purchase T^ = purchase T^ = purchase T n = purchase Pr e d i c t o r s P^ = l o c a t i o n a l v a r i a b l e s = b i o g r a p h i c a l v a r i a b l e s P = d i s p o s i t i o n a l v a r i a b l e s d P = combined v a r i a b l e s c t r i p 1 t r i p 2 t r i p 3 t r i p (norm) 134. concise i n i t s d e t a i l . ( i i ) S t a t i s t i c a l Procedures. Previous reference has been made to the appropriateness o f d i s c r i -minant analysis to s i t u a t i o n s where a c r i t e r i o n v a r i a b l e i s nominal and the p r e d i c t o r s are i n t e r v a l or r a t i o - s c a l e d (see Chapter 3). These con-d i t i o n s held i n the p r e t e s t , where one objec t i v e was to assess the pre-d i c t i v e v a l i d i t y of the preliminary d i s p o s i t i o n a l s c a l e s , and they apply equally to the c r i t e r i o n and p r e d i c t o r measures obtained i n the major data c o l l e c t i o n phase. The following purposes of discriminant analysis were seen to corres-pond with the research o b j e c t i v e s : to determine whether or not the dif f e r e n c e s i n p r e d i c t o r score p r o f i l e s f o r two or more c r i t e r i o n groups are s t a t i s t i c a l l y s i g n i f i c a n t ; to maximize the d i s c r i m i n a t i o n among groups by d e r i v i n g l i n e a r combinations of p r e d i c t o r v a r i a b l e s ; to assign o r i g i n a l subjects to groups based on derived scores and to assess the accuracy of d i s c r i m i n a t i o n by comparing p r e d i c t e d and actu a l group membership. In geometrical terms, i t i s po s s i b l e to conceive of each of a set of subjects being represented by a p o i n t l o c a t i o n i n a k-space based upon a k-variable score p r o f i l e . In ad d i t i o n , the c r i t e r i o n group mem-bership of each subject i s known. The accuracy with which group member-ship can be pre d i c t e d from score p r o f i l e s i s then dependent upon the existence of d i s t i n c t c l u s t e r s within the k-space comprising members of the same c r i t e r i o n group. Where "same group" members are t i g h t l y c l u s -tered and c l e a r l y separated from other group c l u s t e r s , h i g h l y accurate 135. p r e d i c t i o n s are p o s s i b l e . On the other hand, where there i s a near ran-dom d i s t r i b u t i o n of 'same group' members and hence extensive group over-lap, low p r e d i c t a b i l i t y r e s u l t s . In order to maximize d i s c r i m i n a t i o n between groups, weighted combi-nations of p r e d i c t o r v a r i a b l e s are defined which normally assume the following l i n e a r form: Y = a,X, + a„X„ + ... a X where: Y = discriminant scores. X I 2 2 n n X WX_...X = raw scores. 1 2 n a l ' a 2 " * * a n = var^a^-'-'-e weights. The same set of weights i s applied to the scores of each person i n each group, r e s u l t i n g i n a new score f o r each person (Y) which combines the information from X.,X„...X regarding d i s c r i m i n a t i o n of the groups. The 1 2 n optimization r u l e f o r determining the v a r i a b l e weights i s based upon the maximization of between to within-group variances (Fisher, 1936) . Where there are more than two groups, i t i s p o s s i b l e to derive more than one discriminant function, the number being determined by the number of v a r i a b l e s or the number of groups minus one, whichever i s l e s s . The f i r s t discriminant function i s that l i n e a r combination of the v a r i a b l e s which maximizes the r a t i o of between to -.within-group variance. A d d i t i o n a l orthogonal functions are derived which successively account f o r smaller proportions of the variance, and normally very l i t t l e variance i s e x p l a i -ned beyond the second or t h i r d discriminants. There are obvious s i m i l a r i t i e s between discriminant analysis and f a c t o r analysis and, as Nunnally (1967, p. 394) notes, the major computa-t i o n a l step i n d e r i v i n g m u l t i p l e discriminant functions i s to perform a 136. p r i n c i p a l - a x e s f a c t o r a n a l y s i s . But, i n contrast to f a c t o r a n a l y s i s , computations are not based upon the c o r r e l a t i o n s among the v a r i a b l e s , but upon the sums of squared deviations about the means within groups. In the step-wise form of multiple discriminant a n a l y s i s , employed i n t h i s study, the v a r i a b l e s f o r i n c l u s i o n i n the discriminant function are determined i n an i t e r a t i v e manner such that at each step the v a r i a b l e included i s that f o r which group means d i f f e r most s i g n i f i c a n t l y as assessed by un i v a r i a t e analysis of variance. Where there i s c o v a r i a t i o n amongst p r e d i c t o r s the i n c l u s i o n of a l l v a r i a b l e s i n the discriminant function i s often precluded. A predefined F p r o b a b i l i t y tolerance l e v e l (usually .05) serves as the basis f o r deciding whether a v a r i a b l e i s to be added to (or deleted from) the discriminant function. Once a l l v a r i a -bles with F r a t i o s below the tolerance l e v e l have been entered, the l i n e a r discriminant functions f o r the various groups are c a l c u l a t e d , and F r a t i o s between p a i r s o f groups are computed based on discriminant scores to determine how e f f e c t i v e l y the discriminant functions d i s t i n -guish the groups. An a d d i t i o n a l basis f o r assessing the accuracy of group separation i s by the c l a s s i f i c a t i o n of subjects to groups based upon the distances separating subjects and group centroids i n the space defined by the discriminant functions. The p o s t e r i o r p r o b a b i l i t y of membership i n each of the groups i s also c a l c u l a t e d . The degree of coincidence between pred i c t e d and actual group membership i s a fu r t h e r i n d i c a t i o n of the e f f i c a c y of the o r i g i n a l p r e d i c t o r s as di s c r i m i n a t o r s between the c r i -t e r i o n groups. For a d e t a i l e d discussion o f discriminant a n a l y s i s , the reader i s 137. r e f e r r e d to such m u l t i v a r i a t e s t a t i s t i c s texts as Cooley and Lohnes (19 71) and O v e r a l l and K l e t t (1973). ( i i i ) Analyses of Store and Shopping Centre Patronage. The r e s u l t s of sixteen of the s i x t y four p o s s i b l e analyses shown i n Figure 5.10 are reported i n t h i s s e c t i o n . Attention at t h i s p o i n t i s r e s t r i c t e d to two of the four c r i t e r i o n v a r i a b l e s p r e v i o u s l y men-tioned - type of store and type of shopping centre - and to two of the four shopping t r i p s - the most recent and most frequent. In large measure, the r e s u l t s obtained using data r e l a t i n g to the two less recent t r i p s coincided with the r e s u l t s presented, and, furthermore, they are represented i n the analysis summary which appears i n a l a t e r section of t h i s chapter. This sixteen analyses represent d i f f e r e n t combinations of c r i t e r i o n v a r i a b l e , shopping t r i p and p r e d i c t i v e model, and are presented i n the following order: ANALYSIS CRITERION TRIP PREDICTIVE MODEL 1 Store Type Most Recent Locational 2 Store Type Most Recent Biographical 3 Store Type Most Recent D i s p o s i t i o n a l 4 Store Type Most Recent Composite 5 Store Type Most Frequent Locational 6 Store Type Most Frequent Biographical 7 Store Type Most Frequent D i s p o s i t i o n a l 8 Store Type Most Frequent Composite 138. ANALYSIS CRITERION TRIP PREDICTIVE MODEL 9 Shopping Centre Most Recent Locational 10 Shopping Centre Most Recent Biographical 11 Shopping Centre Most Recent D i s p o s i t i o n a l 12 Shopping Centre Most Recent Composite 13 Shopping Centre Most Frequent Locational 14 Shopping Centre Most Frequent Biographical 15 Shopping Centre Most Frequent D i s p o s i t i o n a l 16 • Shopping Centre Most Frequent Composite Accompanying each analysis i s a table summarizing the r e s u l t s using a standard format to ease the comparison of d i f f e r e n t sets of r e s u l t s . Included i n the tables are: the mean scores f o r each group on the pre-d i c t o r v a r i a b l e s , the discriminant c o e f f i c i e n t s f o r each group, the f i n a l between group s i g n i f i c a n c e l e v e l s and the percentage of cases c l a -s s i f i e d i n t o each group on the basis of discriminant scores. The i n t e r -p r e t a t i o n of these sets of fig u r e s i s explained i n reporting the r e s u l t s of Analysis 1. ANALYSIS 1. CRITERION: Store Type (5 groups) PURCHASE TRIP: Most Recent PREDICTIVE MODEL: Locational V a r i a b l e s - estimated distance (origin-destination) - a c t u a l distance (origin-destination) - distance to Downtown Vancouver - distance to the nearest secondary centre - distance to the nearest t e r t i a r y centre - time distance to Downtown Vancouver - s i z e of shopping area patronized - s i z e of the nearest secondary centre - s i z e of the nearest t e r t i a r y centre 139. The mean scores f o r each of the f i v e store patronage groups on each of the nine l o c a t i o n a l v a r i a b l e s included i n the analysis are shown i n the f i r s t s e c t i o n of Table 5.11. Also l i s t e d are the i n i t i a l F probabi-l i t i e s i n d i c a t i n g the s i g n i f i c a n c e of the d i f f e r e n c e s i n group means. Between group dif f e r e n c e s were s i g n i f i c a n t at .05 l e v e l i n the case of only one v a r i a b l e - size of the nearest tertiary centre. I n i t i a l l y therefore t h i s was the only v a r i a b l e to s a t i s f y conditions f o r i n c l u s i o n i n the discriminant function. Two other v a r i a b l e s - distance to the nearest secondary centre and size of shopping area patronized - were subsequently included however. This demonstrates that c o n t r o l l i n g f o r the v a r i a b l e ( s ) previously entered had the e f f e c t of i n c r e a s i n g the s i g n i f i c a n c e of these two a d d i t i o n a l v a r i a b l e s such that t h e i r F proba-b i l i t i e s f e l l below the c o n t r o l value of .05. The F p r o b a b i l i t i e s f o r entry of each of the three v a r i a b l e s i s shown i n the second part of Table 5.11 together with the discriminant c o e f f i c i e n t s f or each group. These l a t t e r f igures represent the weights applied to the raw scores of members i n the respective groups to deter-mine t h e i r discriminant scores. Consistent with the s t a t i s t i c a l model, the weights are c a l c u l a t e d so as to maximize between to within group variance. Their magnitude r e f l e c t s the scale p r o p e r t i e s of the r e s -p e c t i v e v a r i a b l e s and hence the figures are to be i n t e r p r e t e d row-wise. Scanning across the rows i n the t a b l e reveals that the pattern of v a r i a -t i o n i n the group c o e f f i c i e n t s commonly, although not i n v a r i a b l y , c o r r e s -ponds to the pattern of group means. The v a r i a t i o n s i n group means and discriminant c o e f f i c i e n t s show the following major group d i f f e r e n c e s : high-price s p e c i a l t y store shoppers TABLE 5.11 ANALYSIS 1: SUMMARY OF RESULTS i HIGH PRICE SPECIALTY MEDIUM PRICE SPECIALTY BOUTIQUE DEPARTMENT BUDGET SIGNIFICANCE LEVEL GROUP MEANS Estimated d i s t a n c e 3 3.2 4.3 3.4 3.9 3.0 .418 Actual distance' 3 103.9 126.1 105.1 127.4 97.2 .326 Distance downtown 143.5 158.8 149.4 156.6 143.7 .867 Distance to secondary centre 86.4 113.9 113.6 105.5 124.9 .085 Distance to t e r t i a r y centre' 3 29.6 30.2 28.3 29.9 25.8 .931 Time distance downtown0 15.6 16.7 16.1 16.7 15.6 .879 Size of shopping area^ 3.4 3.2 3.5 3.5 3.4 .056 Size of secondary centre 493.1 495.6 483.6 466.8 457.7 .272 Size of t e r t i a r y c e n t r e e 222.1 217.5 256.9 292.2 282.4 .046 DISCRIMINANT COEFFICIENTS Distance to secondary centre .016 .028 .026 .024 .031 .044 Size of shopping area 6.226 5.654 6.324 6.363 6.151 .024 Size of t e r t i a r y centre .010 .011 .012 .013 .013 .046 FINAL BETWEEN GROUP SIGNIFICANCE LEVELS Medium p r i c e s p e c i a l t y .055 Boutique .202 .127 Department .061 .000 .730 Budget .023 .095 .718 .286 PERCENTAGE OF CASES CLASSIFIED i INTO EACH GROUP High p r i c e s p e c i a l t y 44 20 20 16 0 Medium p r i c e s p e c i a l t y 22 44 16 9 9 Boutique 19 16 23 29 13 Department 27 16 13 23 22 Budget 21 34 14 14 17 Figures i n miles ^Figures i n map millimeters c . Figures m minutes d e Thousands of square fe e t Clog transformed) Thousands of square feet 141. were distin g u i s h e d by t h e i r proximity to secondary centres; medium p r i c e s p e c i a l t y store patrons by t h e i r tendency to shop at smaller shopping centres and t h e i r proximity to small t e r t i a r y centres; department store shoppers by t h e i r i n e v i t a b l e a t t r a c t i o n to l a r g e r centres; and budget store patrons by t h e i r distance from secondary centres and t h e i r p r o x i -mity to l a r g e r t e r t i a r y centres. The s i g n i f i c a n c e of between group d i f f e r e n c e s based on discriminant scores i s shown by the set of F p r o b a b i l i t i e s i n the t h i r d section of Table 5.11. E n t r i e s below .05 are considered s i g n i f i c a n t since they demonstrate a l e s s than 5 per cent p r o b a b i l i t y of observed group d i f f e r -ences being a t t r i b u t a b l e to chance v a r i a t i o n s . Only two of the ten e n t r i e s are s i g n i f i c a n t suggesting the l i m i t e d power of the l o c a t i o n a l v a r i a b l e s to account f o r d i f f e r e n c e s i n the respondent's choice of type of store. The only s i g n i f i c a n t d i s t i n c t i o n s to emerge are those between high-price s p e c i a l t y and budget store shoppers and between medium-price s p e c i a l t y and department store shoppers. As previously mentioned (p. 136), a f u r t h e r step i n discriminant analysis i s to c l a s s i f y membership i n the o r i g i n a l groups based on the proximity of a subject's discriminant score to the group centroids. The accuracy of c l a s s i f i c a t i o n serves as an a d d i t i o n a l index of the d i s c r i -minatory power of the s e l e c t e d v a r i a b l e s . The f i n a l s e ction of Table 5.11 shows the c l a s s i f i c a t i o n matrix obtained i n Analysis 1. The p r i n -c i p a l diagonal e n t r i e s of the matrix show the percentage of cases cor-r e c t l y c l a s s i f i e d i n each group, and the o f f - d i a g o n a l e n t r i e s the percent-age of m i s c l a s s i f i e d cases. In the case of p e r f e c t l y accurate c l a s s i f i -cation a l l the diagonal values would be 100 per cent and a l l the o f f -142. diagonal values zero. The large percentage of m i s c l a s s i f i e d cases i n Analysis 1 confirms the f a i l u r e of the l o c a t i o n a l measures to d i s t i n -guish accurately between the various c r i t e r i o n groups. Only i n the cases o f the high-price s p e c i a l t y and medium-price s p e c i a l t y groups d i d the percentage of c o r r e c t assignments exceed 40 per cent. ANALYSIS 2. CRITERION: Store Type TRIP: Most Recent PREDICTIVE MODEL: Biographical Variables - age - sex - m a r i t a l status - number of c h i l d r e n at home - education - occupation of head of household Table 5.12 shows that s i g n i f i c a n t d i f f e r e n c e s i n group means occurred fo r f i v e of the seven b i o g r a p h i c a l v a r i a b l e s , the exceptions being the sex of the respondent and the number of children living at home. Of the f i v e v a r i a b l e s , four were included i n the discriminant functions. Edu-cation, though i n i t i a l l y s i g n i f i c a n t , was not entered due to i t s c o v a r i a -t i o n with v a r i a b l e s previously included. C o n t r o l l i n g f o r p r e v i o u s l y entered v a r i a b l e s had the reverse e f f e c t i n the case of sex of the respondent, which, though i n i t i a l l y not s i g n i f i c a n t , subsequently s a t i s -f i e d conditions f o r entry i n the discriminant equations. Main d i s t i n c t i o n s i n group p r o f i l e s were as follows: high-price s p e c i a l t y store shoppers tended to be older, female, married, more a f f l u -ent, and to come from households where the head t y p i c a l l y had a managerial or p r o f e s s i o n a l occupation; boutique shoppers were commonly younger and TABLE 5.12 ANALYSIS 2: SUMMARY OF RESULTS HIGH PRICE SPECIALTY GROUP MEANS Age Sex M a r i t a l status Number of c h i l d r e n Income Education Occupation DISCRIMINANT COEFFICIENTS 40.8 1.7 1 1 239 4 5 Age .102 Sex 7.200 M a r i t a l status 6.158 Income .018 Occupation 2.646 FINAL BETWEEN GROUP SIGNIFICANCE LEVELS Medium p r i c e s p e c i a l t y Boutique Department Budget .000 .000 .000 .000 MEDIUM PRICE SPECIALTY 36.2 1.5 1.4 0.7 115.6 3.8 4.0 .100 6.178 4.739 .0006 2.415 .001 .010 .331 BOUTIQUE DEPARTMENT BUDGET PERCENTAGE OF CASES CLASSIFIED INTO EACH GROUP High p r i c e s p e c i a l t y Medium p r i c e s p e c i a l t y Boutique Department Budget 72 16 16 11 14 16 23_ 13 14 21 23.4 1.6 1.3 0.6 129.7 3.4 3.8 .044 6.791 4.410 .006 2.144 .000 .001 4 28 52 23 21 36.3 1.5 1.6 0.9 109.8 3.3 3.4 .094 5.984 5.331 .001 2.016 .968 4 12 19 23 31 37.7 1.4 1.6 1.0 102.5 3.2 3.5 .101 5.785 5.296 - .0005 2.093 SIGNIFICANCE LEVEL .000 .121 .000 .453 .000 .002 .000 .000 .038 .033 ,000 ,001 4 20 0 29 14 144. s i n g l e ; department store patrons had the lowest mean job status; and budget store patrons were more l i k e l y to be male and the l e a s t a f f l u e n t . Between group differences i n discriminant scores based on biographi-c a l v a r i a b l e s were s i g n i f i c a n t i n eight out of ten cases. The high-price s p e c i a l t y group was the most c l e a r l y d i s t i n g u i s h e d and the budget store group l e a s t so. The c l a s s i f i c a t i o n matrix confirms t h i s conclusion, whereas 72 per cent of high-price s p e c i a l t y shoppers were c o r r e c t l y assigned, the same was true of only 14 per cent of the budget store group. A reasonable l e v e l of p r e d i c t i v e accuracy (51 per cent) was also achieved i n the case of the boutique group. ANALYSIS 3. CRITERION: Store Type TRIP: Most Recent PREDICTIVE MODEL: D i s p o s i t i o n a l Variables - status scale - convenience scale - fashion scale - q u a l i t y scale Mean scores on a l l f i v e of the d i s p o s i t i o n a l scales d i f f e r e d s i g n i -f i c a n t l y between groups (Table 5.13), but the degree of c o v a r i a t i o n amongst the scales was such that only three were included i n the d i s c r i -minant function: status, fashion and quality. The p r i n c i p a l d i s t i n c t i o n s i n the store group p r o f i l e s were as follows: high-price s p e c i a l t y shoppers were the most status and q u a l i t y conscious and the l e a s t fashion conscious; not s u r p r i s i n g l y , boutique patrons showed the greatest fashion o r i e n t a t i o n ; while budget store shoppers were dis t i n g u i s h e d by t h e i r low scores on the status and q u a l i t y TABLE 5.13 ANALYSIS 3: SUMMARY OF RESULTS GROUP MEANS Status Convenience Fashion P r i c e Q u a l i t y DISCRIMINANT COEFFICIENTS Status Fashion Qua l i t y HIGH PRICE SPECIALTY 42.4 32.2 33.0 28.1 49.2 .661 .469 1.678 FINAL BETWEEN GROUP SIGNIFICANCE LEVELS Medium p r i c e s p e c i a l t y Boutique Department Budget .000 .000 .000 .000 MEDIUM PRICE SPECIALTY 36.0 34.1 36.2 31.4 44.7 BOUTIQUE DEPARTMENT BUDGET PERCENTAGE OF CASES CLASSIFIED INTO EACH GROUP High p r i c e s p e c i a l t y Medium p r i c e s p e c i a l t y Boutique Department Budget 80 23 10 9 3 .429 .599 1.581 .000 .001 .000 4 19 16 14 10 37.0 31.0 42.8 33.8 42.8 .424 .733 1.506 .000 .000 8 27 58_ 23 7 33.2 34.5 34.6 33.8 42.4 .367 .584 1.515 .001 8 12 6 22 17 29.8 37.7 31.9 37.0 39.1 .306 .548 1.407 SIGNIFICANCE LEVEL .000 .002 .000 .000 .000 0 19 10 33 62 ,000 .000 .000 146. s c a l e s . Medium-price s p e c i a l t y and department store shoppers f a i l e d to demonstrate extreme scores on any of the three scales included i n the discriminant function. The table of f i n a l between group F p r o b a b i l i t i e s shows that a l l e n t r i e s are highly s i g n i f i c a n t , demonstrating the effectiveness of the d i s p o s i t i o n a l measures as d i s c r i m i n a t i n g v a r i a b l e s . The c l a s s i f i c a t i o n matrix- reveals that the improvement i n p r e d i c t i v e accuracy over previous analyses was small however, mainly because the two l a r g e s t groups, comprising medium-price s p e c i a l t y and department store shoppers, e x h i b i -ted the greatest within-group heterogeneity. The d i s t i n c t d i s p o s i t i o n a l p r o f i l e s of the high-price s p e c i a l t y , boutique and budget store groups i s borne out by the percentage of correc t assignments: 80 per cent, 58 per cent and 62 per cent r e s p e c t i v e l y . ANALYSIS 4. CRITERION: Store Type TRIP: Most Recent PREDICTIVE MODEL: Combined v a r i a b l e s - l o c a t i o n a l v a r i a b l e s (see Analysis 1) - b i o g r a p h i c a l v a r i a b l e s (see Analysis 2) - d i s p o s i t i o n a l v a r i a b l e s (see Analysis 3) - number of past purchases This analysis was performed to determine the increase i n d i s c r i m i n a -tory power achieved by combining the sets of measures treated s i n g l y i n previous analyses. An a d d i t i o n a l v a r i a b l e was included i n the form of the number of past purchases made at a s p e c i f i e d store. The group means for the l o c a t i o n a l , b i o g r a p h i c a l and d i s p o s i t i o n a l v a r i a b l e s were of course i d e n t i c a l to those previously reported and so are not tabulated TABLE 5.14 ANALYSIS 4: SUMMARY OF RESULTS DICSRIMINANT COEFFICIENTS Past purchases Size of shopping area Size of t e r t i a r y centre Income Occupation Status Fashion Quality HIGH PRICE SPECIALTY 1.242 6.501 .009 - .006 1.743 .560 .395 1.684 FINAL BETWEEN GROUP SIGNIFICANCE LEVELS Medium p r i c e s p e c i a l t y Boutique Department Budget .000 .000 .000 .000 MEDIUM PRICE SPECIALTY 1.143 5.942 .010 - .021 1.538 .385 .522 1.621 BOUTIQUE DEPARTMENT BUDGET PERCENTAGE OF CASES CLASSIFIED INTO EACH GROUP High p r i c e s p e c i a l t y Medium p r i c e s p e c i a l t y Boutique Department Budget 68 16 10 4 3 .000 .000 .000 24 34 19 11 24 432 435 012 017 295 357 656 550 .000 .000 8 23 55 17 0 2. 6. 337 655 013 021 204 279 497 575 .005 0 0 6 43 24 1.822 6.584 .012 - .019 1.340 .230 .465 1.457 SIGNIFICANCE LEVEL .000 .023 .012 .000 .048 .000 .000 .002 0 20 10 25 48 148. again. Table 5.14 shows that eight of the o r i g i n a l twenty two v a r i a b l e s were entered i n t o the discriminant functions. Most notable i s the f a c t that three d i s p o s i t i o n a l scales - status, fashion and quality - were amongst the f i r s t f i v e v a r i a b l e s entered, underlining the comparative power of these measures as discriminators between store patronage groups. Also noteworthy i s the s i g n i f i c a n c e of past purchases; the magnitude of the discriminant c o e f f i c i e n t s f o r t h i s v a r i a b l e shows that whereas depart-ment store patrons were l i k e l y to e x h i b i t frequent repeat purchases, the converse held i n the case of both s p e c i a l t y store groups. This r e l a t i o n -ship confirms the a s s o c i a t i o n noted p r e v i o u s l y (see Table 5.9). Highly s i g n i f i c a n t between group F p r o b a b i l i t i e s were obtained using the combined measures and the group c l a s s i f i c a t i o n matrix reveals, as expected, that the o v e r a l l p r e d i c t i v e accuracy was higher than i n any of the three previous analyses. This i s mainly a t t r i b u t a b l e to the more accurate c l a s s i f i c a t i o n of department store shoppers (43 per cent c o r r e c t l y assigned) which i n turn i s l a r g e l y accounted f o r by the i n c l u -s i o n of two l o c a t i o n a l v a r i a b l e s and the past purchase measure i n the discriminant functions. * ANALYSIS 5. CRITERION: Store Type TRIP: Most Frequent PREDICTIVE MODEL: Loca t i o n a l Variables (as f o r Analysis 1 excluding estimated distance t r a v e l l e d ) * Results of the analyses of store patronage based on 'most frequent 1 t r i p information (Analyses 5 through 8) do not include tabulations of group means and i n i t i a l F p r o b a b i l i t i e s since these figures v a r i e d l i t t l e from the corresponding s t a t i s t i c s reported i n Analyses 1-4. TABLE 5.15 ANALYSIS 5: SUMMARY OF RESULTS DISCRIMINANT COEFFICIENTS Size of shopping area HIGH PRICE SPECIALTY 6.148 MEDIUM PRICE SPECIALTY 6.266 BOUTIQUE DEPARTMENT BUDGET 7.402 7.075 6.968 SIGNIFICANCE LEVEL .0040 FINAL BETWEEN GROUP SIGNIFICANCE LEVELS Medium p r i c e s p e c i a l t y Boutique Department Budget .781 .010 .014 .069 .006 .003 .057 .335 ,306 ,709 PERCENTAGE OF CASES CLASSIFIED INTO EACH GROUP High p r i c e s p e c i a l t y Medium p r i c e s p e c i a l t y Boutique Department Budget 73 55 26 41 37 0 16 0 .5 0 27 29 74 56 59 0 0 0 2_ 0 0 0 0 0 4 150. In t h i s case only one v a r i a b l e - the size, of the shopping area patron-ized - met the conditions for entry into the discriminant function. The discriminant c o e f f i c i e n t s (Table 5.15) show that boutique shoppers tended to patronize l a r g e r shopping areas, whereas the reverse was true f o r spe-c i a l t y store patrons and p a r t i c u l a r l y high-price s p e c i a l t y shoppers. Not s u r p r i s i n g l y , the s i n g l e v a r i a b l e discriminant function proved to be a very weak dis c r i m i n a t o r . Only four of the ten between group F r a t i o s were s i g n i f i c a n t , and there i s a correspondingly low coincidence between pre d i c t e d and actual group membership. These r e s u l t s confirm the conclusions drawn from Analysis 1 i n terms o f the very l i m i t e d power of the l o c a t i o n a l measures to discriminate between store patronage groups. ANALYSIS 6. CRITERION: Store Type TRIP: Most Frequent PREDICTIVE MODEL: Biographical Variables (as f o r Analysis 2) Consistent with the r e s u l t s of Analysis 2, income, age and occupa-tion of head of household again emerged as the main d i s c r i m i n a t i n g v a r i a -b l e s . The discriminant c o e f f i c i e n t s (Table 5.16) confirm the d i s t i n c t i o n s i n the group b i o g r a p h i c a l p r o f i l e s noted previously with the a d d i t i o n a l i n d i c a t i o n that department store shoppers were the most l i k e l y to have ch i l d r e n l i v i n g at home while t h i s was l e a s t l i k e l y i n the case of high-p r i c e s p e c i a l t y shoppers. " Nine of the ten between group F r a t i o s proved to be s i g n i f i c a n t at the .05 l e v e l , the exception being between the budget and department store groups. A r e l a t i v e l y high l e v e l of p r e d i c t i v e accuracy was achieved i n TABLE 5.16 ANALYSIS 6: SUMMARY OF RESULTS DISCRIMINANT COEFFICIENTS Age Number of c h i l d r e n Income Occupation HIGH PRICE SPECIALTY .196 - .328 .031 2.257 MEDIUM PRICE SPECIALTY .181 .225 .005 2.580 BOUTIQUE DEPARTMENT BUDGET .091 - .137 .008 2.322 .153 .353 .004 1.931 .168 .165 - .001 1.900 SIGNIFICANCE LEVEL .000 .011 .000 .001 FINAL BETWEEN GROUP SIGNIFICANCE LEVELS Medium p r i c e s p e c i a l t y Boutique Department Budget .000 .000 .000 .000 .000 .000 .000 .000 .000 .088 PERCENTAGE OF CASES CLASSIFIED INTO EACH GROUP High p r i c e s p e c i a l t y Medium p r i c e s p e c i a l t y Boutique Department Budget 67 13 10 6 4 20 45 0 13 4 7 13 74 21 22 0 6 5 31. 19 7 23 11 30 52 152. c l a s s i f y i n g group membership on the basis of discriminant scores; f o r a l l f i v e groups the modal category appears on the p r i n c i p a l diagonal and, as i n Analysis 2, m i s c l a s s i f i c a t i o n was lowest f o r the high-price s p e c i a l -ty and boutique groups. ANALYSIS 7. CRITERION: Store Type TRIP: Most Frequent PREDICTIVE MODEL: D i s p o s i t i o n a l Scales (as f o r Analysis 3) The status, fashion and quality scales emerged as the d i s t i n g u i s h -in g v a r i a b l e s as they d i d i n Analysis 3, and the group p r o f i l e d i f f e r e n c e s r e f l e c t e d i n the discriminant c o e f f i c i e n t s (Table 5.17) were also con-s i s t e n t with the previous r e s u l t s . A l l ten of the between group F r a t i o s were s i g n i f i c a n t , and the c l a s s i f i c a t i o n matrix r e f l e c t s these d i s t i n c -t i o n s i n the r e l a t i v e l y accurate p r e d i c t i o n of membership i n the high-p r i c e s p e c i a l t y , boutique and budget store groups. I t i s to be noted however that the o v e r a l l p r e d i c t i v e accuracy was i n t h i s case below that achieved on the basis of the bi o g r a p h i c a l measures, demonstrating the a b i l i t y of these l a t t e r v a r i a b l e s to d i s t i n g u i s h medium-price s p e c i a l t y and department store shoppers more p r e c i s e l y . ANALYSIS 8. CRITERION: Store Type TRIP: Most Frequent PREDICTIVE MODEL: Combined Variables (as f o r Analysis 4 excluding estimated distance travelled and past purchases) TABLE 5.17 ANALYSIS 7: SUMMARY OF RESULTS DISCRIMINANT COEFFICIENTS Status Fashion Quality HIGH PRICE SPECIALTY .524 .351 1.615 MEDIUM PRICE SPECIALTY .388 .525 1.508 BOUTIQUE DEPARTMENT BUDGET .299 .752 1.466 .269 .571 1.437 .208 .529 1.326 SIGNIFICANCE LEVEL .000 .000 .001 FINAL BETWEEN GROUP SIGNIFICANCE LEVELS Medium p r i c e s p e t i a l t y Boutique Department Budget .000 .000 .000 .000 .000 .000 ,000 .000 .000 .000 PERCENTAGE OF CASES CLASSIFIED INTO EACH GROUP High p r i c e s p e c i a l t y Medium p r i c e s p e t i a l t y Boutique Department Budget 60 19 11 10 7 33 29_ 11 12 0 0 16 63 23 7 7 16 16 24 19 0 19 0 31 67 154. Of the seven v a r i a b l e s entered i n t o the discriminant function (Table 5.18), f i v e were the same as those included i n Analysis 4, and, i n both cases, income, status orientation and fashion orientation were the p r i -mary d i s c r i m i n a t o r s . Price orientation also s a t i s f i e d the conditions f o r entry, r e f l e c t i n g the d i s t i n c t l y higher, scores of the budget store group on t h i s s c a l e . The i n c l u s i o n of only one l o c a t i o n a l v a r i a b l e -size of shopping area patronized - reemphasises the l i m i t e d power of the l o c a t i o n a l measures to discriminate between the f i v e store patronage groups. The table of F p r o b a b i l i t i e s and the c l a s s i f i c a t i o n matrix demon-st r a t e again that the most e f f e c t i v e between group d i s c r i m i n a t i o n was achieved when the various sets of measures were combined. Between group di f f e r e n c e s i n discriminant scores are a l l highly s i g n i f i c a n t and 46 per cent of the cases were c o r r e c t l y c l a s s i f i e d . Membership i n the high p r i c e specialty, boutique and budget groups was again the: most accurately predicted. An o v e r a l l comparison of the r e s u l t s of Analyses 1 to 4 with those obtained i n Analyses 5 to 8 demonstrates consistency i n the discriminatory powers of the p r e d i c t i v e models and i n the major d i s c r i m i n a t i n g v a r i a b l e s . I t i s c l e a r that the b i o g r a p h i c a l and d i s p o s i t i o n a l measures are f a r superior to the l o c a t i o n a l v a r i a b l e s which i s as might be a n t i c i p a t e d at the disaggregate l e v e l of choice between types of store. There appears to be an equivalence i n the p r e d i c t i v e accuracy achieved on the basis of the b i o g r a p h i c a l and d i s p o s i t i o n a l models. There i s i n general an improvement i n p r e d i c t i v e accuracy f o r analyses based on the 'most frequent' t r i p data confirming the pr o p o s i t i o n that store group p r o f i l e s TABLE 5.18 ANALYSIS 8: SUMMARY OF RESULTS DISCRIMINANT COEFFICIENTS Size of shopping area Number of c h i l d r e n Income Occupation Status Fashion P r i c e HIGH PRICE SPECIALTY 5.871 - .132 .028 2.160 1.968 .0005 2.363 MEDIUM PRICE SPECIALTY 5.955 .539 .003 2.479 1.837 .187 2.310 BOUTIQUE DEPARTMENT BUDGET 6.891 .271 .007 2.218 1.737 .405 2.330 6.821 .722 .005 1.883 1.726 .232 2.370 6.725 .462 .002 1.872 1.728 .174 2.534 SIGNIFICANCE LEVEL .023 .007 .000 .002 .000 .000 .006 FIANL BETWEEN GROUP SIGNIFICANCE LEVELS Medium p r i c e s p e c i a l t y Boutique Department Budget .000 .000 .000 .000 .000 ,000 ,000 .000 .000 .000 PERCENTAGE OF CASES CLASSIFIED INTO EACH GROUP High p r i c e s p e c i a l t y Medium p r i c e s p e c i a l t y Boutique Department Budget 60_ 19 10 3 4 27 42 11 14 7 0 10 6_3 16 7 7 16 5 41 7 7 13 11 24 74 156. are l i k e l y to be more d i s t i n c t when defined by the c h a r a c t e r i s t i c s of a 'regular' c l i e n t e l e . ANALYSIS 9. CRITERION: Shopping Centre Patronage TRIP: Most Recent PREDICTIVE MODEL: Locational Variables * (excluding size of shopping area patronized) The increased importance assumed by l o c a t i o n a l measures i n r e l a t i o n to the analysis of shopping centre patronage i s immediately apparent from Table 5.19) Five of the eight v a r i a b l e s have s i g n i f i c a n t F r a t i o s and a s i x t h {distance to the nearest tertiary centre) i s only s l i g h t l y above the .05 s i g n i f i c a n c e l e v e l . These F r a t i o s s i g n i f y d i f f e r e n c e s i n group means which i n part are consistent with expectations based upon the t r a d i t i o n a l normative assumptions regarding consumer s p a t i a l behaviour. There was, for example, a general tendency f o r respondents to patronize the types of f a c i l i t y nearest to t h e i r t r i p o r i g i n s . Furthermore, the group means corresponding to the actual distances t r a v e l l e d to shop, although not s i g n i f i c a n t l y d i f f e r e n t , suggest a d i r e c t r e l a t i o n s h i p be-tween the s i z e of shopping centre and the t e r r i t o r i a l extent of the trade area. Expectations regarding the r e l a t i o n s h i p s between s i z e s of nearest f a c i l i t i e s and r e t a i l patronage do not appear to be upheld however. Group * This v a r i a b l e was excluded from the analysis of shopping area pa-tronage since by d e f i n i t i o n i t corresponds c l o s e l y with the c r i t e r i o n v a r i a b l e and i s e s s e n t i a l l y a constant f o r the Downtown group. TABLE 5.19 ANALYSIS 9: SUMMARY OF RESULTS DOWNTOWN GROUP MEANS Estimated distance Actual distance Distance downtown Distance to secondary centre Distance to t e r t i a r y centre Time distance downtown Size of secondary centre Size of t e r t i a r y centre 3. 123. 123. 122. 27. 13.8 487.9 248.7 .9 .5 ,3 .0 ,2 SECONDARY CENTRE 3.8 122.5 188.0 86.0 33. 19. 456. 310. .6 ,4 ,2 .0 TERTIARY CENTRE 3.4 111.8 177.8 117.8 26. 18. 486. .5 .5 ,1 225.3 SIGNIFICANCE LEVEL .736 .776 .000 .000 .069 .000 .036 .005 DISCRIMINANT COEFFICIENTS Estimated distance Distance to secondary centre Time distance downtown Size of t e r t i a r y centre -.076 .067 .478 .011 .188 .060 .611 .012 -.228 .071 .622 .010 .002 .000 .000 .043 FIANL BETWEEN GROUP SIGNIFICANCE LEVELS Secondary centre T e r t i a r y Centre .000 .000 .004 PERCENTAGE OF CASES CLASSIFIED INTO EACH GROUP Downtown Secondary centre T e r t i a r y centre 55_ 23 31 29 60 36 16 17 33 means for size of nearest secondary centre and size of nearest tertiary centre are somewhat s u r p r i s i n g l y lowest f o r the secondary and t e r t i a r y centre groups r e s p e c t i v e l y , suggesting that the conventional use of s i z e as a measure of centre a t t r a c t i v e n e s s (Huff, 1962) i s not wholly suppor-ted by these data; However, i t should be recognized that these s i z e measures r e f e r to the nearest secondary and t e r t i a r y f a c i l i t i e s which of course were not n e c e s s a r i l y the ones patronized by the respective groups. Table 5.19 shows that four of the o r i g i n a l eight v a r i a b l e s were entered i n t o the discriminant functions and i t i s i n t e r e s t i n g that the estimated distance travelled to shop was amongst them, r e f l e c t i n g a s i g n i f i c a n t d i r e c t r e l a t i o n s h i p between distance estimates and s i z e o f shopping centre patronized when other entered v a r i a b l e s were c o n t r o l l e d . A l l three of the between group F r a t i o s based on discriminant scores proved to be s i g n i f i c a n t and some 54 per cent of cases were c o r r e c t l y c l a s s i f i e d . The discriminant functions were most e f f e c t i v e i n c l a s s i -f y i n g secondary centre shoppers of which 60 per cent were c o r r e c t l y assigned and l e a s t e f f e c t i v e i n c l a s s i f y i n g t e r t i a r y centre patrons of which only 33 per cent were c o r r e c t l y predicted. ANALYSIS 10. CRITERION: Shopping Centre Patronage TRIP: Most Recent PREDICTIVE MODEL: Biographical Variables (as f o r Analysis 2) Group means on three of the eight b i o g r a p h i c a l v a r i a b l e s - marital status, number of children living at home and occupation - d i f f e r e d TABLE 5.20 ANALYSIS 10: SUMMARY OF RESULTS GROUP MEANS Age Sex M a r i t a l status Number of ch i l d r e n Income Education Occupation DISCRIMINANT COEFFICIENTS M a r i t a l status Occupation DOWNTOWN 34.4 1.5 1.4 0.7 113.7 3.4 3.5 641 786 SECONDARY CENTRE .6 .1 36.8 1.5 1 1 129.0 3.4 3.8 6.527 1.917 TERTIARY CENTRE 35. 1. 1. 0.8 133.2 3, 4. .6 .6 ,6 .8 .2 6.172 2.143 SIGNIFICANCE LEVEL .420 .110 .000 .036 .201 .082 .010 .010 .016 FINAL BETWEEN GROUP SIGNIFICANCE LEVELS Secondary centre T e r t i a r y centre .000 .005 PERCENTAGE OF CASES CLASSIFIED INTO EACH GROUP Downtown Secondary centre T e r t i a r y centre 47 29 40 ,154 34 45 21 19 26 38 160. s i g n i f i c a n t l y (Table 5.20). The d i s t i n g u i s h i n g group c h a r a c t e r i s t i c s were as follows: downtown shoppers were commonly s i n g l e ; secondary centre patrons were more l i k e l y to be married and to have c h i l d r e n l i v i n g at home; while t e r t i a r y centre shoppers were d i s t i n g u i s h e d by the higher job status of household heads. Only two of the three s i g n i f i c a n t v a r i a b l e s - marital status and occupation - were entered i n t o the discriminant function i n d i c a t i n g the extent of the co v a r i a t i o n between m a r i t a l status and the number of ch i l d r e n l i v i n g at home. The between group F p r o b a b i l i t i e s demonstrate that the downtown group was the most c l e a r l y d i s t i n g u i s h e d and the t e r t i a r y centre group l e a s t so. This pattern i s r e f l e c t e d i n the pre-d i c t i o n of group membership whereas 47 per cent of downtown shoppers were accurately assigned the figu r e dropped to 38 per cent f o r the t e r t i a r y centre group. These r e s u l t s are not t o t a l l y consistent with those obtained i n a s i m i l a r study by Bucklin (1967) based on data c o l l e c t e d from residents of the Oakland area of C a l i f o r n i a . He noted, f o r example a tendency f o r f a m i l i e s within the higher income and s o c i a l groups to shop downtown, whereas secondary centres appealed to middle to lower income and s o c i a l groups, and t e r t i a r y centres to shoppers with "better incomes" but "l e s s e r s o c i a l status" (p. 85). In contrast, the r e s u l t s of t h i s a n a l y s i s show that the income and s o c i a l status of downtown shoppers (based on measures of education and occupation) was c h a r a c t e r i s t i c a l l y below that of the other two groups. The inconsistency i s perhaps a t t r i -butable to the strong influence of an ethnic f a c t o r i n Bucklin's study. 161. ANALYSIS 11. CRITERION: Shopping Centre Patronage TRIP: Most Recent PREDICTIVE MODEL: D i s p o s i t i o n a l Variables (as for Analysis 3) Two of the f i v e d i s p o s i t i o n a l scales - fashion and price - show s i g -n i f i c a n t i n i t i a l F - r a t i o s (Table 5.21). The group mean scale scores for these v a r i a b l e s i n d i c a t e that fashion o r i e n t a t i o n was highest f o r the downtown group, followed by the t e r t i a r y centre and secondary centre groups; while, i n the case of p r i c e o r i e n t a t i o n , the highest score was for secondary centre shoppers, followed by downtown and t h i r d l y t e r t i a r y centre patrons. Only the fashion scale s a t i s f i e d the conditions f o r entry i n t o the discriminant function. There was s u f f i c i e n t c o v a r i a t i o n between the fashion and p r i c e scales to r a i s e the F p r o b a b i l i t y of the l a t t e r s l i g h t -l y above the c r i t i c a l .05 l e v e l . A s i g n i f i c a n t d i f f e r e n c e i n the d i s c r i -minant scores of the downtown and secondary centre groups again emerges, whereas the t e r t i a r y centre group was not c l e a r l y d i s t i n g u i s h e d at a l l . The c l a s s i f i c a t i o n matrix confirms t h i s r e s u l t with only 12 per cent of the t e r t i a r y centre shoppers being c o r r e c t l y assigned i n contrast to the figures of 50 per cent and 57 per cent f o r the other two groups. In order to t e s t whether the i n c l u s i o n of the p r i c e scale would improve p r e d i c t i v e accuracy, the F p r o b a b i l i t y tolerance l e v e l was relaxed to .10. The r e s u l t was a considerable increase i n the accuracy of assignment f o r the t e r t i a r y centre group (from 12 per cent to 48 per cent c o r r e c t ) , but at the expense of decreased accuracy f o r the other two groups: the percentage of c o r r e c t assignments dropped from 50 to TABLE 5.21 ANALYSIS 11: SUMMARY OF RESULTS GROUP MEANS DOWNTOWN SECONDARY CENTRE TERTIARY CENTRE SIGNIFICANCE LEVEL Status Convenience Fashion P r i c e Q u a l i t y 34.6 33.6 36.2 33.3 42.7 33.7 35.0 34.0 33.8 43.0 36.1 34.1 36.0 31.3 44.8 .087 .254 .038 .045 .083 DISCRIMINANT COEFFICIENTS Fashion .655 .615 .651 .038 FINAL BETWEEN GROUP SIGNIFICANCE LEVELS Secondary centre T e r t i a r y centre .013 .847 .133 PERCENTAGE OF CASES CLASSIFIED INTO EACH GROUP Downtown Secondary centre T e r t i a r y centre 50 36 36 41 57 52 9 7 12 163. 33 f o r the downtown group and from 56 to 44 f o r the secondary centre group, r e s u l t i n g o v e r a l l i n a reduction i n the percentage of c o r r e c t assignments (from 48 to 39). This r e s u l t appears to confirm the v a l i d i t y of the established p r a c t i c e of s e t t i n g the F p r o b a b i l i t y tolerance l e v e l f or entry of a v a r i a b l e into the discriminant function at .05. ANALYSIS 12. CRITERION: Shopping Centre Patronage TRIP: Most Recent PREDICTIVE MODEL: Combined Variables (as f o r Analysis 4 excluding size of shopping area patronized) When the three sets of measures were combined, seven of the twenty one o r i g i n a l v a r i a b l e s were included i n the discriminant functions (Table 5.22). As i n Analysis 4, the number of past purchases proved to be s i g n i -f i c a n t , d i s t i n g u i s h i n g secondary centre shoppers, amongst whom frequent repeat purchasing was most common, from t e r t i a r y centre patrons of whom i t was l e a s t c h a r a c t e r i s t i c . In contrast to Analysis 4, the l o c a -t i o n a l v a r i a b l e s emerged as the most e f f e c t i v e discriminators between the shopping centre patronage groups: four of the seven entered v a r i a -bles were l o c a t i o n a l measures and only one v a r i a b l e from each of the other two models was included: namely, marital status and price orien-tation. The between group F r a t i o s were a l l hi g h l y s i g n i f i c a n t and the c l a s s i f i c a t i o n of subjects based on discriminant scores was more accurate than that achieved i n any of the analyses described to t h i s p o i n t . I t i s noticeable that p r e d i c t i v e accuracy was highest f o r the t e r t i a r y TABLE 5.22 ANALYSIS 12: SUMMARY OF RESULTS DISCRIMINANT COEFFICIENTS DOWNTOWN SECONDARY CENTRE TERTIARY CENTRE SIGNIFICANCE LEVEL Estimated distance Past purchases Distance to secondary centre Time distance downtown Size of t e r t i a r y centre M a r i t a l status P r i c e .064 3.231 .065 .404 .012 6.101 1.092 - .053 3.276 .058 .535 .014 6.945 1.107 - .104 2.764 .069 .551 .012 6.687 1.010 ,002 .033 .000 .000 .031 .010 .017 FINAL BETWEEN GROUP SIGNIFICANCE LEVELS Secondary centre T e r t i a r y centre .000 .000 .000 PERCENTAGE OF CASES CLASSIFIED INTO EACH GROUP Downtown Secondary centre T e r t i a r y centre 52_ 20 24 26 58_ 14 22 22 62 165. centre group (62 per cent correct) although, without exception, i t was lowest for that group i n the three previous s i n g l e model analyses. This must i n part be simply a t t r i b u t a b l e to the e f f e c t of combining the d i f f e r e n t models, but also important was the i n c l u s i o n of the past pur-chase v a r i a b l e which so c l e a r l y d i s t i n g u i s h e d the t e r t i a r y centre group. * ANALYSIS 13. CRITERION: Shopping Centre Patronage TRIP: Most Frequent PREDICTIVE MODEL: Locational Variables (as f o r A n a l y s i s 1 excluding size of shopping area patronized and estimated distance t r a v e l l e d ) Three of the seven l o c a t i o n a l v a r i a b l e s included i n the a n a l y s i s were entered i n t o the discriminant function, and a l l three were distance measures: actual distance t r a v e l l e d , distance from downtown and distance from the nearest secondary centre. The discriminant c o e f f i c i e n t s shown i n Table 5.23 confirm the between group di f f e r e n c e s noted i n Analysis 9. Consistent with trade area expectations, the distance t r a v e l l e d to shop was highest for the downtown group and lowest f o r the t e r t i a r y centre group. Equally consistent i s the f a c t that distance from downtown showed the lowest c o e f f i c i e n t f or the downtown group, while distance from the nearest secondary centre was lowest f o r the secondary centre group. In Analyses 13-16 group means and i n i t i a l s i g n i f i c a n c e l e v e l s are excluded from the tabulations due to t h e i r close correspondence to the f i g u r e s reported i n Analyses 9-12. TABLE 5.23 ANALYSIS 13: SUMMARY OF RESULTS DOWNTOWN DISCRIMINANT COEFFICIENTS SECONDARY TERTIARY SIGNIFICANCE CENTRE CENTRE LEVEL Actual distance -.002 -.019 -.027 .000 Distance downtown .038 .060 .065 .000 Distance to secondary centre .068 .058 .071 .002 FINAL BETWEEN GROUP SIGNIFICANCE LEVELS Secondary Centre .000 T e r t i a r y Centre .000 .011 PERCENTAGE OF CASES CLASSIFIED INTO EACH GROUP Downtown 88 11 1 Secondary Centre 23 5_3_ 24 T e r t i a r y Centre 19 27 54 167. The table of between group F p r o b a b i l i t i e s i n d i c a t e s highly s i g n i -f i c a n t d i f f e r e n c e s i n the discriminant scores of the downtown group; di f f e r e n c e s between the secondary and t e r t i a r y centre groups were some-what l e s s sharp. A high percentage of cases were c o r r e c t l y c l a s s i f i e d - 72 per cent o v e r a l l - r e f l e c t i n g the exceptionally accurate p r e d i c t i o n of membership i n the la r g e s t group such that 88 per cent of downtown shoppers were c o r r e c t l y assigned. Comparing these r e s u l t s with the corresponding analysis based on the 'most recent' t r i p data (Analysis 9), reveals an e s s e n t i a l s i m i l a r i t y , although i t i s c l e a r that the between group d i f f e r e n c e s were more sharply defined i n the present case. ANALYSIS 14. CRITERION: Shopping Centre Patronage TRIP.: Most Frequent PREDICTIVE MODEL: Biographical V a r i a b l e s (as f o r Analysis 2) The discriminant c o e f f i c i e n t s (Table 5.24) confirm the between group b i o g r a p h i c a l d i f f e r e n c e s noted i n Analysis 10. Downtown shoppers were dis t i n g u i s h e d by being comparatively young and having a lower job status: secondary centre patrons were the most l i k e l y to be male and have c h i l d -ren; while t e r t i a r y centre shoppers were characterized by being older, female, having few i f any c h i l d r e n , and the highest job status. The between group s i g n i f i c a n c e l e v e l s show that the t e r t i a r y centre shoppers were the most c l e a r l y d i s t i n g u i s h e d , and t h i s i s r e f l e c t e d i n Table 5.24 with 81 per cent of that group being c o r r e c t l y assigned com-pared with 56 per cent and 40 percent f o r the downtown and secondary centre groups r e s p e c t i v e l y . Again the r e s u l t s show a more accurate TABLE 5.24 ANALYSIS 14: SUMMARY OF RESULTS DISCRIMINANT COEFFICIENTS Age Sex Number of c h i l d r e n Occupation DOWNTOWN .136 6.226 .106 2.028 FINAL BETWEEN GROUP SIGNIFICANCE LEVELS Secondary Centre T e r t i a r y Centre .001 .000 SECONDARY CENTRE .141 6.117 .481 2.195 .000 TERTIARY CENTRE .175 7.320 .009 2.880 SIGNIFICANCE LEVEL .024 .032 .000 .000 PERCENTAGE OF CASES CLASSIFIED INTO EACH GROUP Downtown Secondary Centre T e r t i a r y Centre 56_ 37 4 20 40 15 24 22 81 between group d i s c r i m i n a t i o n than was achieved i n the corresponding s i n g l t r i p analysis (Analysis 10), the most marked improvement being i n the p r e d i c t i o n of t e r t i a r y centre membership (from 31 per cent to 81 per cent c o r r e c t ) . This improvement i s a t t r i b u t a b l e to the more hig h l y s i g n i f i c a n t s i n g l e v a r i a b l e d i f f e r e n c e s i n the present analysis and the subsequent i n c l u s i o n of more va r i a b l e s i n the discriminant functions. ANALYSIS 15. CRITERION: Shopping Centre Patronage TRIP: Most Frequent PREDICTIVE MODEL: D i s p o s i t i o n a l Variables (as f o r Analysis 3) Group di f f e r e n c e s with respect to the status and fashion scales were s u f f i c i e n t l y s i g n i f i c a n t to r e s u l t i n these two v a r i a b l e s being entered i n t o the discriminant functions (Table 5.25). As was the case f o r the most recent t r i p data (Analysis 11), fashion scale scores were highest f o r the downtown group but there was a r e v e r s a l i n the mean fashion scores for the other two groups with the secondary centre mean exceeding the t e r t i a r y centre mean i n t h i s a n a l y s i s . Scores on the status scale were highest f o r the t e r t i a r y centre group followed by the downtown and secondary centre groups r e s p e c t i v e l y , which i s probably a f u r t h e r i n d i c a t i o n of. the previously noted tendency of s p e c i a l t y store shoppers to patronize t e r t i a r y centres (see Table 5.2). The table of between group F p r o b a b i l i t i e s shows s i g n i f i c a n t d i f f e r enees between t e r t i a r y centre shoppers and patrons of the other two types of centre, but the figures reveal a n o n - s i g n i f i c a n t d i f f e r e n c e between the discriminant scores of the downtown and secondary centre groups. I t follows that group membership was p r e d i c t e d most accurately « TABLE 5.25 ANALYSIS 15: SUMMARY OF RESULTS DOWNTOWN DISCRIMINANT COEFFICIENTS SECONDARY TERTIARY SIGNIFICANCE CENTRE CENTRE • LEVEL Status .648 .635 .725 .050 Fashion .447 .428 .343 .013 FINAL BETWEEN GROUP SIGNIFICANCE LEVELS Secondary Centre .291 T e r t i a r y Centre .002 .008 PERCENTAGE OF CASES CLASSIFIED INTO EACH GROUP Downtown 47_ 20 34 Secondary Centre 34 36_ 30 T e r t i a r y Centre 15 35 50_ o 171. i n the case of t e r t i a r y centre shoppers (50 per cent c o r r e c t ) , and l e a s t so f o r the secondary centre group (36 per cent c o r r e c t ) . Compar-ing these r e s u l t s with the corresponding figures i n Table 5.21 shows a marked improvement i n the p r e d i c t i o n of t e r t i a r y centre membership (from 12 to 54 per cent c o r r e c t ) , but an appreciable reduction i n the accurate assignment of secondary centre shoppers (from 57 to 36 per cent c o r r e c t ) . These changes r e f l e c t the d i f f e r e n t i a l e f f e c t of the i n c l u s i o n of the status scale i n the discriminant equations. When the r e s u l t s of t h i s a n alysis are compared with those obtained using the l o c a t i o n a l and b i o g r a p h i c a l v a r i a b l e s as p r e d i c t o r s , i t i s c l e a r that the d i s p o s i t i o n a l v a r i a b l e s proved the l e a s t e f f e c t i v e d i s -criminators between the three groups. Whereas the o v e r a l l percentage of cases c l a s s i f i e d was 72 per cent for the l o c a t i o n a l model and 53 per cent f o r the b i o g r a p h i c a l model, the fi g u r e dropped to 43 per cent on the basis of the d i s p o s i t i o n a l s c a l e s . ANALYSIS 16. CRITERION: Shopping Centre Patronage TRIP: Most Frequent PREDICTIVE MODEL: Combined Variables (as f o r Analysis 4 excluding estimated distance t r a v e l l e d , size of shopping area patronized and past purchases) The comparative effectiveness of the three sets of measures i s f u r -ther evidenced by the r e s u l t s obtained when the v a r i a b l e s were combined, of the 19 p r e d i c t o r s , 7 s a t i s f i e d the conditions f o r entry i n t o the discriminant functions (Table 5.26); of these, three were l o c a t i o n a l TABLE 5.26 ANALYSIS 16: SUMMARY OF RESULTS DISCRIMINANT COEFFICIENTS Actual distance Distance downtown Distance to secondary centre Sex Number o f ch i l d r e n Occupation Fashion DOWNTOWN - .0006 .038 .061 4.695 .532 1.650 .574 SECONDARY CENTRE - .017 .060 .051 4.760 .917 1.789 .563 TERTIARY CENTRE - .024 .064 .062 6.229 .413 2.475 .467 SIGNIFICANCE LEVEL .000 .000 .002 .006 .003 .000 .024 FINAL BETWEEN GROUP SIGNIFICANCE LEVELS Secondary Centre T e r t i a r y Centre .000 .000 .000 PERCENTAGE OF CASES CLASSIFIED INTO EACH GROUP Downtown Secondary Centre T e r t i a r y Centre 80 21 19 14 63 8 5 15 73 to 173. measures - actual distance travelled to shop, distance from downtown and distance from the nearest secondary centre, three were b i o g r a p h i c a l measures - occupation, number of children at home and sex, and only one was a d i s p o s i t i o n a l measure - the fashion s c a l e . The order i n which the v a r i a b l e s were entered i s also i n s t r u c t i v e : the f i r s t two were l o c a -t i o n a l v a r i a b l e s - distance from downtown and distance travelled, f o l l o -wed by two b i o g r a p h i c a l v a r i a b l e s - occupation and number of children, and the t h i r d l o c a t i o n a l v a r i a b l e - distance from the nearest secondary centre-, s i x t h was the fashion scale and f i n a l l y the sex of the respond-ent. The type and ordering of these v a r i a b l e s underlines the primary importance of the l o c a t i o n a l model and the comparative weakness of the d i s p o s i t i o n a l scales as discriminators between the shopping centre patronage groups. This conclusion i s consistent with the r e s u l t s obtained i n the corresponding analysis based on the 'most recent' t r i p data (Analysis 12) although, i n that case, the b i o g r a p h i c a l measures assumed le s s importance than i s seen here. The between group s i g n i f i c a n c e l e v e l s and the c l a s s i f i c a t i o n matrix show that the composite model was the most e f f e c t i v e of the four i n d i s c r i m i n a t i n g between the three groups, although the improvement i n pre-d i c t i v e accuracy over that achieved using the l o c a t i o n a l v a r i a b l e s alone was noticeably s m a l l i the percentage of cases c o r r e c t l y assigned over-a l l increased from 72.44 to 73.72. Membership i n the downtown group was i n f a c t more accurately p r e d i c t e d by the l o c a t i o n a l model (88 per cent c o r r e c t compared with 80 per cent), but the p r e d i c t i o n of member-ship i n the other two groups was improved when the d i f f e r e n t models were combined. 174. I t i s i n t e r e s t i n g to n o t i c e the close correspondence between the r e s u l t s of t h i s p a r t i c u l a r a n alysis and those reported i n Bucklin's study of shopping centre patronage c i t e d e a r l i e r (Bucklin, 1967, p. 78 f f . ) . He too was concerned to p r e d i c t shopping patterns i n terms of the patron-age of downtown, secondary and t e r t i a r y centre f a c i l i t i e s . Of some 55 o r i g i n a l p r e d i c t o r s , comprising demographic, shopping plan, motivational and distance v a r i a b l e s , 17 were entered i n t o the discriminant equations (although four of those would not have s a t i s f i e d the more s t r i n g e n t con-d i t i o n s f o r entry adopted i n t h i s a n a l y s i s ) . Consistent with the r e s u l t s reported above, the most hig h l y s i g n i f i c a n t of those seventeen were three distance measures, two of which were equivalent to the mea-sures of distance from downtown and distance from the nearest secondary centre used here. The p r e d i c t i v e accuracy i n the case of the Oakland study (72.6 per cent of cases c o r r e c t l y assigned) also c l o s e l y corresponds with the r e s u l t s of Analysis 16, where, as i n d i c a t e d i n the c l a s s i f i c a t i o n matrix, 73.72 per cent of cases were c o r r e c t l y c l a s s i f i e d o v e r a l l . Unfortunately, Bucklin d i d not analyse (or at l e a s t f a i l e d to report) the comparative e f f e c t i v e n e s s of the d i f f e r e n t sets of measures he employed as discriminators between the three c r i t e r i o n groups, although the F r a t i o s l i s t e d f o r each entered v a r i a b l e (p. 84) suggest that h i s data too would have demonstrated only a small increase i n p r e d i c t i v e accuracy using a combination of v a r i a b l e s over that achieved on the ba s i s of distance measures alone. 175. SUMMARY In order to provide a concise summary of the r e s u l t s presented i n the previous sixteen analyses, and, at the same time, to r e f l e c t the r s u l t s of other analyses not reported i n d e t a i l , Tables 5.27 and 5.28 are included. The e n t r i e s i n these tables represent, f o r the store and shopping centre patronage groups r e s p e c t i v e l y , the corre c t p r e d i c t i o n s (in percentages) achieved by the three d i f f e r e n t models plus the com-bined model averaged over the four shopping t r i p s f o r which data were a v a i l a b l e . TABLE 5.27 AVERAGE PERCENTAGE OF CASES CORRECTLY ASSIGNED TO STORE PATRONAGE GROUPS G R O U P HIGH PRICE MEDIUM PRICE MODEL SPECIALTY SPECIALTY BOUTIQUE DEPARTMENT BUDGET LOCATIONAL 52 32 35 20 20 BIOGRAPHICAL 66 30 55 24 39 DISPOSITIONAL 65 19 57 20 58 COMPOSITE 62 36 53 41 56 In the case of store patronage (Table 5.27), the l o c a t i o n a l v a r i a b l e s f a i l e d to d i s t i n g u i s h accurately between shoppers s e l e c t i n g d i f f e r e n t types of store, with the possible exception of the high-price s p e c i a l t y grQup where size of the shopping area patronized was the most e f f e c t i v e d i s c r i m i n a t i n g v a r i a b l e . The bi o g r a p h i c a l v a r i a b l e s , p a r t i c u l a r l y age, occupation, sex and income, performed somewhat b e t t e r e s p e c i a l l y as 176. p r e d i c t o r s of high-price s p e c i a l t y and boutique store shoppers. The d i s p o s i t i o n a l s c a l e s , e s p e c i a l l y status, fashion and quality, comprised the most e f f e c t i v e of the three separate models, y i e l d i n g moderately accurate p r e d i c t i o n s of membership i n the high-price s p e c i a l t y , boutique and budget store groups. O v e r a l l , the composite model achieved the most accurate r e s u l t s , but, even so, the percentage accuracy was r e l a t i v e l y low with respect to the medium-price s p e c i a l t y and department store groups, demonstrating the heterogeneous c h a r a c t e r i s t i c s of the shoppers p a t r o n i z i n g these two classes of store. This heterogeneity was not unexpected i n the case of department stores which, through t h e i r various departments, can, and do, appeal to a broad cr o s s - s e c t i o n of consumers. The f a i l u r e to p r e d i c t medium-price s p e c i a l t y store patronage with greater accuracy was more s u r p r i s i n g however and perhaps c a l l s i n t o question the v a l i d i t y of d i v i d i n g s p e c i a l t y stores i n t o two clas s e s . TABLE 5.28 AVERAGE PERCENTAGES OF CASES CORRECTLY ASSIGNED TO SHOPPING CENTRE PATRONAGE GROUPS G R O U P SECONDARY TERTIARY MODEL DOWNTOWN CENTRE CENTRE LOCATIONAL 70 54 52 BIOGRAPHICAL 46 47 53 DISPOSITIONAL 51 50 22 COMPOSITE 64 57 66 In general, p r e d i c t i o n s with respect to shopping centre patronage 177. (Table 5.28) were more accurate, which, i n p a r t , may simply r e f l e c t the smaller number of groups and the more even group s i z e s . Locational v a r i a b l e s , e s p e c i a l l y the distances separating the consumer's t r i p o r i -gin and f a c i l i t i e s of varying s i z e , were p a r t i c u l a r l y e f f e c t i v e i n d i s -t i n g u i s h i n g between patrons of the three d i f f e r e n t classes of centre. Indeed, the o v e r a l l performance of the l o c a t i o n a l model was only s l i g h t l y l e s s accurate than that of the composite model. This evidence c l e a r l y supports the importance placed upon distance measures i n e x i s t i n g models of shopping behaviour. Furthermore, i t demonstrates that the l o c a t i o n of the consumer r e l a t i v e to r e t a i l f a c i l i t i e s remains a strong influence upon s p a t i a l behaviour even under conditions where a medium to high order good (such as clothing) i s sought i n an urban context o f f e r i n g a range of r e t a i l opportunities within a r e l a t i v e l y small area; a s i t u a t i o n i n which a confounding of distance e f f e c t s might have been a n t i c i p a t e d . The p r e d i c t i v e accuracy of the b i o g r a p h i c a l and d i s p o s i t i o n a l models was s i m i l a r except that the patrons of t e r t i a r y centres were f a r more accurately d i s t i n g u i s h e d by t h e i r b i o g r a p h i c a l c h a r a c t e r i s t i c s . (iv) Analysis of Department Store Patronage. The poor p r e d i c t i o n of department store patronage i n the previous analyses indicates a wide range i n the c h a r a c t e r i s t i c s of department store shoppers. To i n v e s t i g a t e these v a r i a t i o n s , a discriminant analysis was performed where the c r i t e r i o n groups comprised patrons of s p e c i f i c depart-ment stores. Four groups were defined corresponding to the four major * Both Massy (1965) and Bucklin (1967) have noted the e f f e c t which uneven group s i z e s can have upon the r e s u l t s of discriminant a n a l y s i s . department store chains having o u t l e t s i n Greater Vancouver: The Bay, Woodwards, Eatons and Sears. The store l o c a t i o n s patronized by the subjects included i n the analysis are shown i n Figure 5.2. A s i n g l e analysis was performed on the 'most frequent' t r i p data using a combined set of p r e d i c t o r v a r i a b l e s . The r e s u l t s of the analysis confirm s i g n i f i c a n t d i f f e r e n c e s between shoppers at d i f f e r e n t department stores. The four groups d i f f e r e d s i g -n i f i c a n t l y on s i x of the v a r i a b l e s with b i o g r a p h i c a l d i f f e r e n c e s being the most marked (Table 5.29). The Bay patrons were di s t i n g u i s h e d by t h e i r tendency to be s i n g l e and consequently to have few, i f any, c h i l d -ren; Woodwards' shoppers had generally t r a v e l l e d shorter distances, were ol d e r and more l i k e l y to be married; Eatons ' customers tended to be younger and to have higher educational and occupational status; while Sears ' patrons were characterized by having t r a v e l l e d longer distances to shop ( a t t r i b u t a b l e to the p e r i p h e r a l l o c a t i o n of the s t o r e ) , having l a r g e r f a m i l i e s and the lowest educational and occupational status. Seven v a r i a b l e s were included i n the discriminant functions. The primary importance of b i o g r a p h i c a l d i f f e r e n c e s i s underlined by three of the f i r s t four v a r i a b l e s entered being number of children, occupation and age. Covariation amongst the p r e d i c t o r s was such tha t , although none of the d i s p o s i t i o n a l v a r i a b l e s had i n i t i a l F r a t i o s s i g n i f i c a n t at the .05 l e v e l , both the status and quality scales s a t i s f i e d the condi-t i o n s f o r entry i n t o the discriminant equations. C o n t r o l l i n g f o r other entered v a r i a b l e s , The Bay shoppers emerged as the most status o r i e n t e d followed i n descending order by Woodwards', Eatons' and Sears' patrons; whereas, on the q u a l i t y scale the ordering was Sears, The Bay, Eatons TABLE 5.29 ANALYSIS OF DEPARTMENT STORE PATRONAGE: SUMMARY OF RESULTS BAY GROUP MEANS Actual distance 129.5 Distance downtown 134.1 Distance to secondary centre 120.1 Distance to t e r t i a r y centre 30.1 Time distance downtown 14.8 Size of secondary centre 496.0 Size of t e r t i a r y centre 291.3 Age 33.8 Sex 1.6 Ma r i t a l status 1.4 Number of chi l d r e n 0. Income 105. Education 3, Occupation 3. Status 35. Convenience 33. Fashion 36. Pri c e 32. Quality 43. DISCRIMINANT COEFFICIENTS Actual distance .013 Time distance downtown .175 Age .119 Number of chi l d r e n .162 Occupation - .184 Status .825 Quality 1.278 FINAL BETWEEN GROUP SIGNIFICANCE LEVELS Woodwards .000 Eatons .019 Sears .002 PERCENTAGE OF CASES CLASSIFIED INTO EACH GROUP Bay Woodwards Eatons Sears 46 14 21 21 WOODWARDS 112.2 159.8 102.6 28.6 17.3 467.9 271.0 39.3 1.5 1.6 1.3 120.0 3.3 3.6 32.7 35.5 34.3 34.5 41.4 .005 .266 .138 .635 .190 .736 1.213 .009 .000 19 48 23 11 .1 .1 EATONS 113.3 151.1 104.3 29.2 16. 466. 267.0 30.3 1.5 1.5 0.7 112.2 3. 3, 33. 33. .6 .7 .0 .9 35.8 33.7 42.9 .007 .239 .097 .386 .154 .730 1.275 .004 16 22 36 11 SEARS 163.9 177.8 91.0 25.5 18.5 436.6 266.0 32.6 1.6 1.6 1.7 104 2 2 32 33 36.2 33.5 43.2 .016 .224 .095 .966 .479 .715 1.354 SIGNIFICANCE LEVEL .041 .117 .089 .913 .105 .187 .897 .005 .762 .038 .000 .611 .027 .009 .070 .167 .232 .204 .094 .033 .048 .015 .000 .009 .009 .050 19 16 19 58 180. and Woodwards. Two distance measures were also entered i n d i c a t i n g s i g n i -f i c a n t between group dif f e r e n c e s i n distances t r a v e l l e d to shop and i n time distances from downtown Vancouver. A l l between group F r a t i o s were s i g n i f i c a n t beyond the .05 l e v e l with the c l e a r e s t d i s t i n c t i o n being between The Bay and Woodwards and the greatest overlap between The Bay and Eatons. The c l a s s i f i c a t i o n matrix i n d i c a t e s that group membership was most accurately predicted i n the case of Sears (58 per cent c o r r e c t ) , followed i n descending order of accuracy by Woodwards (48 per cent), The Bay (46 per cent) and Eatons (36 per cent). Although the l e v e l of accuracy i s r e l a t i v e l y low, the modal category f o r each store occurs on the p r i n c i p a l diagonal of the c l a s s i f i c a t i o n matrix r e f l e c t i n g the s i g -n i f i c a n t between group d i f f e r e n c e s . A separate analysis was performed to i n v e s t i g a t e p r o f i l e d i f f e r e n c e s between the patrons of the three department stores located i n Downtown Vancouver - The Bay, Eatons and Woodwards - again using the 'most f r e -quent' t r i p data and a combined set of p r e d i c t o r s . In t h i s case, only two of the nineteen v a r i a b l e s were entered i n t o the discriminant func-t i o n s : namely, number of children and the status s c a l e . The Bay shoppers again emerged as having fewer c h i l d r e n and the strongest status o r i e n t a t i o n , while the reverse held for Woodwards' patrons. Between group F p r o b a b i l i t i e s and the c l a s s i f i c a t i o n of subjects on the b a s i s of discriminant scores (Table 5.30) showed a s i g n i f i c a n t d i f f e r e n c e between patrons of The Bay and Woodwards, but the Eatons' group was not s i g n i -f i c a n t l y d i f f e r e n t from e i t h e r of the other two. This r e s u l t i s essen-t i a l l y consistent with that reported by Gayler (1974, pp. 143-4) i n an analysis of the s o c i a l c l a s s d i f f e r e n c e s of Downtown Vancouver department I TABLE 5.30 ANALYSIS OF DOWNTOWN DEPARTMENT STORE PATRONAGE BAY WOODWARDS EATONS FINAL BETWEEN GROUP SIGNIFICANCE LEVELS Woodwards .000 Eatons .337 .110 PERCENTAGE OF CASES CLASSIFIED INTO EACH GROUP Bay 6_1 23 14 Woodwards 27 56_ 17 Eatons 52 45 _3 CO i—• 182. store patrons. He too found, i n the case of c l o t h i n g purchases, that the most pronounced d i f f e r e n c e s i n c l i e n t e l e c h a r a c t e r i s t i c s were between shoppers at The Bay and Woodwards with the former appealing f a r more to the middle and upper s o c i a l c l a s s e s ; a r e s u l t perhaps r e f l e c t e d i n the higher status scale scores for The Bay patrons i n the present case. I t i s noteworthy, however, that between group d i f f e r e n c e s on the income, education and occupation v a r i a b l e s employed by Gayler as measures of s o c i a l c l a s s were not s i g n i f i c a n t i n t h i s a n a l y s i s . The existence of s i g n i f i c a n t d i f f e r e n c e s i n the c h a r a c t e r i s t i c s of s p e c i f i c department store patrons accounts for the proportion of w i t h i n -group variance for the department store group apparent from the r e s u l t s of the previous analyses of store patronage (Analyses 1 to 8). I t would seem to follow therefore that the achievement of more pre c i s e between group d i s c r i m i n a t i o n depends upon a disaggregation of the department store category. (v) Analysis of S p e c i f i c Shopping Centre Patronage A further discriminant a n a l y s i s was performed to determine what d i f f e r e n c e s , i f any, existed between the patrons of two s p e c i f i c shop-ping centres: Oakridge i n Vancouver and Park Royal i n West Vancouver. Both centres are located within the sampling area (see Figure 5.2) and both are planned developments of approximately equal s i z e i n terms of commercial floorspace containing a s i m i l a r range of r e t a i l establishments. Since appropriate l o c a t i o n a l measures were not a v a i l a b l e - s p e c i f i c a l l y , the distances separating respondents from the two centres - the set of p r e d i c t o r s was l i m i t e d to the seven b i o g r a p h i c a l v a r i a b l e s and the f i v e TABLE 5.31 ANALYSIS OF SHOPPING CENTRE PATRONAGE GROUP MEANS Age Sex M a r i t a l status Number of c h i l d r e n Income Education Occupation Status Convenience Fashion P r i c e Q u a l i t y DISCRIMINANT COEFFICIENTS Age .196 OAKRIDGE 41.1 1.5 1.7 1 133 3 3 33 34 34 34.0 42.1 PARK ROYAL 34.1 1.5 1.5 1.2 150.0 3 4 33 6 3 7 34.8 35 :o 33.6 44.1 SIGNIFICANCE LEVEL .028 .936 .100 .376 .395 .046 .111 .797 .823 .608 .711 .083 ,162 .028 FINAL BETWEEN GROUP SIGNIFICANCE LEVEL Park Royal .028 PERCENTAGE OF CASES CLASSIFIED INTO EACH GROUP Oakridge Park Royal 64 44 36 56 oo CjJ 184. d i s p o s i t i o n a l scales. The data base, as i n the previous analysis of department store patronage, comprised the 'most frequent' t r i p responses. The i n i t i a l F p r o b a b i l i t i e s l i s t e d i n Table 5.31 show that for only two of the twelve o r i g i n a l v a r i a b l e s - age and education - were the centre means s i g n i f i c a n t l y d i f f e r e n t . Patrons of Oakridge tended to be older, whereas Park Royal shoppers c h a r a c t e r i s t i c a l l y had higher l e v e l s of edu-catio n . There was a s t r i k i n g s i m i l a r i t y i n the d i s p o s i t i o n a l character-i s t i c s of the two groups. Only i n the case of the q u a l i t y s c a l e , d i d the F r a t i o come near the .05 s i g n i f i c a n c e l e v e l with Park Royal shop-pers e x h i b i t i n g the greater q u a l i t y o r i e n t a t i o n . Age was the only v a r i a b l e to s a t i s f y conditions f o r entry i n t o the discriminant equation and the associated centre c o e f f i c i e n t s are shown i n Table 5.31. Since no other v a r i a b l e s were entered, the f i n a l between group F p r o b a b i l i t y was the same as that f o r the age v a r i a b l e (.0285). The c l a s s i f i c a t i o n of subjects based on the s i n g l e v a r i a b l e discriminant equation shows that approximately 61 per cent of cases were c o r r e c t l y assigned with membership i n the Oakridge group being more accurately p r e d i c t e d than was true f o r Park Royal. I t i s c l e a r therefore that the b i o g r a p h i c a l and d i s p o s i t i o n a l cha-r a c t e r i s t i c s of the patrons of these two centres could not be c l e a r l y d i s t i n g u i s h e d except.in the case of age. The absence of s i g n i f i c a n t d i f f e r e n c e s i n respondent c h a r a c t e r i s t i c s suggests that a shopper's s e l e c t i o n of one or other of the centres was based mainly upon distance considerations. However t h i s must remain a speculative conclusion since the necessary l o c a t i o n a l measures were not a v a i l a b l e . I t i s also unfortunate that group s i z e s were too small to permit the i n c l u s i o n of a l a r g e r number of shopping centres, which may have served to accentuate between group d i f f e r e n c e s . (vi) Regression Analysis of Distance T r a v e l l e d to Shop The analyses reported i n t h i s chapter have centred on the problem of accounting for v a r i a t i o n s i n retail patronage. Attention i n t h i s section s h i f t s to a d i f f e r e n t aspect of consumer s p a t i a l behaviour, namely the distance travelled to shop. The r e s u l t s of a regression analysis are reported i n which distance t r a v e l l e d to shop was the depen-dent v a r i a b l e and the independent v a r i a b l e s comprised measures of the l o c a t i o n a l , b i o g r a p h i c a l and d i s p o s i t i o n a l c h a r a c t e r i s t i c s of the consu-mer. Various studies have sought to account f o r v a r i a t i o n s i n distances t r a v e l l e d to shop i n terms of d i f f e r e n c e s i n consumer c h a r a c t e r i s t i c s . As e a r l y as 1948, Converse, i n a study of buying habits i n southern I l l i n o i s , noted the w i l l i n g n e s s of higher income consumers to t r a v e l longer distances to s a t i s f y t h e i r shopping needs (Converse, 1948). Rich (1963) found a greater propensity among women i n the upper income groups to t r a v e l f u r t h e r to patronize downtown stores i n New York and Cleveland, and s i m i l a r r e s u l t s were obtained by Jonassen (1955), Pahl (1965) and Hess (1966) i n the United States, B r i t a i n and Denmark r e s -p e c t i v e l y . Murdie (1965) has demonstrated the e f f e c t of c u l t u r a l d i f f e r -ences on t r a v e l behaviour i n comparing the shopping patterns of Old Order Mennonites and 'modern' Canadians for c l o t h i n g and footwear i n south-western Ontario. Perhaps of most relevance i n the present con-text are the r e s u l t s of Gayler's study of consumer s p a t i a l behaviour i n Greater Vancouver i n which s i g n i f i c a n t d i f f e r e n c e s were shown i n the 186. distances t r a v e l l e d to purchase various goods by consumers i n d i f f e r e n t s o c i a l c l a s s and family status groups (Gayler, 1974). Upper and. middle cla s s consumers were found to t r a v e l s i g n i f i c a n t l y greater distances for groceries and dress items, while consumers having a low family status t r a v e l l e d s i g n i f i c a n t l y shorter distances for groceries, dress and f u r n i t u r e . In t h i s study, four step-wise regression analyses were performed to determine the r e l a t i o n s h i p between distance t r a v e l l e d to shop and a set of p r e d i c t o r s comprising l o c a t i o n a l , b i o g r a p h i c a l and d i s p o s i t i o n a l v a r i a b l e s . The dependent v a r i a b l e s were the o r i g i n - d e s t i n a t i o n distances for each of the four shopping t r i p s f o r which data were a v a i l a b l e . Step-wise regression i s a form of m u l t i p l e regression i n which only those p o t e n t i a l independent v a r i a b l e s that s i g n i f i c a n t l y contribute to explaining v a r i a t i o n i n the dependent v a r i a b l e are included i n the regression equation. The analysis proceeds i n a step-wise manner such that the v a r i a b l e introduced at each stage i s the one which contributes most s i g n i f i c a n t l y to the explanation of the dependent v a r i a b l e , c o n t r o l -l i n g f o r the v a r i a b l e s previously included i n the regression equation. The analysis i s terminated when no further p o t e n t i a l independent v a r i a -bles make a contribution s i g n i f i c a n t beyond a predetermined l e v e l , u s ually .05. Table 5.32 shows the regression equations derived i n the four analyses. The regression c o e f f i c i e n t s for independent v a r i a b l e s included i n "each equation are l i s t e d and the s i g n i f i c a n c e of the c o n t r i b u t i o n of each o f these v a r i a b l e s i s shown i n parentheses by the F p r o b a b i l i t y . Also tabulated are the i n t e r c e p t values, c o e f f i c i e n t s of determination TABLE 5.32 REGRESSION EQUATIONS ON DISTANCE TRAVELLED TO SHOP POTENTIAL DISTANCE DISTANCE DISTANCE DISTANCE INDEPENDENT TRAVELLED TRAVELLED TRAVELLED TRAVELLED VARIABLES (TRIP 1) (TRIP 2) (TRIP 3) (TRIP 4) DISTANCE DOWNTOWN . .5856 (.0000) .6423 (.0000) .7500 (.0000) .7265 (.0000) DISTANCE SEC. CENTRE .3122 (.0005) .2526 (.0021) .3364 (.0000) DISTANCE TER. CENTRE * SIZE OF SHOPPING AREA 20.2996 (.0024) 33.0470 (.0000) 27.9884 (.0000) 32.2024 (.0000) SIZE OF SEC. CENTRE -.1231 (.0003). SIZE OF TER. CENTRE AGE SEX + -23.4159 (.0052) MARITAL STATUS NUMBER OF CHILDREN INCOME EDUCATION OCCUPATION -5.3096 (.0124) STATUS CONVENIENCE -1.3912 (.0197) -1.1089 (.0162) FASHION PRICE QUALITY 2.2046 (.0024) INTERCEPT -71.5814 -184.1477 -25.8311 -24.5571 R 2 .2369 .3848 .3954 .5233 STAND. ERR. (Y) 84.5709 70.7870 71.5566 56.8276 * Log. Transform + B i n a r y V a r i a b l e 188. (R ) and standard errors of the estimate f o r each equation. The most s t r i k i n g feature of the r e s u l t s i s the small number of p o t e n t i a l independent v a r i a b l e s s a t i s f y i n g the conditions f o r i n c l u s i o n i n the regression equations, and, more p a r t i c u l a r l y , the comparative absence of b i o g r a p h i c a l and d i s p o s i t i o n a l p r e d i c t o r s . The most s i g n i f i -cant p r e d i c t o r s were distance from Damtown Vancouver and the size of shopping area patronized, both of which were included i n a l l four equa-t i o n s . The strong p o s i t i v e r e l a t i o n s h i p between distance t r a v e l l e d and distance from Downtown Vancouver r e f l e c t s the dominance of Downtown Vancouver as a centre for c l o t h i n g purchases, shown pr e v i o u s l y i n the high percentage of shopping t r i p s to downtown locatio n s (see Figure 5.5) . The importance placed upon s i z e considerations as a surrogate mea-sure o f centre u t i l i t y i n t r a d i t i o n a l trade area models i s supported here by the s i g n i f i c a n t p o s i t i v e r e l a t i o n s h i p between distance t r a v e l l e d and s i z e of shopping area patronized. C l e a r l y respondents were prepared to t r a v e l f u r t h e r to shop at l a r g e r centres. The t h i r d most s i g n i f i c a n t p r e d i c t o r was again a l o c a t i o n a l measure i n the form of distance from the nearest secondary centre, which was included i n three of the four equations and was also p o s i t i v e l y r e l a t e d to distance t r a v e l l e d to shop. The i n c l u s i o n of t h i s v a r i a b l e r e f l e c t s the importance of secondary centres as the major t r i p d estinations a f t e r Downtown Vancouver. " The only other l o c a t i o n a l v a r i a b l e to appear i n any of the equations was size of the nearest secondary centre which emerged as a s i g n i f i c a n t p r e d i c t o r of distance t r a v e l l e d to the 'most frequent' d e s t i n a t i o n . The negative r e l a t i o n s h i p i n t h i s case i n d i c a t e s that the smaller the nearest secondary centre the further the respondent tended to t r a v e l . This implies that respondents were prepared to forego the convenience of shopping at nearby, but comparatively small, secondary centres f o r the sake of b e t t e r shopping opportunities at a greater distance; a r e s u l t which i s again consistent with trade area model assumptions regarding the r e l a t i o n s h i p between centre s i z e and a t t r a c t i v e n e s s . In only two instances were bi o g r a p h i c a l v a r i a b l e s entered i n the regression equations. In the case of the t h i r d most recent t r i p , sex of the respondent emerged as a negative p r e d i c t o r of distance t r a v e l l e d , i n d i c a t i n g a tendency f o r men to t r a v e l f u r t h e r than women. This i s a somewhat s u r p r i s i n g outcome i n l i g h t of the previously noted convenience-o r i e n t a t i o n of the male respondents, but i t i s probably a t t r i b u t a b l e to t h e i r preference for department stores, hence n e c e s s i t a t i n g t r i p s e i t h e r to Downtown Vancouver or a secondary centre. The other b i o g r a p h i c a l v a r i a b l e to contribute s i g n i f i c a n t l y to the p r e d i c t i o n of distance t r a v e l l e d was occupational status i n the case of the 'most frequent' t r i p , and again the r e l a t i o n s h i p was negative demonstrating that r e s -pondents with higher status jobs to t r a v e l shorter distances. This i s consistent with the r e s u l t s of the e a r l i e r discriminant analyses which showed that patrons of the more immediately accessible t e r t i a r y centres were most c l e a r l y d i s t i n g u i s h e d by t h e i r higher job status. I t i s how-ever i n c o n s i s t e n t with Gayler's conclusion that consumers of higher s o c i a l status (assessed p a r t l y i n terms of occupation) t r a v e l l e d longer distances to purchase c l o t h i n g goods (Gayler, 1974, p. 123 f f . ) . This inconsistency may i n part be a t t r i b u t a b l e to changes i n the r e t a i l 190. structure i n Greater Vancouver subsequent to Gayler's study, and i n p a r t to d i f f e r e n c e s i n the s p a t i a l extent of the sampling area. Only two of the f i v e d i s p o s i t i o n a l scales appeared i n the regression equations: convenience orientation and quality orientation. The conve-nience scale was included as a negative p r e d i c t o r i n two instances, con-firming the expected inverse r e l a t i o n s h i p between convenience o r i e n t a t i o n and distance t r a v e l l e d . The q u a l i t y scale on the other hand emerged i n a s i n g l e instance as p o s i t i v e l y r e l a t e d to the length of the shopping t r i p , suggesting that respondents who p a r t i c u l a r l y valued q u a l i t y were prepared to t r a v e l somewhat furth e r to s a t i s f y that d i s p o s i t i o n . The c o e f f i c i e n t s of determination associated with each equation show that the percentage of v a r i a t i o n i n distance t r a v e l l e d 'explained' by the various sets of p r e d i c t o r s ranged from a low of 24 i n the case of the 'most frequent' t r i p . This pattern confirms the r e s u l t s of the discriminant analyses i n that the l e v e l of p r e d i c t i v i t y was again highest f o r the ' t y p i c a l ' shopping t r i p . I t i s p o s s i b l e that the c o n t r i b u t i o n of these v a r i a b l e s was r e s t r i c t e d by the s p a t i a l l i m i t a t i o n s of the sampling area. Had a more s p a t i a l l y extensive sample been drawn, the s i g n i f i c a n c e of the b i o g r a p h i c a l and d i s p o s i t i o n a l c h a r a c t e r i s t i c s of the respondent may have been increased. But i n the absence of such a data base t h i s conjecture remains l a r g e l y speculative. ( v i i ) Male-Female Differences i n Shopping Behaviour and D i s p o s i t i o n s As i n d i c a t e d i n Chapter 4, the sample was designed so as to provide an approximately equal number of male and female respondents. Figure 5.3 shows that i n f a c t 166 respondents were male and 185 female. The primary reason f o r c o n t r o l l i n g the sample i n t h i s way was to obtain a 191. data base for comparing male and female shopping patterns. Studies of shopping behaviour have commonly d e a l t e x c l u s i v e l y with female consumers, and yet t h i s emphasis i s becoming i n c r e a s i n g l y questionable i n the con-temporary North American c i t y where the marketing of men's goods i s r a p i d l y achieving equal prominence. This trend i s perhaps most e x p l i c i t i n the r e t a i l d i s t r i b u t i o n of c l o t h i n g , the purchasing of which formed the focus of attention i n t h i s study. Two contingency te s t s were performed to determine the r e l a t i o n s h i p between the respondent's sex and store and shopping centre patronage. Responses were aggregated over the four shopping t r i p s f o r which data were obtained, and hence the u n i t of analysis was a 'respondent t r i p ' rather than a respondent. This procedure has the e f f e c t of i n c r e a s i n g the c e l l s i z e s i n the tabulations which i s advantageous i n computing chi-square s t a t i s t i c s (Blalock, 1960, p. 214 f f . ) , but i t i s recognised that t h i s b e n e f i t i s gained at the expense of compromising the assump-t i o n of independent random sampling. The most s t r i k i n g d ifferences between male and female shoppers emerge i n the patronage high-price s p e c i a l t y , boutique and department stores (Table 5.33) . whereas female shoppers were the more l i k e l y to patronize high-price s p e c i a l t y and boutique stores, male shoppers were the more l i k e l y to s e l e c t department stores. While only 5 per cent of male t r i p s were to boutique stores the corresponding fig u r e (9.62 per cent) was almost double f o r female t r i p s , and the d i f f e r e n c e was almost as great i n the case of high-price s p e c i a l t y store patronage. On the other hand, department store shopping was some 9 per cent higher f o r men (65 per cent) than for women (56 per cent). These male-female d i f f e r e n c e s were s u f f i -c i e n t to produce a highly s i g n i f i c a n t chi-square value showing that the TABLE 5.33 DISTRIBUTION OF SHOPPING TRIPS BY SEX AND STORE TYPE HIGH PRICE MEDIUM PRICE ^ o ^ r m , , „„„„„„,.„,„ BOUTIQUE DEPARTMENT BUDGET TOTAL SPECIALTY SPECIALTY Male 27 ( 4%) 94 (15%) 31 ( 5%) 401 (65%) 66 (11%) 619 Female 53 ( 7%) 116 (16%) 68 (10%) 394 (56%) 76 (11%) 707 TOTAL 80 ( 6%) 210 (16%) 99 ( 7%) 795 (60%) 142 (11%) 1326 Chi Square = 19.6 Degrees of Freedom = 4 S i g n i f i c a n t at .001 l e v e l TABLE 5.34 DISTRIBUTION OF SHOPPING TRIPS BY SEX AND SHOPPING CENTRE SECONDARY TERTIARY „™,» T DOWNTOWN TOTAL CENTRE CENTRE Male 326 (55%) 220 (37%) 52 ( 9%) 598 Female 373 (55%) 213 (31%) 98 (14%) 684 TOTAL 699 (55%) 433 (34%) 150 (12%) 1282 Chi Square = 11.66 Degrees of Freedom = 2 193. v a r i a t i o n s i n store patronage were greater than could be ascribed to change alone. In the case of shopping centre patronage, a v i r t u a l l y i d e n t i c a l proportion of male and female t r i p s were to downtown locatio n s (Table 5.34). Secondary centre destinations accounted f o r a somewhat higher percentage of male t r i p s , whereas a greater proportion of female t r i p s were to t e r t i a r y centres. Indeed, nearly two-thirds of t r i p s to t e r t i a r y centres were made by females, although i t must be noted that t h i s f r a c -t i o n i s somewhat i n f l a t e d by the l a r g e r t o t a l number of female t r i p s . This r e s u l t probably r e f l e c t s the greater i n c l i n a t i o n among women to patronize s p e c i a l t y stores which previous f i g u r e s have shown to be associated with t r i p s to t e r t i a r y centres (see Table 5.2). The c h i -square value i s again s i g n i f i c a n t showing that the male-female d i f f e r -ences i n shopping centre patronage exceeded chance v a r i a t i o n s . In a d d i t i o n to these contingency t e s t s , u n i v a r i a t e analyses of variance were performed to e s t a b l i s h how male and female respondents d i f f e r e d i n t h e i r scores on the f i v e d i s p o s i t i o n a l s c a l e s . The r e s u l t s obtained are shown i n Table 5.35 which l i s t s f o r each scale: group means, F r a t i o s and F p r o b a b i l i t i e s . TABLE 5.35 MALE-FEMALE GROUP MEANS ON DISPOSITIONAL SCALES AND F STATISTICS MALE FEMALE F RATIO F PROB. STATUS 33 .1 35 .4 12. 06 .0007 CONVENIENCE . 35 .9 32 .8 18. 81 .0000 FASHION 33 .9 36 .4 9. 85 .0020 PRICE 33 .4 33 .0 0. 63 .4323 QUALITY 42 .2 43 .7 6. 39 .0115 194. Male-female di f f e r e n c e s were s i g n i f i c a n t beyond the .05 l e v e l f o r four of the f i v e scales. In general, the d i r e c t i o n of the d i f f e r e n c e s corresponds with i n t u i t i v e expectations. Between group v a r i a t i o n s were most s i g n i f i c a n t with respect to convenience o r i e n t a t i o n . Not s u r p r i s i n g l y scores were higher f o r the male group suggesting a stronger d e s i r e on the part of men to complete the shopping task as expediently as p o s s i b l e . I t follows that women were more l i k e l y to regard shopping as a l e i s u r e a c t i v i t y and hence be w i l l i n g to expend time and e f f o r t i n shopping around p r i o r to making a purchase. Scores on the status scale were s i g n i f i c a n t l y higher f o r women, again, i n l i g h t of previous r e s u l t s , r e f l e c t i n g t h e i r greater i n c l i n a t i o n toward s p e c i a l t y store shopping. S i m i l a r l y , the demonstrated preference of women f o r boutique stores i s r e f l e c t e d i n the s i g n i f i c a n t l y higher scores for the female group on the fashion scale. F i n a l l y , the higher mean score f o r women on the quality scale suggests that they tended to place greater emphasis than men on considerations of q u a l i t y and r e l i a b i l i t y when choosing where to shop. F. CONCLUSION The r e s u l t s of the analyses presented i n t h i s chapter lead to seve-r a l conclusions about the shopping behaviour of the sample group. In the f i r s t place i s was shown that a dec i s i o n to patronize a p a r t i c u l a r type of store or shopping centre was r e l a t e d to other aspects of the shopping t r i p ; i n p a r t i c u l a r , the o r i g i n of the t r i p , the mode of t r a v e l , and the main purpose of the t r i p . In a d d i t i o n , a r e l a t i o n s h i p between the pattern of past purchases and r e t a i l patronage was demonstrated. 195. The r o l e of such va r i a b l e s i n the a n a l y s i s of shopping behaviour i s however e s s e n t i a l l y d e s c r i p t i v e rather than explanatory. In seeking to explain v a r i a t i o n i n the shopping behaviour o f the sample group, three approaches were compared. Two of these, the l o c a t i o n -a l and b i o g r a p h i c a l , have been employed i n previous studies using s i m i l a r v a r i a b l e s to those included here. The t h i r d , the d i s p o s i t i o n a l approach, i s e s s e n t i a l l y the product of t h i s study as described i n the previous chapters. The a b i l i t y of the d i f f e r e n t models to d i s t i n g u i s h accurately between the predefined patronage groups was found to vary dependent upon the c r i t e r i o n . In the case of store patronage, the d i s p o s i t i o n a l v a r i a -b l e s comprised the most e f f e c t i v e of the three models,.yielding moderately accurate p r e d i c t i o n s of membership i n three of the f i v e groups. i n general, the discriminatory power of the p r e d i c t i v e models was l i m i t e d by the heterogeneity of the l a r g e s t group comprising department store shoppers. The within-group v a r i a t i o n s i n t h i s case were demonstrated i n a separate a n a l y s i s . With respect to shopping centre patronage, the l o c a t i o n a l model was the most e f f e c t i v e i n d i s c r i m i n a t i n g between patrons of the three classes of centre, supporting the importance placed upon measures of the consumers' l o c a t i o n r e l a t i v e to f a c i l i t i e s of varying s i z e i n e x i s t i n g models of s p a t i a l behaviour. Results of a regression analysis also showed l o c a t i o n a l v a r i a b l e s as making the most s i g n i f i c a n t c o n t r i b u t i o n t o the explanation of variance i n distance t r a -v e l l e d to shop. Separate consideration was given to male-female d i f f e r e n c e s i n shopping behaviour and d i s p o s i t i o n s which seems to have been a neglected area of in q u i r y h i t h e r t o . S i g n i f i c a n t d i f f e r e n c e s emerged i n both the the r e t a i l choices and d i s p o s i t i o n a l c h a r a c t e r i s t i c s of the two sexes. CHAPTER 6 SUMMARY AND CONCLUSIONS Summary A major objective of t h i s study has been to compare three approaches to the analysis of consumer s p a t i a l behaviour. These approaches were based upon measures of the l o c a t i o n a l , b i o g r a p h i c a l and d i s p o s i t i o n a l c h a r a c t e r i s t i c s of the consumer. The e x i s t i n g l i t e r a t u r e provides many examples of the a p p l i c a t i o n of the f i r s t two approaches and the emphasis i n much of t h i s study has therefore been placed upon the development of a d i s p o s i t i o n a l model of consumer s p a t i a l behaviour. The review of consumer behaviour models i n Chapter 1 demonstrated the importance of an i n d i v i d u a l ' s a t t i t u d e s and values as mediating v a r i a -bles i n the sequential decision-making process leading to purchasing behaviour. Following from that review, a model of consumer s p a t i a l behaviour was proposed wherein a dec i s i o n to patronize a s p e c i f i c store on shopping centre was seen to r e l a t e back to the consumer's d i s p o s i t i o n s . These were defined as psychological dimensions, underlying s p e c i f i c a t t i -tudes and behaviour, used by the i n d i v i d u a l to describe, comprehend and evaluate a set of environmental objects. In the following two chapters, the methodology employed to i d e n t i f y and measure consumer d i s p o s i t i o n s w i t h respect to c l o t h i n g purchases was described. Five major d i s p o s i t i o n s , were i d e n t i f i e d on the basis of interview responses and a f a c t o r analysis of questionnaire data. These - 196 -197. were l a b e l l e d : status-orientation, convenience-orientation, fashion-orientation, price-orientation, and quality orientation. The measure-ment of consumer d i s p o s i t i o n s required the development of a p s y c h o l o g i c a l s c a l i n g instrument. Following the lead of environmental psychologists, the L i k e r t s c a l i n g technique was selected as an appropriate measuring device. A preliminary set of scales was developed and a p r e t e s t conduc-ted to assess t h e i r r e l i a b i l i t y and v a l i d i t y i n accordance with e s t a b l i -shed psychometric methods. On the basis of the r e s u l t s minor modifica-tions were made to produce the f i n a l set of s c a l e s . The major data c o l l e c t i o n phase comprised a questionnaire survey of a s t r a t i f i e d random sample of households drawn from the C i t y of Vancouver and the adjacent m u n i c i p a l i t i e s of North and West Vancouver. The questionnaires were administered by t r a i n e d student interviewers and 351 completed returns were obtained. The data provided information on the respondent's d i s p o s i t i o n s , shopping behaviour and b i o g r a p h i c a l c h a r a c t e r i s t i c s . Location r e l a t i v e to r e t a i l f a c i l i t i e s of varying s i z e was determined from the address of each respondent. Various types of analysis were performed using the questionnaire data but the major emphasis was placed upon a s e r i e s of discriminant analyses. Respondents were grouped based on the type of store and type of shopping centre patronized on each of four shopping t r i p s f o r c l o t h i n g . The a b i l i t y of l o c a t i o n a l , b i o g r a p h i c a l and d i s p o s i t i o n a l measures to discriminate between these r e t a i l patronage groups was compared. In the case of store patronage, the r e s u l t s showed that the: most s i g n i f i c a n t between group dif f e r e n c e s and the most accurate p r e d i c t i o n s of group membership were achieved using the d i s p o s i t i o n a l v a r i a b l e s . The status, 198. fashion and quality scales emerged as p a r t i c u l a r l y e f f e c t i v e d i s c r i m i n a -t o r s . The r e s u l t s from combining the three d i f f e r e n t sets of p r e d i c t o r s confirmed the s i g n i f i c a n c e o f the d i s p o s i t i o n a l v a r i a b l e s and the impro-vement i n p r e d i c t i v e accuracy a t t r i b u t a b l e to t h e i r i n c l u s i o n . The r e s u l t s of the analyses of shopping centre patronage were i n marked contrast. The type of centre selected by the respondents was c l e a r l y best predicted by measures of t h e i r l o c a t i o n r e l a t i v e to f a c i l i -t i e s of varying s i z e . In general, the r e s u l t s provided strong confirma-t i o n of the importance placed upon l o c a t i o n a l v a r i a b l e s i n e x i s t i n g models of shopping behaviour. More s p e c i f i c a l l y , they confirmed the v a l i d i t y of regarding s i z e of centre as a surrogate measure of a t t r a c -tiveness, while, demonstrating the w i l l i n g n e s s of consumers to t r a v e l f u r t h e r to patronise l a r g e r centres. The c o n t r i b u t i o n of the d i s p o s i -t i o n a l and b i o g r a p h i c a l v a r i a b l e s to the explanation of c r i t e r i o n v a r i -ance was small by comparison as evidenced by the r e l a t i v e l y inaccurate p r e d i c t i o n s of group membership and the small number of v a r i a b l e s o f e i t h e r type entered i n t o the discriminant equations. The r e s u l t s of the regression analysis on distance travelled to shop provided f u r t h e r evidence of the strong i n f l u e n c e of l o c a t i o n a l v a r i a b l e s i n consumer s p a t i a l behaviour. In t h i s case, i t was almost i n e v i t a b l e that s i g n i f i c a n t r e l a t i o n s h i p s would emerge between distance t r a v e l l e d and l o c a t i o n r e l a t i v e to r e t a i l f a c i l i t i e s , s i n c e , of n e c e s s i t y , consumers located f u r t h e r from major r e t a i l opportunities have to t r a v e l further to shop. More s u r p r i s i n g however, was the n e g l i g i b l e c o n t r i b u -t i o n made to the regression equations by e i t h e r b i o g r a p h i c a l or dispo-s i t i o n a l v a r i a b l e s . I t was suggested that the r e l a t i v e l y confined 199. sampling area may have been a contributory f a c t o r i n suppressing the s i g n i f i c a n c e of non-locational v a r i a b l e s . Consumer Dispositions and S p a t i a l Behaviour The r e s u l t s of the analyses provide p a r t i a l support f o r the d i s -p o s i t i o n a l model of consumer s p a t i a l behaviour presented i n Chapter 1, V e r i f i c a t i o n of the model was dependent upon demonstrating that consu-mers d i f f e r i n g s i g n i f i c a n t l y i n t h e i r d i s p o s i t i o n s e x h i b i t d i f f e r e n t spa-t i a l behaviour. This r e l a t i o n s h i p was very c l e a r l y shown to hold i n the case of store patronage. Highly s i g n i f i c a n t d i f f e r e n c e s were found i n the d i s p o s i t i o n a l p r o f i l e s of consumers s e l e c t i n g d i f f e r e n t types of c l o t h i n g store. P r o f i l e d i f f e r e n c e s were most pronounced on the status, fashion and q u a l i t y s c a l e s . The r e l a t i o n s h i p between consumer d i s p o s i t i o n s and s p a t i a l behaviour was not c l e a r l y demonstrated i n the case of shopping centre patronage. The d i s p o s i t i o n a l p r o f i l e s of downtown, secondary and .tertiary centre shoppers were, i n many cases, not s i g n i f i c a n t l y d i f f e r e n t , and the p r e d i c t i o n of group membership was correspondingly poor. The r e s u l t s suggest that the r e l a t i o n s h i p between consumer d i s p o s i -tions and r e t a i l patronage may be scale dependent. A t the macro s c a l e , where the consumer i s choosing between d i f f e r e n t shopping centres the influence of d i s p o s i t i o n a l v a r i a b l e s i s not pronounced. The l o c a t i o n of the consumer r e l a t i v e to r e t a i l f a c i l i t i e s of varying s i z e assumes primary importance. At the micro s c a l e , when choosing between d i f f e r e n t stores, the consumer's d i s p o s i t i o n s assume f a r greater s i g n i f i c a n c e and i n f a c t emerge as the most accurate p r e d i c t o r s of r e t a i l patronage. 2 0 0 . The varying s i g n i f i c a n c e of d i f f e r e n t v a r i a b l e s at d i f f e r e n t scales suggests tha t a patronage d e c i s i o n may be the r e s u l t of a two stage search and evaluation process: the f i r s t stage, at a macro scale i n v o l v i n g the i s o l a t i o n of one or more f e a s i b l e a l t e r n a t i v e shopping areas p r i m a r i l y on the basis of l o c a t i o n a l considerations; the second stage, at a micro scale i n v o l v i n g the s e l e c t i o n of a store within one of the preselected shopping areas with d i s p o s i t i o n a l factors being fundamental i n d i r e c t i n g the evaluation of a l t e r n a t i v e s and the choice of a p r e f e r r e d store. C l e a r l y , an examination of t h i s process and of the r e l a t i o n s h i p between store and shopping centre choices i s an import-ant area f o r future research. Implications f o r Market Research In an applied context, the r e s u l t s of t h i s study have im p l i c a t i o n s f o r market research. In chapter 1 (p. 43), i t was suggested that pre-vious attempts to p r e d i c t consumer behaviour on the basis of consumer values have been unproductive because of a naive a p p l i c a t i o n of stand-ard p e r s o n a l i t y t e s t s . A more f r u i t f u l approach was seen to be through the development of t e s t instruments s p e c i f i c a l l y designed f o r consumer research. This approach was followed i n t h i s study i n developing f i v e scales to measure consumer d i s p o s i t i o n s . The question then a r i s e s of the usefulness of these v a r i a b l e s i n market research. The r e s u l t s r a i s e some doubt as to the u t i l i t y of d i s p o s i t i o n a l measures for d e f i n i n g r e t a i l segments. The f i v e s c a l e s f a i l e d to d i s -criminate accurately between the shopping centre patronage groups, and, with respect to the p r e d i c t i o n of store patronage, t h e i r performance was 201. only s l i g h t l y superior to that achieved using b i o g r a p h i c a l v a r i a b l e s . When consideration i s given to the r e l a t i v e l y complex task of develop-ing r e l i a b l e and v a l i d d i s p o s i t i o n a l measures, the appeal of the more e a s i l y obtainable l o c a t i o n a l and bi o g r a p h i c a l v a r i a b l e s i s accentuated. I t i s not s u r p r i s i n g therefore that market researchers, faced with the need to employ the most expedient method f o r d i s t i n g u i s h i n g between consumer groups, have commonly r e l i e d upon l o c a t i o n a l and b i o g r a p h i c a l v a r i a b l e s i n developing trade area and market-segmentation models. Given these strong reservations, the f a c t remains that d i s p o s i t i o n a l measures c o n s t i t u t e a p o t e n t i a l l y r i c h source of information f o r the r e t a i l e r . They are d i s t i n c t from l o c a t i o n a l and bi o g r a p h i c a l v a r i a b l e s i n providing information which has d i r e c t a p p l i c a t i o n i n the formula-t i o n of a pr e c i s e marketing strategy. The c r u c i a l issue i s whether the b e n e f i t gained from t h i s type of information more than o f f s e t s the high costs of data a c q u i s i t i o n . The incr e a s i n g emphasis being placed upon "psychographics" i n marketing research suggests that i n many cases the costs are being accepted. Limitations of the Present Study and Proposals f o r Future Research I t i s important to recognize the l i m i t a t i o n s of the present study i n seeking to generalise on the basis of the r e s u l t s . One p o s s i b l e l i m i t a t i o n a r i s e s from the c l a s s i f i c a t i o n of the c r i t e r i o n v a r i a b l e s i n the main analyses. The need to achieve adequate group s i z e s required that the ana l y s i s be based upon classes of store and shopping centre rather than i n d i v i d u a l stores and centres. I t follows that the v a l i d i t y of the r e s u l t s depends upon the v a l i d i t y of the c l a s s i f i c a t i o n schemes 202. employed. How f a r t h i s represents a serious l i m i t i n g f a c t o r i s uncert-a i n . I t would c e r t a i n l y seem necessary i n subsequent studies to d i s -aggregate the analysis and examine group c h a r a c t e r i s t i c s at the l e v e l of s p e c i f i c stores and centres i n more d e t a i l then the present data permitted. In t h i s study, the r e l a t i o n s h i p s between consumer c h a r a c t e r i s t i c s and s p a t i a l behaviour were examined with respect to one type of pur-chase, namely c l o t h i n g . Further research i s needed to e s t a b l i s h the g e n e r a l i t y of the findings with respect to other goods. I t i s hoped that the f i v e d i s p o s i t i o n s i d e n t i f i e d here w i l l prove a p p l i c a b l e i n other purchasing s i t u a t i o n s . A systematic approach to s c a l e construc-t i o n was adopted with the s p e c i f i c aim of developing a research i n s t r u -ment which might have a p p l i c a b i l i t y beyond a s i n g l e study. In t h i s regard, subsequent research could be p r o f i t a b l y d i r e c t e d toward the development of a consumer response inventory designed to f u l f i l l a function i n consumer research analogous to the r o l e performed by e x i s t i n g p s y c h o l o g i c a l inventories i n p e r s o n a l i t y research. Another important area f o r f u r t h e r research i n v e s t i g a t i o n centres on the r e l a t i o n s h i p between consumer d i s p o s i t i o n s and s p e c i f i c a t t i -tudes toward the set of s p a t i a l a l t e r n a t i v e s . A subsequent attempt should be made to e x p l i c a t e t h i s r e l a t i o n s h i p by obtaining appropriate a t t i t u -d i n a l measures, hence permitting an examination of the mediating function of the a t t i t u d i n a l component. As defined i n t h i s study, consumer d i s p o s i t i o n s represent middle l e v e l p s y c h o l o g i c a l v a r i a b l e s which are on the one hand more glo b a l then a t t i t u d e s and on the other l e s s pervasive than p e r s o n a l i t y t r a i t s . As 2 0 3 . i n d i c a t e d , one area f o r future research i s to explore the l i n k s between d i s p o s i t i o n s and behaviour i n the form of a t t i t u d e s . Another equally i n t e r e s t i n g avenue of enquiry would be to move i n the opposite d i r e c t i o n and trace the connections between consumer d i s p o s i t i o n s and p e r s o n a l i t y t r a i t s . BIBLIOGRAPHY Adler, L. and I. Crespi (Eds.) (1966) Attitude Research at Sea. American Marketing A s s o c i a t i o n . Adler, L. and I. Crespi (Eds.) (1968) Attitude Research on the Rocks. American Marketing A s s o c i a t i o n . Armstrong, J . (1967) "Derivation of Theory by Means of Factor Analysis or Tom Swift and His E l e c t r i c Factor Analysis Machine", The American Statistician, v o l . 25, no. 5, pp. 17-21. Bannister, D. and J. Mair (1968) The Evaluation of Personal Constructs. Academic Press. Baumol, W. and E. Ide (1956) "Variety i n R e t a i l i n g " , Management Science, v o l . 3, no. 1, pp. 93-101. Becker, W. (1967) Retail Store Image: A Conceptual Approach. M.B.A. Thesis, U n i v e r s i t y of B r i t i s h Columbia. Bern, D. (1970) Beliefs, Attitudes and Human Affairs. Belmont: Brooks/ Cole. Berry, B. (1963) Commercial Structure and Commercial Blight. Univ-e r s i t y of Chicago, Department of Geography Research Paper No. 85. Berry, B. (1967) Geography of Market Centres and Retail Distribution. Englewood C l i f f s : Prentice H a l l . Berry, B. and W. Garrison (1958) "Recent Development of Central Place Theory", Papers and Proceedings of the Regional Science Association, 4, pp. 107-20. Berry, L. (1968) The Components of Department Store Image, D . B . A . Thesis, Arizona State U n i v e r s i t y . Blalock, H. (1960) Social Statistics. New York: McGraw H i l l . Briggs, R. (1973) "Urban Cognitive Distance", i n Downs, R. and D. Stea (Eds.) Image and Environment. Chicago: Aldine, pp. 361-88. Brbdy, R. and S. Cunningham (1968) "Personality Variables and the Consumer Decision Process", Journal of Marketing Research, v o l . 5, pp. 50-57. - 204 -Brown, F. and G. F i s k (1965) "Department Stores and Discount Houses: Who Dies Next?", Journal of Retailing, v o l . 41. Brown, s. (1974) An Investigation of the Determinants of Consumer Spatial Behaviour. Doctoral D i s s e r t a t i o n , U n i v e r s i t y of A l b e r t a . Brush, J . and H. Gauthier (1968) Service Centres and Consumer Trips. Research Paper #13, Department of Geography, U n i v e r s i t y of Chicago. Bucklin, L. (1967) Shopping Patterns in an Urban Area, i n s t i t u t e of Business and Economic Research, U n i v e r s i t y of C a l i f o r n i a , Berkeley. Bucklin, L. (1967a) "The Concept o f Mass i n Intra-Urban Shopping", Journal of Marketing, v o l . 31, pp. 37-42. Bush, R. and F. Mosteller (1955) Stochastic Models for Learning. New York: Wiley. Carman, J . (1965) The Application of Social Class in Market Segmenta-tion. I n s t i t u t e f o r Business and Economic Research, U n i v e r s i t y of C a l i f o r n i a , Berkeley. C h r i s t a l l e r , W. (1966) The Central Places of Southern Germany (trans. by C. Baskin). Englewood C l i f f s : Prentice H a l l . Clark, W. and G. Rushton (1970) "Models of Intra-Urban Consumer Behaviour and Their Implications f o r Central Place Theory", Economic Geography , 46, pp. 486-97. Claus, J . and W. Hardwick (1972) The Mobile Consumer. Don M i l l s : C o l l i e r Macmillan. Cohen, J . , Fishbein, M and Ahtola, 0. (1972) "The Nature and Use of Expectancy-Value Models i n Consumer A t t i t u d e Research", Journal of Marketing Research, v o l . i x , pp. 456-60. Converse, P. (1948) Consumer Buying Habits in Selected South Central I l l i n o i s Communities. Business Studies, No. 6, Urbana: Un i v e r s i t y of I l l i n o i s . Cooley, W. and P. Lohnes (1971) Multivariate Data Analysis. New York: Wiley. Coembs, C. (1950) "Psychological S c a l i n g Without a Unit of Measurement", Psychological Review; v o l . 57, pp. 148-58. Craik, K. (1969) "Assessing Environmental D i s p o s i t i o n s " , Paper presented at the Annual Meeting of the American Psychological Asso-c i a t i o n , Washington. Craik, K. (1970) "The Environmental Dispositions of Environmental Decision-Makers", Annals of the Ameviean Academy of P o l i t i c a l and Social Science, 389, pp. 87-94. Cronbach, L. and P. Meehl (1955) "Construct Validity in Psychological Tests", Psychological B u l l e t i n , 52, pp. 281-302. Cyert, R. and J. March (1963) A Behavioral Theory of the Firm. Englewood C l i f f s : Prentice Hall. Dalrymple, D. and D. Thompson (1969) R e t a i l i n g , An Economic View. New York: Free Press. Day, G. (1972) "Evaluating Models of Attitude Structure", Journal of Marketing Research, vol. 9, pp. 279-86. Doddridge, P. (1972) "Dimensions of Alternatives in Spatial Choice Behavior", Proceedings of International Geographical Union, pp. 538-41. Dornoff, D. and D. Tatham (1972) "Congruence Between Personal Image and Store image", Journal of the Market Research Society, vol. 14, no. 1. Downs, R. (1970) " C o g n i t i v e Structure of an Urban Shopping Centre", Environment and Behavior-, vol. 2, no. 1, pp. 13-39. Downs, R. (1970a) "Geographic Space Perception", in Board, C., et al. (Eds.) Progress in Geography, 2, London: Edward Arnold, pp. 65-108. Downs, R. and D. Stea (Eds.) (1973) Image and Environment: Cognitive Mapping and Spatial Behavior. Chicago: Aldine. Edwards, A. (1957) Edwards' Personal Preference Schedule Manual. Psychological Corporation. Engel, J. Kollat, D. and R. Blackwell (1968) Consumer Behavior. New York: Holt, Rinehart and Winston. Evans, F. (1959) "Psychological and Objective Factors in the Prediction of Brand Choice", Journal of Business, vol. 32, pp. 340-69. Evans, F. (1968) "Ford versus Chevrolet: Park. Forest Revisited", Journal of Business, vol. 41, pp. 445-59. Farley, J. and L. Ring (1970) "An Empirical Test of the Howard-Sheth Model of Buyer Behavior", Journal of Marketing Research, vol. 7, pp. 427-438. F i s h b e i n , M. (1963) "An Investigation of the•Relationships Between B e l i e f s about an Object and the At t i t u d e Toward that Object", Human Relations, v o l 16, pp. 233-40. Fishbein, M. (1965) "A Consideration of B e l i e f s , A t t i t u d e s and Their Relationship", i n Ste i n e r , I. and M. Fishbein (Eds.) Current Studies in Social Psychology. New York: Holt, Rinehart and Winston, pp. 107-20. Fishbein, M. (1967) "A Consideration of B e l i e f s and Their Role i n At t i t u d e Measurement", i n Fishbein, M. (Ed.) Readings in Attitude Theory and Measurement. New York: Wiley, pp. 257-66. Fishbein, M. (1967a) "A Behavior Theory Approach to the Relations Between B e l i e f s About an Object and the Att i t u d e Toward the Object", i n Fishbein, M. (Ed.) Readings in Attitude Theory and Measurement. New York: Wiley, pp. 389-400. Fishbein, M. (1972) "The Search f o r A t t i t u d i n a l - B e h a v i o r a l Consist-ency", i n Cohen, J. (Ed.) Behavioral Science Foundations of Consumer Behavior. New York: Free Press,pp. 242-52. F i s h e r , R. (1936) "The Use of M u l t i p l e Measurements i n Taxonomic Problems", Ann. Eugenics, v o l . 7, pp. 179-99. F i s k , G. (1961) "A Conceptual Model f o r Studying Customer Image", Journal of Retailing, v o l . 37, no. 4, pp. 1-8. Gayler, H. (1974) Consumer Spatial Behavior and Its Relation to Social Class and Family Status in Metropolitan Vancouver. Doctoral D i s s e r t a t i o n , U n i v e r s i t y of B r i t i s h Columbia. Golledge, R. (1967) "Conceptualizing the Market Decision Process", Journal of Regional Science, 7 (no. 2 Supplemental), pp. 239-258. Golledge, R. (1969) "Geographical Relevance of Some Learning Theories", i n Cox, K. and R. Golledge (Eds.) Behavioral Problems in Geography. Evanston: Northwestern U n i v e r s i t y . Golledge, R. (1970) "Process Approaches to the Analysis of Human Sp a t i a l Behavior". Ohio State U n i v e r s i t y , Department of Geography, Discussion Paper #16. Golledge, R. and L. Brown (1967) "Search, Learning and the Market Decision Process", Geografiska Annaler, 49B1, pp. 116-24. 208. Golledge, R. and Rushton, G. (1970) "Multidimensional Sc a l i n g : Review and Geographical A p p l i c a t i o n s " , Paper presented f o r I.G.U. Commission on Quantitative Geography. Green, P. and F. Carmone (1970) Multidimensional Scaling and Related Techniques in Marketing Analysis. Boston: A l l y n and Bacon. Green, P. and V. Rao (1972) Applied Multidimensional Scaling: A Comparison of Approaches and Algorithms. New York: Holt, Rinehart and Winston. G u i l f o r d , J . (1965) Psychometric Methods. New York: McGraw H i l l . Haines, G. (1969) Consumer Behavior: Learning Models of Purchasing. New York: Free Press. Haley, R. (Ed.) (1972) Attitude Research in Transition. American Marketing A s s o c i a t i o n . Hansen, F. (1972) Consumer Choice Behavior. New York: Free Press. Harrison, J . and P. Sarre (1971) "Personal Construct Theory i n the Measurement of Environmental Images", Environment and Behavior, v o l . 3, no. 4, pp. 351-374. Harvey, D. (1969) "Conceptual and Measurement Problems i n the Cogni-tive-Behavioral Approach to Location Theory", i n Cox, K. and R. Golledge (Eds.) Behavioral Problems in Geography: A Symposium. Evanston: Northwestern Uni-v e r s i t y , pp. 35-68. Hess, C. (1966) "Geographical Aspects of Consumer Behavior i n the R e t a i l Service Area o f Greater Alborg", Geografisk Tiddsskrift, v o l . 65, pp. 1-26. Howard, J . and J . Sheth (1969) The Theory of Buyer Behavior. New York: Wiley. Hudson, R. (1970) "Personal Construct Theory, Learning Theories and Consumer Behavior", U n i v e r s i t y of B r i s t o l , Department of Geography, Seminar Paper A21. Hudson, R. (1971) "Towards a Theory of Consumer S p a t i a l Behavior", U n i v e r s i t y of B r i s t o l , Department of Geography, Seminar Paper A22. Huff, D. (1962) Determination of Intra-Urban Retail Trade Areas. Los Angeles: Real Estate Research Program, U n i v e r s i t y of C a l i f o r n i a . H u l l , C. (1962) A Behavior System. New York: Wiley. Hurst, M. (1972) A Geography of Economic Behavior. Duxbury Press. J a n i s , I. (1959) "Motivational Factors i n the Resolution of D e c i s i o n a l C o n f l i c t s " , i n Jones, R. (Ed.) Nebraska Symposium on Motivation. L i n c o l n : Nebraska U n i v e r s i t y Press. Jonassen, c. (1955) The Shopping Centre versus Downtown: Motivation Research on Shopping Habits and Attitudes in Three Cities. Bureau of Business Research, Ohio State Univ. Ke l l e y , E. (1956) Shopping Centres: Locating Controlled Regional Centres. Saugatuck, Conn.: The Eno Foundation f o r Highway T r a f f i c C o n t r o l . K e l l y , G. (1955) The Psychology of Personal Constructs. New York: Norton. Kenny-Levick, C. (1969) "Consumer Motivations: Examples from the Grocery Trade", British Journal of Marketing. (Spring) pp. 2-8. Ker l i n g e r , F. (1964) Foundations of Behavioral Research. New York: Holt, Rinehart and Winston. King, C. and D. T i g e r t (Eds.) (1971) Attitude Research Reaches New Heights. American Marketing A s s o c i a t i o n . Koponen, A. (1960) "Personality C h a r a c t e r i s t i c s of Purchasers", Journal of Advertising Research, v o l . 1, pp. 6-12. Kunkel, J . and L. Berry (1969) "A Behavioral Conception of R e t a i l Image", Journal of Marketing, v o l . 32, pp. 21-27. Leigh, R. (1965) Specialty Retailing: A Geographic Analysis. Vancouver: Tantalus. Levy, S. (1966) " S o c i a l Class and Consumer Behavior", i n Newman, J . (Ed.) On Knowing the Consumer. New York: Wiley, pp. 146-60. Lewin, K. (1936) A Dynamic Theory of Personality. New York: McGraw H i l l . Lewin, K. (1936a) Principles of Topological Psychology. New York: McGraw H i l l . Lewin, K. (1951) Field Theory and Social Science. New York: Harper. L i k e r t , R. (1932) "A Technique f o r the Measurement of A t t i t u d e s " , Archives of Psychology, no. 140, pp. 44-53. Margulis, H. (1972) "A Note on Consumer Choice, B e l i e f Perception and the Intra-Urban Journey-to-Shop", Proceedings of the International Geographical Union, pp. 566-69. Martineau, P. (1958) " S o c i a l Classes and Spending Behavior", Journal of Marketing, v o l . 23, no. 2, pp. 121-30. Martineau, P. (1958a) "The Pe r s o n a l i t y of the R e t a i l Store", Harvard Business Review, v o l . 36, no. 1, pp. 47-55. Massy, W. (1965) "Discriminant Analysis of Audience C h a r a c t e r i s t i c s " , Journal of Advertising Research, v o l . 5, pp. 39-48. Massy, W., Frank, R., and T. Lodahl (1968) Purchasing Behavior and Personality Attributes. P h i l a d e l p h i a : U n i v e r s i t y of Pennsylvania Press. McKehnie, G. (1972) A Study of Environmental Life-Styles. Doctoral D i s s e r t a t i o n , U n i v e r s i t y of C a l i f o r n i a , Berkeley. Murdie, R. (1965) " C u l t u r a l Differences i n Consumer T r a v e l " , Economic Geography, v o l . 41, pp. 211-33. Murphy, P. (1970) A Study of the Influence of Attitude As a Behavioral Parameter on the Spatial Choice Patterns of Consumers. Doctoral D i s s e r t a t i o n , Ohio State U n i v e r s i t y . Murphy, P. and R. Golledge (1972) "Attitude as a V a r i a b l e i n Urban Geographic Research", Ohio State U n i v e r s i t y , Department of Geography Seminar Paper. Ni c o s i a , F. (1966) Consumer Decision Processes. Englewood C l i f f s : P r e ntice H a l l . Nunnally, J . (1967) Psychometric Theory. New York: McGraw H i l l . Osgood, C., Suci, G. and Tannenbaum, P. (1957) Measurement of Meaning. Urbana: U n i v e r s i t y o f I l l i n o i s Press. O v e r a l l , J . and c. K l e t t (1972) Applied Multivariate Analysis. New York: McGraw H i l l . Pahl, R. (1965) "Urbs i n Rure: The Metropolitan Fringe i n Hert f o r d -s h i r e " , London School of Economics Geographical Papers no. 2. Patterson, M. (forthcoming) Factorial Ecology of Metropolitan Vancouver. M.A. Thesis, U n i v e r s i t y of B r i t i s h Columbia. 211. R e i l l y , W. (1931) The Law of Retail Gravitation. Austin: U n i v e r s i t y of Texas. Rich, S. (1963) Shopping Behavior of Department Store Customers. Boston: Harvard U n i v e r s i t y Press. Rich, S. and S. J a i n (1968) " S o c i a l Class and L i f e - C y c l e as Predictors of Shopping Behavior", Journal of Marketing Research, v o l . 5, pp. 41-49. Rogers, C. (1965) Client-Centred Therapy. Boston: Houghton M i f f l i n . Rokeach, M. (1968) Beliefs, Attitudes and Values. San Francisco: Jossey-Bass Inc. Rokeach, M. (1969) "The Role of Values i n P u b l i c Opinion Research", Public Opinion Quarterly, v o l . 32,pp. 547-59. Romney, A., Nerlove, S. and Shepart R. (1972) Multidimensional Scaling: Theory and Applications in the Behavioral Sciences. (2 vols.) New York: Seminar Press. Rosenberg, M. (1956) "Cognitive Structure and A t t i t u d i n a l A f f e c t " , Journal of Abnormal and Social Psychology, v o l . 53, pp. 367-72. Rosenberg, M. (1960) "A S t r u c t u r a l Theory of A t t i t u d e Dynamics", Public Opinion Quarterly, v o l . 24, JJ>. 319-40. Rosenberg, M. (1965) "Inconsistency Arousal and Reduction i n A t t i t u d e Change", i n Stainer, I. and M. Fishbein (Eds.) Current Studies in Social Psychology. New York: Holt, Rinehart and Winston, pp. 121-34. Rushton, G. (1969) "Temporal Changes i n Space Preference Structures", Proceedings of the Association of American Geographers, 2, pp. 129-132. Rushton, G. (1970) "Preference and Choice i n D i f f e r e n t Environments", Proceedings of the Association of American Geographers , 3, pp. 146-50. Rushton, G. (1971) "Postulates of Central Place Theory and the Proper-t i e s of Central Place Systems", Geographical Analysis, 8, pp. 140-56. Sampson, P. (1972) "Using the Repertory G r i d Test", Journal of Marketing Research, v o l . 9, pp. 78-81. Sheth, J. (1967) "A Review of Buyer Behavior", Management Science, v o l . 13, no. 12, B-718-B-786. 212. Spence, K. (1951) "Theoretical Interpretations of Learning", i n Stevens, S. (Ed.) Handbook of Experimental Psychology. New York: Wiley. Stephenson, P. (1969) " I d e n t i f y i n g the Determinants of R e t a i l Image", Journal of Marketing, v o l . 33, no. 3, pp. 57-61. Stone, G. (1954) "C i t y Shoppers and Urban I d e n t i f i c a t i o n " , American Journal of Sociology, v o l . 59, pp. 36-45. Thomas, K. (Ed.) (1972) Attitudes and Behavior. Harmondsworth: Penguin Books. Thompson, D. (1966) "Future D i r e c t i o n s i n R e t a i l Area Research", Economic Geography, v o l . 42, no. 1, pp. 1-18. T i t t l e , C. and R. H i l l (1967) "Attitude Measurement and P r e d i c t i o n of Behavior: An Evaluation of Conditions and Measurement Techniques", So dome try, v o l . 30, pp. 199-213. Tolman, E. (1932) Purposive Behavior in Animals and Man. New York: Appleton-Century-Crofts. Tolman, E. (1952) "A Cognition-Motivation Model", Psychological Review, v o l . 59. Warner, L. Meeker, M. and K. E e l l s (1949) Social Class in America. Chicago: S o c i a l Research Inc. Wells, W. (1966) "General P e r s o n a l i t y Tests and Consumer Behavior", i n Newman, J . (Ed.) On Knowing the Consumer. New York: Wiley. White, I. (1966) "The Perception of Value i n Products", i n Newman, J. (Ed.) On Knowing the Consumer. New York: Wiley, pp. 90-106. Wi l k i e , W. and E. Pessemeir (1973) "Issues i n Marketing's Use of Mu l t i - A t t r i b u t e A t t i t u d e Models", Journal of Marketing Research, v o l . x, pp. 428-41. Winkel, G. (1971) "Theory and Method i n Behavioral Geography", Paper presented at Canadian Association of Geographers Annual Meeting. Waterloo, Ontario. Zaltman, G., Paison, C , and R. Angelmar (1973) Metatheory and Consumer Research. New York: Holt, Rinehart and Winston. APPENDIX A: CONSUMER SHOPPING SURVEY CONSUMER SHOPPING SURVEY INTERVIEW NO. . . INTERVIEWER . . . DATE OP INTERVIEW SECTION 1. DISPOSITIONS The following 60 statements are designed to find out the basis on which you choose the stores where you buy your major items of clothing (value over $10.00). Read each statement and indicate how i t corresponds to your personal opinion using the response categories provided. SD - Strongly Disagree D - Disagree N - Neutral or Don't Know A - Agree SA - Strongly Agree Record your answers by circling the category below the statement which seems to reflect your opinion most accurately. Don't be concerned i f some statements seem similar to ones you have previously answered. Remember throughout that your answers should reflect your opinions when buying clothes for yourself. 216. SD»Strongly Disagree, D=Dieagree, N=Neutral, A=Agree, SA=Strongly Agree 1. I enjoy buying expensive clothes. SD D N A SA 1 2 3 4 5 2. I shop as infrequently as possible. SD D N A SA 1 2 3 4 5 3. Lighting effects a.nd rock music create a. store atmosphere which appeals to me. SD D N A SA 1 2 3 4 5 4. I dori't patronise stores which have a reputation for low prices. SD D N A SA 5 4 3 2 1 5. I'm more i n c l i n e d to buy expensive items infrequently than to make frequent cheap purchases. SD D N A SA 1 2 3 4 5 6. I l i k e to shop i n elegant surroundings. SD D N A SA 1 2 3 4 5 7. I regard shopping as a necessity rather tha.n a pleasure. SD D N A SA 1 2 3 4 5 8. I . l i k e to dress i n the latest fashions. SD D N A SA 1 2 3 4 5 9. I enjoy bargain hunting. SD D N A SA 1 2 3 4 5 10. I take care to ensure tha,t the. garments I buy are well-made. 11. I regularly read high fashion magazines. SD D N A SA 1 2 3 4 5 12. I'm w i l l i n g to expend s. l o t of time and miles to get what I want. SD D N A. SA 5 4 3 2 1 1 3 . I would not consider shopping i n a trendy boutique. SD D N A SA 5 4 3 2 1 14» I huy at stores which undersell t h e i r competitors. SD D N A SA 1 2 3 4 5 15. I'm w i l l i n g to sa.crifice quality for low prices. SD D N A SA .5 4 3 2 1 16. Shopping at exclusive stores i s beyond my means. SD D N A SA 5 4 3 2 1 17. I'm quite w i l l i n g to travel to the other side of the c i t y to shop. SD D IT A SA 5 4 3 2 1 18. I don't see myself as one of the younger set. SD D 11 A SA 5 4 3 2 1 19* It doesn't worry me to know tha.t other people own clothes i d e n t i c a l to mine. SD D N A SA 1 2 3 4 5 SD D N A SA 1 2 3 4 5 (1-10) (11-19) 217. STtaStrongly Disagree, ]>=Disagreet N=Neutral, A=Agreef SA=Strongly Agree 20. I'm put o f f by l o t s of signs ?nd advertisements i n a s t o r e . SD D N A SA 1 2 3 4 5 21. I f e e l uneasy i n an e x c l u s i v e s t o r e . S3) D N A SA 5 4 3 2 1 22. I don't regard distan.ce as ve r y import°nt when choosing where to shop. SD I) N A SA 5 4 3 2 1 23. I shop at s t o r e s which appeal mainly to the youncer age group. SD D N A SA 1 2 3 4 5 24. I'm not a t t r a c t e d to s t o r e s l o c a t e d i n low income areas. SD D N A SA 5 4 3 2 1 25. I'm always s u s p i c i o u s o f low p r i c e s . SD D IT A SA 1 2 3 4 5 26. I'm not i n t e r e s t e d i n what happens i n the world of high f a s h i o n . SD D N A SA 5 4 3 2 1 2 7 . Shopping forms one of my main l e i s u r e l e i s u r e a c t i v i t i e s . SD D N A SA 5 4 3 2 1 28. I a v o i d s t o r e s which make use of l i g h t i n g and sound e f f e c t s . SD D N A SA 5 4 3 2 1 29. I don't mind shopping i n the poorer parts o f the c i t y . SD D N A SA 1 2 3 4 5 30. I would d e s c r i b e myself as a d i s c r i m i n a t i n g shopper. SD D N A SA 1 2 3 4 5 31. I avo i d s t o r e s which look or sound expensive. SD D N A SA 5 4 3 2 1 32. I only p a t r o n i s e s t o r e s which ?re e a s i l y a c c e s s i b l e t o my home. SD D TJ A SA 1 2 3 4 5 33. I p r e f e r s t o r e s which ere qu i e t and p e a c e f u l . SD D H A SA 5 4 3 2 1 34- I l i k e to f e e l t h a t my patronage i s r e a l l y a p p r e c i a t e d . SD D N A SA 5 4 3 2 1 35* I'm a t t r a c t e d by t a s t e f u l d i s p l a y s and decor. SD D N A SA 1 2 3 4 5 36. I a p p r e c i a t e h i g h l y p e r s o n a l i s e d s e r v i c e . S3) D N A SA 1 2 3 4 5 37. The n i c e t i e s of s t y l e end. f a s h i o n are not important t o me. SD D N A SA 1 2 3 4 5 38. S t o r e s which c a t e r f o r the avant garde appeal to me. SD D N A SA 1 2 3 4 5 39» I co n s i d e r i t e s s e n t i a l that the merchandise be n e a t l y arranged. SD D N A SA 5 4 3 2 1 (20-29) / ™ , M rn n rn 1 n— 1 — 1 SD=.Strongly Disagree, D=Disagree, N=Neutral, A=Agree, SA=Strongly Agree 218. 40. My f i r s t concern i s with the quality of what I "buy. SD D N A SA 1 2 3 4 5 41. I patronise specialty stores even though I could possibly buy similar goods more cheaply elsewhere. SD D N A SA 1 2 3 4 5 42. I rarely shop outside my home neighbourhood. SD D N A SA 1 2 3 4 5 43. My taste i n clothes i s conservative. SD D N A SA 5 4 3 2 1 44. 45. 46. 47. 43. 49. 50. I f i n d i t hard to r e s i s t sales offers. SD D N A SA 1 2 3 4 5 The stores I patronise base th e i r reputation on quality. SD D N A SA 1 2 3 4 5 Pew people can afford to shop where I do. SD D N A SA 1 2 3 4 5 I enjoy looking around stores even though I have no intention of "buying anything. SD D N A SA 5 4 3 2 1 I l i k e to shop i n bright and colourful surroundings. SD D N A SA 1 2 3 4 5 It's essential to me that the sales s t a f f be well-trained. SD D N A SA 5 4 3 2 1 It doesn't worry me to buy mass-produced clothes. SD D N A SA 51. I l i k e to shop where the sales s t a f f know me end appreciate my needs. SD D N A SA 1 2 3 4 5 52. I spend considerable time shopping around before making a purcha.se. SD D N A SA 5 4 3 2 1 53. I would enjoy shopping on Carnaby Street. SD D N A SA 1 2 3 4 5 54« I f garments i n the store window aren't priced I wont go i n . SD D N A SA 1 2 3 4 5 55« I'ro not deterred by impolite personnel, SD D N A SA 5 4 3 2 1 56. I l i k e to know that what I buy can't be bought for less elsewhere. SD D N A SA 5 4 3 2 1 57* I would be quite content to do a l l my shopping by phone. SD D N A SA 1 2 3 4 5 58. I'm attracted to stores which make shopping exciting. SD D N A SA 1 2 3 4 5 59« I prefer stores which "base their appeal on low prices. SD D N A SA 1 2 3 4 5 60. I prefer to buy several cheaper items than one expensive one. SD D N A SA 5 4 3 2 1 (40-50) (51-60) SECTION 2 . PAST CLOTHING PURCHASES 219. PLEASE TELL ME ABOUT YOUR MOST RECENT CLOTHING PURCHASE. Pla AT WHAT STORE WAS THE PURCHASE MADE? P2» WHAT TYPE OP STORE IS THAT? HIGHER PRICED SPECIALTY ( l ) MEDIUM PRICED SPECIALTY (2) "BOUTIQUE" (3) REGULAR DEPARTMENT STORE (4) DEPARTMENT STORE - BUDGET FLOOR (5) BUDGET PRICED STORE (6) P3a WHERE IS THE STORE LOCATED? P4a WHERE DID THIS PURCHASING TRIP ORIGINATE? HOME (l) WORK (2) OTHER (specify) . P5a WHAT IS THE HUNDRED BLOCK ADDRESS OP THAT PLACE? P6a HOW PAR IS IT PROM THERE TO THE STORE? MILES P7» HOW DID YOU TRAVEL TO THE STORE? CAR BUS WALK OTHER ( 1 ) (2) (3) (4) P8a WHAT WAS THE MAIN PURPOSE OP THIS TRIP? TO PURCHASE CLOTHING TO MAKE OTHER SHOPPING PURCHASES BUSINESS PURPOSES SOCIAL PURPOSES RECREATIONAL PURPOSES P9a HOW MANY CLOTHING PURCHASES VALUED OVER $10.00 HAVE YOU MADE AT THIS STORE IN THE LAST 2 YEARS? ONE ( 1 ) TWO TO FIVE (2) SIX TO TEN (3) MORE THAN TEN (4) (1) (2) (3) (4) (5) • • • • • 1— i — r SECTION 2. PAST CLOTHING PURCHASES PLEASE TELL ME ABOUT YOUR NEXT MOST RECENT CLOTHING PURCHASE. Plb AT WHAT STORE WAS THE PURCHASE MADE? . P2b WHAT TYPE OP STORE IS THAT? HIGHER PRICED SPECIALTY (l) MEDIUM PRICED SPECIALTY (2) "BOUTIQUE" (3) REGULAR DEPARTMENT STORE (4) DEPARTMENT STORE - BUDGET FLOOR (5) BUDGET PRICED STORE (6) P3b WHERE IS THE STORE LOCATED? P4b WHERE DID THIS PURCHASING TRIP ORIGINATE? HOME (1) WORK (2) OTHER (specify) P5b WHAT IS THE HUNDRED BLOCK ADDRESS OP THAT PLACE? P6b HOW PAR IS IT PROM THERE TO THE STORE? MILES P7b HOW DID YOU TRAVEL TO THE STORE? CAR (1) BUS (2) WALK (3) OTHER (4) P8b WHAT WAS THE MAIN PURPOSE OP THIS TRIP? TO PURCHASE CLOTHING ^_ (1) TO MAKE OTHER SHOPPING PURCHASES (2) BUSINESS PURPOSES (3) SOCIAL PURPOSES (4) RECREATIONAL PURPOSES (5) P9V--HOW MANY CLOTHING PURCHASES VALUED OVER $10.00 HAVE YOU MADE AT THIS STORE IN THE LAST 2 YEARS? ONE (1) TOO TO FIVE (2) SIX TO TEN (3) MORE THAN TEN (4) SECTION 2. PAST CLOTHING PURCHASES PLEASE TELL ME ABOUT YOUR THIRD MOST RECENT CLOTHING PURCHASE. Pic AT WHAT STORE WAS THE PURCHASE MADE? P2C WHAT TYPE OP STORE IS THAT? HIGHER PRICED SPECIALTY (l) MEDIUM PRICED SPECIALTY (2) "BOUTIQUE" (3) REGULAR DEPARTMENT STORE (4) DEPARTMENT STORE - BUDGET FLOOR ( 5 ) BUDGET PRICED STORE ( 6 ) P3c WHERE IS THE STORE LOCATED? P4c WHERE DID THIS PURCHASING TRIP ORIGINATE? HOME (l) WORK (2) OTHER (specify) P5c WHAT IS THE HUNDRED BLOCK ADDRESS OF THAT PLACE? P6c HOW FAR IS IT FROM THERE TO THE STORE? MILES P7c HOW DID YOU TRAVEL TO THE STORE? CAR ( l ) BUS (2) WALK (3) OTHER (4) P8c WHAT WAS THE MAIN PURPOSE OF THIS TRIP? TO PURCHASE CLOTHING (l) TO MAKE OTHER SHOPPING PURCHASES ( 2 ) . BUSINESS PURPOSES (3) SOCIAL PURPOSES (4) RECREATIONAL PURPOSES _ (5) P9c HOW MANY CLOTHING PURCHASES VALUED OVER $10.00 HAVE YOU MADE AT THIS STORE IN THE LAST 2 YEARS? ONE (1) TWO TO FIVE (2) SIX TO TEN (3) MORE THAN TEN (4) P10. AT WHICH STORE DO YOU MAKE MOST OP YOUR CLOTHING PURCHASES? P l l . WHAT TYPE OP STORE IS THAT? HIGHER PRICED SPECIALTY MEDIUM PRICED SPECIALTY "BOUTIQUE" REGULAR DEPARTMENT STORE DEPARTMENT STORE - BUDGET FLOOR BUDGET PRICED STORE P12. WHERE IS THE STORE LOCATED? SECTION 3. BIOGRAPHICAL DATA. Bl. HOW OLD ARE YOU? B2. SEX OP RESPONDENT: MALE (l) FEMALE (2) B3. ARE YOU SINGLE OR MARRIED? SINGLE (1) MARRIED (2) WIDOWED ( 3 ) DIVORCED (4) B4. HOW MANY CHILDREN DO YOU HAVE LIVING AT HOME? B5. WHAT IS YOUR BEPO] TAXES ANNUAL HOUSEHOLD INCOME? below $4 ,000 (1) $4f ooo - 7,999 _ (2) 8,000 - 11,999 _ (3) 12,000 - 15,999 (4) 16,000 - 19,999 _ (5) 20,000 - 23,999 _ (6) 24,000 - 27,999 (7) 28,000 - 31,999 _ (8) over $32,000 (9) B6. WHAT IS THE HIGHEST LEVEL OP EDUCATION YOU HAVE REACHED? ELEMENTARY SCHOOL ONLY (l) SOME HIGH SCHOOL (2) HIGH SCHOOL GRADUATE ( 3 ) SOME POST SECONDARY (4) UNIVERSITY DEGREE ( 5 ) POST GRADUATE ( 6 ) B7. WHAT IS THE OCCUPATION OP THE HEAD OP HOUSEHOLD? B8. HUNDRED BLOCK ADDRESS OP RESPONDENT APPENDIX B: LETTER INTRODUCING THE SURVEY T H E UN IVERS ITY O F BRITISH C O L U M B I A V A N C O U V E R 8, C A N A D A DEPARTMENT OF GEOGRAPHY NoVeilber 1973, I would l i k e to i n v i t e you to take part i n a study of shopping patterns i n Vancouver which i s the basis of ray doctoral dissertation i n the Department of Geography at the University of B r i t i s h Columbia. The study involves c o l l e c t i n g information frora Vancouver and North Shore residents to determine where they shop f o r certain goods and for what reasons. Your residence may be selected as part of a sample of $00 households forming the basis of the interview programme. I f so, an authorized interviewer w i l l c a l l at your home within the next few days. I hope you w i l l be able to spare about 20 minutes to speak to hira. A l l information obtained w i l l remain s t r i c t l y c o n f i d e n t i a l . No names or s p e c i f i c addresses w i l l be used and the data w i l l never appear i n any way that could possibly i d e n t i f y you as an i n d i v i d u a l . I w i l l be happy to answer any questions you may have. My telephone number i s 434-7867. Yours sincerely, Martin Taylor APPENDIX C: INTERVIEWER'S MANUAL INTERVIEWER'S MANUAL 227. This manual has "been prepared to direct the oonduct of interviews and is designed to follow the questionnaire sequence of questions. This manual should he read over at least once before going into the field, and should be read in conjunction with the questionnaire. GENERAL POINTS - be polite at a l l times. - do not show offence at refusal to participate or to answer a particular question. - complete interview as quickly and efficiently as possible. - as far as possible ensure that a l l questions have been answered. - sincerely thank respondents for their cooperation. - remember you are the only contact with the respondents; the success of the project therefore depends on how well you handle the task of face-to-face interviewing. FIELD PROCEDURES - in each sample area deliver letters to the f i r s t ten addresses listed on your assignment sheet. If addresses include an apartment building to which entrance can be gained, deliver a letter to as many apartments as the number of remaining letters (for that area) permits. - return to areas within three days, preferably during the early evening or on the weekend. - use random household number on assignment sheet to determine f i r s t c a l l in each sample area. - introduce yourself eg.: "Hullo, I'm a geography student at U.B.C. and I'm an interviewer for the shopping survey described in a letter you recently received. Is there a man/woman (required sex of respondent in each sample area is shown on the assignment sheet) in the household - over 17 who is willing to help in this study?" - i f suitable respondent iss found, complete interview and proceed to next sample axea. - i f suitable respondent ie^ not found proceed to the next household at which letter was l e f t ; continue until respondent is found; i f respondent not found among ten households which have received a letter, phone me (Martin Taylor) at 434-7867. — repeat procedure in other sample areas. — i f at any point the respondent asks questions about the survey which you are unable to answer, suggest that they phone me at 434-7867. GUIDE TO THE QUESTIONNAIRE  SECTION 1. DISPOSITIONS This section should be self-administered by the respondent. He/ she should be encouraged to give his/her f i r s t reaction to each state-ment, and not to linger and consider them more deeply. The section should take no longer than 10 minutes to complete; no statement should be left unanswered. Avoid being drawn into debates arising from the statements, Introduction might be: "Would you please read each of these statements and indicate how i t corresponds to your opinion when buying clothes for yourself. Just circle the appropriate response category beneath each statement." Provide a pencil. Point out and explain the response categories (SD — SA). Stress that answers must reflect opinions held when buying clothes. SECTION 2. PAST CLOTHING PURCHASES Ask the respondent to recall the last three clothing purchases he/she has made for himself/herself valued over $10.00. Record details of each purchase beginning with the most recent. Note carefully the following points of clarification: PI — record name of store. P2 - 'HIGHER PRICED SPECIALTY1 eg. 'Jay David' and 'Mr Roberts' for women; 'Edward Chapmans' for men. *MEDIUM PRICED SPECIALTY' eg. 'Jerraaines' for women; 'Fred Asher1 for men. *BOUTIQUE' eg. HGP and 'Bootlegger'. 'REGULAR DEPARTMENT STORE' = 'The Bay', 'Woodwards', 'Eatons' and 'Simpson Sears' (excluding Budget Floor). 'DEPARTMENT STORE BUDGET FLOOR' « Budget Floors of 'Regular Department Stores*. 229. 1BUDGET PRICED STORE' eg. 'Hamilton Harvey', 'Array and Navy', 'Fields'. N.B. I f purchase was made at a 'REGULAR DEPARTMENT STORE' determine whether i t was from the Budget Floor. P3 - determine hundred block address of store i f possible or name of shoppingoentre waere relevant (eg. 'Oakridge'). P4 - 'OTHER' origins are most l i k e l y to occur when the main purpose of the t r i p was something other than to purchase clothing (Q. P 8 ) . Write i n 'OTHER' origins (eg. home of r e l a t i v e or f r i e n d ) . P5 - ensure that location recorded refers to o r i g i n specified i n P4. P6 - distance estimates w i l l inevitably be of varying accuracy; don't worry, accurate measurements w i l l be made prior to coding. P7) P8> - (>/) t i c k the appropriate category. P9J Questions PI to P9 are repeated twice before PIO to P12 are asked. PIO} Pll\- record as for PI, P2 and P3. SECTION 3. BIOGRAPHICAL DATA Suggested introduction: "In order to help interpret shopping habits and opinions, I'd l i k e to f i n d out some b r i e f background information." This section w i l l probably require the most tact on your part; remind the the respondent i f necessary that a l l information i s anonymous. Bl — i f respondent i s reluctant to give exact age, p o l i t e l y request an approximation (eg. ' t h i r t i e s ' ) . B2 - by observation! B3 - only record 'WIDOWED' or 'DIVORCED' i f information i s volunteered. B4 - record number. B5 - hand card to respondent to t i c k the appropriate income category and then transfer information to questionnaire. Income recorded should represent a l l contributing members of the household. B6 - 'SOME POST SECONDARY' includes academic, technical and professional ~ t r a i n i n g . B7 — 'head of household* » husband i n nuclear family; = respondent i n single person households, and i n situations where friends are sharing accommodation. Record precise job description. B8 — by observation. APPENDIX D: INTERVIEWER'S ASSIGNMENT SHEET 231. INTERVIEWER'S ASSIGNMENT SHEET NAME; NUMBER: SAMPLE AREA E.D. E.A. ADDRESSES RANDOM SEX OF INTERVIEW HOUSEHOLD RESPONDENT NUMBER NUMBER 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0100000/manifest

Comment

Related Items