UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Interpretation of the field-independence dimension : the effect of variations in stimulus input on the… Smith, June Makins 1970

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

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
831-UBC_1970_A1 S55.pdf [ 4.93MB ]
Metadata
JSON: 831-1.0302460.json
JSON-LD: 831-1.0302460-ld.json
RDF/XML (Pretty): 831-1.0302460-rdf.xml
RDF/JSON: 831-1.0302460-rdf.json
Turtle: 831-1.0302460-turtle.txt
N-Triples: 831-1.0302460-rdf-ntriples.txt
Original Record: 831-1.0302460-source.json
Full Text
831-1.0302460-fulltext.txt
Citation
831-1.0302460.ris

Full Text

INTERPRETATION OF THE FIELD-INDEPENDENCE DIMENSION: THE EFFECT OF VARIATIONS IN STIMULUS INPUT ON THE PERFORMANCE OF FIELD-INDEPENDENT, INTERMEDIATE AND FIELD-DEPENDENT SUBJECTS. by JUNE MAKINS SMITH M.A., University of British Columbia, 1959 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF Doctor of Philosophy In the Department of Psychology We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA June 1970 I n p r e s e n t i n g t h i s t h e s i s i n p a r t i a l f u l f i l m e n t o f t he r e q u i r e m e n t s f o r an advanced degree a t t h e U n i v e r s i t y o f B r i t i s h C o l u m b i a , I a g r e e t h a t t h e 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 t u d y . I f u r t h e r ag ree 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 p u r p o s e s may be g r a n t e d by t h e Head o f my Depar tment o r by h i s r e p r e s e n t a t i v e s . I t i s u n d e r s t o o d t h a t 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 no t be a l l o w e d w i t h o u t my w r i t t e n p e r m i s s i o n . Depar tment o f Psychology 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 301 June 1970 i i ABSTRACT This study was designed to test the theoretical interpretation of the Field-Independence Dimension of intellectual functioning. Witkin (1962, 1964, 1965) and Silverman (1968) proposed that consistent individual differences, which were found when subjects responded to the Rod and Frame Test (RFT) and the Embedded Figures Test (EFT), reflected differences in a b i l i t y to analyse a complex perceptual f i e l d and to resist the attraction of irrelevant elements of the f i e l d . Field-independent (FID) subjects were thought to be highly analytic and highly resistant to distraction whereas field-dependent (FD) individuals were thought to be overwhelmed by the complexity of the perceptual f i e l d . E l l i o t t (1961) suggested that the responses of field-independent and field-dependent persons were determined by brain function and were not affected by the nature of the perceptual f i e l d . A discriminant identification task was designed so that the effect of variations in stimulus input could be studied i n relation to FID, intermediate (N) and FD groups of subjects. The subjects were 96 male volunteers who were divided into FID, N and FD groups on the basis of their combined scores on the RFT and EFT. They were presented with complex visual arrays and were asked to focus upon a designated part (or item) of each array. Their task was to name a feature of the designated item which distinguished that item from a l l other items on the array. The stimuli were altered in three specific ways: f i r s t l y there were either three or six separate items; secondly each item carried either three or six attributes (e.g. shape, size and i i i colour of the central figure, colour of the background, number and design of borders); thirdly the subject's chance of finding a correct response by "luck" (chance) could be either one in three (high) or one in six (low). Thus the amount of input that the arrays carried was varied in two ways (items, attributes) and the need to analyse complexity was also varied (probability of chance success). The response measures were latency of response and the number of errors. On the basis of Witkin's and Silverman's interpretations, increases i n the amount of stimulus input or in the need for stimulus analysis should cause d i f f e r e n t i a l changes in the response latencies and error scores of FID, N and FD subjects. The FID-Group should show least increase in latencies of response and in error scores and the FD-Group should show the greatest increase on these measures. The response latencies and error scores of the N-Group should increase to an intermediate degree. The alternate interpretation which E l l i o t t proposed generated the prediction that changes in stimulus parameters would not cause dif f e r e n t i a l changes among FID, N and FD groups. According to this author, a l l changes on response measures should be of the same magnitude and in the same direction. The results of the study supported E l l i o t t ' s interpretation of Field Independence. There was no interaction between increased stimulus input and level of f i e l d dependence, or between decreased probability of success and level of f i e l d dependence. The results showed that there were significant differences among the FID, N and FD groups with respect to latency of response (.025) and error scores (.05). The differences were in the predicted direction. Changes in the amount of stimulus input iv and in the probability of success by chance were also reflected in significant differences in response latencies (items, p<,001; attributes, p<.001; probability of success, p<.001). Error scores were a less sensitive measure of stimulus manipulations. There was a significant effect, in the predicted direction, which was due to probability of success (p<.001), but the other significant effects (items, p<.05; attributes, p<.001) were in the opposite direction to that which had been predicted. Fewer errors were made when the amount of stimulus input was increased. When the amount of stimulus input was varied, the number of errors appeared to be a function of the latency of response. Less than 10% of responses were error responses. It was concluded that differences between FID, N and FD groups do not reflect d i f f e r e n t i a l responsiveness to external stimulation, and i t was suggested that more attention should be paid to consideration of the manner in which stimulus inputs are processed. V TABLE OF CONTENTS Page CHAPTER I INTRODUCTION 1 Field-Articulation: Theoretical Considerations and Definitions 1 Field-Articulation: Research Considerations 8 Field-Articulation: Interpretation and Implications 16 CHAPTER II METHOD 22 Apparatus and Stimuli 22 Apparatus . 22 Stimulus Arrays 22 Procedure 28 Testing and Assignment of Subjects . . 28 Experimental Design 29 The Experimental Session 30 a) Instructions to the Subjects: Introductory Slides 31 b) Instructions to the Subjects: Experimental Slides 33 Recording and Scoring 34 Treatment of Data 34 Additional Methodological Consider-ations 35 a) Order of Slide Presentation . . 35 b) Distribution of Attribute Values 35 Pilot Studies 36 Experiment 1 37 Experiment II 37 Field-Articulation in Male and Female Subjects 38 v i Page CHAPTER III RESULTS . . . . 39 Pilot Studies 39 Experiment 1 39 Experiment 2 40 Field-Articulation in Male and Female Subjects 41 Major Research Project 42 Latencies 42 Errors 53 CHAPTER IV DISCUSSION OF RESULTS 65 REFERENCES 73 APPENDIX A The Hidden Figures Test 79 APPENDIX B The Rod and Frame Test: Procedure 84 APPENDIX C Sample Record Sheet: Low Attribute (A_) Condition 85 APPENDIX D Sample Record Sheet: High Attribute (A,) Condition 88 o APPENDIX E Counterbalancing of the Stimulus Arrays . . . . 92 APPENDIX F Scores of F.I.D., N. and F.D. Groups on the Rod and Frame Test and the Embedded Figures Test 94 v i i LIST OF TABLES Page TABLE 1 The Effects of Sampling Influences on the Correlation Coefficient between the Rod and Frame Test and the Embedded Figures Test 7 TABLE 2 Attributes and Attribute Values used in the Construction of Stimulus Arrays 24 TABLE 3 Analysis of Variance of Response Latencies . . . . 44 TABLE 4 Correlations Between the Selection Tests and the Response Measures in the Low Attribute (A..) Condition 45 TABLE 5 Correlations between the Selection Tests and the Response Measures in the High Attribute (A &) Condition 46 TABLE 6 Analysis of Variance for Response Latencies with Low Probability of Success Scores Excluded 48 TABLE 7 Analysis of Variance for Error Scores 55 TABLE 8 Analysis of Variance for Error Scores with Low Probability of Success Scores Excluded . . . . 57 v i i i FIGURE 1 FIGURE 2 FIGURE 3 FIGURE 4 FIGURE 5 FIGURE 6 FIGURE 7 FIGURE 8 FIGURE 9 FIGURE 10 FIGURE 11 LIST OF FIGURES Page Sample Three and Six Item Arrays which show the Distribution of Attribute Values . . . . 23 Sample Three Item, Three Attribute Array. The arrow indicates the designated item 25 Sample Six Item, Six Attribute Array. The arrow indicates the low probability designated item. The other arrow (P-) points to the high probability of success, designated item 27 Mean Latency of F.I.D., N. and F.D. Groups i n Response to Low (A^) and High (A^) Attribute Stimulus Arrays 50 Mean Latency of F.I.D., N. and F.D. Groups in Response to Low (I.-) and High (1^) Item Stimulus Arrays 52 Mean Latency of F.I.D., N. and F.D. Groups i n Response to High Attribute Arrays, under Conditions of High (P.j) and Low (Pfe) Probability of Success 54 Association between Mean Latency and Mean Error Scores in Low (I») and High (I,) Item Conditions and in Low (A..) and High (A &) Attribute Conditions 59 Association between Mean Latency and Mean Error Scores i n High (P 3) and Low (Pg) Probability of Success Conditions 60 Mean Errors of F.I.D., N. and F.D. Groups in Response to Low (A-) and High (A^) Attribute Stimulus Arrays 61 Mean Number of Errors of F.I.D., N. and F.D. Groups in Response to Low (I-) and High (1^) Item Arrays 63 Mean Number of Errors of F.I.D., N. and F. D. Groups in Response to High (P„) and Low (P^) Probability of Success Conditions 63 ix Acknowledgement The author would like to thank members of the Dissertation Committee—Dr. G. Plum, Dr. J. Johnson, Dr. D. Papageorgis, Dr. S. Blank and Dr. D. Kenny—who gave advice and encouragement during the course of this research. Special gratitude i s due to Dr. Plum who devoted many hours to consultation, and to Dr. Johnson whose advice on s t a t i s t i c s was invaluable. The author would also lik e to thank Professor E. Belyea who made the colour transparencies which were such an essential part of the research material. 1 CHAPTER I INTRODUCTION1 Field-Articulation: Theoretical Considerations and Definitions. For many years psychologists have searched for consistent response patterns which would account for differences in individual behaviour, and in recent years an increasing number of theoretical and research articles have appeared which have dealt with the subject. These publications have been concerned with the identification and delineation of individual patterns of responding, and when such patterns appeared, they came to be known as "individual difference dimensions." Interest in the study of these dimensions has shown growth in recent years. This has been due, i n part, to the introduction and development of improved methods of analysing data. Whereas in the past, one or a few pieces of behaviour could be studied at a time, i t i s now possible to include many varied items of behavioural data in a single study. Such multivariate methods of analysis have been chosen by a number of authors (Cattell, 1959; Eysenck, 1960, 1961; Gardner, Holzman, Klein, Linton & Spence, 1959; Gardner, Jackson & Messick, 1960; Holtzman & Klein, 1950, 1954; Kagan, Moss & Sigel, 1963; Kagan, Rosman, Day, Albert & P h i l l i p s , 1964; Pettigrew, 1958; Schooler & Silverman, 1969; Witkin, Lewis, Hertzman, Machover, Meissner & Wapner, 1954; Witkin, Dyk, Faterson, Goodenough & The author wishes to thank Dr. G. Plum of the Department of Psychology, University of British Columbia, who made equipment and funds available for this research. 2 Karp, 1962). When multivariate analysis is used, a wide variety of data are collected and correlations among the items are noted. The studies are then replicated on new groups of subjects and "factors" are extracted from among data items which have shown consistently high intercorrelations. Several of the authors who are referred to above have paid special attention to intellectual factors and have uncovered a number of hypothesized response modes. For example, Holzman and Klein (1950, 1954) have discussed "levellers" and "sharpeners;" Pettigrew (1958) has worked on "category width" in an attempt to identify subjects who used broad as opposed to narrow categories; broad versus narrow scanners have been discussed by Gardner et a l . (1959) and researched by Silverman (1964, 1967, 1969), and f i n a l l y Kagan and his colleagues (1963, 1964) have discussed and researched subjects who were described along a dimension of analytic to non-analytic. Probably the dimension which has attracted the greatest amount of interest, however, is that having to do with field-articulation (Witkin et a l . , 1954, 1962). This dimension is known by a number of equivalent names such as "field-independence" and "analytic f i e l d approach." The individual differences, which were described by the authors who were mentioned above, may be seen as specific, consistent, individual modes (or styles) of responding in a number of different situations. They refer to the "styles" by which people habitually respond to stimuli, discriminate between cues or process information. Individuals may d i f f e r along some continuum in terms of each of the dimensions which have been mentioned and in terms of how they perform the tasks of responding, discriminating and processing. Conceptualizations of the response styles are based on implications 3 which are drawn from data. These conceptualizations are at best inferences, and frequently they are l i t t l e more than descriptive. It might be worthwhile to note, however, that while the conceptual delineations are f a i r l y ambiguous, the operational definitions are usually based on clear, relatively standardized, test procedures. In the case of f i e l d articulation, a f a i r l y general interpretation has been made by Witkin and his colleagues (1954, 1962, 1964, 1965). According to Witkin some individuals, who are called f i e l d independent (FID), are more able than others to analyse a perceptual situation and to withstand a complex, or misleading background. These persons can separate a complex stimulus into i t s component parts and can perceive an object, which i s embedded in a complex f i e l d , in relation to i t s background. Other people, who are called f i e l d dependent (FD), are thought to be over-whelmed by the prevailing f i e l d . They do not analyse the f i e l d into i t s components and they are more influenced by a misleading background which embeds a stimulus object. At one extreme (Field Dependence) there is a consistent tendency for experience to be global and diffuse; the organization of the f i e l d as a whole dictates the manner in which i t s parts are experienced. At the other extreme (Field Independence) there i s a tendency for experience to be delineated and structured. Parts of a f i e l d are experienced as discrete and the f i e l d as a whole is organized. (Witkin, 1965; page 319) Witkin's interpretation was rephrased by Silverman (1968), who included f i e l d - a r t i c u l a t i o n among attentional variables. Silverman constructed a model which attempted to account for attention in terms of both a response dimension and an hypothesized response mechanism. This model contained three response dimensions. These were stimulus intensity control, scanning control and f i e l d articulation control. Stimulus intensity 4 c o n t r o l r e f e r r e d to the s e n s i t i v i t y of the subject to stimulation and to h i s tendency to e i t h e r augment or reduce the i n t e n s i t y of s t i m u l a t i o n . Scanning c o n t r o l described the extensiveness with which a perceptual f i e l d was sampled. F i e l d a r t i c u l a t i o n included two response f a c t o r s : f i r s t l y subjects were considered to d i f f e r i n t h e i r a b i l i t y to respond to d i s c r e t e segments of a stimulus f i e l d as opposed to the f i e l d as a whole; secondly they were considered to d i f f e r i n t h e i r s u s c e p t i b i l i t y to d i s t r a c t i o n by i r r e l e v a n t stimulus inputs. According to t h i s formulation a highly f i e l d independent person would be both able to analyse a perceptual f i e l d and able to r e s i s t d i s t r a c t i o n by i r r e l e v a n t elements. An extremely f i e l d dependent subject, on the other hand, would respond to the f i e l d as a whole and would also be unable to r e s i s t d i s t r a c t i o n . Silverman wrote as follows: One of the response factors concerned with selectiveness of a t t e n t i o n , describes responses to stimulus configurations on a dimension which ranges from passive responsiveness to an o v e r a l l configuration to responsiveness to d i s c r e t e segments i n a configuration . . . The weaker the d i s p o s i t i o n to segmentalize g e s t a l t s , the more l i k e l y i s an i n d i v i d u a l to experience stimulus elements i n a configuration as r e l a t e d to each other In some concrete manner or as otherwise le s s d i f f e r e n t i a t e d . A second selectiveness of a t t e n t i o n response f a c t o r describes d i f f e r e n c e s i n the degree to which i n d i v i d u a l s are susceptible to interference from c o n f l i c t i n g stimulus elements i n a c o n f i g u r a t i o n . The more int e r f e r e n c e prone one i s , the greater i s one's d i f f i c u l t y i n maintaining d i s t i n c t figure-ground a r t i c u l a t i o n s and accurate perceptions and cognitions i n the presence of d i s t r a c t i n g s t i m u l a t i o n . (Silverman, 1968; page 1204) The i n t e r p r e t a t i o n s which have been offered by Witkin were based on observations which he and h i s colleagues (1954) made while they were doing research on the perception of the upright. In the course of t h e i r studies, they noted that they obtained consistent i n t e r c o r r e l a t i o n s on several per-ceptual t e s t s . Among these were the Rod and Frame Test (RFT) and the 5 Embedded Figures Test (EFT). Scores on these measures have since been used to o p e r a t i o n a l l y define f i e l d independent and f i e l d dependent groups of subjects. These procedures w i l l therefore be described i n the following paragraphs. The RFT apparatus c o n s i s t s of a luminous frame and a luminous rod which i s contained w i t h i n the frame. Both are pivoted at the centre and they can be rotated independently of each other. The subject views the apparatus i n a completely darkened room and the luminous rod and frame are the only objects that can be seen. Various s e t t i n g s of the rod and the frame are made by the experimenter, who i s seated behind the apparatus. The subject's task i s to adjust the rod to v e r t i c a l regardless of the p o s i t i o n of the frame. His accuracy i s recorded by the experimenter who watches an illuminated disk, at the back of the apparatus and reads o f f the number of degrees that the subject's s e t t i n g was i n e r r o r . Some subjects perceive the rod as v e r t i c a l only when i t i s f u l l y aligned with the t i l t e d frame. These persons are considered f i e l d dependent since t h e i r p o s i t i o n i n g of the rod i s thought to have been d i c t a t e d by the context provided by the t i l t e d frame. Other subjects are able to adjust the rod more or l e s s to true upright regardless of the p o s i t i o n of the frame and are c a l l e d f i e l d independent. The EFT i s a timed paper and p e n c i l t e s t . I t requires the subject to l ocate a simple f i g u r e which i s concealed w i t h i n a complex design. The design i s drawn i n such a way that i t breaks up the f i g u r e and includes parts of i t i n other g e s t a l t s . The score of the subject i s the t o t a l number of f i g u r e s located within the time l i m i t . For some persons, who are c a l l e d f i e l d independent, the f i g u r e stands out d i s t i n c t l y . For 6 other subjects the design i s d i f f i c u l t to locate and as a r e s u l t they i d e n t i f y fewer f i g u r e s i n a given amount of time. These l a t t e r i n d i v i d u a l s are labeled f i e l d dependent. Since both these te s t s of f i e l d a r t i c u l a t i o n are used i n the present study, some mention should be made about the r e l a t i o n s h i p between them. P o s i t i v e c o r r e l a t i o n s between the RFT and the EFT have occurred c o n s i s t e n t l y , b u t have been found to be quite v a r i a b l e . For instance, Witkin et a l . (1962) reported c o r r e l a t i o n s of .64 f o r college males and .63 f o r male h o s p i t a l employees. Gruen (1955) found that the RFT and the EFT co r r e l a t e d .73 f o r male college students. In general though, the magnitude of the c o e f f i c i e n t s have been lower. Adevai, Silverman and McGough (1968) reported c o r r e l a t i o n s of .49 for male subjects when the experimenter operated the RFT apparatus and of .35 when the subjects operated i t . Some other co-e f f i c i e n t s f or male subjects were: Cancro and Voth (1969) .42; C r u t c h f i e l d , Woodworth and Albrecht (1958) .27; E l l i o t t (1962) .40; Goodman (1960) .32; Gruenfeld and Arbuthnot (1969) .29. Differences i n test administration and sampling may account f o r some of the i n c o n s i s t e n c i e s . An excel l e n t commentary on methodological defects and i n c o n s i s t e n c i e s i n the use of the RFT has been provided by Lester (1968). The RFT procedures are not we l l standardized and there are several forms of the EFT (French, 1954; Jackson, 1956; French, Ekstrom & P r i c e , 1963; Witkin, 1950). When a l l p o s s i b l e v a r i a t i o n s of the RFT procedure are taken together with a l l p o s s i b l e v a r i a t i o n s of the EFT format, i t i s obvious that many poss i b l e combinations e x i s t . I t i s perhaps s u r p r i s i n g , not that c o r r e l a t i o n s vary, but 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 occur so r e g u l a r l y . Sampling d i f f e r e n c e s probably contribute even more to di f f e r e n c e s i n 7 the amount of common variance than do variations of methodology. Gruen's (1955) study provided a striking example of this. Gruen used one group which was composed of professional dancers and another which was composed of college students. The results of his study appear in Table 1. TABLE 1 The Effect of Sampling Influences on Correlation Coefficients Between the Rod and Frame Test and the Embedded Figures Test. Males Females Coefficient Number Coefficient Number Dancers .36 30 .62 30 Students .76 52 .26 51 It can be seen that both the sex of the subject and his primary occupation influenced the correlations between the tests. To be more specific, the table gives the visual impression of an interactive relation-ship; male dancers and female students yielded low correlations and male students and female dancers yielded high correlations. Most studies have found that correlations between the RFT and the EFT were lower for females than for males. The Gruen study is a partial exception, as are the studies of Cancro and Voth (1969) and Dickstein (1968). Cancro and Voth found a higher correlation for female community volunteers (.70) than for a similar group of males (.42). Dickstein found no sex differences. In short, even the l i t t l e evidence presented here is somewhat compelling in terms of the influence of sampling on the correlation between the EFT and the RFT. 8 In t h i s s e c t i o n a number of i n d i v i d u a l d i f f e r e n c e dimensions were mentioned and the dimension of f i e l d - a r t i c u l a t i o n was elaborated upon. T h e o r e t i c a l and operational d e f i n i t i o n s were discussed and some mention was made about the d i f f i c u l t i e s of each. Given these d i f f i c u l t i e s , one might w e l l wonder why t h i s area of research has a t t r a c t e d so much att e n t i o n . The appeal i s l a r g e l y r e l a t e d to the research that has been generated and to i t s p o s s i b l e s i g n i f i c a n c e f o r the understanding of complex human behaviours. The next s e c t i o n discusses some of t h i s research i n greater d e t a i l . F i e l d - A r t i c u l a t i o n : Research Considerations In t h i s s e c t i o n studies which deal with the s t a b i l i t y of f i e l d -a r t i c u l a t i o n over time w i l l be r e f e r r e d to. Hypothesized r e l a t i o n s h i p s of f i e l d - a r t i c u l a t i o n to c e r t a i n 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 w i l l be mentioned, and i t s possible s i g n i f i c a n c e i n the area of c l i n i c a l psychology and psychopathology w i l l be discussed. Turning f i r s t to s t a b i l i t y , i t may be noted that Witkin et a l . (1954, 1962) reported that there was s t r i k i n g s t a b i l i t y on a l l t e s t s of the perceptual battery over periods of one to three years. Adevai and McGough (1968) found that the RFT scores of undergraduates were stable over a four year period. F i n a l l y Schimek (1968) has reported that the RFT scores of male subjects showed s i g n i f i c a n t s t a b i l i t y from childhood to adulthood. These studies suggest that f i e l d a r t i c u l a t i o n merits consideration as a p e r s i s t e n t response dimension which does not change gr e a t l y with the passage of years. A second area of research has been s i g n i f i c a n t i n showing that f i e l d -a r t i c u l a t i o n has a number of p e r s o n a l i t y c o r r e l a t e s . The e a r l i e r research 9 in this area has been reviewed by Witkin et a l . (1962). Witkin and his colleagues summarized the personality characteristics of f i e l d dependent and f i e l d independent subjects under three headings. F i r s t l y , f i e l d dependent subjects were said to rely on others for guidance and support more than did f i e l d independent subjects. Evidence for this attribute came from a study of the guidance seeking behaviour of children in a situation where they were required to respond to TAT cards. Field dependent children sought guidance more often than did f i e l d independent children. Further evidence was provided in a study by Gordon (1953). This author found that subjects who were f i e l d dependent tended to rate themselves as socially dependent on a rating scale. These subjects were also rated as socially dependent by others. A similar result was found by Pemberton (1952) who also used personality inventories. Beller (1958) found that f i e l d dependent nursery school children tended to look at the teacher more than did f i e l d independent children and, again, this was interpreted as a need for guidance. The second heading that Witkin used referred to a tendency for f i e l d dependent persons to be more susceptible to influence by external standards in the formation and maintenance of attitudes and judgements. Linton (1952) found that f i e l d dependent individuals were more l i k e l y to yield to social pressure and to change previously expressed opinions, judgements or attitudes. Witkin et a l . (1962) l i s t e d several other studies which tended to confirm Linton's findings. Thirdly, Witkin et a l . (1962) reviewed studies which seemed to show that f i e l d independent subjects were more stable in the ways in which they described themselves. From the above i t is apparent that the field-articulation dimension 10 shows promise as a tool for the investigation of personality factors, and in particular as a key to patterns of individual responding. Turning to the role of field-articulation in psychopathology, i t has been suggested (Schooler and Silverman, 1969; Silverman, 1964, 1967, 1968; Witkin, 1965) that the symptom patterns of schizophrenic patients may be related to s t y l i s t i c response modes. Silverman, whose attentional model has already been described in some detail, has been a pioneer in this f i e l d . He has attempted (1967, 1969) to account for differences among schizophrenic sub-types in terms of attentional mechanisms. He suggested, for instance, that paranoid patients with a good premorbid history should be high scanners, whereas non-paranoid patients with poor premorbid histories should be restricted scanners. Some studies (Silverman, 1964, 1967) have supported this hypothesis. Stimulus augmentation, another of Silverman's attentional variables, has also been related to the presence of a paranoid symptom pattern and to the premorbid history of schizophrenic patients (Silverman, 1967, 1969). Field articulation has not been researched as an attentional variable under Silverman's model, but some studies (Jannucci, 1964; Powell, 1964; Witkin, 19s4) have suggested that f i e l d independence may be associated with paranoid symptomatology. In a recent study, Schooler and Silverman (1969) attempted a comprehensive test of the attentional model, with schizophrenic patients. They wished to establish that scanning, stimulus augmentation and f i e l d -articulation would emerge as attentional factors among schizophrenic patients. They also wished to confirm earlier studies which had suggested that there is a relationship between attentional modes and patterns of symptomatology. Specific predictions were made which related scanning and 11 stimulus augmentation to diagnostic patterns, but no predictions were made concerning the relationship of field-articulation to symptoms. In the Schooler and Silverman study there were scores on seven perceptual tests (which had been found to reflect the three attentional constructs) and these served as the dependent variables. There were three independent variables; a) psycho-social functioning as indicated by length of hospitalization, schizophrenic sub-type and scores on rating scales; b) intellectual functioning as demonstrated on WAIS sub-test and c) personal history variables such as age at f i r s t hospitalization, education, and social class. Two additional behaviour rating scales were included to provide a f u l l e r picture of current symptomatology and behaviour. Results of the study supported the tridimensional model of perceptual functioning, but suggested that modifications may be needed i f this model is to be applied to hospitalized schizophrenics. Scanning, stimulus augmentation and f i e l d articulation emerged as separate factors, as had been expected, but two additional dimensions (perceptual disorganization and anchoring) were also found. These additional factors had not been observed in studies in which normal subjects were used and i t was thought that they might be specific to schizophrenic patients. One factor, perceptual inefficiency, accounted for a large amount of the variance and, as this dimension appears to resemble field-articulation, confusion might be expected to arise, between perceptual inefficiency and f i e l d -dependency. The hypothesized relationship of perceptual performance to diagnostic sub-type was only partially supported. A cluster analysis of perceptual 12 modes confirmed the presence of groups of patients who functioned in ways which paralleled expected differences between paranoid and non-paranoid schizophrenics. However, these clusters were not related to the symptom patterns which are considered to reflect the diagnostic categories. This finding raised many questions, but these questions relate to diagnostic categorization, to scanning and to stimulus aug-mentation rather than to field-articulation. No predictions were made regarding the diagnostic expectations for high and low field-independent subjects. Another approach to the problem of d e f i c i t in psychosis was taken by Chapman (1966). This author has done detailed c l i n i c a l studies of early schizophrenia. He did not employ any particular model, but the attentional model which was described by Silverman (1968) might perhaps be applied to his findings. According to Chapman: Visual adaptation to the environment i s normally achieved by the individual developing the capacity to select from the diffuse mass of visual stimuli impinging upon him, that information which i s relevant and necessary for him to function e f f i c i e n t l y . In this process sensory information i s automatically organized during the act of perception i t -s e l f . . . . . . schizophrenic patients experience from time to time transient but severe disturbances i n visual perception. At these times, perceptual s t a b i l i t y appears to be lost and the patients are unable to reduce, organize and interpret visual information in a normal fashion. . . . Much of the c l i n i c a l data obtained i n the present study suggests that i n schizophrenia the breakdown in perceptual functioning results in an increase in the volume of information with which the patient has to deal but which he is unable to process. . . (Chapman, 1966; pages 229-234) If viewed in terms of Silverman's attentional model, these descriptive comments could apply to either a state of increasing f i e l d dependency in early schizophrenia, or to perceptual breakdown, or to both. 13 On the one hand, Chapman's comments seem to describe a condition i n which there i s diminished a b i l i t y to analyse a perceptual f i e l d and to resist distraction. On the other hand, the author seems to be suggesting that established perceptual organization i s , for some reason, lost. Better understanding of d e f i c i t i n schizophrenia might be gained i f the response processes which are thought to be involved i n both normal and abnormal responding were more clearly understood. The three response dimensions which were described by Silverman were a l l conceptualized on the basis of inferences from data. The inferences may have been correct and the response dimensions may, indeed involve the processes which they were hypothesized to involve. However, data may have been interpreted incorrectly and i n this case the conceptualization of the response dimension would be faulty. The work of Silverman (1969) and of Chapman (1966) suggests that there is an immediate need for experimental verification of the processes which are hypothesized to operate i n the field-articulation dimension. Since lack of conceptual c l a r i t y may lead to confusion between two different response processes, i t is necessary to determine whether schizophrenic behaviour involves an extreme, but normal, type of responding, or whether i t reflects intrusion of abnormal processes. Cla r i f i c a t i o n of the processes which are involved in the responses of schizophrenics may have prognostic as well as diagnostic usefulness. Chapman (1966) found that almost half of the patients who showed disturbances of visual perception progressed to severe degrees of dementia. Cancro and Sugarman (1969) made a specific attempt to relate f i e l d -14 articulation to prognosis. They tested process and reactive schizo-phrenics on the RFT and they related field-dependency scores to diagnostic sub-type (process or reactive) and to length of stay in hospital. They found that the reactive cases tended to have intermediate scores on the RFT, while process patients scored at the extremes of the f i e l d articulation dimension. Those patients who had intermediate scores on the RFT spent least time i n hospital, while those who were most highly f i e l d independent were hospitalized longest. This relationship between the level of field-articulation and hospital stay was observed at four measured time periods (6 months, 12 months, 24 months and 36 months), but i t was significant only at 12 months and 24 months. The authors noted that rate of discharge did not necessarily predict to continuing health. Highly f i e l d dependent patients, who had intermediate stays in hospital, were most l i k e l y to be readmitted. This study suggests that the f i e l d -articulation dimension may have prognostic usefulness, but that the relationship between f i e l d dependency and prognosis is a complex one. The results were not particularly clearcut, but they provided some evidence for the hypothesis that there i s an association between f i e l d dependency and diagnostic sub-type and between f i e l d dependency and prognosis. Field dependency has also been studied in non-psychotic c l i n i c a l conditions. For instance Bailey, Hustmeyer and Kristofferson (1961), Jacobson (1968), Karp, Poster and Goodenough (1963) and Witkin, Karp and Goodenough (1959) found that alcoholics were unusually f i e l d dependent. There were similar findings for asthmatics (Fishbein, 1963) and for gastic ulcer cases (Gordon, 1953). These results seem to suggest that specific symptoms may be associated with specific levels of f i e l d -15 articulation and, since excessive dependency on other people forms part of the c l i n i c a l picture of alcoholism, asthma and gastric ulcer cases, the studies were interpreted as supporting Witkin's (1962, 1964, 1965) conceptualization of f i e l d articulation. It was said (Witkin, 1965) that a l l these conditions reflected the patient's i n a b i l i t y to analyse experience and to resist irrelevant stimulus inputs. Other studies (Jacobson, 1966, 1968; Kristofferson, 1966; Goldstein and Chotlos, 1966; Astrup, 1968) suggested however that an individual's level of f i e l d dependency, as demonstrated by perceptual tests, could be changed to a minor, but significant degree by alterations in his surroundings or in his internal state. For instance, a significant decrease in f i e l d dependence has been found in college students (Jacobson, 1966) and in alcoholics (Jacobson, 1968) following an hour of sensory deprivation. Astrup (1968) discovered a similar effect in young manual laborers after a two and a half hour shift underground. These studies raised the question of whether personality as a whole was involved or whether only some aspect of the perceptual or the intellectual processes was affected. Two of the studies (Kristofferson, 1966; Goldstein and Chotlos, 1966) relate specifically to alcoholism. Kristofferson (1966) found that non-alcoholic, male college students became more f i e l d dependent when they were given alcohol in an experimental situation. Goldstein and Chotlos (1966) found that alcoholics were less f i e l d dependent after three months of hospitalization and treatment. During this time, presumably, they had been deprived of alcohol. These results point even more directly to the conclusion that performance on tests of f i e l d dependency 16 may be related to the intellectual, or possibly to the neurological condition of a subject. From this brief review i t w i l l be apparent that there may be interesting associations between the level of f i e l d articulation and the social, c l i n i c a l or psychopathological characteristics of an individual. There is a great deal of potential for the use of the f i e l d articulation dimension in the study of interpersonal processes, psychopathological and c l i n i c a l conditions. Most of the studies which have been cited lent themselves to the interpretation that field-dependency i s a pervasive individual difference variable. Some authors (Astrup, 1968; Goldstein & Chotlos, 1966; Jacobson, 1966, 1968; Kristofferson, 1966), however, suggested that i t might be worthwhile to study this variable i n more det a i l . In the next section, the general theoretical formulations of Witkin and his associates (1954, 1962, 1964, 1965) w i l l be considered and an alternative formulation, which was proposed by E l l i o t t (1961), w i l l be discussed. Field Articulation: Interpretation and Implications Witkin (1954, 1962, 1964) emphasized the importance of "cognitive styles" i n behaviour, and field-articulation was, of course, included among cognitive styles. He wrote as follows: . . . cognitive styles are salient yet specific dimensions of behaviour, rather readily identified and measured, and also tying in with broad networks of psychological characteristics. (Witkin, 1964; page 173) According to this author the conceptualiztion of f i e l d dependency, which has been described in an earlier section, can be generalized to non-perceptual situations. People who are dependent when responding to 17 perceptual tasks are expected to rely on inputs which are external to themselves in other situations also. Linton (1955), who studied f i e l d dependency in relation to perception, attitudes and judgement, described these expectations: The tendency for behaviour to be modified by an external stimulus, regardless of whether the external stimulus is personal or Impersonal in nature, is a function of enduring attributes of the person; consequently, Ss whose performance in perceptual tasks i s highly affected by the perceptual f i e l d w i l l be those whose behaviour in other situations i s most l i k e l y to be modified so as to conform to an external standard. (Linton, 1955; page 502) The tenet that some individuals are more susceptible to external inputs than others is crucial to Witkin's interpretation of the f i e l d -articulation dimension. These authors view dependency, whether in a perceptual or a social situation as a pervasive personality t r a i t . This interpretation has, however, been questioned by some writers. E l l i o t t (1961) suggested that f i e l d dependency might have nothing to do with dependency in personality and behaviour. He noted that studies which were cited as evidence in support of Witkin's conceptualization that there is a broard personality t r a i t 'dependency,1 were not conclusive. Elliottnoted that in those studies which used social situations individuals who were f i e l d dependent on the RFT and the EFT did not always behave dependently in other situations. He also pointed out that the relationship of c l i n i c a l symptomatology i s open to other interpretations than those given by Witkin and his colleagues. E l l i o t wrote as follows: Most central in assessing the general result of the Witkin report is the study of the relationships between f i e l d dependence and the dependent personality. Among c l i n i c a l groups, alcoholics are more frame dependent than any other diagnostic group except patients with definite 18 brain damage. . . . We cannot know as yet whether this result is accounted for by the psychodynamic dependence said to be involved in the etiology of alcoholism, or by diffuse brain damage resulting from alcoholism. ( E l l i o t t , 1961; page 27) E l l i o t t suggested that i t might be more profitable and economic to change the focus of investigation from personality to intellectual function. Thus, f i e l d dependence would be viewed primarily as a kind of intellectual d e f i c i t and secondarily as a correlate of personality attributes. According to E l l i o t ' s formulation f i e l d independence would be associated with higher brain function and the level of f i e l d -articulation which was shown by a subject would be relatively independent of outside stimulus control. The question of whether or not d i f f e r e n t i a l performance on the criterion measures of field-articulation (the RFT and the EFT) is due to differing susceptibility to outside stimuli seems to be crucial to the interpretation of the field-articulation dimension. Witkin and his colleagues (1954, 1962, 1964, 1965) and Silverman (1968) have stated clearly that some subjects are better able than others to analyse a complex stimulus f i e l d and to resist interference from distracting elements. Elliott(1961) has said that the nature of the stimulus f i e l d is relatively unimportant, and that d i f f e r e n t i a l performance among individuals reflects d i f f e r e n t i a l effectiveness of brain function. Thus, the position taken by Witkin and others would predict that as the stimulus became increasingly complex, or contained more distracting elements, f i e l d dependent subjects would display a proportionately greater decrement in accomplishing embeddedness tasks than would f i e l d independent subjects. E l l i o t t (1961) would predict that f i e l d dependent and f i e l d independent 19 subjects would display equal amounts of performance decrement. The conceptualization of the response of subjects to the RFT and the EFT, which was described by Witkin and Silverman, has not been tested. No attempt has been made to present f i e l d independent and f i e l d dependent subjects with perceptual fields in which the amount of stimulus input i s varied and in which the need for analytic activity is manipulated. It seems necessary that this should be done before broader interpretations of the field-independence dimension can be made. The study which i s to be described here was designed to provide such a test. The object of this study was to investigate whether or not di f f e r e n t i a l performances could be obtained from f i e l d independent (FID), intermediate f i e l d independent-dependent (N) and f i e l d dependent (FD) subjects i n response to changes in the perceptual f i e l d . A number of visual stimulus arrays were designed and used in order to alter the nature of the stimulus input. The subject was required to focus upon a designated item of each array and to discover a feature which discriminated that item from a l l others in the array. The stimuli could be altered in specific ways to increase the number of stimulus inputs and the need for a subject to analyse the task. For example, the number of items on each array could be varied, and so could the number of attributes which was carried by each item. In addition, the subject's chance of finding a correct response by "luck" was varied. These manipulations of the stimulus increased or decreased the amount of material that competed for attention. They also made the task more or less d i f f i c u l t and so, presumably called for differing amounts of analytic a b i l i t y . 20 Some pilo t work was required in order to establish the effective-ness of the stimulus arrays, and to answer certain questions related to methodology. This pilot work indicated that the manipulation of items, attributes of items, and probability of chance success (luck) in identifying a discriminant, produced significant response differences among subjects who were unselected in terms of f i e l d independence-dependence. Thus the task had been established as being valid for the purposes of this research project. It was expected that when groups containing f i e l d independent, neutral and f i e l d dependent individuals were formed, there would be significant differences between them in their performance on the research task. Field independent subjects were expected to perform most effectively and f i e l d dependent subjects least effectively. The opposed conceptualizations of Witkin (1964, 1965) and Silverman (1968) on the one hand and of E l l i o t t (1961) on the other, generated di f f e r e n t i a l predictions with regard to the interaction effects which were expected to occur. These predictions have already been mentioned, but w i l l be stated again with the hope that redundancy w i l l provide c l a r i t y . According to Witkin and Silverman, there should be an interaction between the level of f i e l d articulation (FID, N and FD) and the d i f f i c u l t y level of the perceptual variables. When many items or attributes are present, or when the probability of a chance success i s low, the f i e l d dependent group should be relatively more disadvantaged than the neutral (intermediate) group, and neutral subjects should be more disadvantaged than f i e l d independent subjects. That is to say, i f the c r i t e r i a of good or poor performance were to be latency of response and the number of 21 errors, FID subjects should give the fastest responses and make least errors. Field dependent subjects should take longest to respond and should make most errors. Neutral subjects should be intermediate with respect to latency and error scores. E l l i o t t would predict that changes i n the stimulus parameters would f a c i l i t a t e or impede the performance of a l l groups to the same degree. In other words FID, N, and FD groups should show similar changes on response measures (latency of response and error scores) in response to experimental manipulations. 22 CHAPTER II METHOD 2 Subjects The subjects were 96 volunteers enrolled in the summer session at Simon Fraser University. Apparatus and Stimuli Apparatus The apparatus consisted of a Sawyer's Rotary Projector, model 707 AQ, which was used to project the stimuli onto a 24" x 36" viewing screen. Light from the projector activated a Knight Photoelectric c e l l . This in turn closed a c i r c u i t which started a Lafayette ADW semi-digital 1/100 second stop clock. The ci r c u i t was re-opened by a voice key which was activated by the subject's verbal response. With the opening of the ci r c u i t the Lafayette clock was automatically stopped. The voice key was a non-commercial product which was developed by the departmental shop. Stimulus Arrays Subjects viewed an array which consisted of either three (I^) or six (Ig) different complex stimuli. An example of a three item stimulus array i s presented in Figure 1(a) and an example of a six item array i s shown in Figure 1(b). In the arrays in Figures 1(a) and 1(b) there i s only one attribute (or dimension), namely that of shape. The stimulus arrays which the subjects viewed, however, consisted of items which contained either three 2 The author wishes to thank Dr. R. Blackman, of the Department of Psychology, and other staff members of Simon Fraser University, who made i t possible for subjects to be contacted. 23 F i g u r e 1 (a ) A T h r e e i t e m a r r a y f i g u r e 1 ( b ) D i s t r i b u t i o n of a t t r i b u t e - v a l u e s f o r a s i x f t e m a r r a y Figure 1. Sample Three and Six Item Arrays which show the Distribution of Attribute Values 24 (A^) or six (Ag) attributes. The different attributes and attribute-values may be seen in Table 2. TABLE 2 Attributes (dimensions) and Attribute-Values Used in the Construction of Stimulus Arrays. Attributes Shape Size Colour (central figure) Colour (background) Border (number) Border (design) Attribute-Values Cross, Square, Triangle, Circle Large, medium, small Red, blue, brown, grey, green, yellow Same as above One, two, three Dash, dot, line, wave (squiggle) It can be seen that the attributes were shape, size, and colour of the central figure and the colour of the background, the number of borders and the type of border. Each attribute had a number of values which were distributed over the items of the array and, as mentioned, these also appear in Table 2. In Figure 1(a) only two attribute-values appear (triangle and square) In Figure 1(b) there are three attribute-values (triangle, square and cross). These figures provide examples of the distribution of a single attribute on a three or a six item array. The arrays were designed so that one value of each attribute appeared once, one twice and, on six item arrays, one three times. Figure 2 provides an example of how the distribution of three attributes might have appeared on a three item Figure 2. Sample Three Item, Three Attribute Array. The arrow indicates the designated item. 26 stimulus array of the type used in the study. Figure 3 shows the distribution of the attribute-values on a highly complex array which has six attributes and six items. There were two additional types of array. One had six items and three attributes; the other had three items and six attributes. When the subject responded to a stimulus array his task was to identify an attribute-value which was single. He was directed to look at a designated item on an array and was told to discover a feature which distinguished that item from a l l others. The attribute-values were distributed in an orderly manner so that the probability of the subject discovering a discriminant by chance could be varied. The items (three or six) of a three attribute stimulus never carried more than one discriminant each. Those (three or six) of the six attribute arrays might contain one or two discriminants. This meant that on the simple (A^) arrays, the probability that the subject would discover a discriminant by chance was always one in three (high). On complex (A^) arrays the probability of chance discovery of a discriminant might be one i n three (high) or one in six (low). The notation P- and P, i s used to refer to the high and low probability J o conditions. In Figure 2, which shows an A^I-P-j array, the arrow indicates the designated item and the discriminant i s shape. In Figure 3, the array might be an A^I^P^ stimulus i f the item indicated by the arrow which i s labelled P^ were designated; or i t might be an AglgP^ stimulus i f the item which is indicated by the other arrow (P^) were designated. In the f i r s t case the discriminants would be shape and the number of Figure 3. Sample Six Item, Six Attribute Array. The arrow P, indicates the low probability designated item. The other arrow (P_) points to the high probability of success, designated item. 28 borders, while in the second case there would be only one discriminant, namely size. A l l high attribute (A^) arrays were presented twice during the course of the study. In the pages which follow i t w i l l often be necessary to refer to the type of stimulus array that the subjects viewed. There were six possible ways in which the number of item's (I^ and 1^), the number of attributes (A^ and Ag) and the probability of success by chance (P^ and Pg) could be combined. These combinations w i l l be cited in order to specify the particular type of slide. Thus an array which contained three attributes, three items and a high probability of chance success would be referred to by the notation A.-1-jP-, while an array which contained six attributes, six items and a low probability of success by chance would be an Agl^P^ stimulus. The f u l l l i s t i n g of stimulus arrays is as follows: A 3I 3P 3, A 3I 6P 3, A.I 3P 3, A ^ , A - I ^ , and A ^ P g . Two sets of 48 slides each had been prepared and each set contained either low attribute (A_) or high attribute (A,) stimulus arrays. The low attribute set contained 48 different arrays, but the high attribute set consisted of 24 different arrays, each of which had been duplicated. Fourteen introductory slides were also used. Ten of these were generally applicable and could be used equally well to explain either A 3 or A^ arrays. Of the remaining four, two showed typical examples of A 3 arrays ( A 3 I 3 P 3 and A^I^P^) and two showed typical examples of Ag slides (A^I^P^ or A 6 I 3 P 6 and A g l ^ or A ^ ) . Procedure  Testing and Assignment of Subjects The Embedded Figures Test from the Cognitive Testing Kit of French, 29 Ekstrom and Price (1963) was administered to a l l subjects as a group test. The instructions were printed on the booklet and this booklet has been reproduced in Appendix A. A l l subjects were tested individually on a standard Rod and Frame apparatus which was manufactured by Marietta Company Ltd. The details of this procedure are presented in Appendix B. Since there is an inverse relationship between scores on the RFT and the scores on the EFT, i t was necessary to convert one of the tests in order to arrive at a total score on the field-articulation dimension. Scores on the EFT ranged from 0 - 3 0 and scores on the RFT from 7-74. A conversion was accomplished by subtracting a l l RFT scores from 75 and dividing the result by 2.5. Thus both scales scored in the same direction and both had the same range. The scores from the RFT and EFT were then simply added together in order to arrive at a compound score for the f i e l d independence-dependence dimension. High scores indicated f i e l d independence, low scores indicated f i e l d dependence. Groups were obtained by a r b i t r a r i l y dividing the sample into equal groups of high (FID), neutral (N) and low (FD) scoring subjects. Those subjects who f e l l on the borderline between the FID and N, or N and FD groups were randomly assigned to one group or the other. The proportion of subjects going to each group was determined by the need to ensure that each group contained 32 members. The subjects were given appointments for the experimental task following the completion of the screening tests. Experimental Design There were two between-subjects variables. These were field-articulation 30 groups (FID, N and FD) and attributes. Of the 32 subjects in each field-articulation group, 16 viewed low attribute, A^, arrays and 16 responded to high attribute, A^, arrays. Two within subjects variables were also present. One of these was items. This was a variable in both the A^ and A^ conditions, so that a l l subjects responded to both the I^ and the 1^ arrays. The other was the probability that the subject would succeed by chance in identifying a discriminant. This variable had two levels (P_ and Pg), but only the high probability condition (P^) was present in A^. Both P^ and Pg were represented in A^. In the A- condition the subjects responded to 48 slides, in each of which there was a high (P^) probability that success in identifying the discriminant would be achieved by chance. In the Ag condition the subjects viewed the same number of slides, but two presentations were given of each slide. On one presentation they responded to a high probability item and on the other they responded to a low probability item. Pilot work had indicated that complex slides could be presented a second time without response measures being influenced by memory effects. As a special precaution against memory effects, at least three other slides were interposed between the two presentations of an Ag slide. The Experimental Session The following is a description of a typical experimental session. The subject was seated at a table with his back to the experimenter and to the projection and timing apparatus. He viewed a display panel which was approximately 5' distant. The voice relay was on the table in front of him and this was adjusted to a convenient height. The subject was told 31 that he must speak into the relay and that his voice must be loud enough to stop the clock on the experimenter's desk. The apparatus was arranged so that the subject's response latencies could be recorded. As the stimuli appeared, light from the projector activated the stop clock and this device continued in operation u n t i l the subject responded verbally. When he spoke, his voice activated a voice relay and this broke the c i r c u i t and stopped the clock. The subject was introduced to the task through a series of 12 introductory slides. The f i r s t seven slides were shown in order to introduce him to the attributes which would appear. The procedure which was followed ensured that the subject would have names readily available for each attribute-value when he was presented with the experimental task. When the eighth slide was reached, the experimenter for the f i r s t time, required the subject to speak loud enough to operate the timing apparatus. This usually involved some readjustment of the subject's posture or of the sensitivity of the voice relay. The subject was asked to repeat responses u n t i l he found an effective level of vocalization. This procedure was followed for the remaining five introductory slides. If his voice dropped during the presentation of the experimental slides, the subject was merely asked to speak more loudly. Most subjects complied well with the vocalization requirements. Since pi l o t work had shown that many individuals preferred to give a double attribute (blue square) rather than a single attribute (blue or square) response, specific training on single attribute responding was included in the introductory procedure. a) Instructions to the subject: Introductory slides 32 The subject was told "You w i l l be looking at colour slides with designs on them. This is an example of the sort of slide you w i l l see." (Slide 1) Some slides w i l l have three items l i k e this one. Others w i l l have six. Each time I shall name one of the items and w i l l ask you to name something that makes i t different from a l l other items on the array. Items can dif f e r in various ways and we w i l l now run through the differences that can occur. Items can dif f e r as to colour. This i s an example of the colours that w i l l be present. (Slide 2) Will you please name a l l the colours on this slide . . . If i t i s colour that i s different you w i l l name the colour just as you did then. Shape can also d i f f e r . This i s an example of the shapes that may appear. (Slide 3) Please name each shape. There are 3 sizes. (Slide A) Would you name each size. When doing this task you have to be careful to name one characteristic only. For instance i f you said 'blue square' you would be naming two differences. 'Blue' i s a colour difference. 'Square' is a shape difference. On the next slide would you name only the shapes of the figures. (Slide 5) Now would you name only the sizes." (If the subject made a double attribute response further explanation was given.) "These items also have borders. These are the types of border which w i l l appear. (Slide 6) Would you name each border. There w i l l be different numbers of borders. (Slide 7) Would you name the numbers. The slides that you w i l l see w i l l have either three or six items. These next slides with two items are for practice. . 33 (Slide 8) Look at the l e f t hand item and name something that i s different. (Slide 9) Look at the l e f t hand item and name something which i s different. (Slide 10) Look at the l e f t hand item and name something which i s different. Now you w i l l look at examples of the type of three or six item slides that you w i l l be seeing. These two are for practice. (Slide 11) Look at the l e f t hand item and name something that i s different. (Slide 12) Look at the top l e f t hand item and name something that i s different. Now would you look at the bottom l e f t item and name something which is different." If the subject gave a double response this was pointed out and further explanation was given i f necessary. If he named an attribute value which did not distinguish the item, the experimenter said "No there is another that is . . . " and encouraged the subject to find i t . b) Instructions to the subject: Experimental slides When the introductory slides had been completed, the carousel was removed and the low attribute (A^) or high attribute (A^) slides were set in place. The experimenter said "Now you w i l l be seeing either three or six item slides. I w i l l t e l l you which item to look at and you w i l l name the thing which distinguishes i t from a l l the others. The f i r s t is a . . . item slide. Look at the . . . and name the thing which i s different." If the subject named a dimension which did not distinguish, the experimenter said "No there was another which was . . . . " The next 34 slide was then presented. If a double response was given, the experimenter asked "Was i t . . . or . . . . " Recording and Scoring Protocols had been prepared which l i s t e d the slides in the order in which they would appear. Examples of an A- and an A^ protocol are reproduced in Appendix C and Appendix D. The protocols enabled the experimenter to Inform the subject as to how many items he should expect and which one he should look at. They also carried a f u l l description of the designated item and li s t e d the correct response(s). They provided space for the experimenter to record the subject's verbal response and his response latency. Incorrect responses were noted as they occurred and the subject was informed when he made an error. Protocols were rechecked later to ensure that no errors had been missed. Treatment of Data Latency and error scores were treated separately, but both were handled by the same procedures. The raw scores of each subject were transferred from the original protocol to a summary sheet and were list e d under the condition to which they belonged. That is to say in the A^ condition each subject's latencies and errors were l i s t e d under either the three item (A^I-jP-j) o r the six item (A_I,P_) condition. In the A, condition they were l i s t e d under one 3 D j o of the four conditions which occurred in A^. These were low item/high probability, low item/low probability, high item/low probability and high item/high probability ( A g l ^ , A^I^y A ^ P g and AglgPg). Once they were li s t e d , latency and error scores in each of the A, 35 and Ag conditions were summed and averaged. Since there were 24 scores in each of the two A^ conditions, these scores were added and divided by 24. As the four A^ conditions contained only 12 scores each, these scores were summed and divided by 12. Analysis of variance was applied to the averaged scores. Additional methodological considerations a) Order of slide presentation The slides were presented in a counterbalanced order so that possible effects of the order of presentation could be controlled for. There were two types of stimulus array in the low attribute (A_) condition (A^I-P^ and A-I^-P-) and four types of stimulus array i n the high attribute (A,) j O 3 o condition (A^I^P^, A^I-P^, A^I-P^ and A^I^P^). However, both types of array would appear to the subject to di f f e r only with respect to the number of items on each slide. It was therefore arranged that both the subjects who viewed the A^ slides and the subjects who responded to the Ag slides would experience similar sequences of three and six item arrays. Counterbalancing was, i n fact, determined by the need to vary the order of presentation of A^ slides and these variations are described in Appendix E. There were, of course, eight groups of three slides each in both of the A^ conditions and four groups of three slides i n each of the four A^ conditions. b) Distribution of attribute-values On each designated item either three or six attributes were free to vary. The subject's latencies and possible error scores were sure to be affected by his "luck" i n choosing the particular attributes that he f i r s t 36 responded to. On an A^I^P^ array for instance, the discriminant might be shape. If the subject immediately reviewed the shape dimension, his latency would be short. If, however, he checked shape last, his latency would be appreciably longer. Nothing was known about the preferences of the subjects for choosing one dimension over another, but i t was thought necessary to control for subject response bias. It was therefore arranged that discriminants should be varied systematically and that each attribute should appear as a discriminant an equal number of times in each condition. Forty-eight arrays were presented in the condition and 48 i n the Ag condition. In the A^ condition shape appeared four times as a discriminant on the 1^ arrays and four times on the 1^ arrays. The same was true for each of the figure colour, background colour, size, border number and border type. A similar arrangement occurred on the arrays in Ag condition. Thus when the probability of a success by chance was low (Pg), each attribute appeared twice as a discriminant in I^ and twice in 1^. When the probability of a chance success was high (P-j)> there were two discriminants on each designated item and so each attribute appeared four times as a discriminant in each of I^ and 1^. Pilot Studies A certain amount of preliminary work needed to be done before the main study could be planned. To begin with i t was necessary to ensure that the experimental variables (number of items, number of attributes, and the probability of a success by chance) would produce differences in the response measures of unselected subjects. Also, since a repeated exposure of the same slide was desirable i f two levels of probability (P- and ?/) were to be included, i t was necessary to see whether significant 37 memory effects would appear at the second presentation. In addition, the question of selecting a mixed or a single sex group of subjects had to be decided. The literature suggested that males and females would d i f f e r significantly in their levels of field-articulation and that the correlation between the RFT and the EFT would be higher for males than for females. Two pilo t studies were planned in order to gain information about the effects of the experimental variables and of memory on the response measures, and normative data were collected on the levels of f i e l d articulation in male and female subjects. Experiment 1 In the f i r s t p i l o t study, 24 slides were shown to each of seven male and five female subjects. The six conditions which were used in the main study were also present in this experiment, but in the pilot study each subject responded to each condition. Thus there were no between-subjects variables. As in the main study the subject was asked to view a designated item and to name a discriminant. Response latencies and the number of errors were recorded. Experiment 2 Five male and five female subjects participated in the second pil o t study. As previously, their task was to identify a discriminant of a designated item, but instead of responding once to each slide, they responded twice to each slide. An item was designated before the slide appeared and when the subject responded, the array was automatically with-drawn. While the viewing screen was blank, the experimenter recorded the subject's verbal response and response latency. As soon as this had 38 been done, the subject was told that the same slide would reappear and he was given a new designated item. As before his verbal response and response latency were recorded. As in Experiment 1 a l l subjects viewed a l l of the stimulus arrays but, as only a few slides were available at this time, only the A^I-P-j, A_I,P_ and A,I,P, conditions could be studied. Since this experiment was 3 o 3 b o o designed to investigate the effect of memory on the response measures, i t was necessary that the probability of the subject obtaining a correct response by chance be the same for each of the two designated items on the same slide. This constraint could be satisfied only in the three conditions which are list e d above. In A ^ I - J P - J and A^I^P-j a l l items which carried a discriminant had a probability of P-. On the six item, six attribute arrays there were two items with probabilities of P^, but there was usually only one item with a probability of P_. On the three item, six attribute arrays, no level of probability was duplicated. It was therefore impossible to include A,I,P 0, A,I_P- and A,I_P,. Field-articulation ln male and female subjects The correlations which have been reported between the RFT and the EFT are quite variable and they seem to be affected by the nature of the sample which i s used. It therefore seemed advisable to obtain data on the levels of field-articulation and of correlation, between the EFT and the RFT, that could be expected among Canadian university students. The RFT and the EFT were therefore administered to 48 male and 52 female students. The same procedure was used as was followed in the main study. That i s to say, the French, Ekstrom and Price (1963) form of the EFT was used and the RFT procedure which i s outlined in Appendix B was followed. 3 9 CHAPTER III RESULTS Pilot Studies Two pilot studies were done using the stimulus arrays which were included i n the main study. Both of them used subjects who were un-selected with regard to f i e l d articulation. These studies w i l l be described below. Pi l o t work also included the collection of data on the comparability of male and female college students with regard to field-articulation. Experiment 1 The primary purpose of this experiment was to test the effectiveness of the attribute (A^ vs A^), item (Ij vs Ig) and probability (P.- vs P &) of correct response variables. Due to the fact that there were very few error scores, only the response latencies were analysed. The mean latencies for the A^ and Ag conditions, respectively, were 3.70 seconds and 4.55 seconds. The probability of correct response variable resulted in mean latencies of 3.39 seconds (P-,) and 6.03 seconds (P^). The number of items (I^ and 1^) yielded mean response latencies of 4.04 seconds and 4.49 seconds. The effects of the variables (A, I and P) was assessed by means of a within subject's (AxIxPxS) analysis of variance. The analysis of variance showed that there was a significant effect due to attributes (p<.001) and to probability of correct response (p<.001). The effect of the items variable did not result in s t a t i s t i c a l significance, but, there was a significant interaction between attributes and items (p<.05). 40 The results of this study Indicated that, as expected the effect of increasing the number of attributes from three to six and of decreasing the probability of a correct response from one in three to one in six, was to lengthen response latencies. Had a significant main effect for the items variable been present, the significant interaction between items and attributes would have been i n accordance with the hypothesis that increase in the complexity of the stimuli should be accompanied by longer response latencies. There is no really satisfactory explanation for the occurrence of an interaction in the absence of a main effect. Since two of the predicted main effects appeared, and since the interaction was in accordance with expectations, i t was decided that the task would provide a satisfactory test of the experimental hypotheses. Experiment 2 This experiment was designed to study the effects of presenting the same slide twice on response latencies and error scores. It was expected that response latencies might be shorter and errors fewer when the slides were presented for the second time. Only slides which contained the A^I^Py A^IgP-j and AglgPg a r r a y s w e r e used. As in Experiment 1 there were very few errors and only response latencies were analysed. A t_-test was used to compare latencies for the responses to the f i r s t and second presentation of the slides. It was found that latencies were significantly shorter on the second presentation under the A^I^P^ condition (p<.001) and under the A^I^P^ condition (p<.05). There were no significant differences in the latencies for the f i r s t and second presentations under the low probability A,I,P, 41 condition. On the basis of the results i t was concluded that the arrays with six attributes and six items could be presented twice in succession without memory effects occurring. Since no memory effects were observed on low probability (Agl^Pg) slides, which subjects responded to after long latencies, none should occur with high probability (AgI_Pg) arrays which the subjects viewed for a shorter period. No data were available for six attribute, three item slides. However when data from the other three conditions were considered i t seemed that memory effects tended to diminish as the complexity of the arrays increased. Six attribute, three item arrays were complex and i t was presumed that i f memory effects occurred, they would be of marginal significance only. If other slides were interposed between the f i r s t and second presentations of slides, no effects of memory should be evident. It was therefore decided that six attribute (Ag) slides should be presented twice, so that the effect of varying the probability of success could be observed in f i e l d independent, neutral and f i e l d dependent groups, but that at least three other slides should be interposed between the f i r s t and second presentations of any array. Field-articulation in male and female subjects The mean scores of 52 college females and 48 college males on the EFT were 11.6 and 13.0 respectively. The mean scores of the same subjects on the RFT were 26.4 for females and 20.3 for males. The average male subject was more f i e l d independent than the average female subject on both tests. This suggested that i t might be impossible to combine male and female scores for the purposes of conducting an analysis of variance and 42 therefore only males were included in the main study. Major Research Project An analysis of variance (ANOVA) was carried out in order to determine whether or not f i e l d independent, neutral and f i e l d dependent groups of subjects would be d i f f e r e n t i a l l y affected by manipulation of the stimulus variables (A^ vs A^, I- vs 1 ^ and P^ vs P^). If Witkin's interpretation of the field-articulation dimension were correct, then an interaction should occur between the field-articulation groups (FID-group, N-group and FD-group) and the variables of the experimental task (A, I and P). If, on the other hand, E l l i o t t ' s conceptualization were supported, then no interactions should occur between field-articulation groups and stimulus variables since, according to E l l i o t t , the factors which cause high, medium and low field-articulation are not under stimulus control. In this ANOVA main effects were of interest because their presence would indicate a) that the experimental task produced the same type of separation between field-articulation groups as did the selection tests (RFT and EFT) and b) that the manipulation of task variables caused changes in the response measures of the subjects when the groups were combined. Main effects which can be attributed to field-articulation groups w i l l be discussed f i r s t ; main effects which were due to attributes, items and probability of chance success w i l l be discussed next. The analyses of response latencies and of error scores w i l l be presented separately. Latencies The analysis of variance for response latencies in seconds i s presented 43 in Table 3. As was expected, there was a significant main effect (p<.025) which was attributable to groups (levels of field-articulation). The mean latency of the FID group was 10.35 seconds, that of the N group was 12.08 seconds, while that of the FD group was 12.38 seconds. As predicted the mean latency of the FID group was the shortest while that of the FD group was the longest. The presence of a significant main effect due to Groups suggests that similar factors may have contributed toward differences on both the measures of field-articulation and the response latencies. This assumption is supported by the fact that there were a number of significant correlations between the EFT and the combined EFT/RFT scores on the one hand and the response latencies on the other. Table 4 presents the correlations between the RFT, the EFT and the combined RFT/EFT scores and the response latencies in the low attribute condition. The corresponding correlations for the high attribute condition are presented in Table 5. It w i l l be noted (Table 4 and Table 5) that a l l the correlations are in the predicted direction. That is to say, those subjects who achieved high scores on the EFT and the RFT tended to have short latencies of response. Table 4 shows that the combined RFT/EFT score, which was used to separate the subjects into f i e l d independent, neutral and f i e l d dependent groups, i correlated significantly (p<.01) with response latency in both of the low attribute conditions (A_I 0P~ and A_I,P_). Correlations between the EFT 3 3 3 J o J scores and latencies were also significant (.005), but the RFT scores did not correlate significantly with the response measure. Table 5 shows that the combined RFT/EFT score correlated significantly with latencies in the A,I_P- (p<.025) and A-I,P„ (p<.05) conditions. Significant correlations 0 3 3 o b J 44 Table 3 Analysis of Variance of Response Latencies Source SS df MS F Between Subjects 413.99 95 Groups (G) 25.66 2 12.83 4.39** Attributes (A) 122.14 1 122.14 41.82**** G x A 3.31 2 1.66 .57 Ss/GxA 262.87 90 2.92 Within Subjects 585.69 192 Items (I) 91.98 1 91.98 74.06**** I x A 13.00 1 13.00 10.46*** I x G 1.86 2 .93 •73 I x A x G .17 2 .08 .07 I x Ss/A x G 111.78 90 1.24 Probability (P)/A 6 209.89 1 209.89 87.05**** P x G/Aft 6.88 2 3.44 1.43 P x S/G/A6 108.51 45 2.41 I x P/A6 10.75 1 10.74 16.33**** I x P x G/A6 1.27 2 .64 .97 I x P x Ss/G/A6 29.59 A-> .66 Totals 999.68 287 **p <.025 ***p <.005 ****p <.001 45 Table 4 Correlations between the Selection Tests and Response Measures in the Low Attribute (A„) Condition RFT EFT COMB A3 T3 P3 A3 I6 P3 RFT 1.000 EFT .3492*** 1.000 A3 r3 P3 -.2086 -.4480**** -.3915*** 1.000 A 3 X 6 P 3 -.1737 -.3989**** -.3426*** .8268**** 1.000 ***p<.01 ****p<.005 46 T a b l e 5 C o r r e l a t i o n s b e t w e e n t h e S e l e c t i o n T e s t s and Response Measures i n t h e H i g h A t t r i b u t e ( A , ) C o n d i t i o n RFT EFT COMB A 6 I 3 P 3 A 6 I 3 P 6 A 6 I 6 P 3 A 6 I 6 P 6 RFT 1 .000 EFT . 3 5 8 5 * * * 1 .000 V 3 P 3 - . 1 2 1 0 - . 3 5 6 3 * * * - . 3 1 4 3 * * 1 .000 V 3 P 6 - . 0 2 5 1 - . 1 2 4 1 - . 1 0 1 2 . 6 8 8 3 * * * * 1 .000 V 6 P 3 - . 0 4 9 0 - . 3 3 9 7 * * * - . 2 6 7 4 * . 6 4 8 3 * * * * . 4 6 9 1 * * * * 1 .000 V 6 P 6 - . 0 2 1 0 - . 2 2 0 8 - . 1 6 8 6 . 5 4 3 8 * * * * . 5 9 4 1 * * * * . 3 8 7 7 * * * 1 .000 * p < . 0 5 * * p < . 0 2 5 * * * p < . 0 1 * * * * p < . 0 0 5 47 were also obtained between the EFT and the A^I^P-j condition (p<.01) and the A,I,P_ (p<.01). o b i Returning to Table 3 i t can be seen that there were significant main effects for the attribute (p<.001), item (p<.001) and the probability of success (p<.001) variables. The mean latency in A^ and A^ were 2.95 seconds and 4.33 seconds respectively. When 1^ and 1^ were compared, the mean latencies for the two conditions were 3.30 and 4.42 seconds. Probability of success was a variable in A^ only. Within this condition the mean latencies for P^ and P^ were respectively 3.28 seconds and 5.37 seconds. As they stand these results suggest that a l l the experimental manipulations were reflected in the response latencies of subjects in the way that had been predicted. However, the attribute variable was con-founded with the probability of correct response variable. In the A, condition, arrays contained both the low (P,) and the high D O (P^) probability of success conditions, whereas i n the A^ condition only P^ was present. A^ latencies were, in fact, composed of effects due to attributes and effects due to high and low probability of chance success. Since the probability-of-success variable caused significant differences in response latency, i t is to be expected that A^ scores would be considerably increased by the presence of P^. A subsiduary analysis was carried out in an attempt to assess the effects of attributes exclusive of probability of success. The A^I^Pg and A,I,P, conditions were omitted and analysis of variance was carried o o o out as though only A^P.^, A 3 I 6 P 3 » A D 1 2 P 3 a n d A6 I6 P3 c o n d i t i o n s h a d b e e n present. The results are presented in Table 6. It w i l l be seen that a significant main effect (p<.025) due to attributes was present in the 48 Table 6 Analysis of Variance for Response Latencies with Low Probability of Success Scores Excluded Source SS df MS F Between Subjects 90.80 95 Groups (G) 13.70 2 6.85 8.62**** Attributes (A) 5.42 1 5.42 *.82** G x A .16 2 .08 Ss/G x A 71.53 90 .79 Within Subjects 97.52 96 Item (I) 26.53 1 26.53 35.92**** I x G 1.87 2 .93 1.26 I x A 2.20 1 2.20 2.98 I x A x G .43 2 .22 .29 I x S/G x A 66.48 90 .74 Totals 157.32 191 **p<.025 ****p<.001 49 subsiduary analysis as in the main analysis. The mean latency of subjects was 2.95 seconds In and 3.28 seconds in A^. For an il l u s t r a t i o n of the effects of conducting separate main and subsidiary analyses, reference may be made to Figures 4(a) and 4(b). Figure 4(a) shows the mean latencies of the FID, N and FD groups as they appeared in the A^ and A^ conditions In the main analysis. Figure 4(b) shows these latencies in the analysis i n which the A-I^P^ and A^I^P^ conditions were omitted. ' To return to the main analysis, Table 3 shows that there was a significant interaction between items and attributes. As the arrays became more complex the latencies of subjects became longer. In the simplest arrays which had only three Items and three attributes the mean latency was 2.68 seconds. In the intermediate arrays which had three attributes and six items or six attributes and three items, the mean latencies were 3.21 seconds and 3.61 seconds respectively. The longest mean latency, 5.01 seconds, occurred on the most complex six attribute, six item arrays. This effect provided additional evidence as to the effectiveness of the experimental manipulations. There were no interactions between level of f i e l d articulation and the levels of the experimental variables. As is shown in Table 3, the interaction between Groups (FID, N and FD) and attributes ( A - j vs A^), items (I^ vs 1^ ) and probability of success (P. vs P^) were a l l non-significant. Witkin's hypothesis that increases in the complexity of a visual task would have dif f e r e n t i a l effects on f i e l d independent, neutral and f i e l d dependent subjects was not supported. The mean scores for the three groups of subjects In the low (A_) 50 Mean l a t e n c y i n seconds 5 . 0 ^ 4.0-3 . 0 2 . 0 . 1.0. • F.I. D. • N . • F.D. A 3 A 6 Number of a t t r i b u t e s (a) Mean latencies include latency due to attributes and latency due to probability of success. 5 . 0 M e a n l a l e n c y i n seconds 4 . 0 , 3 .0 . 2.0, 1.0. • F. I. D. A N . • F. D. Km I » A . A 3 6 N u m b e r o f a l t r i b u t e s (b) Mean latencies with latency due to probability of success excluded. Figure 4. Mean Latency of F.I.D., N. and F.D. Groups in Response to Low (A 3) and High (A&) Attribute Stimulus Arrays. 51 and high ( A ^ ) attribute conditions are shown in Figures 4(a) and 4(b). Figure 4(a) illu s t r a t e s the mean scores of the experimental groups in the main analysis. In the low attribute (A-.) condition the mean latencies of the FID, N and FD groups were respectively 2.67 seconds, 2.90 seconds and 3.27 seconds. In the high attribute (A,) condition, D these means were 3.83 seconds, 4.59 seconds and 4.55 seconds respectively. Again i t may be noted from Table 3 that the Groups effect was significant (p<.025) while the Groups x Attributes interaction was not significant. Thus, the fie l d - a r t i c u l a t i o n groups may be ranked in terms of their efficiency, with the FID-group displaying the best and the FD-group displaying the poorest performance, regardless of the number of attributes present. Figure 4(b) shows the mean latencies of the groups when, as in the subsiduary analysis, Agl^Pg and A^I^P^ are excluded. The mean latencies i n A^ were, of course, unchanged. In A^ the mean latencies were 2.93 seconds for the FID-group, 3.28 for the N-group and 3.64 for the FD-group. Reference to Table 6 w i l l show that the effect due to fie l d - a r t i c u l a t i o n groups was again significant (p<.001) whereas the interaction between Groups and Attributes was not. This analysis again suggests that the FID-group performed the task most effectively and the FD-group performed i t least effectively regardless of the stimulus input. Figure 5 shows the mean latencies of the FID, N and FD groups when the low (I-.) and high (1^) item conditions were compared. In the I_ condition the FID, N and FD groups had latencies of 2.89 seconds, 3.53 seconds and 3.44 seconds respectively. In the 1^ condition the latencies were respectively 3.96 seconds, 4.52 seconds and 4.80 seconds for the FID, N and FD groups. The comparison of the mean latencies of the experimental 52 Figure 5 . Mean Latency of F.I.D., N. and F.D. Groups in response to Low (I_) and High (I.) Item Stimulus Arrays. 53 groups in the high ( P ^ ) and low ( P ^ ) probability of success conditions i s shown in Figure 6. These means were 2.93 seconds, 3.28 seconds and 3.64 seconds respectively for the FID, N and FD groups in the P^ condition and 4.76 seconds, 5.91 seconds and 5.47 seconds in the Pg condition. These figures provide a visual i l l u s t r a t i o n of the inter-action results which are presented in Table 3. It w i l l be noted that the three groups appear to have been very similar in the extent to which they were affected by the manipulation of the stimulus variables. There was no interaction between groups and items or between groups and probability of success. Errors The same predictions were made for error scores as for latencies. Again the purpose of the study was to investigate the interaction effects between f i e l d articulation groups and task variables. Since relatively few errors were made in this study, the analysis which i s to be reported involves less than 10% of the responses which were made by the subjects. The analysis of variance of the error scores i s presented in Table 7. There was a significant main effect (p<.05) which was attributable to f i e l d articulation groups. The mean number of errors made by the FID group was 0.68. The N and FD groups made an average of 1.02 and 1.11 errors respectively. As expected most errors were made by the f i e l d dependent subjects and least by the f i e l d independent subjects. Since the average number of errors which were made by the subjects was low, no attempt was made to correlate error scores with the measure of f i e l d articulation (RFT and EFT). As Table 7 shows, there were significant main effects which were 54 P r o b c b i l i t y o f success Figure 6. Mean Latency of F.I.D., N. and F.D. Groups in Response to Low (p ) and High (p ) P r o b a b i l i t y Stimulus Arrays. 55 Table 7 Analysis of Variance for Error Scores Source SS df MS F Between Subjects 141.8A 95 Groups (G) 10.50 2 5.25 3.64* Attributes (A) .31 1 .31 .21 G x A 1.2 2 .60 .41 Ss/GxA 129.83 90 1.44 Within Subjects 182.24 192 Items (I) 2.03 1 2.03 3.53* I x A .31 1 .31 .54 I x G .92 2 .46 .80 I x A x G 1.22 2 .61 1.07 I x Ss/A x G 51.67 90 .57 Probability (P)/Ag 42.19 1 42.19 44.28**** P x G/A& 1.92 2 .99 1.04 P x S/G/A6 42.88 45 .95 1.12 I x P/A, 0 .00 1 .00 I x P x G/A- .85 2 .42 .49 I x P x Ss/G/A6 38.25 45 .85 Totals 323.18 287 | *p .05 ****p .001 56 due to probability of success (p<.001) and to items (p<.05). The subjects made an average of 0.98 errors in the high probability condition and an average of 1.43 errors i n the low probability condition. There were a mean number of 1.02 errors made on three item arrays and a mean of .87 errors made on six item arrays. Thus, though both these effects were significant, only the effects due to probability of correct response were in the expected direction. More errors would have been expected on six item arrays than on three item arrays. The opposite effect occurred. This result w i l l be discussed more f u l l y later. As i s shown in Table 7, the attribute variable was not significant i n the main analysis. The mean number of errors in the A^ condition was 1.76 and the mean number of errors in the A^ condition was 1.94. As with the latency measure the effects of attributes was confounded with the effects of probability of success, in the main analysis. A subsiduary analysis, similar to that which was done for latency scores, was therefore carried out. The results of this ANOVA are presented in Table 8. As Table 8 shows, there was a significant effect (p<.001) which was due to attributes. The mean number of errors in the A^ condition was 1.76 and the mean number in the A^ condition was .98. The direction of this effect i s contrary to that which would have been predicted. It seems appropriate, at this point, to devote some space to the discussion of the item and attribute variables. Increases from three to six items and from three to six attributes resulted in increases in the response latencies of subjects, but in decreases in the number of errors that they made. This seems to suggest that the two measures were related and that when more time was spent reviewing a stimulus array, 57 Table 8 Analysis of Variance for Error Scores with Low Probability Scores Excluded Source SS df MS F Between Subjects 54.08 95 Groups (G) 3.69 2 1.84 4.02** Attributes (A) 7.65 1 7.65 16.70**** G x A 1.48 2 .74 1.61 Ss/G x A 41.26 90 .48 Within Subjects 56.09 96 Items (I) 1.75 1 1.75 3.86* I x G .23 2 .11 .25 I x A .23 1 .23 .51 I x A x G .27 2 .14 .30 I x S/G x A 40.78 90 Totals 110.17 191 *p<.05 **p<.025 ****p<.001 fewer errors were l i k e l y to be made. The relationships between latency and error scores are shown in Figure 7. Figure 7 shows the mean latency of response and mean number of errors which were made in the high and low attribute and in the six and three item conditions. By way of contrast the mean latency and mean error scores for the high and low probability of success conditions i s illustrated in Figure 8. A return to Table 7 shows that, as with the latency scores, interactions between levels of field-articulation and task variables were not significant. This means that the hypothesis that FID, N and FD groups would be di f f e r e n t i a l l y affected by changes In the task variables was not supported. Again the field-articulation groups maintained the same relative levels of efficiency even though the stimulus inputs changed. Figure 9(a) ill u s t r a t e s the comparison of the low (A.) and high (Ag) attribute conditions in the main analysis. In A^ FID, N and FD groups averaged 1.30, 1.68 and 2.32 errors. In the Ag condition these groups averaged 1.36, 2.22 and 2.22 errors. Figure 9(b) shows the mean error score in the subsiduary analysis. The mean score of the experimental groups in the A^ condition remain as before. In the Ag condition FID, N and FD groups made an average of .68, 1.24 and 1.04 errors. It w i l l be noted that the effect due to groups was significant and that the groups-by-attributes interaction was non-significant regardless of whether or not the A^I^Pg and A^I^P^ conditions were included in the analysis. As in the analysis of latency scores, the experimental task was performed most e f f i c i e n t l y by the FID-group and least e f f i c i e n t l y by the FD-group. 5 9 6.0-j . M e a n 1 l a t e n c y II second s 5.0J 4.0. I t ems A t t r i b u t e s 3.0. ---. A. 2-0. 1.0. 0-5 1-0 1-5 Mean e r r o r s 2-0 Figure 7. Association between Mean Latency and Mean Error Scores In Low (I ) and High (I,) Item Conditions and in Low (A_) and High (A,) Attribute Conditions. 60 M e a n e r r o r s Figure 8. Association between Mean Latency and Mean Error Scores in High (P~) and Low (P,) Probability of Success Conditions. 61 / d e a n e r r o r s 2 - 5 2 - 0 . 1-5 1-0, 0 - 5 . • E I. D. • N . • F. D. J A t t r i b u t e s ° (a) Mean Errors include Errors due to Attributes and Errors due to Probability of Success. .Mean e r r o rs 2 - 5 , 2 - 0 j 1-5. 1-0. 0-5 • F. f. D. • N • F. D. A A t t n " b u t e s (b) Errors due to Probability of Success Excluded from Mean Errors in Low (P^) and High (P &) Attribute Conditions. Figure 9. Mean Errors of F.I.D., N. and F.D. Groups in Response to Low (A^) and High (A^) Attribute Stimulus Arrays. 62 When the three item and six item conditions were compared the respective mean error scores of the FID, N and FD groups were .70, 1.09 and 1.29 in the I 3 condition and .65, .96 and .96 in the I f e condition. These effects are illustrated in Figure 10. The probability of success means are shown in Figure 11. This variable appeared In the high attribute condition only, and the means of the three experimental groups in the P_^  condition were .68 (FID-Group), 1.24 (N-Group) and 1.04 (FD-Group). In the P^ condition the mean number of errors were 1.01 (FID-Group), 1.59 (N-Group) and 1.69 (FD-Group). Once again the main effect which was attributable to groups was apparent, but the interactions between groups and items and between groups and probability of success were not significant. Subjects were assigned to FID, N and FD groups on the basis of their combined scores on the RFT and the EFT. The correlations between these tests were .35 in the condition and .36 in the Ag condition. Since these correlations are relatively low there i s some question as to whether the absence of interaction effects was due to the low correlation. It might be argued that one or another of the measures used alone would have produced interaction effects, but that the grouping of subjects in terms of a combined score lead to misplacement in f i e l d -articulation groups. This question can be resolved i f the extreme groups (FID and FD) are considered separately. The scores of a l l subjects on the RFT and the EFT are shown in Appendix F. It w i l l be seen that the mean score on the EFT was 11.61 and that the mean score on the RFT was 20.49. No FID subject scored below the mean on either selection test. Only two FD 63 1 te ims Figure 1 0 . Mean Number of Errors of F.I.D., N. and F.D. Groups in Response to Low (I ) and High (I,) Item Arrays. i i P3 P6 P r o b a b i l i t y o f success Figure 1 1 . Mean Number of Errors of F.I.D., N. and F.D. Groups in Response to High (P^) and Low (P^) Probability of Success Conditions. 64 subjects scored above 11.61 on the EFT and only seven scored above 20.49 on the RFT. The FID and FD groups therefore represent extreme samples on both selection measures. If interaction effects were present they should be discernible when these groups are compared. Reference to Figures 4(b), 5, 6, 10 and 11 indicates clearly that no interactions occurred between field-articulation extreme groups and the experimental variable. Figure 9(b) shows an interaction between FID and FD groups and attributes when errors were measured. However the FD subjects made fewer errors in the high attribute condition not more errors as would have been predicted. This result does not support Witkin's hypothesis. It might have been interesting to conduct a separate analysis using the RFT and EFT measures separately. However i f this were done, the results would be uninterpretable. Even i f the subjects were re-assigned to groups within the A.- and A^ conditions (as would be necessary since attributes was a between subjects variable) i t would be impossible to maintain equivalence in the order of counterbalancing. Effects due to experimental variables and effects due to order of slide presentation would be confounded. 65 CHAPTER IV DISCUSSION OF RESULTS In the study which has been presented the field-articulation dimension was defined (Silverman, 1968; Witkin et a l . , 1962, 1964, 1965) and an experimental task was described. This task has been constructed to test the hypothesis that groups of subjects with differing degrees of f i e l d independence would be d i f f e r e n t i a l l y affected by changes in the complexity of stimulus arrays. According to Silverman (1968) and Witkin (1962, 1964, 1965) f i e l d independent subjects should have shown the least decrement on response measures when the amount of stimulus input was increased and f i e l d dependent individuals should have shown the most decrement. Subjects who were intermediate for f i e l d independence should have shown moderate decrement on response measures. According to the opposed hypothesis of E l l i o t t (1961) however a l l groups should have shown similar changes in response to stimulus arrays of graded complexity and d i f f i c u l t y . The results which have been presented clearly supported E l l i o t t ' s point of view. Increases i n the number of attributes and the number of items and decreases i n the probability of success by chance, resulted in the same amount of change on response measures (latency and error scores) in the FID, N and FD groups. No interaction was observed between the level of field-articulation and the stimulus parameters. It therefore seems l i k e l y that the factors which produced the differences between the FID, N and FD groups operated independently of the stimulus. Similar results have appeared in another study. Davies (1967) presented f i e l d independent and f i e l d dependent subjects with a visual 66 concept learning task, which had three levels of d i f f i c u l t y , and measured the number of errors which occurred before c r i t e r i a was reached. He found that, contrary to the hypothesis, there was no inter-action between the number of distractors, which made the task more d i f f i c u l t , and the levels of f i e l d articulation of the groups of subjects. Field independent and f i e l d dependent subjects were affected equally by changes i n stimulus conditions. The differences between them seemed to exist independently of the nature of the stimulus. The results of the present study give rise to a number of questions. It might be asked whether the experimental task provided an adequate test of the hypotheses of Witkin (1962, 1964, 1965) and Silverman (1968). The adequacy of the response measures might also be questioned. Finally the low correlation of the RFT and EFT needs to be considered. To deal f i r s t with the question of the experimental task. Silverman (1968) stated clearly, in his elaboration of Witkin's hypothesis, that the field-articulation dimension had two components. One component described differences in the degree to which individuals are distracted by conflicting stimulus inputs. The other component referred to a tendency of f i e l d independent subjects to be better able than f i e l d dependent subjects to respond to discrete segments of a configuration. The f i r s t of Silverman's factors seems to have been adequately covered by the task which the subjects were asked to perform. A l l the manipulations of the stimulus conditions involved changes in the amount of input that the subjects must process in order to make a response. In addition a l l these manipulations were reflected in changes in the latencies which preceded responses. This suggests that as was intended, the subjects experienced 67 variations in the processing load and responded more rapidly to low-load arrays than to high-load arrays. It i s more d i f f i c u l t to be sure that the second component of Silverman's formulation was tested adequately. According to Silverman, f i e l d independent, neutral and f i e l d dependent subjects should d i f f e r in their a b i l i t y to respond to discrete segments of a perceptual f i e l d . Unfortunately the distinction between responding to discrete segments and resisting conflicting stimulus inputs is not entirely clear. This means that i t was d i f f i c u l t to define the variables which pertained sp e c i f i c a l l y to the a b i l i t y to break up a configuration, and to be sure that they were included in the study. It was considered that the low probability condition would demand a greater a b i l i t y to segmentalise than the high probability condition. It was thought that in this condition i t would be more necessary for the subject to maintain an awareness of the individual attribute because he would have to review more of them in order to make a response. It was presumed that failure to segmentalise would lead to confusion and that this would be reflected in long latencies and frequent errors. The results of using probability of success as a test of the subject's a b i l i t y to segmentalise are not entirely clear. Both response latencies and error scores were significantly increased in the low probability of success (A^I^Pg. A^I^P^) conditions. However the correlations between the selection measures (RFT and EFT) and these conditions were not significant. This might mean either that the EFT, RFT and combined EFT/RFT scores reflected variables which were not represented in the experimental task, or, that the a b i l i t y to segmentalise configurations is not an 68 essential element of f i e l d articulation. Clarification of this issue waits upon further research and upon a more precise definition of Silverman's conceptualization of f i e l d articulation. The presence of long latencies and many errors suggested that a l l groups found the low probability of success condition extremely d i f f i c u l t . There were two dependent variables in this study: latency of response and error scores. The former seems to have been the more sensitive in that i t reflected a l l the experimental manipulations. Differences could be attributed to levels of f i e l d articulation and to the number of attributes, the number of items and the probability of success. There was an additional advantage in that latency scores were available for a l l of the subject's responses. The second measure, error scores, was less satisfactory. It appeared to be independent of the latency of response measure for only two of the experimental variables. When levels of f i e l d articulation were compared, short latencies appeared to be associated with few errors and longer latencies w i t h more errors. The same was true when the high and low probability conditions were considered. It was apparent, from the scores on both of the dependent variables, that the f i e l d dependent group had most d i f f i c u l t y performing the discriminant identification task while the f i e l d independent group had least d i f f i c u l t y . Both response measures also indicated that the low probability of success condition was more d i f f i c u l t than the high probability of success condition. When the response measures were compared for the high and low item and attribute variables however, short latencies were associated with high error scores. This suggested that, when the independent variable was either the number of attributes or the number of 69 items, error scores might be a function of response latency. Turning now to the question of the relatively low correlation between the selection measures (RFT and EFT), i t is obvious that each test measured both factors which were shared and factors which were not shared. The shared factors have been the subject of the present study. The other factors have received some discussion in the literature. Witkin (1962, page 36) wrote of the RFT that in order to perform effectively a subject must "'extract' the rod from the t i l t e d frame through reference to body position." There was no suggestion, however, that body position played an important part in response to the EFT, and intuitively i t seems unlikely that this would occur. Awareness of body position may therefore be a factor which specifically affects response to the RFT but which does not affect response to the EFT. With regard to specific factors affecting the EFT, E l l i o t t (1961) pointed out that this test i s timed, whereas the RFT is not. He suggested that timing might introduce an element of stress which was not present when subjects responded to the RFT. E l l i o t t noted that the EFT tends to be more highly correlated with a b i l i t y factors than does the RFT. In the present study intuitive comparisons would suggest that the experimental task was more similar to the EFT than to the RFT. Both the EFT and the discriminant identification task presented the subjects with visual stimuli, both were timed and neither required dependence on bodily cues. Study of the correlations between the combined EFT/RFT, the EFT and the RFT and response latencies tended to confirm the intuitive deduction. The combined EFT/RFT scores and the EFT scores correlated more highly with the response measure than did the RFT scores. It is therefore evident that 70 allowance nraist be made for specific test factors when conceptualizations about the f i e l d - a r t i c u l a t i o n dimension are made. Since this study provided no support for the hypothesis (Witkin, 1962, 1964, 1965; Silverman, 1968) that dependence on outside inputs exists as a personality t r a i t , an alternative explanation must be found for the differences which are observed between f i e l d independent and f i e l d dependent subjects. E l l i o t t (1961) suggested that the focus of investi-gation should be changed from the study of personality to the study of brain function. He proposed that some of the theories of intelligence (Werner, 1957; Piaget, 1950; Goldstein, 1953; or Pribram, 1960) which are now current might provide f r u i t f u l hypotheses. Insofar as these theories deal with the manner in which stimulus inputs are processed, the present results would support such a change of emphasis. In the present study the relevant factor appears to have been the amount bf input that the stimulus arrays contained. The nature of the input seems to have been of l i t t l e importance. The three stimulus variables (number of items, number of attributes, probability of success) seem to have been equivalent in terms of their effects on response latency. When the processes by which normal individuals deal with complex stimulation are better understood, then i t w i l l be easier to investigate the disruption of these processes which appears to occur in abnormal groups. In the case of schizophrenic patients, for instance, Silverman (1969) isolated f i e l d independence and two factors (Anchoring and Perceptual Disorganization) which seemed to be specific to schizophrenia. It might be useful for the purposes of treatment and prognosis to know more about 71 field-articulation and about the way in which this attentional factor interacts with the other, presumably psychotic factors. If f i e l d independence indicates more effective processing of stimulation than f i e l d dependence, then perhaps the f i e l d independent person has a better chance of recovery once the schizophrenic process i s halted or reversed. On the other hand, i t might be, as Witkin (1965) suggested that individuals with differing levels of f i e l d articulation react in different ways to the experience of schizophrenic disruption. If this were true, then some basis might be provided for understanding the different patterns of schizophrenic symptomatology which have been described. The appearance of high levels of field-dependence in non-psychotic c l i n i c a l groups, notable alcoholics, poses a different problem. Alcoholics are not generally considered to suffer from intellectual dysfunction in the way that schizophrenics do. Traditionally dependence on alcohol has been regarded as a manifestation of the personality t r a i t , dependency. When f i e l d dependence was considered to reflect extreme susceptibility to external stimulation, then this level of field-articulation was regarded as being yet another instance of the gross personality t r a i t , dependency. The results of the present study suggest that alcoholics cannot be regarded as being especially dependent on external stimuli, therefore some other explanation must be found for their poor performance on the RFT and the EFT. Perhaps, as is suggested by the studies of Kristofferson (1966) and of Goldstein and Chotlos (1966), their relative d e f i c i t might have been caused by the ingestion of alcohol and by a consequent disruption of brain function. 72 Conclusions On the basis of the present study, in which FID, N and FD subjects responded to complex visual stimuli, i t was concluded that f i e l d -dependence does not reflect differing degrees of dependency on external stimulus inputs. Instead differences among FID, N and FD Groups seemed to be unaffected by the characteristics of the external stimulus f i e l d . It i s therefore suggested that the focus of future investigation should be changed and that more attention should be paid to the study of the means by which the processing of incoming stimulation Is carried out. 73 References Adevai, G., Silverman, A. J. & McGough, W. E. Perceptual correlates of the Rod and Frame Test. Perceptual and Motor S k i l l s , 1968, 26, 1055-1060. Astrup, A. M. In 'Reduction of f i e l d dependence in chronic alcoholic patients.' Author G. R. Jacobson. Journal of Abnormal Psychology, 1968, 73, 547-549. Bailey, W., Hustmeyer, F., & Kristofferson, A. Alcoholism, brain damage and perceptual dependency. Quarterly Journal Studies on Alcohol, 1961, 22, 387-393. Beller, E. A study of dependency and perceptual orientation. Paper read at American Psychological Association, Washington, 1958. Cancro, R., & Sugarman, A. Psychological differentiation and process-reactive schizophrenia. Journal of Abnormal Psychology, 1969, 74, 415-419. Cancro, R., & Voth, H. M. Autokinesis and psychological differentiation. Perceptual and Motor S k i l l s , 1969, 28, 99-103. Cattell, R. B. Personality and Motivation, Structure and Measurement. New York, 1957. Chapman, J. The early symptoms of schizophrenia. Psychiatry, 1966, 112, 22-251. Crutchfield, R. S., Woodworth, D. G., Albrecht, R. E. Perceptual performance and the effective person. Lackland AFB, Texas Personnel  Lab. Report WADC-RN-58-60. ASTIA No. AD 151 039. Davies, J. K. Concept identification as a function of cognitive style, complexity and training procedures. Unpublished doctor's dissertation, 74 University of Wisconsin, 1967. Microfilm, U.B.C. Dickstein, L. S. Field independence in concept attainment. Perceptual and Motor S k i l l s , 1968, 27, 635-642. E l l i o t t , R. Inter-relationships among measures of field-dependence, a b i l i t y and personality t r a i t s . Journal of Abnormal and Social Psychology, 1961, 63, 27-36. Eysenck, H. J. Experiments in Personality: Volume II, Routledge & Kagan Paul, 1960. Eysenck, H. J. Handbook of Abnormal Psychology: An Experimental Approach. Basic Books, 1961. Fishbein, G. M . Perceptual modes and asthmatic symptoms: An application of Witkin's hypothesis. Journal of Consulting Psychology, 1963, 27, 54-58. French, J. W. (Ed.) Manual for k i t of selected tests for reference aptitude and achievement factors. Educational Testing Service, Princeton, New Jersey, 1954. French, J. W., Ekstrom, R. B., & Price, L. A. Manual for k i t of reference tests for cognitive factors. Educational Testing Service, Princeton, New Jersey, 1963. Gardner, R. J., Holzman, P. S., Klein, G. S., Linton, H. B., & Spence, D. Cognitive control: A study of individual consistency in cognitive behaviour. Psychological Issues, 1959, 1(4), New York, International Universities Press. Gardner, R. W., Jackson, D. N., & Messick, S. J. Personality organization in cognition, controls and intellectual a b i l i t i e s . Psychological  Issues, 1960, 2(8), New York, International Universities Press. Goldstein, K. The significance of psychological research in schizophrenia. 75 In S. S. Tomkins (Ed.), Contemporary Psychopathology. Cambridge: Harvard University Press, 1943, pages 302-318. Goldstein, S., & Chotlos, J. Stability of f i e l d dependence in chronic alcoholic patients. Journal of Abnormal Psychology, 1966, 71, 420. Goodman, B. (1960) Field dependence and the closure factors. In Witkin (Ed.) Psychological Differentiation, John Wiley & Sons, New York, 1962. Gordon, B . An experimental study of dependence-independence in a social and a laboratory setting. Unpublished doctor's dissertation, University of Southern California,1953. Gruen, A. Relationship of dancing experience and personality to perception. Psychological Monographs, 1955, 69, No. 9 (Whole No. 399). Gruenfeld, L., & Arbuthnot, J. Field independence as a conceptual frame-work for prediction of v a r i a b i l i t y in ratings of others. Perceptual  and Motor S k i l l s , 1969, 28, 31-44. Holzman, P. S., & Klein, G. S. Cognitive system-principles of levelling and sharpening: Individual Differences in assimilation effects in visual time-error. Journal of Psychology, 1954, 37, 105-122. Holzman, P. S., & Klein, G. S. Perceptual attitudes of "Leveling" and "Sharpening": Relation to individual differences in time-error. American Psychologist, 1951, 6, 257. Jackson, D. N. A short form of Witkin's Embedded Figures Test. Journal  of Abnormal and Social Psychology, 1956, 53, 254-255. Jacobson, F. R. Effect of brief sensory deprivation on f i e l d dependency. Journal of Abnormal and Social Psychology, 1966, 71, 115-118. 76 Jacobson, G. R. Reduction of field-dependence in chronic alcoholic patients. Journal of Abnormal Psychology, 1968, 73, 547-549. Jannucci, G. Size constancy i n schizophrenia: A study of subgroup differences. Unpublished doctoral dissertation, Rutgers State University, 1964. Kagan, J., Rosman, B. L., Day, D., Albert, J., P h i l l i p , W. Information processing in the child: The significance of analytic and reflective attitudes. Psychological Monographs, 1964, 78 (1 Whole No. 578). Kagan, J., Moss, H. A., Sigel, I. E. Psychological significance of styles i n conceptualising. Monograph of the Society for Research ln Child Development. Karp, S., Poster, D., & Goodman, A. Differentiation in alcoholic women. Journal of Personality, 1963, 31, 386-394. Kristofferson, M. Effects of alcohol on perceptual f i e l d dependence. Journal of Abnormal and Social Psychology, 1968, 73, 387-391. Lester, G. The Rod and Frame Test: Some comments on methodology. Perceptual and Motor S k i l l s , 1968, 26, 1307-1314. Linton, H. B. Relations between mode of perception and tendency to conform. Unpublished doctor's dissertation, Yale University. In Witkin et a l . , Psychological Differentiation, John Wiley and Sons, 1962. Linton, H. B. Dependence on external influence: Correlates in perception, attitudes, and judgement. Journal of Abnormal and Social  Psychology, 1955, 51, 502-507. Pemberton, C. L. The closure factors related to temperament. Journal  of Personality, 1952, 21, 159-175. 77 Pettigrew, T. F. The measurement and correlates of category width as a cognitive variable. Journal of Personality, 1958, 26, 532-544. Piaget, J. The Psychology of Intelligence. New York, Harcourt, Brace, 1950. Powell, B. J. A study of the perceptual f i e l d approach of normal subjects and schizophrenic patients under conditions of an oversize stimulus. Unpublished doctor's dissertation, Washington University, 1964. Pribram, K. H. A review of theory in physiological psychology. Annual Review of Psychology, 1960, 11, 1-40. Schimek, J. G. Cognitive style and defenses: A longitudinal study of intellectualization and f i e l d independence. Journal of Abnormal Psychology, 1968, 73, 575-580. Silverman, J. The problem of attention in research and theory in schizophrenia. Psychological Review, 1964, 71, 352-397. Silverman, J. Scanning control mechanisms and cognitive f i l t e r i n g in paranoid and non-paranoid schizophrenia. Journal of Consulting Psychology, 1964a, 28, 385-393. Silverman, J. Variations in cognitive control and psychophysiological defence in schizophrenia. Psychosomatic Medicine, 1967, 19, 225-251. Silverman, J. A paradigm for the study of altered states of conscious-ness. British Journal of Psychiatry, 1968, 114, 1201-1218. Schooler, C , & Silverman, J. Perceptual styles and their correlates among schizophrenic patients. Journal of Abnormal Psychology, 1969, 74, 459-470. Taylor, J. M. A comparison of delusional and hallucinatory individuals 78 using f i e l d dependency as a measure. Unpublished doctoral dissertation, Purdue University, 1956. Werner, H. The concept of development from a comparative and organismic view. In D. B. Harris (Ed.), The Concept of Development: An issue i n the study of human behaviour. Minneapolis: University of Minnesota Press, 1957, 125-148. Witkin, H. A . Individual differences in ease of perception of embedded figures. Journal of Personality, 1950, 1-15. Witkin, H. A . , Karp, S. A., & Goodenough, D. R. Dependence in alcoholics. Quarterly Journal of Studies in Alcohol, 1959, 20, 493-504. Witkin, H. A . , Lewis, H. B., Hertzman, M., Machover, K., Meissner, P., & Wapner, S . Personality through perception. New York, Harper, 1954. Witkin, H. A . , Dyk, R. B., Faterson, H. F., Goodenough, D. R., & Karp, S. A . Psychological Differentiation. John Wiley & Sons, Inc. New York, 1962. Witkin, H. A . Origins of cognitive style. In Scheerer (Ed.) Cognition: theory, research, promise. Harper & Row, 1964. Witkin, H. A . Psychological differentiation and forms of pathology. Journal of Abnormal Psychology, 1965, 70, 317-336. Young, H. H. A test of Witkin's f i e l d dependence hypothesis. Journal of Abnormal and Social Psychology, 1959, 59, 188-192. APPENDIX *A' / y Usmsi HIDDEN FIGURES TEST — Cf-I. This i s a t e s t of your a b i l i t y to t e l l which one of f i v e simple figures can be found i n a more complex pattern. At the top of each page i n t h i s t e s t are ftvo simple figures lettered A, B, C, D, and E. Beneath each row of figures ta a page of patterns. Each pattern has a row of let t e r s beneath i t . Indicate your answer by putting an X through the l e t t e r of the figure which you f i n d i n the pattern. , NOTE; There Is only one of these figures In each pattern, and t h i s figure w i l l always be righ t side up and exactly the same size as one of the f i v e lettered figures* Now t r y these 2 examples. < A B O D E A B O D E The figures below show how the figures are included in the problems. Figure A is in the first problem and figure D in the second. X B C D E A B C X E. Your score on this test will be the number marked correctly minus a fraction of the number marked incorrectly. Therefore,, i t will not be to your advantage to gueee unless you are able to eliminate one or more of the answer choices as wrong. You will have 10 minutes for each of the two parts of this test. Each part has 2 pages. V?hen you have finished Part 1, STOP. Please do not go on to Part 2 until you are asked to do so. DO NOT TURN THIS PAGE UNTIL ASKED TO DO SO. Copyright (c) I962 by Educational Testing- Service. A l l rights reserved. Developed under NIMH Contract M-i»l66 8 0 r a t - ::: Part 1 ( 1 0 minutes) I. A B O D E A B O D E A B O D E A B O D E A B O D E GO ON TO THE NEXT PAGE Page 3 81 Part 1 (cont/liaie-) / 7 < r V 7 A B C A B C D E I I . x i v > : "^i A B C D E 12. r — 7 \ • A B C D E 13. 14. \ AA / A B O D E A B O D E 15. 16. A B C D E A B C D E DO NOT TURN THIS PAGE UNTIL AGKED TO DO SO. Page k Part P. (10 minutes) 82 / V 7 V 8 \ . / Z _ _ J c D 17. A B O D E A B O D E 19. / / \ A B O D E SO. A B O D E 21. A B O D E 24. 22. / y \ \ A B C D E 2S. A B O D E A B C D E A B O D E GO ON TO THE NEXT PAGE Page 5 Part 2 (continued) 83 0 28, X/ 1/ \ 2?. A B O D E A B O D E 28. A B O D E 20. \ ! / ,„i 20 . A B O D E A B C •31. 32. A B O D E A B O D E DO NOT GO BACK TO PART 1 , Al© DO NOT GO ON TO ANY OTHER TEST UliTIL ASKED TO DO SO. STOP. 8 4 Appendix 'B'  The R.F.T.: Procedure Ss were seated approximately eight feet from the apparatus in a totally dark room. The ultra violet lamp which illuminated the RFT was switched on and they were told that they would be presented with various placements of the rod and the frame and that each time their task would be to set the rod to v e r t i c a l . It was stressed that the rod should be perpendicular to the ground. They were handed a small e l e c t r i c a l switch and told that they could adjust the rod by pressing the control nob either to l e f t or to the right. A very brief practice period was permitted. There were eight settings of the RFT. These were: Frame - Right t i l t of 3 5 ° , rod right t i l t 4 5 ° it II II 2 5 ° n II II 2 0 ° " " " 3 0 ° II II n 1 0 ° J Left t i l t of 3 5 ° , rod l e f t t i l t 4 5 ° II II II 2 5 ° „ .. i. 2 o ° , „ „ „ 3 Q o .1 i. .1 1 Q o The subject's score was the average number of degrees that his setting deviated from v e r t i c a l . Appendix 'C' Sample Record Sheet: Low Attribute (A_) Condition The following abbreviations have been used: One border BN(1) Two borders BN(2) Three borders BN(3) Yellow Y e l l Square Q Cross -4 -Triangle Circle O Slide No Item Item Dimen Prop Characteristics Discrim Response Latency 1 1 1 3/3 1 Large • y e l l BN(1) [ <. i i 2 1 II 1 Large ^\ y e l l y e l l i 1 3 1 II 1 Green 1 line Green I 4 1 6/3 1 Grey 2 wave BN(2) j 5 1 it 1 Large Q red wave Red j i | 6 . 1 ii 1 | Small Q brown Small j | 7 3/3 1 | Brown 2 wave BN(2) | j i 8 1 | " 1 ! Blue small line > Small < \ 1 | " 1 | • 1 line j BN(1) ] j 110 1 | 6/3 1 j Med blue wave j Wave i 1 1 1 1 1 II 1 j Y e l l 1 wave Y e l l | 12 1 II 1 Green Q large j Square i ! 13 1 3/3 1 1 Large green red J Green 14 1 II 1 Large c i r c l e blue Large 15 1 II 1 Green 1 line Line 86 Slide No Item No y Item Dimen Prop Characteristics Discrim Response Latency 16 1 6/3 1 Red 1 dash Dash 17 1 II 1 Brown A blue Blue 18 1 II 1 Small O y e l l Small 19 1 3/3 1 Large O r e c* Red 20 1 II 1 Large blue Cross 21 1 II 1 Green A 1 BN(D 22 1 6/3 1 Green O * BN(1) 23 1 II 1 Green D large Green i 24 1 II 1 Red 1 Cross i 25 1 3/3 1 Large wave y e l l Large 26 1 | | 1 Large brown Square i 1 27 i | - | i Red 3 line Red i j 28 1 j 6 / 3 ! 1 Large ^ red Triangle i i 29 i ! - 1 > Green "f* brown Green i ! 30 i j 1 Green 1 line Green j l s i ! 31 1 3/3 | 1 f Grey O Red 1 line Grey 32 i " i 1 t Large 0 line Square ! i 1 \ 33 n 1 Blue 1 line Line ! \ 34 1 j 6/3 1 Small 1 wave Small ! 35 1 II Brown 2 line Line 36 1 II 1 Small O 2 line 37 1 3/3 1 Large /\ red Trianglt 38 1 M 1 Blue O brown Blue 87 S l i d e No Item No Item Dim en Prop C h a r a c t e r i s t i c s Discrim Response Latency 39 1 3/3 1 Large *f* y e l l Large 40 1 6/3 1 Large O r e d Square 41 1 II 1 £ 2 l i n e Line 42 1 II 1 Blue O red Blue 43 1 3/3 1 Green 2 dash Dash 44 1 ti 1 D 1 l i n e BN(1) 45 1 II 1 Green CJ red Red 46 1 6/3 1 Y e l l 3 l i n e BN(3) 47 1 II 1 Red D y e l l Y e l l 48 1 II 1 Large A y e l l Large 88 Appendix 'D' Sample Record Sheet: High Attribute (A,) Condition The following abbreivations have been used: One border BN(1) Two borders BN(2) Three borders BN(3) Yellow Y e l l Square {3 Cross HF-Triangle A Circle O Slide No Item No Item Dimen Prob Characteristics Discrim * Response Latency DI 37 4 6/6 2 Small D y e l l red 3 line small line DI 38 4 II 2 Large & green brown 1 dash Dash BN(1) DI 39 4 ti 2 Small *f» green y e l l 2 wave small Al 1 2 3/6 1 Small A y e l l blue 2 wave blue Al 2 1 II 1 Large "f" red blue 2 dash dash Al 3 2 it 1 Large ^ y e l l blue 2 dash BN(2) CI 25 1 6/6 1 Large -f* y e l l blue 2 dot blue CI 26 1 II 1 Med Q y e i l green 1 dash green CI 27 1 11 1 Med O brown red 3 wave BN(3) Bl 13 1 3/6 2 Small +• red green 2 wave 4-red 89 Slide No Item No Item Dimen Prob Characteristics Discrim Response Latency Bl 14 1 II 2 Large O red blue 1 line red o Bl 15 1 ti 2 Large ^ green red 1 line l i n e D2 40 4 6/6 2 Med © brown grey 3 dash BN(3) brown D2 41 1 ii 2 Large + y e l l red 3 line + BN(3) D2 42 1 II 2 Small O red y e l l 2 line small y e l l C3 33 1 6/6 1 Large O red y e l l 2 line • B3 19 1 3/6 2 Small D red y e l l 1 line small BN(1) B3 20 2 j " 2 Large blue y e l l 1 line A BN(1) B3 21 1 it 2 Small A blue brown 2 wave BN(2) brown D4 46 1 6/6 2 Large D red brown 2 dash brown o D4 47 4 II 2 Small $ green y e l l 2 dash 2 small D4 48 4 II 2 Med Q grey red 2 line grey red A4 10 1 3/6 1 Large O blue grey 1 dot grey A4 11 2 II 1 Large Q red grey 1 line o A4 12 1 II 1 Large O blue Brown 1 dot BN(1) C4 34 4 6/6 1 Med O green blue 1 dash BN(1) 90 Slide No Item No Item Dimen Prob Characteristics Discrim Response Latency CA 35 1 6/6 1 Large & brown y e l l 3 line brown CA 36 5 II 1 Large JX blue y e l l 1 line A BA 22 2 II 2 Large 0 green brown 2 line green line BA 23 1 II 2 Large & y e l l red 2 dash y e l l dash BA 2A 1 ti 2 Large .A blue green 1 wave wave A A2 A2 2 3/6 1 Large O blue red 1 wave blue A2 5 2 II 1 Large D red y e l l 2 dash dash 1 A2 6 1 ti 1 Small & brown red 2 line small ! i C2 28 A 6/6 1 Large O red grey 1 wave wave I C2 29 A ti 1 Small D grey blue 2 dash small C2 30 A 3/6 1 Med O green blue 1 line green B2 16 1 it 2 Small O blue y e l l 1 wave small y e l l B2 17 1 ti 2 Large green grey 1 wave green wave B2 18 2 it 2 Large A brown blue 2 line brown A D3 A3 A 6/6 2 Med + grey brown 2 wave brown wave D3 AA 1 11 2 Large + y e l l green 2 dash y e l l 91 Slide No Item No Item Dimen Prob Characteristics Discrim Response Latency D3 45 4 6/6 2 Small + blue grey 1 wave small , grey A3 7 2 3/6 1 Large O blue red 1 wave blue A3 8 1 II 1 Large + blue brown 1 wave H-A3 9 1 II 1 Large Q red y e l l 2 line red C3 31 1 6/6 1 Large + y e l l blue 1 dash large C3 32 1 II 1 Large A green y e l l 1 line 92 Appendix 'E'  Counterbalancing of the stimulus arrays There were four types of stimulus arrays (A^I^P-j, A6 I6 P3' A6*3 P6 and AglgPg) i n the high attribute (Ag) condition and each array contained 12 slides. These slides were randomly divided into four groups of three slides each (1, 2, 3, 4) and there were thus four equivalent subgroups of slides within each stimulus condition. The three slides which constituted each subgroup were always presented i n the same order. The order of presentation of the stimulus conditions (Agl^P^* Ag*6 P3 A ^ I o P , . A,I,P,) was counterbalanced. It was arranged that every condition o 3 o b o o should both precede and follow every other condition. This resulted i n four different orders of presentation (shown in Table 9). These orders were repeated u n t i l a l l of the subgroups of slides had been used. That i s to say, each order of presentation was repeated four times. Table 9 Orders of presentation of the experimental conditions in A, E3 h P3 P6 P3 V Order 1 1 2 3 4 Order 2 3 1 4 2 Order 3 4 3 2 1 Order 4 2 4 1 3 1, 2, 3, 4 represent groups of three slides each. 93 It w i l l be noted that in order 1, two groups of three item slides are repeatedly followed by two groups of six item slides. In order 2, groups of three and six item slides alternate. Order 3 is the reverse of order 1 with two groups of six item slides being followed by two groups of three item slides. Order A i s the reverse of order 2. Six item and three item slides alternate. This same type of ordering was used i n the low attribute (A^) condition also. Thus the stimulus presentations in A^ and A^ resembled each other as closely as possible. 94 Appendix 'F' Scores of FID, N and FD subjects on Rod and Frame Test (RFT) and the Embedded Figures Test (EFT) Group RFT EFT RFT EFT 23 16 26 12 27 16 26 20 21 16 25 14 19 16 23 20 27 12 25 13. 26 28 25 20 20 19 23 20 FID 24 18 28 1 25 15 21 22 22 19 24 23 23 18 25 . 12 27 12 26 30 24 16 24 13 28 20 21 24 26 15 24 13 26 11 26 18 20 4 12 5 1 14 23 4 10 3 24 0 18 10 21 5 FD 19 8 8 4 95 Group RFT EFT RFT EFT FD (cont'd) 15 5 14 5 14 8 21 6 19 8 18 7 11 5 11 6 13 7 18 0 7 4 15 5 18 4 22 4 24 5 16 4 17 4 14 9 1 2 20 1 15 13 21 2 Mean RFT score is 20.49 Mean EFT score is 11.61 

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}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            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:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0302460/manifest

Comment

Related Items