@prefix vivo: . @prefix edm: . @prefix ns0: . @prefix dcterms: . @prefix skos: . vivo:departmentOrSchool "Business, Sauder School of"@en ; edm:dataProvider "DSpace"@en ; ns0:degreeCampus "UBCV"@en ; dcterms:creator "Wolf, Frank Michael"@en ; dcterms:issued "2010-03-26T03:43:48Z"@en, "1980"@en ; vivo:relatedDegree "Doctor of Philosophy - PhD"@en ; ns0:degreeGrantor "University of British Columbia"@en ; dcterms:description """In this study an attempt is made to better understand the audit process and the characteristics of the decisions effected on the audit by focusing on the work of individuals integrally involved with the audit. Specifically, attention is directed to the work of audit managers in large public accounting firms and two fundamental questions are: (1) What do audit managers do? (2) Why do audit managers do what they do on the job? In order to address the first question, a specification of managerial work in large public accounting firms is provided. For this purpose Flanagan's (1954) critical incident methodology is adopted. Flanagan (1954) has indicated that the functional description of the work of any group of individuals can be derived by gathering incidents critical to the effective or ineffective completion of their work. An extensive series of interviews was conducted in five public accounting firms in a large city in Western Canada for the purpose of implementing the critical incident methodology. Partners, managers and senior field staff participated in the study and provided a total of 582 critical incidents. These incidents were summarized and abstracted to yield 88 unique critical task requirements. Various validation procedures were implemented to ensure that the description of managerial work derived in this study was complete. In order to address the second question, attention is directed to the economic incentives which exist within the audit environment and the structural characteristics of the audit firm which may influence decision making on the audit. An assessment is made as to the way in which these factors influence the decision based behaviour of managers on the job. A positive theory of managerial decision making is developed. Such a theory will help us better understand the source of pressures which drive decision making in an audit context and the way different audit problems affect different groups of individuals in the audit firm. A positive theory is in many ways a precondition for answering normative questions about the audit process and decisions effected on the audit. The basic hypothesis investigated is that relative to what their partners would want them to do, managers will give a higher priority to conformance to those rules which can be readily monitored by their superiors. Conversely, managers will give a lower priority to conformance to those rules which cannot be readily monitored, or which can only be monitored at a prohibitive cost. Behaviour of this nature is entirely rational from the viewpoint of managers within public accounting firms. In addition, it is entirely rational for their partners to allow such behaviour. The test of the basic hypothesis makes use of a Q-sort methodology and draws extensively on the descriptive foundation which was provided by specifying what audit managers do. A Q-sort methodology is the general name used to characterize a set of philosophical, statistical, and psychometric ideas oriented towards research, on individuals. The Q procedure may be viewed as essentially a kind of rating procedure for rank ordering stimuli. In the context of this study, a Q-sort was developed which characterizes the work of the audit manager in terms of behaviours critical to the completion of his task. The Q-sort is then used to investigate whether audit partners and managers have similar perceptions about the nature of managerial work in specific decision contexts. Perceptions about the nature of the audit manager's work are expressed by rank ordering the cues in the Q-sort. Four Eastern Canadian offices of large public accounting firms participated in the Q-sorting phase of this study. One hundred and twenty-seven completed questionnaires were received from the partners and managers who participated at this stage of the research. Managers indicated how they typically respond to three broad decisional contexts. Partners indicated how they would want their managers to respond in those contexts. Significant differences were observed in partners' and managers' rank assignments with respect to the decisional cues in the Q-sort. Relative to what their partners would want them to do, managers give a higher priority to conformance to organizational rules which can be readily monitored within the firm. Conversely, managers give a lower priority to conformance to those rules which cannot be readily monitored or which can only be monitored at a prohibitive cost."""@en ; edm:aggregatedCHO "https://circle.library.ubc.ca/rest/handle/2429/22604?expand=metadata"@en ; skos:note "A POSITIVE THEORY OF MANAGERS' DECISIONAL BEHAVIOUR IN PUBLIC ACCOUNTING FIRMS by FRANK MICHAEL WOLF B.A. Hons., Macquarie University, 1976 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Faculty of Commerce and Business Administration) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA March 1980 © Frank Michael Wolf, 1980 In presenting this thesis in partial fulf i lment of the requirements for an advanced degree at the University of Brit ish Columbia, I agree that the Library shall make i t freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the Head of my Department or by his representatives. I t is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Faculty -H&pKXM&KX of Commerce and Business Administration The University of Brit ish Columbia 2075 Wesbrook Place Vancouver, Canada V6T 1W5 Date March 19, 1980 i i . ABSTRACT In this study an attempt is made to better understand the audit pro-cess and the characteristics of the decisions effected on the audit by focusing on the work of individuals integrally involved with the audit. Specif ically, attention is directed to the work of audit managers in large public accounting firms and two fundamental questions are: (1) What do audit managers do? o (2) Why do audit managers do what they do on the job? In order to address the f i r s t question, a specification of managerial work in large public accounting firms is provided. For this purpose Flanagan's 0954) c r i t i ca l incident methodology is adopted. Flanagan (1954) has indicated that the functional description of the work of any group of individuals can be derived by gathering incidents c r i t i ca l to the effective or ineffective completion of their work. An extensive series of interviews was conducted in f ive public accounting firms in a large c i ty in Western Canada for the purpose of implementing the c r i t i ca l incident methodology. Partners, managers and senior f ie ld staff participated in the study and provided a total of 582 c r i t i ca l incidents. These incidents were summarized and abstracted to yield 88 unique c r i t i ca l task requirements. Various validation procedures were implemented to ensure that the description of managerial work derived in this study was complete. In order to address the second question, attention is directed to the economic incentives which exist within the audit environment and the structural characteristics of the audit firm which may influence decision making on the audit. An assessment is made as to the way in which these i i i . factors influence the decision based behaviour of managers on the job. A positive theory of managerial decision making is developed. Such a theory wi l l help us better understand the source of pressures which drive decision making in an audit context and the way different audit problems affect dif ferent groups of individuals in the audit f i rm. A positive theory is in many ways a precondition for answering normative questions about the audit process and decisions effected on the audit. The basic hypothesis investigated is that relative to what their partners would want them to do, managers wi l l give a higher pr ior i ty to conformance to those rules which can be readily monitored by their superiors. Conversely, managers wi l l give a lower pr ior i ty to conformance to those rules which cannot be readily monitored, or which can only be monitored at a prohibitive cost. Behaviour of this nature is entirely rational from the viewpoint of managers within public accounting firms. In addition, i t is entirely rational for their partners to allow such behaviour. The test of the basic hypothesis makes use of a Q-sort methodology and draws extensively on the descriptive foundation which was provided by specifying what audit managers do. A Q-sort methodology is the general name used to characterize a set of philosophical, s ta t i s t i ca l , and psychometric ideas oriented towards research, on individuals. The Q procedure may be viewed as essentially a kind of rating procedure for rank ordering st imul i . In the context of this study, a Q-sort was developed which character-izes the work of the audit manager in terms of behaviours c r i t i ca l to the completion of his task. The Q-sort is then used to investigate whether iv-. audit partners and managers have similar perceptions about the nature of managerial work in specific decision contexts. Perceptions about the nature of the audit manager's work are expressed by rank ordering the cues in the Q-sort. Four Eastern Canadian offices of large public accounting firms par t i -cipated in the Q-sorting phase of this study. One hundred and twenty-seven completed questionnaires were received from the partners and managers who participated at this stage.of the research. Managers indicated how they typical ly respond to three broad decisional contexts. Partners indicated how they would want their managers to respond in those contexts. Significant differences were observed in partners' and managers' rank assignments with respect to the decisional cues in the Q-sort. Relative to what their partners would want them to do, managers give a higher pr io r i ty to conformance to organizational rules which can be readily monitored within the f i rm. Conversely, managers give a lower pr ior i ty to conformance to those rules which cannot be readily monitored or which can only be monitored at a prohibitive cost. V. TABLE OF CONTENTS Page Abstract i i List of Appendices v i i List of Figures ix List of Exhibits x List of Tables xi Acknowledgments- x i i CHAPTER 1: INTRODUCTION 1 CHAPTER 2: THE NATURE OF MANAGERIAL WORK: AN INVESTIGATION OF THE WORK OF THE AUDIT MANAGER 6 Introduction °. 6 The Work of Audit Personnel: A Review of the Extant Literature 7 The Crit ical Incident Methodology 9 The Application of the Crit ical Incident Methodology 11 Data Collection 12 Interpreting the Crit ical Requirements of Managerial Work 16 Discussion 22 Validating the List of Cri t ical Requirements 24 Conclusions 28 Appendices 31 CHAPTER 3: ECONOMIC INCENTIVES AND DECISION MAKING IN AUDITING - STEPS TOWARDS A POSITIVE THEORY 43 Introduction 43 Empirical Studies of Judgment in Auditing 44 The Contractual Relationship in Public Accounting Firms 50 v i . Paje Research Objectives 55 Imp!ementati on 56 Research Design and Procedures 60 Implementation of the Q-Sort Methodology 63 The Substantive Hypotheses 65 The Research Material 70 Appendices 74 CHAPTER 4: THE COLLECTION AND ANALYSIS OF THE Q-SORT DATA . . . . 101 Data Collection 101 Stat ist ical Analysis 101 Discussion of the Contextual Hypotheses I l l tests of the Substantive Hypotheses 112 Discussion of the Substantive Hypotheses 119 Appendices 121 CHAPTER 5: FURTHERING THE DEVELOPMENT OF POSITIVE THEORY IN AUDITING: SOME CONCLUDING REMARKS 133 Appendix 139 Bibliography 140 vi i . List of Appendices Page Appendix I Applying the Cri t ical Incident Methodology: An I l lust rat ion of the Structured Interview Procedure 31 Appendix I I Checklist of Cri t ical Requirements of Managerial Work . . . 38 Appendix I I I The Crit ical Task Requirements in The Q-Sort . . . . 74 Appendix IV The Research Instruments 76 Appendix V Test of Hypothesis Cl Communal i t y of Partners' Rank Assignments 121 Appendix VI Test of Hypothesis C2 Communal i t y of Managers' Rank Assignments 122 Appendix VII : Test of Hypothesis C3 Rank Order Correlation.Coefficients. . 123 Appendix VI I I Test of Hypothesis C4 Communality of Managers' Rank Assignments 124 Appendix IX Test of Hypothesis C5 Communality of Rank Assignments Across Partner Groups 125 Appendix X Test of Hypothesis C6 Communality of Rank Assignments Across Manager Groups 126 Appendix XI Test of Hypothesis SI The Manager as a Resource Allocator. Results by Role and Firm 127 Appendix XII Test of Hypothesis SI Analysis of Variance of Q-Sort Data. The Manager as a Resource Allocator 128 Appendix XI I I Test of Hypothesis S2 The Manager as a Negotiator. Results by Role and Firm 129 Appendix XIV Test of Hypothesis S2 Analysis of Variance of Q-Sort Data. The Manager as a Negotiator 130 Appendix XV Test of Hypothesis S3 The Manager as. an Entrepreneur. Results by Role and Firm 131 v i i i . List of Appendices (continued) Page Appendix XVI Test of Hypothesis S3 Analysis of Variance of Q-Sort Data. The Manager as an Entrepreneur 132 Appendix XVII Correlation Between the Respondents' Years of Experience and Rank Assignment With Respect to Specific Q-Sort Items 139 ix. List of Figures Page Figure 1 The Manager's Roles 18 Figure 2 An I l lus t ra t ive Flowchart of the Work Undertaken on an Audit 57 X. List of Exhibits Page Exhibit 1 A Description of Mintzberg's Managerial Roles. . . . 20 Exhibit 2 I l lust rat ion of the Q-Sort Methodology 62 xi.. List of Tables Page Table I Participants in the Interview Program Designed for the Implementation of the Crit ical Incident Methodology 13 Table I I Frequency with Which Crit ical Task Requirements Pertained to Each Managerial Role 23 Table I I I Table of the Substantive Hypotheses to be Tested at Each Audit Stage 70 Table IV Pattern of Responses to the Q-Sort 102 Table V Partner-Manager Differences in the Rank Assignment of Resource Allocative Cues 116 Table VI Partner-Manager Differences in the Rank Assignment of the Negotiation Cue 117 Table VII Partner-Manager Differences in the Rank Assignment of the Entrepreneurial Cues 118 x i i . Acknowledgments This study develops a positive theory of managerial behaviour in public accounting firms. A great deal of assistance was provided in the development of this -research. I would l ike to formally acknowledge the assistance I have received from the audit personnel who participated at the various stages of this study. Members of the public accounting profession in Canada very wi l l ing ly co-operated with this research through i t s various phases. A great deal of assistance and guidance in the planning and execu-tion of this research was provided by my dissertation committee at The University of Bri t ish \"Columbia, Vancouver. I would l ike to thank Jerry Feltham, Danny Kahneman, Ron Taylor and Don Wehrung for a l l their help. A particular vote of thanks must go to my thesis supervisor, Mike Gibbins, who provided much needed guidance throughout the research and meticulously read and re-read the many drafts of this work. Financial support from the Clarkson Gordon Foundation and the Accounting Development Fund at The University of Bri t ish Columbia are grateful ly acknowledged. Furthermore, the assistance that I received from Colleen Colclough in typing and presenting this research is acknowledged. A particular debt of gratitude must go to my wife Karen, who con-tinuously encouraged my efforts throughout the doctoral program and provided a great deal of enthusiasm and support throughout this research. 1. Chapter 1 A. INTRODUCTION An auditor's task is to evaluate evidence and make judgments and decisions about the c l ient 's state of affairs so that a professional opinion may be issued. Ultimately, such an opinion wi l l be valued by society only to the extent that i t is supported by soundly based and reliable decisions and evaluations. The external auditor's decisions and evaluations have come under close scrutiny in recent years. There has been a growing trend for l i t i -gation against business professionals and the courts, part icularly in the United States, but also elsewhere, have continued to extend the audi-tor 's l i a b i l i t y for third party negligence as well as for his own third party l i a b i l i t y . Audit firms have always been concerned about the quality and character of their members' decisions and evaluations. Comments in the popular press and at recent audit symposia indicate that audit firms are at least as interested as their c r i t i cs in improving decision making on the audit. More than ever i t seems that the auditor is searching for new tools to deal with the challenges with which he is confronted. This study focuses upon the characteristics of the decisions effected by audit personnel in large public accounting firms and hence is of a descriptive nature. The need for descriptive research in auditing has been widely documented. Mautz (1976) suggested that we need to gather a great deal of information about what is being done in auditing and Kaplan (1977) called for the devotion of signif icant resources towards the develop-ment of what auditors are actually doing now. In addition, Gibbins (1977) 2. in a review of empirical research in the audit area, drew attention to' the paucity of descriptive research which te l l s us what an auditor does and why he does i t . In this study, an attempt is made to better understand the audit pro-cess and the characteristics of the decisions effected on the audit by focusing on the work of individuals integrally involved with the audit. In public accounting firms there are four basic staff positions - partners, managers, senior f ie ld staff and juniors or staf f accountants. This study is specif ical ly directed towards the work of audit managers and two fundamental questions are addressed: (1) What do audit managers do? (2) Why do audit managers do what they do on the job? An understanding of what audit managers do on the job is important for understanding the characteristics of the output from the audit pro-cess (Peat, Marwick & Mitchell and Co., 1976, p. 8), and for the evalua-tion of positive and normative aspects of the audit. Such an understanding wi l l provide a descriptive foundation which is necessary for the purpose of investigating why managers do what they do on the job. In order to provide a specification of managerial work in large public accounting firms, Flanagan's (J954) c r i t i ca l incident methodology is adopted. Flanagan (1954) indicated that the functional description of the work of any group of individuals can be derived by gathering incidents c r i t i ca l to the effective completion of their task. The c r i t i ca l incident tech-nique consists of a set of procedures for collecting direct observations of human behaviour. By an incident we refer to any observable human activ-3. i t y that is suf f ic ient ly complete in i t se l f to permit inferences and pre-dictions to be made about the person performing the act. A c r i t i ca l inc i -dent is one which occurs in a situation where the purpose or intent of the act seems clear to the observer and where i ts consequences are such that there is l i t t l e doubt as to i ts effects. A detailed specification of the work of audit managers\"is presented in Chapter 2. In that chapter the c r i t i ca l incident methodology is de-scribed and the application of that methodology in the audit context is elaborated. A checklist of c r i t i ca l task requirements which define what managers do on the job is presented. The third chapter examines why managers do what they do on the job and in so doing provides the rudiments of a positive theory of managerial behaviour in the public accounting environment. Attention focuses on the decision based behaviour of audit managers. Hence a positive theory of decision making is developed. The development of a positive theory of decision making in auditing wi l l enable a better understanding of the sorts of pressures which drive decision making in the audit context and the way in which different audit problems affect different,groups of individuals within the audit f i rm. Jensen (1976, p. 13) indicated that positive theory building is a precon-dit ion for addressing normative questions which may be of interest to us. Tradit ionally, research in accounting and auditing has been norma-tive and def in i t ional . L i t t l e attention has been directed towards the explanation of observable phenomenon or to the veri f icat ion of theory by observation. Normative theories are prescriptive in nature and hence accounting and auditing research has tended to focus on questions of a 4. \"what ought to be\" nature. Such theories do not set out to explain exist-ing phenomena. They are concerned with propositions of the following form: (1) How should inventories be valued? (2) Should interim statements be audited? (3) Should auditors audit for efficiency? (4) How should leases be treated on the balance sheet? L i t t l e attention has been directed towards answering positive questions about actual practice. A positive theory is descriptive in nature. I t attempts to describe phenomena which are observable and hence a positive theory might address questions of the following form: (1) Why do firms change auditors? (2) Why do different audit firms charge different rates for comparable services? (3) How has accounting research influenced actual practice? (4) Why do audit managers do what they do on the job? Recently there has been a growing interest in the development of positive theories in accounting and auditing (Jensen, 1976; Ng, 1978; Watts and Zimmerman, 1978, 1979; Zimmerman, 1979). Ng has indicated that: \" unless we understand why auditing is what i t i s , why auditors do what they do, and what effects these phenomena have on people and resource u t i l i z a -t ion, the question 'how much auditing is enough', cannot be sat isfactor i ly answered.\" (1978, p. 2) emphasis added In Chapter 3, a positive theory which explains why audit managers do what they do on the job is developed. The theory stems from an examina-tion of the incentive system in large public accounting firms and an eval-uation of the way in which that system influences the decision based behaviour of managers on the job. In that chapter the characteristics 5. of the incentive system in public accounting firms are described and the way in which that system is seen to influence the decision based behaviour of audit managers explicated. Specific hypotheses subjected to investiga-tion are detailed and the research methodology described. The results which relate to the hypothesis testing stage of this study are presented in Chapter 4. Some discussion of the results follows. In Chapter 5 the major conclusions from this study are drawn together and certain l imitations of the research are discussed. Possible direc-tions for further work in this area are alluded to. The implications of the findings from this study are examined in relation to some of the extant empirical research in auditing. 6. Chapter 2 THE NATURE OF MANAGERIAL WORK: AN INVESTIGATION OF THE WORK OF THE AUDIT MANAGER A. Introduction: The task of describing any job and discovering what i t calls for in employee behaviour is d i f f i c u l t , but i t is especially d i f f i c u l t when a description of managerial work is required. A manager's job is subject to change from one setting to another: there are time determined changes, person determined changes and situation determined changes, a l l of which influence the nature of managerial work. Lewis and Stewart stated that: \"We know more about the motives, habits, and most intimate arcana of the primitive peoples of New Guinea or elsewhere, than we do of the denizens of the execu-tive suites in Unilever House \" (Lewis and Stewart, 1958, p. 17). More recently Mintzberg indicated: \"Although an enormous amount of material has been pub-lished on the manager's job, we continue to know very l i t t l e about i t . Much of the l i terature is of l i t t l e use, being merely endless repetit ion of vague statements. \" I t must be admitted that most managers both manage and do, but at the times they are doing, they are not manag-ing,\" an academic wrote recently Descriptions l ike this - abstract generalities devoid of hard data of empirical research - persist despite repeated warnings over the years that we know almost nothing about what managers do.\" (Mintzberg, 1973, p. 7). In the public accounting environment, some attention has been directed towards describing the work of audit personnel (Baker, 1977, 1979; Brooks and Schroeder, 1979; Montagna, 1974). However, the extant research has not fu l l y specified the work of the audit manager and\"indeed has not adequately described what is being done in auditing (Mautz, 1976). In 7. this chapter attention is directed to the work of audit managers in large public accounting firms and a detailed description of the nature of managerial work in those firms is provided. The specification of mana-gerial work provided wi l l form a descriptive foundation for the remainder of the research. This chapter wi l l proceed as follows. Fi rs t , a review of the extant research which describes the nature of work in public accounting firms wi l l be provided. Certain problems with that research wi l l be alluded to and the need for alternate methodologies for the description of managerial work within the public accounting environment wi l l be advanced. The research methodology to be used in this chapter w i l l be described and the application of that methodology wi l l be discussed. The method of data collection wi l l be delineated. A detailed checklist of c r i t i ca l require-ments of managerial work wi l l be provided. That checklist w i l l be used in the development of a positive theory of decision making in auditing . which wi l l be the focus of Chapter 3. B. The Work of Audit Personnel: A Review of the Extant Literature In the auditing area there has been l i t t l e descriptive research which has attempted to describe the nature of managerial work within that environment. Montagna (1974), in a sociological examination of the accounting pro-fession, devoted one chapter of his book to the career pattern of a large-firm accountant. Casual evidence is reported about the nature of the work performed by juniors, seniors, managers and partners within large public accounting firms. However, no systematic description is developed. 8. Terkel (1975, pp. 351-355) provided a brief description of l i f e and work in public accounting firms on the basis of a single interview with a staff accountant from one such f irm. However, the description provided has l i t t l e to say about the nature of managerial work in that environment. More recently, Baker (1977, 1979) set out to provide a detailed description of managerial strategy in large public accounting firms. He indicated that previous research had l i t t l e to offer in the way of informa-tive knowledge about the area he sought to investigate (Baker, 1979, p. 225), and as a result conducted a detailed empirical investigation of the work of partners and managers using a participant observation of methodology. Baker's study was limited by a number of methodological considera-tions which seriously mitigate the value of the description actually pro-vided (Brooks and Schroeder, 1979). A participant observation methodology is a suitable research technique for the description of the work of any group of individuals i f the group of individuals subject to observation are representative of the population to which the study is to relate and i f the time period during which the.observations-are undertaken reasonably represents the sort of work undertaken by those individuals. Baker (1977) selected a sate l l i te off ice of the metropolitan and regional headquarters of a large public accounting firm as his research site and chose to observe managerial strategy over a three month period which began in mid-spring. The choice of the sate l l i te off ice as his research site was problematic i f a general description of managerial strategy was to be derived as sate l l i te offices tend to specialize in the services they provide. In addition, the selection of a three month period 9. for the implementation of the methodology selected was problematic,.as much of the work in public accounting firms is of a highly seasonal nature (Brooks and Schroeder, 1979, pp. 221-222). Consequently, the de-scription of managerial work derived is l ike ly to be contingent upon the time period during which the observations were undertaken unless a longi-tudinal or act iv i ty sampling procedure is implemented. Baker's study is also limited because his observations were conducted in the offices of the audit firm alone and hence observations of inter-actions between the audit personnel and their clients would be severely 1imi ted. C. The Cri t ical Incident Methodology The adoption of direct observational techniques for the purpose of describing managerial strategy in public accounting firms has been shown to be problematic. Hence, in this chapter a description of managerial work wi l l be presented which makes use of a research technique of a less direct nature. A number of research methodologies have been used to provide a descrip-tion of managerial work (Mintzberg, 1973, pp. 221-229). In this study, a methodology or iginal ly proposed by Flanagan (1954) which assesses task related behaviours in terms of incidents c r i t i ca l to the completion of those tasks is adopted. Campbell et al (1970) see the c r i t i ca l incident method as one of the best search techniques for sampling many jobs and for focusing on the more important aspects of managerial behaviour. Unlike more direct observational techniques, the use of c r i t i ca l incidents to describe job behaviour does not result in the specification of a set of 10. behaviours which are necessarily contingent upon the time period during which the incidents are gathered. Flanagan describes the c r i t i ca l incidents of a job as those behaviours which are crucial in making a difference between doing the job effectively and doing i t ineffect ively. Cri t ical incidents are simply reports by qualif ied observers of things people did that were especially effective or ineffective in accomplishing part of their job. Such incidents are actual behavioural accounts that are recorded as anecdotes or stories and obtained from those who are qualified to observe the job being studied. Observers may be incumbents of the job themselves or either their superiors or subordinates. The usefulness of this approach l ies in the fact that i t focuses on concrete examples, and thus allows observers of the job subject to description, or the incumbents of the job themselves, to focus on what they are most familiar with (actual events) and leaves the inter-pretation of those events to the researcher. In order to form a composite view of managerial work, a large number of c r i t i ca l events may need to be collected and abstracted. Cascio (1978) indicated that the c r i t i ca l incident technique wi l l typical ly yield descriptions of managerial behavior that are both static and dynamic in nature. By gathering many behavioral incidents i t should be possible to discover important time, person, or situation determined changes that are seen to be c r i t i ca l to the effective completion of the job being studied. Mintzberg (1973, p. 223) indicated that the main disadvantage of the c r i t i ca l incident methodology is that one can never be sure that impor-tant aspects of the job are not missing from the description. He suggests n . that there may be a tendency to ignore act iv i t ies which are routine, compl or sensitive. Tb..ensure that the l i s t of c r i t i ca l requirements derived using the methodology outlined above is comprehensive, two validation procedures are adopted. F i rs t , the l i s t of c r i t i ca l requirements identif ied is com-pared to the requirements identif ied in secondary sources made available to the researcher by participant firms. These include policy guidelines, evaluation forms and job descriptions. Second, a small group of auditors were presented with the l i s t of c r i t i ca l task requirements and asked to comment on the completeness of the l i s t . These procedures were designed to ensure that the specification of managerial work derived using the c r i t i ca l incident methodology is compl ete.', D. The Application of the Crit ical Incident Methodology The c r i t i ca l incident technique does not consist of a single set of rules governing the data collection procedure and should be though of as a f lexible set of principles which can be modified and adapted to meet the situation at hand (Flanagan, 1954, p. 355). This methodology requires the observer of a particular job or the incumbent of that job himself to make only very simple types of judgements about the act iv i ty subject to description. A prerequisite for the formulation of a functional description of an act iv i ty is a basic orientation in terms of the general aims of that act iv i ty . In i ts simplest form, the functional description of an act iv i ty requires the specification of what must be done and what must not be done i f participation in that act iv i ty is to be judged success f u l . 12. In this study, incidents c r i t i ca l to the effective or ineffective completion of the work of the audit manager were gathered. In order to define more clearly what effective or ineffective behavior in a public accounting firm might involve, audit managers and other qualif ied observers were asked to provide actual incidents of behavior in which an audit manager was involved and in which he did something that was seen to be either favorable or unfavorable to his promotion decision within the audit f irm. Favorable incidents are those which enhance the manager's l i k e l i -hood of promotion in the firm whilst unfavorable incidents are those which impair the likelihood that the manager w i l l be promoted within the f i rm. To ensure that judgements about the criticalness of the actual incidents provided were made, i t was suggested that favorable incidents may lead to promotion at a l i t t l e faster rate than is usual whilst unfavorable incidents may lead to a delay in promotion. E. Data Collection For the purpose of gathering c r i t i ca l incidents, f ive of the 'big 14' public accounting firms in a large Western Canadian c i ty were selected at random. The participant firms were asked to provide twenty subjects: f ive audit partners, ten audit managers and five senior f ie ld s taf f , so that a broad range of incidents could be collected. There was no firm that fu l l y satisfied these subject requirements and in many firms al l available staff in each of the categories specified participated in the study. Cri t ical incidents were collected using a structured interview procedure and often repeated v is i ts to a particular firm were required to ensure that al l available staff actually participated in the study. In 13. Table 1, the subject composition from each firm is set out. Each subject was encouraged to provide ten incidents involving the audit manager. Five incidents were requested which described actions taken by the audit manager which were perceived to favourably impact his promo-tion decision. Conversely, respondents were requested to provide f ive further incidents in which the actions of the manager would unfavourably impact his promotion decision. TABLE I PARTICIPANTS IN THE INTERVIEW PROGRAM DESIGNED FOR THE IMPLEMENTATION OF THE CRITICAL INCIDENT METHODOLOGY Firm A Firm B Firm C Firm D Firm E Total Partners 5 3 5 4 2 19 Managers 6 4 9 4 2 25 Senior Field Staff 5 6* 5 4 5 25 TOTAL 16 13 19 12 9 69 * In Firm B six senior staff were made available for the interview program and al l respondents were included in the study. For each incident provided, respondents were asked to describe the general circumstances leading up to the incident and to detail just what i t was that the audit manager did in that context that the respondent f e l t was signif icant in terms of the manager's promotability within the f i rm. By identifying just what the manager did that was c r i t i ca l to his promo-tion decision, i t was possible to identify the c r i t i ca l requirements of 14. his work. To aid respondents an example of the c r i t i ca l incident methodology was provided from.an area unrelated to auditing. The nature of this moti-vating example and the specific structure of the interview conducted, are presented in Appendix I to this chapter. In to ta l , 582 incidents were collected. On average, each respondent provided 8.4 incidents. Each incident was summarized and abstracted so that the c r i t i ca l requirements of managerial work could be ident i f ied. From the set of 582 incidents reported, 88 unique cr i t i ca l task require-ments were identif ied using counting procedures. To.ensure that the set of incidents provided a comprehensive view of an audit manager's work, a running count was kept of the new c r i t i ca l task requirements added with the addition of each 100 incidents. Flanagan (1954) indicated that an adequate coverage is achieved when the addition of 100 incidents to the sample adds only two or three additional c r i t i ca l requirements. The i n i t i a l collection of 582 incidents satisfied that cr i ter ion and no;further incidents were collected. While a l l subjects were asked during the interview to describe each incident in enough detail so that al l the relevant factors associated with that incident are given, the degree of detail provided by each subject varied.enormously. To some extent the degree of detail was contingent upon the recency of particular incidents mentioned, but more often than not, the degree.of detail seemed to ref lect. the crit icalness of the indi -dent being cited from the viewpoint of the subject and the extent to which that incident seemed to impact the subject. In general, interviews were conducted within a twenty minute session but whilst the interview process was i t se l f structured, subjects used their own discretion to determine 15. just how much information to impart and some sessions extended to one hour. Each interview was tape-recorded and most subjects were readily able to recall incidents of both a favourable and unfavourable nature, though a number of subjects did suggest that i t was easier to think of unfavourable incidents. Following are some examples of the kinds of inc i -dents reported: - Prior to departing for a business t r i p , a partner drew the attention of a particular manager to a series of amendments that had to be made to a f i l e . The partner.asked the manager to raise the issues with the c l ient . When the partner returned a week later nothing had been done. The cl ient wondered what was going on and the manager made no attempt to communicate to the partner that he was unable to attend to the problem. The least he should have done was communicate his inab i l i ty to meet the commitment to the partner so other arrangements could have been made. - A senior submitted an audit f i l e ' to his manager. In the opinion of the senior, the f i l e was 90% complete and only limited follow up was required. When the senior checked the statements six months later he found that they were substantially altered. The senior f e l t that feed-back should have been provided to him. - A manager did not get along with one of his clients part icularly wel l , and was nervous in his dealings with them. As a result he wanted to make the cl ient happy. For this purpose the manager made rash promises about delivery to the c l ient . However, this strategy was entirely inappropriate. When the commitments to the cl ient were not met (and they simply could not be met) the manager had an unhappy cl ient -16. and this led, in turn, to an unhappy partner. This incident impaired the promotability of the manager in the f irm. - A senior identi f ies a potential scope l imitat ion and confers with his manager. The manager draws the attention of the senior to a similar problem which thereby created a precedent for dealing with the problem at hand. - A senior on the job of a medium sized cl ient has done a poor job in com-pleting the audit f i l es so that the partner to whom he reported direct ly has to re-do much of the work. A particular manager heard about this and offered to complete the job for the partner. F. Interpreting the Crit ical Requirements of Managerial Work The collection of 88 c r i t i ca l requirements of managerial work in an audit f irm provides a functional description of the work of an audit manager in terms of specific behaviours. Flanagan (1954, p. 343) indicated that i f the sample is representative, the judges well qual i f ied, and the types of judgements well defined, the stated requirements can be expected to be comprehensive, detailed and valid in their present form. However, further analysis of the data was carried out in order to relate more direct ly the work of an audit manager to the work of managers on a more general level. For this purpose, the 88 c r i t i ca l requirements of managerial work were categorized in terms of the ten managerial roles identif ied by Mintzberg (1973). 17. Mintzberg (1973, pp. 54-99) provided a theoretical description of managerial work in terms of ten fundamental roles which he argued are common to the work of a l l managers. The statement of managerial roles provided by Mintzberg was based on his own doctoral dissertation in which he observed the work of f ive chief executives, each for a period of one week. However, Mintzberg indicated that there exists considerable empirical evi-dence to support the generality of his role prescriptions. The ten roles are seen to form a gestalt - an integrated whole. The delineation of the roles is simply a categorizing process - a somewhat arbitrary part i t ion of the characteristics of managerial work into a f f in i t y groups. The development of the ten managerial roles stems from the perception of the manager as an input-output system (Mintzberg, 1973). Formal authority gives rise to'interpersonal relationships that lead to inputs, often of an informational nature. Such inputs lead to outputs in the form of information and decisions. In Figure 1, the ten managerial roles are presented. Managerial act iv i t ies are divided into three groups - those concerned with inter-personal relationships, those concerned with the transfer of information and those concerned with the decision making act iv i t ies of the manager. The ten managerial roles are divided into three groups - three interper-sonal roles, three informational roles and four decisional roles. Each of the 88 c r i t i ca l requirements of managerial work was assigned to the managerial role which best encapsulated i t . The assignment of c r i t i ca l requirements to role descriptors was based on an extensive ver i -f icat ion procedure which involved four audit personnel and three academics. 18. FIGURE 1 THE MANAGER'S ROLES (Mintzberg, 1973, p. 59) Formal Status and Authority I Interpersonal Roles: Figurehead Leader Liaison T Information Roles: Monitor Disseminator Spokesman Decisional Roles: Entrepreneur Disturbance Handler Resource Allocator Negotiator Each judge assigned each c r i t i ca l requirement to that role description which he f e l t was most appropriate. The c r i t i ca l requirements were pre-sented to judges on a randomly arranged l i s t and in order to fac i l ia te the assignment process, each judge was provided with a summary l i s t describ-ing the intent of each of the ten managerial roles. The l i s t of role descriptions is presented in Exhibit 1. Cri t ical requirements were assigned to managerial roles on the basis of a simple majority voting rule. Where no majority existed, the problem of assigning the c r i t i ca l requirement to a. particular role was resolved by 19. discussing the assignments effected with some of the judges. The judges drew attention to the d i f f i cu l t ies associated with their task due to the interrelationships which exist in the audit process and the possibi l i ty that the c r i t i ca l requirements may be best classified under one role description: in one situation but under another role description under more constrained circumstances. In view of the obvious d i f f i cu l t ies associated with any taxonomizing procedure, the degree of category re l i ab i l i t y obtained in this study was encouraging. Category r e l i a b i l i t y was assessed using a procedure developed by Spiegelman, Terwill iger and Fearing (1953). The procedure involves ranking the patterns of agreement among judges on each item. With seven judges, 15 patterns of agreement are possible.* The r e l i a b i l i t y of the category set is reported as the mean rank order of al l items. In this analysis, scores could range from 1.0 (complete agreement on a l l items) to 15.0 (no agreement on any items). The mean rank order obtained was 5.95. The median score was 6.0 and the modal pattern of agreement occurred when six out of the seven judges were in complete agreement as to the managerial role to which a particular task, requirement should be assigned. The f i f teen patterns of agreement can be i l lustrated in the Table below: PATTERN OF AGREEMENT RANK ORDER 7 (complete agreement) 1 6 .1 2 5 2 3 5 1 1 4 1 1 1 1 1 1 1 (no agreement) 15 20. Exhibit 1: A Description of Mintzberg's Managerial Roles INTERPERSONAL ROLES The FIGUREHEAD is the most basic and most simple of al l managerial roles. Because of his formal authority the manager\"is required to perform a number of routine duties of a legal or social nature. All involve inter-personal act iv i t ies but none involve signif icant information processing or decision making. As LEADER the manager is responsible for the motivation and activation of subordinates; their s taf f ing, training and associated ac t iv i t ies . The LIAISON role deals with the signif icant web of relationships that the manager maintains with individuals and groups who provide information and favors. (This defini t ion is somewhat broader than that of Mintzberg, 1973.) INFORMATIONAL ROLES The manager as MONITOR continually seeks and receives information that enables him to understand what is taking place in his organization and i ts environment. He seeks information in order to detect changes, iden-t i f y problems and opportunities and in order to be informed when infor-mation must be disseminated and decisions;made. The manager as DISSEMINATOR sends information received from outsiders or from others in the organization to those within the organization to whom the information is most pertinent. As a SPOKESMAN, the manager transmits information outside the organiza-tion and upward within the organization. He may be a spokesman for the organization or for members within the organization. He transmits in-formation about the organization's actions, plans, policies or products. DECISIONAL.ROLES As an ENTREPRENEUR, the manager searches the organization and its.environ-ment for opportunities and problems so that he can act as an in i t ia to r and designer of changes that wi l l further the objectives of the organiza-t ion. As a DISTURBANCE HANDLER, the manager is responsible for in i t ia t ing cor-rective action when the organization faces unexpected disturbances. The manager is a RESOURCE ALLOCATOR. This decisional role is at the heart of the organization's decision making system. Resource allocation is considered in i ts broadest context. Among organizational resources are money, time, material, equipment, manpower and reputation. As NEGOTIATOR, the manager takes charge when his organization must en-gage in important negotiation act iv i t ies . He represents the interests of the organization in important conf l ict situations. 21. More than 40% of the c r i t i ca l task requirements subject to c lass i f i -cation had at least f ive of the judges in complete agreement as to which managerial role that requirement should be assigned, and over 70% of the c r i t i ca l task requirements were assigned to a single managerial role by at least four of the seven judges. There was thus substantial agreement among the judges as to which managerial role best encapsulated a given task requirement and thus a most satisfactory degree of category re l ia -b i l i t y for a study of this nature. Associated with each of the 582 c r i t i ca l incidents is a c r i t i ca l task requirement. The distr ibution of c r i t i ca l task requirements among mana-gerial roles was sensitive to the characteristics of the voting rule applied. When al l task requirements were assigned to a given managerial role on the basis of a majority voting rule, regardless of the nature of that majority, the distr ibut ion was signif icantly different from that which results when a majority voting rule which requires 4, 5, 6 or 7 out of seven judges is imposed (X = 19.76, d.f. = 9, p. < 0.02; Ferguson, 1959, pp. 161-165, p. 308). This result is somewhat disturb-ing, as i t reduces the interpretabi l i ty of the managerial roles in the audit setting under investigation. However, some solace can be taken in the fact that the significant chi-square result occurs primarily because the twelve c r i t i ca l task requirements assigned to the disturbance handler role using a majority voting rule were not assigned to that role when a more stringent voting rule was applied. I f these task requirements are not included in the study, then the assignment of task requirements to managerial roles is not sensitive to the application of either of the vot-ing procedures discussed above (X = 10.14, d.f. = 8, p < 0.30; Ferguson, 22. 1959, pp. 161-165, p. 308). These results suggest that some caution should be used in interpreting the disturbance handler act iv i t ies of the audit manager. I t may well be that such act iv i t ies can be better described in terms of other managerial roles. In Appendix I I to this chapter, a detailed l i s t ing of the c r i t i ca l task requirements of managerial work is presented. These requirements are categorized in terms of Mintzberg's ten managerial roles using a majority voting procedure. For each group of subjects who participated in the interview program, the relative frequency with which each c r i t i ca l task requirement was reported is presented. A summary of the frequency data presented in Appendix I I is repro-duced in Table I I . G. Discussion Table I I summarizes the frequency with which c r i t i ca l task require-ments reported by partners, managers and senior f ie ld staff were assigned to each of the ten managerial roles or iginal ly identif ied by Mintzberg (1973). Most frequently c r i t i ca l task requirements were reported which per-tained to the decisional roles of the audit manager. In t o ta l , 55% of a l l the requirements reported described some aspect of the manager's decisional behaviour. Cri t ical requirements pertaining to the manager's interpersonal and informational responsibil it ies together accounted for 45% of a l l requirements reported. When al l the c r i t i ca l incidents and the task requirements associated with them are considered, signif icant differences were observed in the 23. TABLE I I FREQUENCY WITH WHICH CRITICAL TASK REQUIREMENTS PERTAINED TO EACH MANAGERIAL ROLE FREQUENCY TASK REQUIREMENTS PERTAINED TO EACH MANAGERIAL ROLE MANAGERIAL Over Task Requirements Reported By ROLE Over All ! Requirements n=582 Partners n=l 78 Managers n=216 Senior Field Staff n=188 Figurehead Leader** Liaison* 7.7% 11.7% -10.5% •6.3% 9.6% 7.4% . 9.2% 7.9% 8.9%. 7.3% 18.1% 15.4% All Interpersonal Roles 29.9% 23.3% 26.0% 40.8% Monitor Disseminator** Spokesman*** 4.4% 2.1% 8.6% 5.1% 0.6% 14.5% 3.3% 0.9% 7.8% 5.3% 4.8% 3.8% All Informational Roles 15.1% 20.2% 12.0% 13.9% Entrepreneur*** Disturbance Handler Resource Allocator Negotiator 28.5% 2.1% 21.3% 3.1% 28.0% 2.8% 24.0% 1.7% 36.6% 1 .4% 20.8% 3.2% 19.7% 2.1% 19.2% 4.3% All Decisional Roles 55.0% 56.5% 62.0% 45.3% All Roles 100.0% 100.0% 100.0% 100.0% * Differences in the frequency with which partners, managers\\ahd f ie ld staff reported task requirements pertaining to a given managerial role signif icant at p = 0.05. '* Differences in the frequency with which partners, managers and f ie ld staff reported task requirements pertaining to a given managerial role signif icant at p = 0.01. Differences in the frequency with which partners, managers and f ie ld staff reported task requirements pertaining to a given managerial role signif icant at p = 0.001. 24. pattern with which requirements assigned to specific managerial roles were reported by partners, managers and senior f ie ld staff (X = 56.06, d.f. = 18, p < 0.001; Ferguson, 1959, pp. 165-169, p. 309). The position of an individual reporting a c r i t i ca l requirement was therefore not independent of the characteristics of the task requirement that was reported. Table I I draws attention to specific managerial roles with respect to which signif icant differences were observed in the frequency of the incidents reported by partners, managers and senior f ie ld staf f . In par t i -cular, signif icant differences in the frequency with which dif ferent groups of auditors reported task requirements of a given characteristic were found to exist with respect to the leadership, l ia isonal , disseminator, spokesman and entrepreneurial role of the audit manager. The analysis undertaken in this chapter does not allow us to assert that managers, partners and senior f ie ld staff perceived incidents c r i t i ca l to the effec-t ive or ineffective completion of managerial work d i f ferent ia l ly . At best, we are able to assert that there were signif icant differences in the pattern with which those incidents, and the task requirements associated with them, were reported by dif ferent groups of individuals within the f i rm. Such differences may be suggestive of perceptual differences. In subsequent chapters of this study attention wi l l specif ically be directed to the way in which partners and managers within public account-ing firms perceive the importance of a number of the c r i t i ca l requirements of managerial work in alternate decision contexts. H. Validating the List of Cri t ical Requirements In the preceding discussion, two procedures were outlined for the validation of the l i s t of c r i t i ca l requirements identif ied from the 582 25. c r i t i ca l incidents reported during the interview program. First , a review of the secondary sources made available to the researcher by the par t i c i -pant firms was undertaken to identify whether the methodology employed in this chapter led to a specification of managerial work which was con-sistent with firm job descriptions, appraisal forms, and so for th. Second, a small group of auditors were asked to verify the completeness of the l i s t of c r i t i ca l task requirements. In general, the l i s t of c r i t i ca l requirements derived using Flanagan's (1954) c r i t i ca l incident methodology was more comprehensive than the secondary material examined. During the interviews, subjects were asked to indicate just what the manager did that they f e l t was part icularly c r i t i ca l to his promotion decision within the f i rm. The degree of speci-f i c i t y that resulted from the interviews tended to be greater than that which appeared on most of the secondary documents examined. However, some subjects summarized the c r i t i ca l incident which they reported in rather broad terms and at times there was a deviance in the way in which particular behavioural requirements were summarized on the checklist and on the secondary sources made available to the researcher. The 88 c r i t i ca l requirements were f e l t by the researcher to adequately portray the work of..the manager. From the secondary material available i t seemed that the checklist derived was more comprehensive than the specifications provided by a single f irm. However, i t was le f t to a small group of auditors to make the f inal assessment. Four senior audit personnel from two public accounting firms were called upon to validate the l i s t of c r i t i ca l requirements. Two auditors were drawn from one of the f ive firms that participated in the interview 26. program and two were drawn from another firm not previously involved in the research. Each auditor was presented with the l i s t of c r i t i ca l requirements that is reproduced in Appendix I I . In addition, a specification of each of Mintzberg's (1973) role descriptions was provided. The validating personnel were asked to check the l i s t for completeness, to add any c r i t i ca l requirements not already included, and to delete any items that were inappropriate. This validation procedure was designed to identify any biases in the description derived as a result of the methodology employed. The validating auditors reported that the l i s t seemed complete. One commented that he couldn't think of anything he'd l ike to do to i t and another indicated that i t looked extremely comprehensive and well thought through. Neither of these auditors altered the l i s t presented to them. Nonetheless, some of the comments provided about specific requirements by the other auditors do lend some insight into the nature of managerial work in the audit context. One validating auditor suggested that the manager, as a disseminator, keeps clients informed about changes in the technical and professional matters affecting the c l ient 's affairs and keeps superiors informed about the work of subordinates. This lat ter task requirement draws attention to one broad area of managerial work that another auditor suggested was absent from the l i s t presented: that area relates to the work of the manager as a personnel manager. In this role, the manager may be involved in the hiring and f i r ing of subordinates, the discussion of their remuner-ation, and their overall management. Act iv i t ies of this sort may be broadly 27. described as part of the managers' leadership act iv i t ies and whilst such act iv i t ies were referred to during the interview sessions (see Appendix I I ) , one of the validating auditors f e l t that a more specific orientation was required in terms of the personnel management function. In his 1iaisonal role i t was suggested that the manager may be required to act as a go-between between the partner and the c l ient . As such, the manager often communicates problems to the c l ient , expresses concerns to them, and may even be responsible for dismissing a cl ient on behalf of the f i rm. Two further points of c lar i f icat ion were called for by the validating auditors. Two of:the auditors questioned thejiaisonal act iv i ty 'subordi-nating personal'interests' and were not convinced that i t constituted a characteristic of managerial work. Further, one of the validating auditors emphasized that the responsibil ity of 'ensuring the financial statements are error free' implies only freedom from spelling errors, addition errors, transfers between statements and agreements with accounting records. Freedom from error is not an assurance of absolute correctness. The validation of the c r i t i ca l requirements of managerial work derived using the c r i t i ca l incident methodology was both insightful and reassur-ing. Basically, the l i s t of 88 c r i t i ca l task requirements developed using an extensive interview program was comprehensive - a detailed specification of managerial work was provided. The assistance of the validating auditors confirmed i n i t i a l observations about the completeness of the l i s t . Two aspects of managerial work, however, did not appear to be included: one relates to the personnel management functions of the audit manager and the other relates to the function of the manager as a 28. go-between between the partner and the c l ient , and his responsibil ity to deal effectively with the cl ient in this capacity. These facets of a manager's job have not been included in the checklist provided in the Appendix, as the methodology used for the identi f icat ion of these factors was not consistent with that used for the identi f icat ion of other dimen-sions on the l i s t . The exclusion of these factors does introduce a bias in the description provided, but the nature of the bias is known. Con-versely, the inclusion of items on the checklist which are not consistently derived, introduces a bias of a less unambiguous nature. I. Conclusions In this chapter the work of audit managers in large public accounting firms was analysed using the c r i t i ca l incident methodology. Eighty-eight c r i t i ca l requirements of managerial work were identif ied from the 582 c r i t i ca l incidents that were reported by the 19 audit partners, 25 audit managers and 25 senior f ie ld staff who participated in this study. Cri t ical requirements were assigned to one of ten managerial roles or iginal ly identif ied by Mintzberg (1973) by a panel of seven judges and a majority voting rule was applied. Some analysis of the pattern with which the c r i t i ca l requirements were reported by dif ferent groups of auditors was undertaken. Significant differences were found to exist in the pattern with which partners, managers, and senior f ie ld staff reported c r i t i ca l requirements and the role of the individual within the firm thus seemed to be related to the character of requirements actually reported. Chapter 2 has provided a descriptive foundation for the remainder of this study. The analysis undertaken in the subsequent chapters could not 29. have been executed without fu l l y specifying what audit managers do in the f i r s t instance. The application of the c r i t i ca l incident methodology in this study has led to a detailed description of the work of audit managers. Various validation procedures were adopted to ensure the com-prehensiveness of the description actually derived and these procedures provided some assurance that no major functional responsibil i t ies of the audit manager were absent.from the description presented in Appendix I I . The c r i t i ca l incident methodology provided a useful research tool for the purpose of identifying what audit managers do, and could be used to further describe the work of other groups of individuals within audit firms. In the analysis undertaken in this chapter, some attention was directed to the wide discrepancies which existed in the frequency with which partners, managers and senior f ie ld staff reported specific require-ments of managerial work. I t was emphasized that the analysis undertaken in this chapter does not enable us to assert that such differences are indicative of perceptual differences on the part of each of these three groups of individuals. Indeed, partners, managers and senior f ie ld staff may share identical perceptions about what constitutes effective behaviour within the audit f irm but may have simply provided i l lustrat ions of that behaviour in signif icantly different ways. In Chapter 3, partners' and managers' perceptions about the impor-tance of a number of c r i t i ca l requirements of managerial work are investi-gated in alternate problem contexts. Systematic and predictable discrep-ancies between the way in which partners and managers perceive various decision related requirements wi l l be hypothesized. A positive theory wi l l be developed which wi l l explain the nature of these discrepancies. The theory wi l l be tested empirically by drawing extensively upon the descriptive foundation developed in this chapter. 31. Appendix I Applying the Crit ical Incident Methodology: An •Il lustration of the Structured Interview Procedure 1. The oral script used to motivate each interview session. First , le t me thank you for agreeing to participate in this study. This study wi l l form a portion of the research I am undertaking for my Ph.D. dissertation at the University of Bri t ish Columbia. I have divided the work into two stages: the information I w i l l be asking you to pro-vide today wi l l form the basis of the f i r s t stage of the research. For this stage of the study, I need to assess just what i t is that determines the c r i t i ca l aspects of an audit manager's work. For this purpose I have adopted a technique which has been used extensively in psychology called the c r i t i ca l incident method. The c r i t i ca l incident technique is based on the premise that in order to characterize the work of any group of individuals you need to ask those who are familiar with his work or the incumbent of the job himself just what constitutes the effective and ineffective completion of his task. An incident associated with bus driving may i l lus t ra te the point more clearly. An incident associated with good bus'driving might be that the bus driver was courteous to his passengers, or that he drove the bus careful ly, or that he knew the name of the street you were looking for. These examples represent actual incidents of behaviour or knowledge which are associated with good bus driving. On the other hand, such incidents as the bus driver trying to beat a red l igh t , or a bus driver having unusual d i f f i cu l t y replacing a dis-placed t ro l ley pole, or a bus driver not being able to provide correct Appendix I (continued) 32. street information might describe poor bus driving. Once again, the examples involve actual behavioural accounts associated with the work of a bus driver. I t is easy to think of countless other i l lustrat ions of what might constitute good and bad bus driving behaviour but in this study we shall focus on incidents which relate to the audit manager. In part icular, I want to focus on incidents in which the audit manager was involved and in which he did something that was either favourable or un-favourable to his promotion decision. Favourable incidents may lead to promotion at a l i t t l e faster rate than is usual whilst unfavourable inc i -dents may lead to a delay in promotion. Before we proceed, le t me assure you that your responses wi l l be s t r i c t l y confidential. In no way wi l l your name or the name of your firm be mentioned or associated with any of the information you supply. At the same time, I would appreciate i t i f you did not discuss any questions which I might ask with others since they might be interviewed later and any knowledge of what information is being sought might bias the results. By the way, do you have any questions you wish to ask? Would you mind i f I record the'interview so that I can refresh my memory later? Appendix I (continued) 33. 2. The Specific Interview Questionnaire* QUESTION 1: For this question I would l ike you to indicate a number of incidents you have participated in or observed in which the audit manager did some-thing which would favourably impact his promotion decision. Such incidents .may lead to promotion at a l i t t l e faster rate than is usual. For the f i r s t incident: What were the general circumstances leading up to the incident? What did the audit manager do that made the incident c r i t i ca l to his promotion decision? For the second incident: What were the general circumstances leading up to the incident? What did the audit manager do that made the incident c r i t i ca l to his promotion decision? For the third incident: What were the general circumstances leading up to the incident? * Only Questions 1, 2 and 5 were used in every interview session. Appendix I (continued) 34. What did the audit manager do that made the incident c r i t i ca l to his promotion decision? For the fourth incident: What were the general circumstances leading up to the incident? What did the audit manager do that made the incident c r i t i ca l to his pro-motion decision? For the f i f t h incident: What were the general circumstances leading up to the incident? What did the audit manager do that made the incident c r i t i ca l to his pro-motion decision? QUESTION 2: For this question I would l ike you to indicate a number of incidents you have participated in or observed in which the audit manager did something which would unfavourably\"impact his.promotion .decision. Such'incidents may lead to a delay in promotion. For the f i r s t incident: What were the general circumstances leading up to the incident? Appendix I (continued) 35. What did the manager do that made the incident c r i t i ca l to his promotion decision? For the second incident: What were the general circumstances leading up to the incident? What did the manager do that made the incident c r i t i ca l to his promotion decision? For the third incident: What were the general circumstances leading up to the incident? What did the manager do that made the incident c r i t i ca l to his promotion decision? For the fourth incident: What were the general circumstances leading up to the incident? What did the manager do that made the incident c r i t i ca l to his promotion decision? Appendix I (continued) 36. For the f i f t h incident: What were the general circumstances leading up to the incident? What did the manager do that made the incident c r i t i ca l to his promotion decision? QUESTION 3: I would l ike to focus on exceptional performance that could account for dismissal or promotion. Please think of someone in your firm who has had a spectacular career and has become a partner very rapidly: why do you think he has achieved that success? QUESTION 4: Please think of someone who has been dismissed from a managerial position recently or who you feel would not be promoted from a managerial position to partnership status: why do you think that is so? QUESTION 5: You have been associated with this f irm for some time. I f a new manager came to you and asked you what he had to do to get promoted, what advice would you give him? Appendix I (continued) 37. QUESTION 6: What other rewards/penalty influence your behaviour on the job? QUESTION 7: How is the promotion decision determined? 38. APPENDIX I I CHECKLIST OF CRITICAL REQUIREMENTS OF MANAGERIAL WORK % of All Requirements Reported By Partners n=178 Managers n=216 Field Staff n=188 INTERPERSONAL ROLES OF THE AUDIT MANAGER 1. Figurehead: Involves himself in the pro-fessional environment Develops a prof i le in community Appears to be well organized Develops a prof i le in the office Conforms to Civil Standards Abides by professional code Keeps abreast of professional developments Maintains professional appearances Maintains cl ient confidential i ty Adheres to organizational pro-cedure Meets personal commitments Displays loyalty to the firm Develops a prof i le with clients Maintains independence Leader: Accepts responsibil ity for work of subordinates Creates meaningful learning experiences for subordinates Provides supervision for sub-ordinates Recognizes capabilit ies of subordinates 1.1 0.6 0.0 0.0 0.6 0.0 0.6 0.6 0.0 1.6 0.0 0.0 0.6 0.6 6.3 0.0 3.4 1.7 1.1 0.0 0.9 0.5 0.9 0.9 0.0 0.0 0.5 0.0 1.8 0.9 1.8 0.5 0.5 9.2 0.9 1.8 0.9 1 .4 0.0 0.0 0.5 1.6 0.5 1.1 0.0 0.5 0.5 0.5 0.5 0.0 0.5 1 .1 7.3 0.0 1 .1 1.1 1.6 Appendix I I (continued) 39. Provides technical assistance to subordinates Assists subordinates with any problems that arise Communicates effectively with subordinates Imposes real is t ic requirements on subordinates Provides instruction to subordinates Adopts an outgoing position Motivates subordinates Manages subordinates Recognizes and accepts individual differences in subordinates 3. Liaison: Fosters c l ient -s ta f f relationships Develops a rapport with subordinates Interacts with subordinates in the f ie ld Maintains contact with cl ient through year Works as a member of a team Develops good relationships with cl ient Maintains existing c l ient good w i l l Develops team relationships Adapts to superiors Subordinates personal interests Relates effectively to staf f in front of cl ient Plays the 'organizational pol i t ics game' 0.0 0.0 1.6 0.0 0.5 3.2 1.7 0.9 3.7 0.0 0.5 2.7 1.1 0.5 0.5 0.0 0.0 0.5 0.6 0.0 0.0 0.0 0.5 0.5 0.0 0.0 1.6 9.6 7.9 18.1 0.6 0.5 3.7 0.6 0.5 2.7 0.0 0.0 0.5 1.1 0.5 0.5 0.6 0.5 0.0 2.7 3.2 ' 4.3 0.0 0.0 1.1 0.6 0.9 0.5 0.6 0.9 0.0 0.6 1.4 0.0 0.0 0.0 1.6 0.0 0.5 0.5 7.4 8.9 15.4 Appendix I I (continued) 40. INFORMATIONAL ROLES OF THE AUDIT MANAGER 1. Monitor: Ensures the completeness of '• _ ~ n A n n audit f i l e Reviews audit f i les promptly 0.0 0.0 3.2 0.6 1.4 0.5 Ensures that the financial statements are error-free Involves the partners only in n fi n n n cr i t i ca l problems Develops a fami l iar i ty with n -n n n , fi audit f i l e u ' u u , u 1 l b Ensures that the financial state-ments meet professional require- 0.6 0.0 0.0 ments 5.1 3.3 5.3 Disseminator: Provides feedback .to subordinates 0.0 0.9 3.7 Keeps subordinates informed about developments of clients 0.6 0.0 1.1 during the year 0.6 0.9 4.8 3. Spokesman: Advises superiors about poten- 6 y 2 2 0 0 t i a l problems Completes the management le t ter 0.6 0.5 0.0 1.1 0.0 1.1 Informs the cl ient about the execution of the audit Provides just i f icat ions for Q Q Q g Q Q actions taken Communicates effectively with 2 2 1 8 0 5 superiors Communicates effectively with 0 0 , „ •, , c l ient 2 - 8 K 4 1 J Makes real is t ic commitments to clients 1.1 0.5 0.0 Stands up for his own beliefs 0 0 0 5 n n and opinions Maintains the reputation of n n n n , , the firm u > u u - u ' • ' 14.5. 7.8 3.8 Appendix I I (continued) DECISIONAL ROLES OF THE AUDIT MANAGER 1. Entrepreneur: Displays imagination on the audit 0 .6 0 .0 0 .5 Sells additional .services to V. existing clients 2 .2 1 .4 2 .7 Is geographically mobile 0 .0 2 .3 0 .0 Displays technical competence 3 .3 2 .8 2 .7 Acts independently of his partner 1 .1 1 .4 0 .5 Assumes the responsibil ity of partner 2 .8 3 .2 1 .6 Assumes additional responsi-b i l i t i e s 2 .8 1 .4 2 .1 Pursues challenging assignments 0 .6 0 .5 0.0 Provides management type advice and service to the cl ient 5 .6 4 .6 2 .1 Displays an ab i l i t y to identify problems 6 .7 6 .5 1 .6 Brings in new clients 0 .6 5 .1 3 .7 Develops technical sk i l l s in a specialist area 1 .1 5 .1 1 .1 Adapts to the individual needs of clients 0 .6 2 .3 1 .1 28 .0 36 .6 19 .7 Disturbance Handler: Proposes solutions for problems identif ied 2 .8 1 .4 2 .1 2 .8 1 .4 2 .1 Resource Allocator: Meets c r i t i ca l deadlines 0 .6 0 .9 0 .5 Meets commitments to partners 2 .2 0 .5 0 .0 Makes real is t ic commitments to partners 1 .1 0.0 0 .0 Prepares an appropriate audit package for c l ient 0 .6 0 .0 0 .5 Attains a high personal u t i l i z a -tion 1 .1 1 .8 1 .1 Plans audit requirements 4 .4 3 .2 4 .8 Appendix I I (continued) 42. 2.8 0.5 2.7 Makes a time commitment to the firm.(puts in his time) Meets commitments made to clients 1.7 1.4 0.0 Completes the job in a timely c n ^ 7 t\\~i manner 5 , u 6 ' ' Completes the job cost n 0 , , c effectively 2 ' 2 5 ' 6 ' - 6 Accommodates the needs of others in the firm 0.0 0.5 1.6 Effectively ut i l izes available n fi n Q n ,-personnel u - b u , y u , b Delegates authority to subor- , , n n , c dinates ' • ' ° - 9 ' - 6 Prepares a coherent audit n fi package 4. Negotiator: Reconciles fee differences with cl ient 0.9 0.0 24.0 20.8 19.2 0.0 1.4 1.6 Develops an audit plan with n n n n n K cl ient u ' u u > u u ' b Reconciles disagreements about the scope and function of the 0.6 0.0 1.1 audit with cl ient Reconciles technical differences , , n 0 , with cl ient 1 J K 8 1 J 1.7 3.2 4.3 43. Chapter 3 ECONOMIC INCENTIVES AND DECISION MAKING IN AUDITING - STEPS TOWARDS A POSITIVE THEORY A. Introduction In the previous chapter a specification of managerial work in Targe public accounting firms was provided. In this chapter we draw on that specification in order to develop a positive theory of decision making in auditing. Tradit ional ly, research in accounting and auditing has been normative and definit ional and as a result there is a dearth of theory or evidence which bears on positive issues ( c f . , Jensen, 1976). The available l i t e r -ature focuses, in the main, on questions of a 'what ought to be nature' and has not tended to examine positive questions concerned with the ex-planation of observable phenomena. In this chapter a positive theory of decision making in auditing wi l l be developed by focusing on the decision based behaviour of audit managers and by specif ical ly examining why the decisions that they take on the job are what they are. Such a theory wi l l help us better under-stand the pressures that drive decision making in an audit context and the way in which different problems affect different groups of individuals within the audit f i rm. The next section of this chapter reviews the extant empirical studies of judgment in auditing and emphasizes that whilst such studies are of a descriptive nature, they have not attempted to explain why the pattern of decisions.observed were what they were. Indeed, whilst most studies have drawn attention to the substantial var iab i l i ty in the judgments of 44. different auditors, given the same information, there has been l i t t l e attempt to explain just why those discrepancies do in fact arise. I t w i l l be argued that discrepancies in the decisional responses of auditors within a given firm may be attributed to differences in the rewards which they face as a result of decisions undertaken. The characteristics of the incentive function confronting non-partners in public accounting firms wi l l be examined and the implications of that function on the nature of the decisions effected by the audit manager wi l l be considered. Specific hypotheses which wi l l be subjected to investigation wi l l be detailed and the research design wi l l be elaborated. The principal results from this section of the analysis wi l l be presented in Chapter 4. B. Empirical Studies of Judgment in Auditing There has been l i t t l e judgemental research in the audit area. How-ever, the extant research does, in general, point to a wide variation in the decisional responses of audit personnel confronted with a given decision problem. In this section, a review of relevant empirical l i t e r -ature in auditing wi l l be provided. The American Inst i tute of Accountants1'Committee on Auditing Pro-cedures (1955) provided eight auditors with a case description of an actual business and found substantial differences in the audit program formulated by respondents. Aly and Duboff (1971) investigated the rela-tionship between stat is t ica l and judgmental samples as applied to the audit of accounts receivable balances for a hypothetical reta i l store. They reported wide variations in the size of the judgment samples recommended by the 158 I l l i no i s CPA's who responded to their mail questionnaire. The 45. size of the judgment samples recommended varied from a minimum of less than 10% of the accounts receivable to a maximum which entailed a fu l l investigation. Corless (1972), in a third study in this area, examined the ab i l i t y of auditors to specify information from which prior distributions could be calculated. A questionnaire methodology was adopted for the purpose of e l i c i t i ng information pertinent to the construction of prior d i s t r i -butions. Corless (1972) observed considerable var iab i l i ty among prior distributions assessed by dif ferent auditors in each of the two audit cases which he presented to them. He suggested that because an auditor \"may typical ly consider a relat ively small error rate as being material, the variation among prior distributions appears to be great\" (Corless, 1972, p. 560). Corless also observed considerable var iab i l i ty among the sample.sizes that different auditors suggested should be taken for each audit case. Joyce (1976) reported the details of a study by Kinney and Ritts (1973) in which auditors from the 'Big 8' public accounting firms were pro-vided with a detailed case description of a plumbing supply f i rm. Respondents were asked to assess the inventory sampling plan required for the audit of this f i rm. A preliminary analysis of the data indicated that there were considerable differences in the plans proposed by respon-dents. Each of the studies referred to above draws attention to the between subject variation in the responses provided by audit personnel. Only Corless (1972) attempted to explain that variation in terms of the per-sonal background of the participants in his study. However, he found no 46. signif icant relationships between the characteristic values of the prior distributions and either the level of audit experience or the stat is t ica l background of participants in the study. A major exception to the above studies reporting substantial indi -vidual differences in auditors' judgments was the study by. Ashton (1974). Ashton asked a total o°f 63 auditors from four public accounting firms to evaluate the quality of internal control in a hypothetical company. The evaluation was based upon a six-item internal control questionnaire for payroll. Ashton systematically manipulated the pattern of responses to each of the internal control questions which he posed and asked respondents to evaluate the extent of control within the f i rm, given the pattern of responses generated. Each auditor was required to evaluate 32 different stimulus combinations. Ashton reported a fa i r l y high level of consistency in the internal control evaluations of dif ferent auditors given the same stimulus combin-ations. Ashton also indicated that, on average, agreement among auditors within firms was the same as agreement among auditors between firms. Differences in the experimental tasks required of subjects may have attributed to the discrepancy which exists between the results which Ashton derived and the results of earl ier empirical research in the area. However, these discrepancies may also be attributable to differences in the actual research design adopted. The work of Aly and Duboff (1971), Corless (1972), and Kinney and Ritts (1973) made use of mail questionnaire procedures and the results of those studies were assessed on the basis of judgments made with respect to one or two cases only. Ashton (1974), On the other hand, administered the research material personally, and obtained 47. 32 responses from each respondent. A later study by Joyce (1976)'adopted a similar methodology to that used by Ashton (1974) for the purpose of assessing the consistency of audit judgments across respondents. Joyce focused upon the consistency of audit program plans assessed to be necessary in the audit of a t i r e whole-saler. Thir ty- f ive practising auditors participated in the study. A high variation in the responses of the participant auditors was reported (Joyce, 1976, p. 53). Joyce suggests that the high across subject con-sistency observed by Ashton (1974) may have resulted from the nature of the task that they were called upon to complete. Felix'(1976) provided some evidence about auditors' ab i l i t y to assess prior probability distributions from a case description of an order-receiving, shipping and b i l l i ng system. His study is closely related to the earl ier work of Corless (1972) reported above. Ten audit personnel participated in the study. Felix (1976, p. 805) found considerable agree-ment amongst the subject-auditors' responses both in terms of the loca-tion and dispersion of their responses. However, he cautions the reader from drawing any inferences from this result because i t may have resulted from an unintentional bias in the content of the case study. More recently, Weber (1978) investigated auditors' judgments about the impact of certain internal control weaknesses on the dollar error the system can produce. Weber indicated that amongst the respondents there was only a low level of consensus in their judgments about error sensi t iv i ty. Weber correlated a number of demographic and psychological measures obtained from each respondent with the specific decisions effected, but found, in general, only a few signficiant relationships 48. (Weber, 1978, p. 385). Weber goes so far as to suggest that the constructs derived from psychology may be inappropriate in the audit context (1978, p. 386). A number of studies have investigated how auditors make materiality judgments. Much of this research has focused on the consistency of the judgments amongst audit professionals. Moriarity and Barron (1976, p. 320) indicated that the findings of this research have demonstrated that there is no consensus in the profession. Bernstein (1967).summarizes much of the l i terature with the observation that auditors' decision functions with respect to materiality judgments appear to be of a highly personal nature since the actual decisions effected can vary signif icantly on what are the same or similar sets of facts. ' Studies by Dyer (1975), Pat t i l lo (1975) and Woolsey (.1954) reveal that dif ferent materiality thresholds are used by different auditors and the survey work of Pat t i l lo and Siebel (1974) and Woolsey (1973) reveal discrepancies amongst audit personnel in their judgments about the variables relevant to the materiality decision. Boatsman and Robertson (1974) examined the materiality judq-ments of a group of audit personnel and security analysts, however, their focus was not on the nature of the individual:judgments effected, but rather on the development of a model to predict those judgments. Moriarity and Barron (1976) extended the previous judgmental research which focused upon the materiality problem and examined the form of the individual auditor's decision model and the actual scaling of levels of contributing variables. Their work revealed that there were some differences in the form of the auditor's decisions models. Furthermore, they indi -cated that audit personnel scaled the contributing variables di f ferent ly 49. (Moriarity and Barron, 1976,.p. 338). The extant research points to substantial var iab i l i ty among auditors' responses to decisional problems in alternate settings. This finding is particularly signif icant because i t results from studies using alternate methodologies, subjects and problem orientations. However, amongst the studies reported above, l i t t l e attention has been directed towards explaining why such discrepancies persist. The existence of such dis-crepancies is of some concern, expecially when i t is considered that audit personnel, regardless of their firm a f f i l i a t i o n , educational back-ground or experience, share a common body of knowledge, are subject to s t r i c t professional accreditation procedures and share an allegiance to professional objectives. Attempts to explain differences in the judg-ments of individual auditors using psychological and demographic measures have not been successful. In the area of audit judgment these variables have not had a pervasive influence. This chapter.sets out to investigate the decision based behaviour of audit managers in large public accounting firms. Decisional responses of audit managers are assessed in terms of the economic incentives con-fronting those individuals. Systematic discrepancies are hypothesized to exist between the decisional responses that partners would want their managers to adopt and those that would typical ly be adopted by the managers. Such discrepancies are predictable and ver i f iable. They can be explained in terms of the economic incentives confronting managers on the job and the assumption that man is a resourceful, maximizing, evaluative individual (Jensen, 1976; Meckling, 1976). In the following section, attention is directed to the characteristics of the incentives confronting managers on the job. 50. C. The Contractual Relationship in Public Accounting Firms L i t t l e work has been undertaken for the purpose of investigating the nature of the contractual relationship in public accounting firms. However, there is some recognition that the reward system confronting audit personnel is l ike ly to be a major determinant of task related behaviour (Lawler and Rhode, 1976), turnover (Rhode, Sorensen, and Lawler, 1977) and interorganizational conf l ic t (Sorensen and Sorensen, 1974). Lawler and Rhode (1976), in their observations of the audit environ-ment, have stressed the need to consider the way in which the reward system in public accounting firms motivates behaviour on the job. For this purpose an agency framework wi l l be adopted. There is a growing l i terature in economics concerned with the derivation of optimal contracts between principals and agents under condi-tions of uncertainty. This l i terature has been broadly categorized under the heading of agency relationships. According to Jensen and Meckling, an agency relationship can be defined as a \"contract under which one or more persons (the principal(s)) engage another person (the agent) to perform some service on their behalf which involves delegating some decision making authority to the agent\" (1976, p..308). There exists a plethora of relationships which involve the elements of the agency relationship described above. However, our concern in this analysis wi l l be with the employer-employee relationship which exists amongst members of the audit f i rm. Ross has indicated that \"any situa-tion where labour services are hired gives rise to a meaningful agency relationship to the extent to which the employee possesses some decision making authority\" (1974, p. 215). 51. In this chapter, we focus on partners and managers in the audit firm who are seen to represent the principal and the agent respectively. I f we assume that al l individuals are resourceful, evaluative, maxi-mizing men (REMM's)(Meckling, 1976; Jensen, 1976), then we would expect that the agent w i l l attempt to improve his own welfare by taking actions which are not necessarily in the best interests of the principal. The agent may divert resources from the principal by shirking his responsi-b i l i t i e s , consuming perquisites, consuming on the job leisure, stealing the principal 's property and so forth (Alchian and Demsetz, 1972). The principal in such a setting wi l l typical ly instal l monitoring devices (such as security guards, supervisory personnel, and time clocks) and incentive systems (such as piece rates, bonuses, and promotions) to discourage the agent from selecting actions which enhance the agent's welfare at the principal 's expense. For his part the agent is able to enter into bonding agreements which reduce the likelihood that he wilUselect an action that wi l l reduce the principal 's welfare or which wi l l compensate the principal for any loss in welfare the principal incurs as a result of the agent's actions. Jensen and Meckling (1976, p. 308) indicate that i t is generally impossible for the principal or the agent to costlessly ensure that the agent w i l l make optimal decisions from the principal 's point of view. Indeed, they suggest that the principal and the agent :wi11 usually incur positive monitoring and bonding costs and that even given optimal monitor-ing and bonding act iv i t ies by the principal and the agent, there wi l l be some divergence between the sort of behaviour that the agent w i l l adopt and the sort of behaviour that the principal would want him to adopt in 52. a specific-problem context.. The dollar equivalent of the reduction in welfare experienced by the principal due to this divergence is referred to as the residual loss associated with the agency relationship. Together with the costs of monitoring and bonding the agent, the residual loss is a cost of agency. In this chapter of the study, a positive theory wi l l be developed to explain the manner in which the actions of audit manager d i f fer from the actions that their partners would want them to adopt in a specific problem context. For this purpose we focus on the characteristics of the reward system in public accounting firms and assess the way in which that system may influence behaviour of managers on the job. Managers in public accounting firms operate within a negative sanc-tion environment and.face a conditional investigation contract. Lawler and Rhode have indicated that i f non-partners \"do the job well i t is the expected result, however, i f the work is not done particularly wel l , punishment is rather immediate. Non-partners may, at the extreme, be discharged, and at a minimum receive a verbal dressing down\" (Lawler and Rhode, 1976, p. 120). Negative sanctions of this nature may be a signal that future rewards wi l l be reduced since the only permanent posi-tions in public accounting firms are those of partners. Staff accountants and managers have temporary positions; they wi l l either become partners or leave the f i rm. The internal promotion policy adhered to is one which emphasizes up or out selection (Lawler and Rhode, 1976, p. 126; Terkel, 1975, p. 354). There is some evidence to suggest that a reward system of this nature may induce behaviour on the part of audit personnel which deviates 53. from the sort of behaviour that their f irm's partners would want them to adopt in a specific problem context. Rhode and Gr i f f in (1978, pp. 32-33) in a questionnaire study found that staff accountants and seniors, especially those with six or fewer years of audit experience, indicated that they actually signed off incomplete audit procedures in order to ensure that they met the audit time budget. The tendency to sign off in-complete procedures was more prevalent amongst junior vis a vis senior audit personnel. Presumably, the evaluation of junior staff is more sensi-t ive to deviations from budgeted time allocations than is the evaluation of more senior personnel. There is a signif icant body of l i terature which draws attention to the dysfunctions associated with performance controls (Ridgway, 1956; Jasi.nski, 1956; Henderson and Dearden, 1966) and the evidence provided by Rhode and Gr i f f in (1978) is i l lus t ra t ive of much of that l i terature. However, a positive theory of managerial behaviour would enable us to predict the nature of the discrepancies which may arise between the behaviour that managers typical ly adopt in a given problem context and the behaviour that their firm would want them to adopt in that context, as well as the circumstances under which those discrepancies may arise. A promotion system which emphasizes up or out selection is l ike ly to induce, on the part of a rational manager, a s t r i c t conformity to organ-izational rules which influence his promotability within the.organiza-t ion. Conformity to organizational rules in organizational contexts which emphasize up or out selection has been well documented in the l i t -erature of modern organizational theory (Merton, 1957, pp. 54-55; Thompson, 1969, pp. 17-22; 1976, p. 124; Warwick, 1975, p. I l l ) and even 54. though not a l l managers may seek promotion internal ly, and may use the accounting firm only as a vehicle for the enhancement of their external marketability, conformance to organizational rules and regulations en-hances both internal and external opportunities for the manager. Rules are usually selected by the organization as a mechanism for performance evaluation when outcome measures are not readily available (Benveniste, 1977) or as a basis for defining which outcome measures wi l l be subject to organizational attention. In addition, rules form a basis for the defini t ion of appropriate behaviour within the organization. How-ever, within any organization, superiors are not usually able to monitor the performance of- their subordinates with respect to each organizational rule, and performance with respect to some rules can only be monitored at a signif icant (often prohibitive) cost. In this study, attention wi l l be directed towards a broad range of managerial behaviours and an assessment wi l l be made as to which of those behaviours are subject to s t r i c t conformance by audit managers within public accounting firms. A rational manager wi l l ensure that his behaviour conforms to those rules which can be readily monitored by his superiors. Conversely, such a manager may not always conform to those rules which cannot be readily monitored or which can only be monitored at a prohibitive cost. A fai lure to conform to these la t ter rules may not impair the manager's e l i g i b i l i t y for hierarchical positions within the organization. Whilst there is some probability that non-conformance to any organ-izational rule wi l l be detected, the size of that probability and the nature of the penalty imposed, i f detected, wi l l be inf luential in deter-mining which rules a rational subordinate wi l l comply with. For every 55. rule there is some probability of detection, given nonconformance - the size of that probability is variant across rules. A rational subordinate wi l l ensure s t r i c t conformance to those rules where the probability that non-conformance can be detected exceeds some threshold level. D. Research Objectives This study sets out to develop a positive theory of managerial behaviour in public accounting firms by focusing on the work of audit managers in those firms and assessing their behaviour in terms of pre-scriptions of the behaviour that their partners would want them to adopt in a specific problem context. Attention is specif ical ly directed to the decisional act iv i t ies of the audit manager i,n order to investigate whether his decisions and judg-ments in a given problem context d i f fer systematically from the decisions and judgments that his partners would want him to adopt in that context. Discrepancies between the decisional responses of partners and managers wi l l be examined in terms of the ease with which partners are able to monitor the performance of their subordiates on those dimensions. The basic hypothesis investigated is that relative to what their partners would want them to do, managers wi l l give a higher pr ior i ty to conformance to those rules which can be readily monitored by their superiors. Conversely, managers wi l l give a lower pr ior i ty to conformance to-those rules which cannot be readily monitored, or which can only be monitored at a prohibitive cost. Behaviour of this nature is entirely rational from the viewpoint of managers within public accounting firms. In addition, i t is entirely rational for their partners to allow such behaviour. 56. In order to test this basic hypothesis, i t is necessary for partners to provide prescriptions as to the way in which they would want their managers to respond to a given problem context and for managers to indi-cate how they would typical ly respond in that context. The decision con-texts that form the focal point of this analysis are defined in terms of the three operational stages of the audit - the design stage, the f i e ld -work stage, and the evaluation stage. The audit is a complex process -by sampling that process at three stages, i t is possible to assess the generality of the basic hypothesis as the course of the audit progresses. At the design stage the audit manager.'is involved in planning the audit, at the fieldwork stage he is involved in the execution of the audit and the management of personnel involved with the job. Finally, at the eval-uation stage the manager has a control function in which he evaluates the work of his subordinates. In Figure 2 a flowchart is presented which outlines the various stages of the annual audit. This flowchart is repre-sentative of the flow of work on the audit. I t is based on an inter-., national f irm's audit manual, oh the o f f i c ia l C.I.C.A. Handbook (C.I.C.A'., 1978) and on auditing texts (especially, Anderson, 1977). E. Implementation For the purpose of investigating the decisional behaviour of audit managers in large public accounting firms, we draw on the specification of managerial work presented in Chapter 2 and use that specification to examine partners' and managers' perceptions about the nature of managerial work in each of the decision contexts under investigation. The specific hypotheses that are examined are elaborated below. Figure 2: An I l lus t ra t ive Flowchart of the Work Undertaken on an Audit Preliminary Stage • Knowledge of business from past audits • Terms of this year's engagement • Update of knowledge of business • Update of systems and controls descriptions Design Stage • Evaluation of al l of above • Design of this year's audit Field Work Stage • Continuing update of knowledge of business • All compliance and substantive tests • All reviews and analyses of statements • All other interim and year-end procedures Evaluation Stage / Evaluation of al l of above • Clearing of f i l e queries, etc. • Drafting of audit report T Finalization Stage • Issuance of audit report •Preparation of f i l e notes for next year • Preparation of fee b i l l i ng • Writing of management le t ter to cl ient 58. The f i r s t six hypotheses (Hypotheses Cl - C6) are of a contextual nature and describe the environment within which the more substantive hypotheses (Hypotheses SI - S3) are investigated. The contextual hypo-theses are of supplementary interest in this study. They fac i l i ta te the interpretation and evaluation of the substantive hypotheses. The substantive hypotheses examine the decisional roles of the audit manager in terms of the ease with which performance with respect to those roles can be monitored. The specific content of these hypotheses is determined empirically on the basis of the specification of managerial work presented in Chapter 2 and a number of measures subsequently adopted to enhance the r e l i a b i l i t y and val id i ty of the measurement procedure. In general, the contextual hypotheses posit a high degree of agree-ment between and among partner and managers groups as to the behaviour managers wi l l typical ly adopt and the behaviour the firm would want them to adopt at a given stage of the audit. A high degree of agreement is also posited between partner and manager groups from different firms. The high degree of consensus hypothesized is indicative of the accredita-tion procedure operative within the audit environment and the careful system of monitoring and evaluation which exists at al l levels within audit firms. Support for such hypotheses wi l l enhance the interpreta-b i l i t y of the substantive hypotheses, which propose systematic discrep-ancies in the decisional responsesof partners and managers. These dis-crepancies are hypothesized to exist even though the intent of the incen-t ive system and the nature of the environmental context is understood by the partners and managers themselves. Whilst the basic hypothesis wi l l be tested across accounting firms, 5 9 . using the firm as a control device, the contextual hypotheses are examined within each f irm. The six contextual hypotheses are elaborated direct ly below. The f i r s t three hypotheses below focus on the explication of the in-centive system within a given firm in terms of the managerial behaviours i t is perceived to induce. In general, these hypotheses suggest that the intent of the incentive system is well explicated within a given f irm. HYPOTHESIS Cl: There wi l l be agreement amongst partners within a given firm as to the behaviour that their managers should take at a given-stage of the audit i f the managers' behaviour is to be in accord with firm policy. HYPOTHESIS C2: There wi l l be agreement amongst managers in a given firm as to the behaviour that they perceive their firm would want them to adopt at a given stage of the audit. HYPOTHESIS C3: There wi l l be agreement between partner and manager groups within a given firm as to the behaviour that they perceive the firm would want their managers to adopt at a given stage of the audit. The fourth contextual hypothesis focuses on what managers actually do at a given stage of the audit. I t suggests that among the managerial group there wi l l be a'high degree of agreement as to just what i t is that managers within a given firm do at a given stage of the audit. HYPOTHESIS C4: There wi l l be agreement amongst managers in a given firm as to the behaviour that .they would typical ly adopt at a given stage of the audit. The next two hypotheses examine the intent of the incentive system in different firms. A high degree of consensus is hypothesized amongst audit firms and support for these hypotheses would provide tentative evidence 60. about the simi lar i ty of':the incentive system in different firms. HYPOTHESIS C5: There wi l l be agreement among partners from different firms as to the behaviour that they perceive their firm would want their managers to adopt at a given stage of the audit. HYPOTHESIS C6: There wi l l be agreement among managers from different firms as to the behaviour that their firm would want them to adopt at a given stage of the audit. In order to specify the content of the substantive hypotheses subject to investigation, attention must f i r s t be directed to the research design and the methodology used for hypothesis testing. F. Research Design and Procedures Both the contextual hypotheses described above and the substantive hypotheses elaborated below examine perceptions about behaviour in a given context. Kogan and Wallach (1967), Bowers (.1973), Bern and Allen (1974), Endler and Magnusson (1976), Magnusson and Endler (1977) and Bern and Funder (1978), amongst others, have indicated that behaviour is a function of the person and the situation and have stressed that i t is the inter-action of the person and the situation that supplies most of the psycho-logically interesting variance in behaviour. In this study, the question examined is a question of person-situation interactions. In part icular, this study focuses on the work of the audit manager at various stages of the audit. For the purpose of investigating this question, we adopt the assessment language proposed by Bern and Funder (1978) which draws on the work of Stephenson (1953) and Block (.1961). That language is generally described as the Q-sort methodology. 61. A Q-sort methodology is the general name used by Stephenson (1953) to characterize a set of philosophical, s ta t i s t i ca l , and psychometric ideas oriented towards research on individuals. Recent reviews of the Q methodology point to i ts extensive use (Brown, 1968; Kerlinger, 1972). The Q procedure may be viewed as essentially a kind of rating procedure for rank ordering st imul i . I t has grown out of a general methodology for the study of verbalized att i tudes, individual preferences and self descrip-tions. This methodology has been widely used to study a broad range of issues (Wittehborn, 1961). In the accounting and auditing l i terature, the Q procedure has been applied to the investigation of the information pro-cessing characteristics of sales managers and supervisors (Dermer, 1973) and for the purpose of investigating the materiality construct in audit-ing (Ward, 1976). In general, the Q technique centres on the arrangement of decks of cards called Q sorts. A set of objects - verbal statements, words, phrases, pictures, or other stimuli - is given to an individual to sort into a set of piles according to some cr i ter ion. For the purpose of s tat is t ica l analysis, the sorter is instructed to put varying numbers of cards into each p i le , the whole making up a predetermined distr ibution which is typ i -cally normal or quasi-normal. In this study a Q-sort is developed which characterizes the work of the audit manager in terms of behaviours c r i t i ca l to his task. The Q-sort is used to investigate whether the audit manager and the firm have similar perceptions about the nature of managerial work at various stages of the audit. Perceptions about the nature of the audit manager's work are ex-pressed by rank ordering the cues in the Q-sort. 62. In this study, the Q-sort is modified to fac i l i ta te data collection. Subjects received randomly ordered sets of c r i t i ca l task requirements and ranked each set by completing a chart at the foot of the page on which each set of requirements appeared. A brief i l lus t ra t ion is presented in Exhibit 2. ILLUSTRATION OF THE Q-SORT METHODOLOGY Task Descriptors (randomly ordered) 1. Manage subordinates 2. Inform cl ient about the execution of the audit 3. Provide instruction to subordinates 4. Develop a fami l iar i ty with audit f i l e 5. Try to meet c r i t i ca l deadlines 6. Provide technical assistance for subordinates Chart to Aid Ranking Highest Rank (Most Important Item) 4 Next most important i terns 5 i 2 ' 1 6 Lowest Rank (Least Important Item) 3 Exhibit 2 63. In the exhibit , six task requirements which describe some of the things an audit manager does at the design stage of the audit are presented. A partial rank ordering of these task requirements is presented in the chart in the exhibit. To i l lus t ra te the Q-sorting procedure, I have ordered the six task requirements in terms of my perceptions about their impor-tance at the design stage of the audit. Nunnally (1967, p.. 555) sees the chief advantage of the Q-sort pro-cedure in i ts ab i l i t y to e l i c i t relat ively precise comparative responses amongst a large number of stimuli in a relat ively short period of time. This is important to the study at hand, because the hypothesis tests require the co-operation of senior audit personnel who give their time voluntarily from otherwise busy schedules. Kerlinger (1973, p. 598) indicated that the Q-sort methodology is able to make a valuable contribution to behavioural research. Using a Q-sort procedure: \"...one tests theories on small sets of individuals care-fu l l y chosen for their \"known\" or presumed possession of some signif icant characteristic or characteristics. One explores unknown and unfamiliar areas and variables for their ident i ty, their interrelation and their function-ing. Used thus, Q is an important and unique approach to the study of psychological, sociological and educational phenomena\". (Kerlinger, 1973, p. 598). F. Implementation of the Q-Sort Methodology The investigation of the hypotheses in this study requires the develop-ment of a Q-sort which characterizes the work of audit managers in terms of a number of cues descriptive of their task related behaviour. For the Q-sort to be descriptively val id, the cues must be unambiguous to those who wi l l sort them and relevant to the context under investigation. 64. For this purpose, cues were selected from the l i s t of 88 c r i t i ca l task requirements identif ied using the c r i t i ca l incident methodology. The task requirements are recorded using the phrases provided by inter-viewees and should be unambiguous to those who wi l l subsequently sort them. However, in order to ensure that the cues selected are relevant to the context under investigation and descriptive of the managerial roles subject to investigation, two additional validation procedures were imple-mented. These procedures enhance thewal id i ty of the cues used in the Q-sort and enhance the r e l i a b i l i t y of the role constructs which wi l l form the test of the substantive hypotheses to follow. In Chapter 2, each of the 88 c r i t i ca l task requirements was assigned by a panel of seven judges to one of the managerial roles which best encapsulated i t . A majority voting rule was applied and where no majority existed, the assignment of task requirements to role descriptors was effected by discussing the assignment with some of the judges. To enhance the r e l i a b i l i t y of the role constructs in this phase of the research, only cues for which the role assignment was agreed to by f i ve , six or seven out of the seven judges were selected for further analysis. This greatly reduced the number of cues upon which i t was necessary to focus but ensured that there was substantial agreement among the judges as to the managerial role which was most descriptive of a given task require-ment. To ensure that the task requirements were descriptively valid at each stage of the audit under investigation, i t was necessary to ta i lo r the Q-sort items to each audit stage. This was accomplished by having three further audit personnel review the l i s t of task requirements in terms of their descriptive relevance at each stage of the audit. Only 65. those task requirements which were perceived to be relevant by two out of the three auditors were selected for further analysis. This procedure further reduced the number of cues that would be investigated at each stage,of the audit but enhanced the face val id i ty of the cues to be used. The research instrument was thus tailored to each stage of the audit sub-ject to investigation. This should make the Q-sorting task more interest-ing and challenging for respondents. The set of Q-sort items at each stage of the audit (see Appendix I I I ) varies s l ight ly . At the design stage there are nineteen cues in the Q-sort whilst at the fieldwork and evaluation stages there are twenty-four and twenty-three items respectively. Each respondent was asked to express his comparative preferences amongst the cues at each stage by sorting the items into seven stacks defined by the chart at the foot of the page on which the cues were pre-sented. The number of items placed in each stack was specified so that a bell shaped distr ibut ion resulted from the sort of the Q-sort items at each stage of the audit. This form of distr ibution has proved to be useful in previous studies (Wittenborn, 1961, p. 140). However, there is some evidence to suggest that the choice of distributional form has l i t t l e effect on the kinds of analyses which are usually made from the Q-sort data (Nunnally, 1967, p. 547). G. The Substantive Hypotheses The basic hypothesis that is subjected to investigation examines whether there are systematic discrepancies between the way in which partners would want their managers to respond to a given problem context and the 66. way in which managers typical ly respond to that context. Specifically, i t is hypothesized that relative to what their partners would want their managers to do, managers wi l l give a higher pr ior i ty to conformance to those rules which can be readily monitored by their superiors. Con-versely, managers wi l l give a lower pr io r i ty to conformance to those rules which cannot be readily monitored by their superiors or which can only be monitored at a prohibitive cost. In order to operationalize the basic hypothesis, four auditors from the regional offices of the firms that participated in the Q-sort were asked to classify the Q-sort items in terms of their f irm's ab i l i t y to monitor subordinates' performance on those dimensions. I f the firm can monitor the performance of their staf f with respect to specific task requirements, then they are also able to monitor performance with respect to rules which relate to those task requirements. In Chapter 2, the manager was seen to have four decisional roles. (1) A negotiator role (2) A resource allocative role (3) An entrepreneurial.role . (4) A disturbance handler role When the procedures designed to ensure the construct r e l i a b i l i t y and face val id i ty of the Q-sort items were applied, no c r i t i ca l task require-ments were assigned to the disturbance handler act iv i t ies of the audit manager. Consequently, tests of the basic hypothesis are only based on tests pertaining to the f i r s t three decisional roles l isted above. The Q-sort items which describe the manager's resource allocative act iv i t ies were each seen to be readily monitorable by the four auditors 67. who provided this information. As a resource allocator, the manager is concerned with the allocation of a broad range of organizational resources. His resource allocative act iv i t ies w i l l determine the extent to which he is able to meet c r i t i ca l deadlines, satisfy budgetary constraints and meet timing-considerations. Resource allocative act iv i t ies are monitored within public accounting firms using output measures which are readily available within'the organization. Of the three decisional roles under investigation, each of the four auditors interviewed indicated that perfor-mance with respect to the manager's resource allocative act iv i t ies was the easiest to monitor within the organization. I t is therefore hypothe-sized that: HYPOTHESIS SI: Managers wi l l give a higher pr ior i ty to conformance to resource allocative act iv i t ies than their partners would want them to give. As a negotiator, the manager represents the firm in important con-frontations with their c l ient . One c r i t i ca l task requirement which was assigned to the negotiator role of the manager appears amongst the Q-sort items. Specif ically, as a negotiator the manager reconciles technical differences with their c l ient . The four audit managers generally f e l t that the performance of personnel can be readily monitored in terms of their ab i l i t y as a negotiator within the organization, though performance evaluation in terms of the negotiation role was seen to be more subjective than the evaluation of personnel on the basis of their resource alloca-tive decisions. I t is hypothesized that: HYPOTHESIS S2: Managers wi l l give a higher pr ior i ty to conformance to negotiation act iv i t ies than their partners would want them to give. 68. F ina l l y , a t tent ion is directed to the ro le of the manager as an entre-preneur wi th in the organizat ion. The Q-sort items (see Appendix I I I ) indicate that as an entrepreneur the manager'is concerned with the main-tenance and furtherance of the ex is t ing pract ice (by providing management type advice and services to ex is t ing c l ien ts and se l l i ng addit ional services to those c l i en ts ) and with the acceptance of respons ib i l i t i es associated with his posi t ion and status wi th in the organization (by d is -playing an a b i l i t y to i den t i f y problems and by developing technical s k i l l s in a spec ia l i s t area). The entrepreneurial a c t i v i t i e s of audit personnel were generally perceived to be more d i f f i c u l t to monitor than e i ther the negotiat ion or resource a l locat ion a c t i v i t i e s of those personnel. Amongst the entre-preneurial a c t i v i t i e s there was a high degree of agreement amongst the four auditors that those a c t i v i t i e s concerned with maintaining or extend-ing the pract ice were more d i f f i c u l t to monitor than those pertaining to the a b i l i t y of the auditor to i den t i f y problems or develop technical s k i l l s in a spec ia l i s t area. Whilst object ive measures are avai lable which re la te to the manager's a b i l i t y to maintain and develop the ex is t ing pract ice , performance evalua-t ion in terms of these dimensions is d i f f i c u l t to monitor. A f a i l u r e to provide management type advice and services to ex is t ing c l i en ts or a f a i l u r e to se l l addit ional services to them may not be ind icat ive of poor entrepreneurial performance. Furthermore, the fac t that such services are provided may not indicate that the auditor concerned displayed entre-preneurial s k i l l s . Performance evaluation in terms of these a c t i v i t i e s was perceived to be d i f f i c u l t by each of the four auditors because no 6 9 . norms were available upon which to base such evaluations. I t is hypothe-Q sized that: HYPOTHESIS S3: Managers wi l l give a lower pr ior i ty to conformance to entrepreneurial act iv i t ies than their partners would want them to give. The preceding discussion has emphasized that managerial performance can be more readily monitored with respect to resource allocative decisions than i t can with respect to either entrepreneurial or negotiation decisions. On the other hand, i t is more d i f f i c u l t to monitor entrepreneurial decisions than i t is to do so with respect to either resource allocative or negotiation decisions. The directional differences hypothesized above between the pr ior i ty assigned to specific managerial roles by partners and managers, are based on the ease with which partners are able to moni-tor the performance of subordinates on those dimensions. The nature of the direction hypothesized is less ambiguous with respect to entrepreneurial and resource allocative decisions than i t is with respect to negotiation decisions. Since the Q-sort has been specif ical ly tailored to each stage of the audit, the test of each of the substantive hypotheses may be based on varying cue combinations at alternative audit stages. Not al l hypotheses wi l l be tested at each stage of the audit because specific task require-ments assigned to a particular managerial role were not seen to be descrip-t ive of the work of the audit manager at that stage.of the audit by the auditors involved in the development of the Q-sorts. In the Table below, the hypotheses which are tested at each stage of the audit are ident i f ied. 70. TABLE I I I TABLE OF THE SUBSTANTIVE HYPOTHESES AT EACH AUDIT STAGE TO BE TESTED Design Fieldwork Stage Stage Evaluation Stage Resource Allocation Hypothesis / (Hypothesis SI) / / Negotiation Hypothesis (Hypothesis S2) / / Entrepreneurial Hypothesis / (Hypothesis S3) / / H. The Research Material Each participant in the study received a booklet containing the research material. Separate booklets were prepared for partners and managers. Copies of the research material are reproduced in Appendix IV. The Q-sort material was preceded by an introductory let ter explaining in very broad terms the nature of the study and e l i c i t ing the co-operation of each participant. Respondents were guaranteed anonymity and were asked not to discuss the research with others. To enhance the val id i ty of the study, subjects were provided with a general description of the stages of the audit and the sort of company that they should consider throughout the ranking process. An example of the Q-sort methodology was provided which was unrelated to the work of the audit manager. This example was designed to expose subjects to Q-sorting and to train them in the completion of the chart at the bottom of each page. Care was taken to ensure that subjects were not biased by the example in the way they responded to Q-sorting the work of the audit manager. For this purpose, 71. the i l lus t ra t ive example was developed on a much smaller scale than the actual Q-sorting task would entail and used different notation for the completion of the actual chart presented than that which would be used in the subsequent Q-sorts. The research instrument was extensively pretested with audit personnel not previously involved in any other phase of the study and with account-ing academics. Debriefing procedures followed each of the pretest sessions to ensure that the Q-sorting task was perceived to be meaningful to this group of individuals. Certain psychometric properties of the research instrument were investigated. Four audit managers not involved in the data collection phase of the study completed the research instrument twice, over an interval of two to four weeks. Test-retest stat ist ics were computed using the data provided by these managers. Over 24 matched Q-sorts (six Q-sorts from each of four auditors), the average Spearman rank order correlation coefficient was .64 which was f e l t to be most sat-isfactory for a study of this nature. Each subject participating in the study was asked to sort the c r i t i ca l task requirements at each of the three stages of the audit. Test of the hypotheses presented requires partners to sort the items once only at each stage of-the audit. Managers are required to complete two sorts at each stage of the audit: one in which they indicate which of the cues l isted they would be expected to comply with i f their behaviour is in accord with firm policy, and another in which they indicate which of the cues l isted they normally adhere to at a given stage of the audit. To reduce problems associated with order effects which may arise in any rank ordering procedure, subjects were told that the Q-sort items 72. were presented in random order when l isted at a given stage of the audit. In addition, at each stage of the audit, four versions of the applicable Q-sort l i s t were prepared so that no item was always at the top or bottom of the l i s t . To reduce any reactive effects that may arise as a result of the order in. which Q-sorts:;were elicited,,-the order in which managers were asked to provide the two sets of rankings was randomized across sub-jects in each f irm. Once subjects have rank ordered the st imul i , values are assigned to ranks as follows: one is assigned to the item in the upper most cell in the Q-sort, two is assigned to the items in the next set of ce l ls , three is assigned.to the items in the third set of ce l ls , and so on unti l seven is assigned to the item in the bottom most ce l l . This procedure for the assignment of numerals to Q-sort data is the conventional one (Block, 1961, pp. 77-82) and typical ly results in consensus orderings which are quite equivalent to those derived using more complex schemes (Block, 1961, p. 36; Dunnette and Hoggatt, 1957, pp. 425-434). The numerals assigned to the Q-sort data are used in the tests of the hypotheses described above. Each of the substantive hypotheses is evalu-ated by comparing the value assignments to each decisional role by partners and managers. Similarly, the contextual hypotheses are evaluated by com-paring the rank d i s t r i b u t i o n provided by each subject within the f i rm. Whilst the data collected in this study using the Q-sort methodology is of a perceptual nature, the Q-sort is used as a mechanism for the pre-diction of behaviour in a specific context. The Q-sort provides a hypo-thetical-deductive methodology for the prediction of the behaviour of audit managers in large public accounting firms. Behaviours ranked highly 73. in the Q-sort are behaviours which managers should adhere to i f their behaviour is in accord with firm policy or behaviours which managers always adhere to during the execution of a given audit stage. Ranking items in the Q-sort provides a mechanism for defining the nature of mana-gerial strategy within audit firms. In the following chapter, the hypothesis tests wi l l be outlined and the test results presented. 74. APPENDIX I I I The Crit ical Task Requirements in the Q-Sort (Items asterisked appear in the Q-sort at the stage of the audit indicated) DESIGN STAGE FIELDWORK STAGE EVALUATION STAGE Resource Allocation Decisions Try to u t i l i ze available personnel effectively Try to complete the job in a timely manner Try to meet c r i t i ca l deadlines Try to complete the job cost effectively Negotiation Decisions Reconcile technical differences with the cl ient Entrepreneurial Decisions Display an ab i l i t y to identify problems Develop technical sk i l l s in a specialist area Provide management type advice and services Sell additional services to existing clients Other Managerial Act iv i t ies Impose real is t ic requirements on subordinates Develop team relationships Maintain professional appearances Create meaningful learning experience for subordinates Inform the cl ient about the execution of the audit Appendix I I I (continued) 75. DESIGN FIELDWORK EVALUATION STAGE STAGE STAGE Provide instruction to * * ^ subordinates Provide technical assistance to * * ^ subordinates Communicate effectively with * * * cl ients Manage subordinates * * * Provide supervision for * * * subordinates Communicate effectively with * * ^ superiors Motivate subordinates * * * Complete the management le t ter * Accept responsibil ity for the work * * * of others Develop a fami l iar i ty with the * * * audit f i l e Adapt to superiors * Abide by professional code of * * ethics Display loyalty to the firm * * * Ensure the financial statements * ^ are error free 76. APPENDIX IV THE RESEARCH INSTRUMENTS 1. Proposed Memo to Participants Frank Wolf, a doctoral student at U.B.C. has asked for our help in the completion of his research on the work of audit managers in large public accounting firms. The research has been conducted in two stages. The f i r s t stage was completed in Vancouver during the summer. Frank has already provided feedback to each of the participant firms and has indicated that he wi l l also provide feedback to us on the completion of the research. During the f i r s t stage of the work, c r i t i ca l incidents pertaining to the work of audit managers were e l ic i ted. These were analysed to yield a number of c r i t i ca l task requirements. Some stat is t ica l tests were carried out. During this stage you wi l l be asked to rank various subsets of these c r i t i ca l task requirements at a number of stages of the audit. The rank-ing task has been simplified by call ing for only a partial rank ordering of each set of requirements. I have discussed the research material and the entire project with Frank and fu l l y endorse i t . I feel that the results which he hopes to derive wi l l be worth the investment of our time in i ts completion. The research material has been carefully debugged by a number of practitioners in Vancouver. They have assured Frank that auditors would have no problems in the completion of the material. Please devote some time to the completion of the research material and return i t to me in the envelope provided. I t would be appreciated i f you could do this within a week so that we can forward the material to Vancouver for analysis. MANY THANKS FOR YOUR COOPERATION. Appendix IV (continued) 78. THE SETTING FOR THE RESEARCH In completing the questionnaire, please focus on the context descrip-tions provided. These refer to: (1) The Nature of the Client (2) The Flow of Work on the Audit ( 1 ) THE NATURE OF THE CLIENT The company that w i l l be the focus of this study is a large public manufacturing company which is incorporated in a Canadian jur isd ic t ion. The company has been your f irm's cl ient for several years and the audit manager is assumed to have been responsible for the audit of this cl ient last year. The audit is a normal statutory audit - there are no special reporting requirements or major international implications. Appendix IV (continued) 79. (2) THE FLOW OF WORK ON THE AUDIT The flowchart presented outlines the various stages of the audit. J_n this study attention wi l l focus on the DESIGN, FIELDWORK and EVALUATION STAGES. Preliminary Stage * Knowledge of business from past audits * Terms of this year's engagement * Update of knowledge of business * Update of systems and controls descriptions DESIGN STAGE * Evaluation of al l of above * Design of this year's audit FIELDWORK STAGE * Continuing update of knowledge of business * All compliance and substantive tests * All reviews and analyses of statements * All other interim and year-end procedures EVALUATION STAGE * Evaluation of a l l of the above * Clearing of f i l e queries * Drafting of audit report Finalization Stage * Issuance of audit report * Preparation of f i l e notes for next year * Preparation of fee b i l l i ng [i * Writing of management le t ter to cl ient FOCUS OF THIS STUDY Appendix IV (continued) 80. COMPLETING THE QUESTIONNAIRE The completion of the questionnaire wi l l involve ranking a randomly ordered subset of c r i t i ca l task requirements at each stage of the audit. A s l ight ly dif ferent l i s t of task requirements is presented at each stage. These requirements are descriptive of some of the things managers do at a given stage; they may not be comprehensive. You wi l l be asked to provide three sets of rankings; one at each stage of the audit. Each set of rankings wi l l indicate which task require- ments your firm would want managers to comply with at that stage. Ranking;can be completed by f i l l i n g in the chart on each page. Below an example is provided of the ranking procedure. I t relates to the work of a bus driver. I have ranked the six task requirements l isted accord-ing to what I believe the Bus Co. would want their drivers' to comply with on the job. EXAMPLE: Ranking the task requirements of a bus driver. Items with a high rank assigned to them are items that the Bus. Co. would want their drivers to comply with. Items with lower ranks are not as c r i t i ca l to the execu- tion of their work. Presumably the Co. would not be too concerned i f they were not always complied with. Task Requirements: Chart to aid ranking: a) Drive within speed l imits b) Try-to meet timetables c) Distribute transfers quickly d) Display courtesy to passengers e) Drive safely f ) Announce street names Highest Rank Next Highest d e b c a Lowest Rank f NOTE: The Bus Co..always wants safe drivers: hence item 'e' is in the top most box. The next 2 items are 1 b' and ' a ' . They are ranked equally. Of least concern is item ' f . Appendix IV (continued) THE RANKING PROCEDURE The charts on each of the following pages are designed to simplify the ranking process. Items at a particular level on the chart are of equal importance and wi l l have the same rank assigned to them. In com-pleting the questionnaire: (1) Please use a lead pencil. This w i l l fac i l i ta te any changes in the rankings you may wish to make. (2) Please do not violate the format provided by each chart. (3) Please f i l l in every box on the chart. (4) Please work through the questionnaire in the order presented. Appendix IV (continued) THE WORK OF THE AUDIT MANAGER AT THE DESIGN STAGE 82. Rankings Based on What the Firm Would Want Managers to Comply With at the Decision Stage..of the Audit Please rank the items l isted below so that: Items with a HIGH RANK assigned to them are items which THE FIRM would ALWAYS WANT MANAGERS TO COMPLY WITH at the DESIGN STAGE. Items with LOWER RANKS assigned to them are NOT AS CRITICAL at this stage of the audit and presumably the firm would not be concerned i f they were NOT ALWAYS COMPLIED WITH at this stage. 1. Try to u t i l i ze available personnel effectively. 2. Inform the cl ient about the execution of the audit 3. Try to complete the job cost effect ively. 4. Develop a fami l iar i ty with the audit f i l e . 5. Display loyalty to the f i rm. 6. Communicate effectively with superiors. 7. Provide motivation for subordinates. 8. Communicate effectively with cl ients. 9. Provide supervision for subordinates. 10. Try to meet c r i t i ca l deadlines. 11. Try to complete the job in a timely manner. 12. Manage subordinates. 13. Develop team relationships. 14. Display an ab i l i t y to identify problems. 15. Provide instruction for subordinates. 16. Create meaningful learning experiences for subordinates. 17. Accept responsibil ity for the work of others. 18. Impose real is t ic requirements on subordinates. 19. Provide technical assistance to subordinates. Highest Rank Next Highest Identi f ication # of items the firm would want their managers to comply with at the DESIGN STAGE of the audit. Lowest Rank Appendix IV (continued) THE WORK OF THE AUDIT MANAGER AT THE FIELDWORK STAGE 83. Rankings Based on What the Firm Would Want Managers to Comply With at the Fieldwork Stage of the Audit Please rank the items l isted below so that: Items with a HIGH RANK assigned to them are items which THE FIRM would ALWAYS WANT MANAGERS TO COMPLY WITH at the FIELDWORK STAGE. Items with LOWER RANKS assigned to them are NOT AS CRITICAL at this stage of the audit and presumably the firm would not be concerned i f they were NOT ALWAYS COMPLIED WITH at this stage. 1. Provide technical assistance to subordinates. 2. Try to meet c r i t i ca l deadlines. 3. Maintain professional appearances. 4. Provide instruction for subordinates. 5. Develop technical sk i l l s in specialist area. 6. Display loyalty to the f irm. 7. Reconcile technical differences with c l ient . 8. Develop a fami l iar i ty with the audit f i l e . 9. Try to complete the job cost effect ively. 10. Ensure the financial statements are error free. 11. Create meaningful learning experiences for subordinates. 12. Display an ab i l i t y to identify problems. 13. Manage subordinates. 14. Try to u t i l i ze available personnel effectively. 15. Try to complete the job in..timely manner. 16. Communicate effectively with cl ients. 17. Abide by the professional code of ethics. 18. Provide supervision for subordinates. 19. Inform the cl ient about the execution of the audit. 20. Communicate effectively with superiors. 21. Provide motivation for subordinates. 22. Accept responsibil ity for the work of others. 23. Impose rea l is t ic requirements on subordinates. 24. Develop team relationships. Highest Rank Next Highest Identif ication # of items the firm would want their managers to comply with at the FIELDWORK STAGE of the audit. Lowest Rank Appendix IV (continued) THE WORK OF THE AUDIT MANAGER AT THE EVALUATION STAGE 84. Rankings Based on What the Firm Would Want Managers to Comply With at the Evaluation Stage of the Audit Please rank the items l isted below so that: Items with a HIGH RANK assigned to them are items which THE FIRM would ALWAYS WANT MANAGERS TO COMPLY WITH at the EVALUATION STAGE. Items with LOWER RANKS assigned to them are NOT AS CRITICAL at this stage of the audit and presumably the firm would not be concerned i f they were NOT ALWAYS COMPLIED WITH at this stage. 1. Accept the responsibil ity for the work of others. 2. Try to complete the job cost effect ively. 3. Display loyalty to the f i rm. 4. Develop technical sk i l l s in a specialist area. 5. Provide management type advice and services. 6. Display an ab i l i t y to identify problems; 7. Provide supervision for subordinates. 8. Complete the management le t ter . 9. Sell additional services to existing cl ients. 10. Provide motivation for subordinates. 11. Try to meet c r i t i ca l deadlines. 12. Communicate effectively with cl ients. 13. Adapt to superiors. 14. Provide technical assistance to subordinates. 15. Provide instruction for.subordinates. 16. Communicate effectively with superiors. 17. Develop a fami l iar i ty with the audit f i l e . 18. Try to complete the job in a timely manner. 19. Create meaningful learning experiences for subordinates. 20. Abide by professional code of ethics. 21. Manage subordinates. 22. Ensure the .'financial statements are error free. 23. Reconcile technical differences with c l ient . Highest Rank Next Highest Identif ication # of [ items the firm would \" want their managers to [ comply with at the EVALUATION STAGE of ' the audit. ] 'Lowest Rank Appendix IV (continued) 85. A FEW CONCLUDING QUESTIONS Please consider each of the following problems and respond in the space provided. The problems are of a general nature and do not necessarily relate to the Manufacturing Co. referred to above. 1. The manager on the job feels somewhat uneasy about the c l ient 's state of affairs but cannot pinpoint any specific problems. The audit has gone well - a l l tests have checked out, a l l deadlines have been met, and the audit is well within budget. The cl ient is not aware of the manager's apprehensions. IF THE MANAGER WERE ACTING IN THE BEST INTERESTS OF THE FIRM, HOW INCLINED WOULD YOU WANT HIM TO BE TO SPEND RESOURCES IN SEARCH OF A PROBLEM WHICH, MAY AFTER ALL, NOT EXIST? Please put a dash (/) on the l ine. Commit Do Not Commit Resources Resources 2. The manager is privy to information from a very good friend who works for the QWER Co. His friend is very reliable and is an accountant by profession. The manager has often been told by his friend that the integ-r i t y of QWER's management is subject to question. Just the other day QWER1 management had a quarrel with their auditors who were apparently not ' f lex ib le ' enough. The manager was advised by his friend that QWER is looking for a change of auditors but was cautioned, at the same time, not to get involved. QWER is a moderate sized company involved in importing a wide range of commodities. IF THE MANAGER WERE ACTING IN THE BEST INTEREST OF THE FIRM, HOW INCLINED WOULD YOU WANT HIM TO BE TO PUT QWER UP FOR CONSIDERATION BY THE FIRM? Recommend Do Not Recommend QWER QWER 3. The senior on the job has come into conf l ict with the c l ient 's accountant regarding the scope of,the audit. The accountant is a d i f f i c u l t person and problems are always experienced on this job. This year, top management has threatened to change auditors. i f any problems arise. The senior has told the accountant that certain preparatory work must be com-pleted by the c l ient . This is firm policy. The accountant refuses to do the work and has threatened to take the problem to top management. There is not a lot of work involved but the senior does not want to establish an undesirable precedent;. The audit manager has been cal led-in. IF THE MANAGER WERE ACTING IN THE BEST INTERESTS OF THE FIRM, WOULD YOU WANT HIM TO INSIST ON THE COMPLETION OF THE PREPARATORY WORK BY THE CLIENT OR WOULD YOU WANT HIM TO NEGOTIATE A COMPROMISE SITUATION? Insist on Negotiate a Firm Policy Compromise Solution Appendix IV (continued) 86. BIOGRAPHICAL QUESTIONS 1. How many years have you been a partner? 2. When did you receive your CA? 3. How long have you been employed in this firm? 4. I f you were offered the opportunity of winning $500 on .the toss of a fa i r coin, how much would you be pre-pared to pay to participate in the gamble, knowing that i f the coin goes against you, you lose what you pay? Please add any comments, suggestions or questions that you think may be helpful in interpreting your answers to the questions asked. Many Thanks For Your Help Appendix IV (continued) 88. THE SETTING FOR THE RESEARCH In completing the questionnaire, please focus on the context descrip-tions provided. These refer to: (1) The Nature of the Client (2) The Flow of Work on the Audit ( 1 ) THE NATURE OF THE CLIENT The company that wi l l be the focus of this study is a large public manufacturing company which is incorporated in a Canadian jur isd ic t ion. The company has been your f irm's cl ient for several years and the audit manager is assumed to have been responsible for the audit of this cl ient last year. The audit is a normal statutory audit - there are no special reporting requirements or major international implications. Appendix IV (continued) 8 9 . (2) THE FLOW QF WORK ON THE AUDIT The flowchart presented outlines the various stages of the audit. In this study, attention wi l l focus on the DESIGN, FIELDWORK and EVALUATION STAGES. FOCUS OF THIS STUDY Preliminary Stage * Knowledge of business from past audits * Terms of this year's engagement * Update of knowledge of business * Update of systems and controls descriptions I DESIGN STAGE If. Evaluation of a l l of above * Design of this year's audit I FIELDWORK STAGE * Continuing update of knowledge of business * All compliance and substantive tests * All reviews and analyses of statements * All other interim and year-end procedures I EVALUATION STAGE * Evaluation of a l l of the above * Clearing of f i l e queries * Drafting of audit report { Finalization Stage t * Issuance of audit report * Preparation of f i l e notes for next year * Preparation of fee b i l l i ng * Writing of management let ter to c l ient Appendix IV (continued) 90. COMPLETING THE QUESTIONNAIRE The completion of the questionnaire w i l l involve ranking a randomly ordered subset of the c r i t i ca l task requirements at each stage of the audit. A s l ight ly different l i s t of task requirements is presented at each stage. These requirements are descriptive of some of the things managers do at a given stage; they may not be comprehensive. The f i r s t 3 sets of rankings you wi l l be asked to provide wi l l indicate which task requirements your firm would want you to comply with at each stage of the audit. To provide these rankings, i t w i l l be necessary to think about just what firm policy would want managers l ike you to do at each stage of the audit. Ranking can be completed by f i l l i n g in the chart on each page. Below an example is provided of the ranking procedure. I t relates to the work of a bus driver. I have ranked the six task requirements l isted according to what I believe the Bus. Co. would want their drivers to comply with on the job. EXAMPLE: Ranking the task requirements of a bus driver. Items with a high rank assigned to them are items that the Bus. Co.wouldwant their drivers to comply with. Items with lower ranks are not as c r i t i ca l to the execution of their work. Presumably the Co. would not be too concerned i f they were not always complied with. Task Requirements: a) Drive within speed l imits b) Try to meet timetables c) Distribute transfers quickly d) Display courtesy to passengers e) Drive safely f ) Announce street names NOTE: The Bus. Co. always wants safe drivers: hence item 'e' is Chart to aid ranking: Highest Rank Next Highest Lowest Rank in the top most box. The next 2 items are 'b' & ' a ' . They are ranked equally. Of least concern is item ' f Appendix IV (continued) THE RANKING PROCEDURE The charts on each of the following pages are designed to simplify the ranking process. Items at a particular level on the chart are of equal importance and wi l l have the same rank assigned to them. In com-pleting the questionnaire: (1) Please use a lead pencil. This wi l l fac i l i ta te any changes in the rankings you may wish to make. (2) Please do not violate the format provided by each chart. (3) Please f i l l in every box on the chart. (4) Please work through the questionnaire in the order presented. Appendix IV (continued) THE WORK OF THE AUDIT MANAGER AT THE DESIGN STAGE 92. Rankings Based on What the Firm Would Want Managers to Comply With at the Design Stage of the Audit Please rank the items below so that: Items with a HIGH RANK assigned to them are items which THE FIRM would ALWAYS WANT MANAGERS TO COMPLY WITH at the DESIGN STAGE. Items with LOWER RANKS assigned to them are NOT AS CRITICAL at this stage of the audit and pre-sumably the firm would not be concerned i f they were NOT ALWAYS COMPLIED WITH at this stage. 1. Try to u t i l i ze available personnel effectively. 2. Inform the cl ient about the execution of the audit. 3. Try to complete the job cost effectively. 4. Develop a fami l iar i ty with the.audit f i l e . 5. Display loyalty to the f i rm. 6. Communicate effectively with superiors. 7. Provide motivation for subordinates. 8. Communicate effectively with cl ients. 9. Provide supervision for subordinates. 10. Try to meet c r i t i ca l deadlines. 11. Try to complete the job in a timely manner. 12. Manage.:siibordinates. 13. Develop team relationships. 14. Display an ab i l i t y to identify problems1:; 15. Provide instruction for subordinates. 16. Create meaningful learning experiences for subordinates. 17. Accept responsibil ity for the work of others. 18. Impose real is t ic requirements on subordinates. 19. Provide technical assistance to subordinates. Highest Rank Next Highest Identif ication # of items the firm would want their managers to comply with at the DESIGN STAGE of the audit. Lowest Rank Appendix IV (continued) THE WORK OF THE AUDIT MANAGER AT THE FIELDWORK STAGE 93. Rankings Based on What the Firm Would Want Managers to Comply with at the Fieldwork Stage of the Audit Please rank the items l isted below so that: Items with a HIGH RANK assigned to them are items which THE FIRM would ALWAYS WANT MANAGERS TO COMPLY WITH at the FIELDWORK STAGE. Items with LOWER RANKS assigned to them are NOT AS CRITICAL at this stage of the audit and presumably the firm would not be concerned i f they were NOT ALWAYS COMPLIED WITH at this stage. 1. Provide technical assistance to subordinates. 2. Try to meet c r i t i ca l deadlines. 3. Maintain professional appearances. 4. Provide instruction for subordinates. 5. Develop technical sk i l l s in specialist area. 6. Display loyalty to the f i rm. 7. Reconcile technical differences with c l ient . 8. Develop a fami l iar i ty with the audit f i l e . 9. Try to complete the job cost effect ively. 10. Ensure the financial statements are error free. 11. Create meaningful learning experiences for subordinates. 12. Display an ab i l i t y to identify problems. 13. Manage subordinates. 14. Try to u t i l i ze available personnel effect ively. 15. Try to complete the job in a timely manner. 16. Communicate effectively with c l ients. 17. Abide by the professional code of ethics. 18. Provide supervision for subordinates. 19. Inform the cl ient about the execution of the audit. 20. Communicate effectively with superiors. 21. Provide motivation for subordinates. 22. Accept responsibil ity for the work of others. 23. Impose real is t ic requirements on subordinates. 24. Develop team relationships. Highest Rank Identi f icat ion # of Next Highest items the firm would want their managers to comply with at the FIELDWORK STAGE of the audit. Lowest Rank Appendix IV (continued) THE WORK OF THE AUDIT MANAGER AT THE EVALUATION STAGE 94. Rankings Based.on What the Firm Would Want Managers to Comply with at the Evaluation Stage of the Audit Please rank the items l isted below so that: Items with a HIGH RANK assigned to ,them are items which THE FIRM would ALWAYS WANT MANAGERS TO COMPLY WITH at the EVALUATION STAGE. Items with LOWER RANKS assigned to them are NOT AS CRITICAL at this stage of the audit and presumably the firm would not be concerned i f they were NOT ALWAYS COM-PLIED WITH at this stage. 1. Accept the responsibil ity for the work of others. 2. Try to complete the job cost effectively. 3. Display loyalty to the f i rm. 4. Develop technical sk i l l s in a specialist area. 5. Provide management type advice and services. 6. Display an ab i l i t y to identify problems. 7. Provide supervision for subordinates. 8. Complete the management le t te r . 9. Sell additional services to existing cl ients. 10. Provide motivation for subordinates. 11. Try to meet c r i t i ca l deadlines. 12. Communicate effectively with cl ients. 13. Adapt to superiors. 14. Provide technical assistance to subordinates. 15. Provide instruction for subordinates. 16. Communicate effectively with superiors. 17. Develop a fami l iar i ty with the audit f i l e . 18. Try to complete the job in a timely manner. 19. Create meaningful learning experiences for subordinates. 20. Abide by professional code of ethics. 21. Manage subordinates. 22. Ensure the financial statements are error free. 23. Reconcile technical differences with c l ient . Highest Rank Next Highest Identi f ication # of items the firm would want their managers to comply with at the EVALUATION STAGE of the audit. Lowest Rank Appendix IV (continued) COMPLETING THE REST OF THE QUESTIONNAIRE 95. The next three rankings you wi l l be asked to provide wi l l indicate what managers actually do at a given stage of the audit. To provide these rankings i t w i l l be necessary to think about just what i t is that managers l ike you do at each stage of the audit. THE BUS DRIVER EXAMPLE AGAIN I f we consider the bus driver example again but this time think about what the driver actually does on the job then I would rank the six items l isted as indicated in the chart. Items with a high rank assigned to them are items that I think bus drivers always do on the job. Items with lower ranks assigned to them are not always done on the job. Task Requirements a) Drive within speed l imits b) Try to meet timetables c) Distribute transfers quickly d) Display courtesy to passengers e) Drive safely f) Announce street names Chart to aid Ranking: Highest Rank Next Highest Lowest Rank NOTE: I have found that bus drivers always drive within speed l imits hence item 'a ' is in the top most box. The next 2 items are 'c 1 and ' f . They are ranked equally. Of the 6 items l i s ted , I f ind that bus drivers t ry to meet timetables least frequently; hence item 'b' is in the bottom most box. Appendix IV (continued) 96. THE WORK OF THE AUDIT MANAGER AT THE DESIGN STAGE Rankings Based on What Managers Actually Do at the Design Stage of the Audit Please rank the items l isted below so that: Items with a HIGH RANK assigned to them are items which managers ALWAYS DO at the DESIGN STAGE. Items with LOWER RANKS assigned to them are NOT\" ALWAYS DONE by managers at this stage. 1. Try to u t i l i ze available personnel effectively. 2. Inform the cl ient about the execution of the audit. 3. Try to complete the job cost effectively. 4. Develop a fami l iar i ty with' the audit f i l e . 5. Display loyalty to the f irm. 6. Communicate effectively with superiors. 7. Provide motivation for subordinates. 8. Communicate effectively • with cl ients. 9. Provide supervision for subordinates. 10. Try to meet c r i t i ca l deadlines. 11. Try to complete the job in a timely manner. 12. Manage subordinates. 13. Develop team relationships. 14. Display an ab i l i t y to identify problems. 15. Provide instruction for subordinates. 16. Create meaningful learning experiences for subordinates. 17. Accept responsibil ity for the work of others. 18. Impose real is t ic requirements on subordinates. 19. Provide technical assistance to subordinates. Highest Rank Next Highest Identi f ication # of items ' managers actually do at the DESIGN STAGE of the \\ audit. Lowest Rank Appendix IV (continued) 97. THE WORK OF THE AUDIT MANAGER AT THE FIELDWORK STAGE Rankings Based on What Managers Actually Do at the Fieldwork Stage of the Audit Please rank the items l isted below so that: Items with a HIGH RANK assigned to them are items which managers ALWAYS DO at the FIELDWORK STAGE. Items with LOWER RANKS assigned to them are NOT ALWAYS DONE by managers at this stage. 1. Provide technical assistance to subordinates. 2. Try to meet c r i t i ca l deadlines. 3. ;. Maintain professional appearances. 4. Provide instruction for subordinates. 5. Develop technical sk i l l s in specialist area. 6. Display loyalty to the f i rm. 7. Reconcile technical differences with c l ient . 8. Develop a fami l iar i ty with the audit f i l e . 9. Try to complete the job cost effectively. 10. Ensure the financial statements are error free. 11. Create meaningful learning experiences for subordinates. 12. Display an ab i l i t y to identify problems. 13. Manage subordinates. 14. Try to u t i l i ze available personnel effect ively. 15. Try to complete the job in a timely manner. 16. Communicate effectively with cl ients. 17. ' Abide by ;the professional code of ethics. 18. Provide supervision for subordinates. 19. Inform the cl ient about the execution of the audit. 20. Communicate effectively with superiors. 21. Provide motivation for subordinates. 22. Accept responsibil ity for the work of others. 23. Impose real is t ic requirements on subordinates. 24. Develop team relationships. Highest Rank Next Highest Identif ication # of items managers actually do at the FIELDWORK STAGE of the audit. Lowest Rank Appendix IV (continued) 98. THE WORK OF THE AUDIT MANAGER AT THE EVALUATION STAGE Rankings Based on What Managers Actually Do at the Evaluation Stage of the Audit Please rank the items l isted below so that: Items with a HIGH RANK assigned to them are items which managers ALWAYS DO at the EVALUATION STAGE. Items with LOWER RANKS assigned to them are NOT ALWAYS DONE by managers at this stage. 1. Accept the responsibil ity for the work of others. 2. Try to complete the job cost effect ively. 3. Display loyalty to the f irm. 4. Develop technical sk i l ls in a specialist area. 5. Provide management type advice and services. 6. Display an ab i l i t y toi identify problems. 7. Provide supervision for subordinates. 8. Complete the management le t ter . 9. Sell additional services to existing cl ients. 10. Provide motivation for subordinates. 11. Try to meet c r i t i ca l deadlines. 12. Communicate effectively with cl ients. 13. Adapt to superiors. 14. Provide technical assistance to subordinates. 15. Provide instruction for subordinates. 16. Communicate effectively with superiors. 17. Develop a fami l iar i ty with the audit f i l e . 18. Try to complete the job in a timely manner. 19. Create meaningful learning experiences for subordinates. 20. Abide by professional code of ethics. 21. Manage subordinates. 22. Ensure the financial statements are error free. 23. Reconcile technical differences with c l ient . Highest Rank Next Highest Identif ication # of items managers actually do at the EVALUATION STAGE of the audit. Lowest Rank Appendix IV (continued) 99. A FEW CONCLUDING QUESTIONS Please consider each of the following problems and respond in the space provided. The problems are of a general nature and do not necessarily relate to the Manufacturing Co. referred to above. 1. The manager on the job feels somewhat uneasy about the c l ient 's state of affairs but cannot pinpoint any specific problems. The audit has gone well - a l l tests have checked out, a l l deadlines have been met, and the audit is well within budget. The cl ient is not aware of the manager's apprehension. PLEASE INDICATE HOW INCLINED YOU ARE AS A MANAGER TO SPEND RESOURCES IN SEARCH OF A PROBLEM WHICH, MAY AFTER ALL, NOT EXIST? Please put a dash (/) on the l ine. Commit Do Not Commit Resources Resources 2. You are privy to information from'.a very good friend who works for the QWER Co. Your friend is very reliable and is an accountant by pro-fession. He has often told you that the integri ty of QWER's management is subject to question. Just the other day QWER's management had a quarrel with their auditors who were apparently not ' f lex ib le ' enough. You are advised by your friend that QWER is looking for a change of audi-tors but he cautioned you, at the same time, not to get involved. QWER is a moderate sized company involved in importing a wide range of commodi-t ies. PLEASE INDICATE HOW INCLINED YOU ARE TO PUT THIS CLIENT UP FOR CONSIDER-ATION BY YOUR FIRM? Recommend Do Not Recommend QWER QWER 3. The senior on the job has come into conf l ic t with the c l ient 's accountant regarding the scope of the audit. The accountant is a d i f f i -cult person and problems are always experienced on this job. This year top management has threatened to change auditors i f any problems arise. The senior has told the accountant that certain preparatory work must be completed by the c l ient . This is firm policy. The accountant refuses to do the work and has threatened to.take the problem to top management. There is not a lot of work involved but the senior does not want to establish an undesirable precedent. The audit manager has been called in. AS THE AUDIT MANAGER CALLED IN, WOULD YOU INSIST ON THE COMPLETION OF THE PREPARATORY WORK BY THE CLIENT OR WOULD YOU ATTEMPT TO NEGOTIATE A COMPROMISE SOLUTION? Insist on Negotiate a Firm Policy Compromise Solution Appendix IV (continued) 100. BIOGRAPHICAL QUESTIONS 1. How many years have you been a manager? 2. When did you receive your CA? 3. How long have you been employed in this f i rm: 4. In f ive years do you expect to * Be working for this firm * Be working for another CA firm * Be working for a non CA firm (Please t ick the most appropriate response) 5. I f you were offered the opportunity of winning $500 on the toss of a fa i r coin, how much would you be prepared to pay to par t ic i -pate in the gamble, knowing that i f the coin goes against you, you lose what you pay? Please add any comments, suggestions or questions that you think may be helpful in interpreting your answers to the questions asked. Many Thanks For Your Help 101. Chapter 4 THE COLLECTION AND ANALYSIS OF THE Q - SORT DATA A. Data Collection In order to test the hypotheses enumerated above, the major Eastern Canadian offices of four large public accounting firms were invited to participate in the study. The firms were not selected randomly, but were i n i t i a l l y chosen because they had a large pool of audit partners and managers who could be called upon to participate in the study. Each firm agreed to participate in the study after i n i t i a l l y discussing the research with the researcher and satisfying themselves that the research procedures were meaningful. Each firm was asked to distribute the research material to 20 partners and 20 managers in their f i rm. A total of 160 questionnaires were d i s t r i -buted. One person in each firm acted as an agent for the researcher and distributed and collected the research material. Responses were anony-mous and were sealed by respondents upon completion in an envelope dis-tributed with the research material. Overall, 127 usable questionnaires were returned. This represents a response rate of over 79% which is high for studies of this nature. The distr ibution and response rate from each firm are shown in Table IV below. B. Stat ist ical Analysis Hypotheses Cl and C2 focus on the degree of agreement which exists within partner and manager groups in a given f i rm. Specif ically, these hypotheses investigate whether amongst members of each of these groups there is a high degree of agreement about what i t is that managers should 102. TABLE IV PATTERN OF RESPONSES TO THE Q-SORT STUDY Firm A Firm B Firm C Firm D Total Partners 17 13 10 20 60 Managers 16 14 17 20 67 Total 33 27 27 40 127 Response Rate 82Jg% 67%% 67%% 100% 79.4% be doing at each stage of the audit i f their behaviour is to be consistent with firm policy. Tests of these hypotheses wi l l make use of Kendall's coefficient of concordance which is designed for use with rank data such as that produced in this study (Siegel, 1956). The coefficient of concordance (W) provides a measure of the commun-a l i t y of the rank assignments in the Q-sorts across subjects. I t is an overall measure of the association between the n items in the Q-sort. I f the rankings are independent, there is no association and W is zero. For complete dependence there is perfect agreement and W is equal to one. The W sta t is t ic may therefore be used to test the null hypothesis that m sets of rankings are independent. In the null case, the ranks assigned to the n Q-sort items are completely random for each of the m sets of rankings. The concordance among m sets of rankings may be described by calculat-ing the Spearman-rank order correlation coefficients (designated r 1 among al l possible pairs of rankings and finding the average value (designated r ) . This la t ter value is l inearly related to the coefficient 103. of concordance (Ferguson, 1959, pp. 186-188). Appendix V at the end of this chapter presents the results which pertain to Hypothesis Cl. Amongst partners in each participant f i rm, the1rankings provided at each stage of the audit were not independent. The null hypothesis that the ranks assigned to the Q-sort items by partner groups at each stage of the audit are completely random was rejected. Each coefficient of concordance in Appendix V was signif icant at the .001 level. Overall, partners in each firm agreed as to what they would want their managers to do at a given stage of the audit. The value of the concordance coefficient is somewhat d i f f i c u l t to interpret direct ly. In order to fac i l i ta te interpretation, the average value of r g over a l l possible pairs of rank orders on which each W coefficient in Appendix V is based, was computed. The average Spearman coefficients (r ) are presented in Appendix V. The average Spearman coefficients were consistently higher at the evaluation stage than they were at either the design or fieldwork stages. At the design stage r g ranged from .183 to .276, at the fieldwork stage r g ranged from .230 to .304 and at the evaluation stage r g ranged from .329 to .496. At each stage of the audit the average Spearman correlation coeffic-ient for al l possible pairwise correlations between the rank assignments of partners from different firms is also presented. The average correla-tions within each firm are substantially the same as those derived from the rankings provided by respondents across firms and the firm does not appear to be an important explanatory variable in the interpretation of the Q-sort data. Whilst the composite rankings of the Q-sort data at each stage of 104. the audit from partner groups in each firm was not a result of the random ordering of the Q-sort items, the average Spearman correlations were not always high, especially at the design stage. Some attention was directed to the variation in partners' rank assignments with respect to individual Q-sort items at each stage of the audit in order to investigate whether disagreements about the rank assignments between partners could be accounted for in terms of disagreements with respect to the rank assign-ment of just a few Q-sort items. Because the average between and within firm Spearman correlations was substantially the same, the rank orderings provided by partners were pooled and the variance in the rank assigned to each Q-sort item was calculated. Q-sort items for which a variance greater than or equal to two was computed, were ident i f ied. Generally, the Q-sort items with respect to which high variations were observed in partners' rank assignments were items which were not subsequently used for the test of the substantive hypotheses. However, at the design stage there was a wide variation in the rank assigned to the Q-sort item 'Try to meet c r i t i ca l deadlines'which is used for the test of the resource allocative hypothesis at that stage. As a result , particular care wi l l be taken in the analysis of the resource allocation hypothesis at the design stage of the audit. Partners' rankings varied widely with respect to the cue 'Develop a fami l iar i ty with the audit f i l e ' at each of the three audit stages. At the design stage the ranks assigned to the cue 'Inform the cl ient about the execution of the audit' varied widely and at the fieldwork stage there was a wide variation in the rank assigned to the cues 'Ensure the financial statements are error free' and 'Manage subordinates'. The cue 105. 'Abide by the professional code of ethics 1 was noted at both the fieldwork and evaluation stages as a cue with a high variation in the rank assign-ment by partners. Disagreements about the rank to be assigned to a given Q-sort item may indicate that a particular cue is inappropriate at a given stage of the audit or poorly defined in the context of that audit stage. Alterna-t ive ly , variations in the ranks assigned to a given cue may indicate that the pr ior i ty that the firm would want their managers to assign to that cue is not well explicated. Some of the cues enumerated above which dis-played a high variation were cues about which respondents raised cer-tain questions at the end of the research material. The cue 'Ensure the financial statements are error free 1 was questioned by a number of respond-ents who sought c lar i f icat ion of the notion of error free. Other cues (e.g. 'Abide by professional code of ethics') were f e l t by respondents to pervade the.entire process but were not subject to more or less con-formity at any particular stage. Even though extensive pretesting pro-cedures were carried out in the development of the Q-sort material and great care was taken to ensure the face val id i ty of the Q-sorts used at each stage of the audit, i t is possible some ambiguities s t i l l remained, and to some extent these ambiguities were reflected in the disagreement with respect to the ranking of some Q-sort items. Some solace can be taken in the fact that generally high variance items were not subject to further analysis. Furthermore, the signif icant W stat is t ics obtained within each firm at each audit stage indicate that the rankings provided by the partner groups in each firm were not independent. The results which pertain to the second contextual hypothesis are 106. presented in Appendix VI. The composite rankings at each stage of the audit are not a result of the random ordering of the Q-sort items by managers. The null hypothesis that the rank ordering of the Q-sort data by respondents was independent was rejected. Each W coefficient was signif icant at the .001 level. Once again, the average Spearman rank order correlation coefficient associated with each W coefficient is pre-sented. At the design stage, the average Spearman correlation coefficients ranged from .177 to .363, at the fieldwork stage the average Spearman coefficients ranged from .237 to .373 and at the evaluation stage the average Spearman coefficients ranged from .375 to .473. Amongst managers in each firm there was agreement as to the behaviour that they perceived their firm would want them to adopt at a given stage of the audit. A comparison of the between and within firm correlation coefficients indicated once again that they were substantially the same. An analysis of the variation in the ranks assigned to specific cues by managers was carried out by f i r s t pooling al l their responses. Cues in which there was a high degree of variation in the rank assignments were noted. These lat ter cues had variances greater than or equal to two in the rank assigned to those cues by managers responding to the questionnaire. The set of items about which managers did not agree in their rank assignments include the cues 'Display loyalty to the f i r m ' , 'Inform.the cl ient about the execution of the audi t ' , 'Develop a fami l iar i ty with the audit f i l e 1 and 'Try to meet c r i t i ca l deadlines' at the design stage and the cues 'Abide by the professional code of eth ics ' , 'Ensure the financial statements are error free' and 'Develop a fami l iar i ty with the audit f i l e ' at the f i e ld -work stage. There were nc items at the evaluation stage about which 107. managers' rankings displayed a high degree of variation in terms of the cr i ter ion above. However, those with the highest variance are the same as those which were identif ied at the design and fieldwork stages. The rank orderings provided by managers which are used for the test of the second contextual hypothesis are not subsequently used for the test of the substantive hypotheses in this study. Nonetheless, because disagree-ments in the rank assignments by partners and managers were noted with respect to a similar set of cues, i t does seem that with respect to those cues, the intent of the incentive system may not be well understood by organizational members. The f i r s t two contextual hypotheses suggest that within partner and manager groups there is a s ta t is t i ca l l y signif icant degree of agreement as to what managers should do at a given stage of the audit i f their be-haviour is to be consistent with f i rm policy. The f i r s t two hypotheses do not examine the agreement between partner and manager groups. This question is examined by the third context hypotheses. Hypothesis C3 investigates whether partner and manager groups agree as to what a manager should be doing at each stage of the audit i f his behaviour is consistent with firm policy. A high degree of between group agreement within a given firm would suggest that the intent of the incen-tive function is well understood by organizational members. To test this hypothesis, two templates are constructed at each stage of the audit; one to summarize the responses of al l partners and another to summarize the responses of a l l managers within a given f i rm. The v iab i l i t y of the tem-plates is enhanced by the support of the f i r s t two hypotheses which point to overall within group agreement as to the rank assignment for each Q-sort item at each stage of the audit. 108. The templates are formed on the basis of the average rank assigned to each cue by each subject (Bern and Funder, 1978; Block, 1961). Block has indicated that consensus scores derived this way are invariably highly rel iable. The kind of r e l i a b i l i t y to which he refers is the kind of correspondence to be expected when this consensus (or average) score is correlated with a consensus score derived from an equivalent set of judges (Block, 1961, p. 37). Hypothesis C3 is tested using Spearman's rank order correlation coeffic-ient (.r ). r is a suitable measure of correlation i f the variables are subject to ordinal measurement (Ferguson, 1959) and hence is appropriate in this study. In Appendix V I I , the detailed results which relate to the test of this hypothesis are presented. In each firm there is strong between group agreement as to just what managers ought to do at a given stage of the audit. The mean correlation coefficient between the two templates constructed in each firm at the design stage was .757, whilst the mean correlation coefficient at the fieldwork and evaluation stages was .691 and .855 respectively. The results presented in Appendix VII indicate that partners and managers are in substantial agreement as to what i t is that managers should be doing at a given stage of the audit. Support of this hypothesis suggests that i f discrepancies exist between the way managers typical ly respond to a given problem context and the way their partners would want them to respond to that context, such discrepancies are not a result of the poor explication of the intent of the incentive system in a given f i rm. Hypothesis C4 focuses upon the degree of agreement between managers within a given firm as to the sort of behaviour that they typical ly adopt 109. in a given problem context. This hypothesis is tested using Kendall's coefficient of concordance. The results of this hypothesis test are pre-sented in Appendix V I I I . Within the manager group in each firm there is substantial agreement as to the sort of behaviour that managers typical ly adopt at each stage of the audit. The coefficients of concordance pre-sented in Appendix VI I I are each s ta t is t i ca l l y signif icant at the .001 level. The rankings assigned to the set of managerial behaviours at each stage of the audit are not a random ordering of those behaviours by man-agers participating in this study. The Spearman coefficient associated with each W sta t is t ic in Appendix VI I I was computed. This data is also presented in that appendix. At the design stage the average r g over.all possible pairs of rank orders in each firm varied from .182 to .204, at the fieldwork stage the average r g •varied from .225 to .361:and at the evaluation stage the average r g varied from .414 to .477. Average Spearman correlation coefficients were computed for a l l possible pairwise rank orderings from managers from different firms and the average between firm correlation coefficients are also presented in Appendix V I I I . Within and between firm correlations were very similar and al l the rankings were pooled in order to investigate whether there are cues in the Q-sort with respect to which managers rankings varied widely in terms of the cr i ter ion identif ied above. At the design stage, there was a high variation in the rank assigned to the cues 'Try to meet c r i t i -cal deadlines', 'Inform the cl ient about the execution of the audit 1 and 'Develop a fami l iar i ty with the audit f i l e ' . At the fieldwork and evalua-tion stages, high variations were observed in the cues 'Abide by the pro-110. fessional code of ethics 1 and 'Develop a fami l iar i ty with the audit f i l e ' and at the fieldwork stage alone a wide variation was found to exist in the cue 'Ensure the financial statements are error f ree ' . These cues are the same as those identif ied previously. The only cue which had a high degree of variation in the rank assigned to i t by managers which wi l l be subject to further analysis is the cue 'Try to meet c r i t i ca l deadlines' and in the evaluation of the hypothesis to which that cue relates care wi l l be taken to assess the implications of the variation with respect to that cue for the hypothesis test. Hypotheses C5 and C6 focus upon the communality which exists in the rank assignments by partners and managers from different firms. Specifi-cal ly , Hypothesis C5 examines whether partner groups from different firms agree as to the way in which managers should respond to a given problem context. A high between firm agreement is hypothesized. Test of this hypothesis makes use of the templates constructed earl ier. The specific results are presented in Appendix IX. The coefficient of concordance computed at each stage of the audit was uniformly high, and each was signif icant at the .001 level. This result suggests that the intent of the incentive system is similarly perceived by partners in different firms. The average Spearman coefficients associated with each W sta t is t ic are presented in Appendix IX. These are uniformly high and range from .785 at the fieldwork stage' to .891 at the evaluation stage. The high across firm agreement between the rankings of partners in different firms suggests once again that the firm is not an interesting explanatory var i -able for the analysis of the data derived in this study. Hypothesis C6 examines whether manager groups from different firms I l l . share similar perceptions about the sort of behaviour they should conform to i f their behaviour is consistent with firm policy. The test of this hypothesis makes use of Kendall's coefficient of concordance. The test is based on the templates of managerial behaviour derived earl ier. Appendix X sets out the results that pertain to hypothesis C6. There is a high degree of agreement amongst managers from different firms as to the sort of behaviour that they perceive their firm would want them to adopt at a given stage of the audit. Average Spearman correlations associated with each W coefficient are presented in Appendix X. The average r g for al l possible pairwise combinations of the templates that are used to test Hypothesis C6 are very high, ranging from .824 at the fieldwork stage to .984 at the evalua-tion stage. The implications of the across firm consistency in respond-ents' rankings have already been alluded to. C. Discussion of the Contextual Hypotheses The six contextual hypotheses were each supported. A significant degree of agreement was found to exist between and among partner and manager groups as to the sort of behaviour that managers should adopt i f their behaviour is perceived to be consistent with firm policy. In addi-t ion, at each stage of the audit there was signif icant agreement amongst the managerial group as to the sort of behaviour they would typical ly adopt at that stage. Some analysis of the variation in the rank assignment by partners and managers was undertaken. In general, cues with respect to which part-ners and managers disagreed about the rank assignment were cues which 112. were not subject to further analysis in this study. However, at the design stage, the cue 'Try to meet c r i t i ca l deadlines' did have a high degree of variation associated with i ts rank assignment and this cue wi l l be used in subsequent analysis. Particular care wi l l be taken in the evaluation of the substantive hypothesis to which this cue relates. When the Q-sort data was compared across firms, there was substantial agreement amongst partner and manager groups as to what partners 'expected their managers to do at each stage. The data derived from the Q-sorts is not a mere random ordering of the Q-sort items by respondents in this study. Respondents understood the research context in which they were placed and were generally in agreement as to how they would respond in that context. In spite of the agreement observed between and among partner and manager groups, the substantive hypotheses posit systematic discrep-ancies between the rank assignment of partners and managers with respect to specific Q-sort items. Such discrepancies result from differences in the incentives confronting partner and manager groups, and not from a fai lure on the part of respondents in this study to understand the environ-mental context which they were asked to consider. D. Tests of the Substantive Hypotheses For the purpose of testing the substantive hypotheses elaborated above, a series of s tat is t ica l tests were applied to various subsets of the Q-sort items. The tests were carried out in order to investigate whether, with respect to specific managerial roles, systematic discrepancies exist between the rank assignment of partners and managers. The substan-tive hypotheses are tested, where applicable, at each of the three stages of the audit. 113. The substantive hypotheses propose relationships which are amenable to analysis of variance tests. The F test used in parametric analysis of variance requires the specification of a number of conditions about the parameters of the population from which the sample is drawn. In particular, i t is assumed that observations are sampled randomly and independently from normally distributed populations with the same variance (Lindman, 1974, p. 27). Lindman (1974, p. 31) has indicated that none of the assumptions neces-sary for an F test are ever' fu l l y satisfied by real data. However, the F test is generally regarded as robust with respect to departures from some of these assumptions. Hays (1963, p. 378) has indicated that infer- . ences about the population means that are valid in the case of normal populations are also valid even when the form of the population distr ibu-tions depart considerably from normal, provided that the number of obser-vations in each sample is relat ively large and Lindquist (1953, pp. 78-86) has indicated that the F distr ibution is \"amazingly insensitive\" to the form of the distr ibut ion,of the parent population. The F test is robust with respect to the homogeneity of variance assumption as long as the number of observations in each sample is the same (Lindman, 1974, p. 33). However, in this study there is some variation in the number of observations in each sample. Lindman (1974, pp. 43-45) has provided a rough index of the effect of unequal variances on the distr ibution of F which is independent of the val id i ty of the null hypothesis i t se l f . The index (x) is based on the weighted and unweighted average of the individual group variances. Lindman has indicated that, in general, i f x is close to one, the F test is not greatly affected by 114. the effects of possibly unequal variances. In this study there are three substantive hypotheses amenable to analysis of variance tests. The three hypotheses are tested, where applicable, at each stage of the audit. In to ta l , there were eight applications of the analysis of variance test pro-cedure. .For each test the x coefficient was computed, x was always close to one. I t ranged in size from 1.02 to 1.11; the mean value over the eight scores was 1.066. In situations such as th is , the F computed using analysis of variance procedures is l ike ly to be s l ight ly larger than i t should be and hence the obtained level of significance somewhat greater than i t should be. The third assumption that is necessary to conduct an.F test used in parametric analysis of variance requires that the data be based on inde-pendent observations both within and across groups. This assumption is not supported when observations are derived using a ranking procedure such as the Q-sort. Kerlinger (1972, pp. 13-14; 1973, p. 595) questions how serious the violation of the independence assumption really is . While i t is clear that a Q-sort with 20 items does not have 19 degrees of free-dom associated with i t , and thus the analysis of variance procedure is v i t ia ted, the number of potential ranking combinations is enormous. A rank order scale with only ten items has 10! or over 3 mil l ion possible rankings (Kerlinger, 1972, p. 14). Kerlinger concludes that whilst i t is true that \"each time a Q item is placed by a subject, a degree of freedom is lost , and thus the independence assumption violated, the possibi l i t ies are so great that i t probably doesn't matter that much\" (Kerlinger, 1972, p. 14). The problems:associated with a lack of independence can be. mitigated 115. further by stressing to subjects that they can make any changes they wish to their Q-sorts. For this purpose, respondents were asked to complete the Q-sort using a lead pencil as this would fac i l i ta te any changes they may wish to make. Even though the Q-sort data obtained in this study does not completely satisfy the assumptions necessary to apply parametric analysis of variance test procedures, the departures from the assumptions noted above were not seen to invalidate the use of such tests in this study. Nonetheless, particular care is exercised in the interpretation of the test results. Hypotheses S1-S3 are analyzed using a two-way fixed effects analysis of variance design at each stage of the audit. Such a design w i l l enable the simultaneous investigation of partner-manager (role) and firm effects. However, only role effects have been hypothesized. In the context of this study, the firm is used as a stat is t ica l control. Hypothesis SI focuses on the resource allocative act iv i t ies of the audit manager. Because such act iv i t ies can be readily monitored within audit firms, i t is hypothesized that managers wi l l give a higher pr ior i ty to conformance to such act iv i t ies than their partners would want them to. Q-sort items ranked highly are items that managers always conform to at a given stage of the audit, or which partners would want their managers to conform to at that stage. Hypothesis SI would be supported i f the rank assigned to resource allocative cues by managers is signif icant ly higher than that assigned to those cues by their partners. In Appendix XI , the results of the partners'and managers' rank assignments are summarized on a firm by firm basis. Appendix XII details the analysis of variance results which pertain to that data. In the Table below the summary F 116. stat is t ics which describe differences * in the rank assignments by partners and managers are presented. TABLE V Partner-Manager Differences in the Rank Assignment of Resource Allocative Cues Design Stage Fieldwork Stage Evaluation Stage F = 6.17 F = 2.34 F = 16.55 p < .025 p < .25 p < .0005 Hypothesis SI is strongly supported at the design and evaluation stages of the audit but only very weakly supported at the fieldwork stage. Because the data on which the test of Hypothesis SI was based did not entirely satisfy the assumptions which underlie the use of the F s ta t i s t i c , one can only confidently assert that this hypothesis was supported in two out of the three test applications. Nonetheless, i t does appear that managers give a higher pr io r i ty to conformance to resource allocative act iv i t ies than their partners would want them to do. In the preceding discussion, attention was directed to the variation in partners' and managers' rank assignments with respect to the cue 'Try to meet c r i t i ca l deadlines'. Such variations were identif ied in rank assignments at the design stage of the audit. Test of Hypothesis SI above includes that high variance cue in the defini t ion of the resource allocative con-struct at the design stage. The analysis of variance test procedure was conducted with the high variance cue specif ical ly excluded when the test of Hypothesis SI was conducted at the design stage. Significant d i f fe r -117. ences in partners' and managers' rank assignments were s t i l l observed, how-ever, the significance of the test s ta t is t ic was s l ight ly reduced (F = 3.69 p < 0.10). Hypothesis S2 focuses on the role of a manager as a negotiator within the f i rm. Like resource allocative ac t iv i t ies , the negotiation act iv i t ies of the manager are relat ively easy to monitor within the organization and as a result i t is hypothesized that managers wi l l give a higher pr ior i ty to conformance to those act iv i t ies than their partners would want them to. The rank assigned to the cue which describes the manager's negotia-tion act iv i t ies w i l l , therefore, be higher when that assignment is by managers than by partners. In Appendix XI I I the results of the partners' and managers' rank assignments are summarized on a firm by firm basis and in Appendix XIV the analysis of variance results which pertain to that data are presented. Below a summary of the F stat is t ics which relate to differences in partners' and managers' rank assignments are presented. TABLE VI Partner-Manager Differences in the Rank Assignment of the Negotiation Cue Fieldwork Stage Evaluation Stage F = 4.39 F = 0.11 p < .05 NS Hypothesis S2 is only supported at the fieldwork stage of the audit. At the evaluation stage no signif icant differences were found to exist in 118. the rank assignment of the negotiation cue by partners and managers. None-theless, differences which did exist were in the hypothesized direction. At the evaluation stage the decisional act iv i t ies of the audit manager which relate to his role as a negotiator within the audit firm become so c r i t i ca l that partners and managers see the execution of these as mandatory to the completion of the audit. Hence there was l i t t l e variation in the rank assignment of the negotiation cue by the two groups of respondents who participated in the study. Hypothesis S3 focuses on the role of the manager as an entrepreneur within the organization. The entrepreneurial act iv i t ies of the manager are more d i f f i c u l t to monitor than either the negotiation or resource alloca-tive act iv i t ies of the manager and hence i t is hypothesized that managers wi l l give a lower pr ior i ty to conformance to entrepreneurial responsibil-i t ies than their partners would want them to. This hypothesis was tested at the three stages of the audit. Detailed results of the rank assign-ment by partners and managers are presented\"in Appendix XV.and the analysis of variance results which pertain to that data are presented in Appendix XVI. A summary of the relevant F stat is t ics is set out below. TABLE VII Partner-Manager Differences in the Rank Assignment of the Entrepreneurial Cues Design Stage Fieldwork Stage Evaluation Stage F = 21.72 F = 6.38 F = 0.85 p < .0005 p < .025 p < .5:' 119. Table VII indicates that Hypothesis S3 is strongly supported at the design and fieldwork stages of the audit, but that this hypothesis does not seem to be supported at the evaluation stage. The differences between partners' and managers' rank assignments at the evaluation stage were, however, in the hypothesized direction. E. Discussion of the Substantive Hypotheses The substantive hypotheses were not unambiguously supported though there was substantial support for them. The hypothesis dealing with the manager's resource allocative ac t i v i -t ies was strongly supported at the design and evaluation stages, the negotiation hypothesis was strongly supported at the fieldwork stage, and the hypothesis dealing with the manager's entrepreneurial act iv i t ies was strongly supported at the design and fieldwork stages. The results presented seem to suggest that there are systematic discrepancies between the way in which managers typical ly respond to a given problem context and the way in which their partners would want them to respond to that context. Such discrepancies are explicable in terms of the economic incentives confronting audit, personnel and the assumption that audit managers are resourceful, evaluative maximizing men. Support for the substantive hypotheses was weakest at the evaluation stage where signif icant differences between the rank assignment of partners and managers were found to exist only with respect to the resource allocative cues in the Q-sort. Whilst differences in the rank assignment by partners and managers with respect to entrepreneurial and negotiation cues were always in the hypothesized direction at the evaluation stage, such d i f fer -ences were not s ta t i s t i ca l l y signif icant. I t seems that at the evaluation 120. stage of the audit, managers have less discretionary decision making authority than they do at the design and fieldwork stages. The results derived in this study support the basic hypothesis that was subjected to investigation. Relative to what their partners would want them to do, managers give a higher pr io r i ty to conformance to those organizational rules which can be readily monitored by their superiors. Conversely, managers give a lower pr ior i ty to conformance to those rules which cannot be readily monitored or which can only be monitored at a prohibitive cost. The rudiments of a positive theory of decision making in auditing has been provided. The decisions taken by managers on the job are explicable in terms of the economic incentives confronting the audit manager and the assumption that the manager is an evaluative, resourceful, maximizing individual. 1 2 1 . APPENDIX V Test of Hypothesis Cl Communality of Partners' Rank Assignments* Design Stage Fieldwork Stage Evaluation Stage Firm A** W=.231 W=.323 W=.503 X2=70.54 X2=l18.86 X2=177.06 m=17 p < .001 p < .001 p < .001 r =.183 s r =.278 s r =.470 s Firm B W=.332 W=.289 W=.381 X2=77.69 X2=86.41 X2=108.97 m=13 p < .001 p < .001 p < .001 rs=.276 rs=.230 r =.329 s Firm C W=.340 W=.374 W=.546 X2=61.11 X2=86.00 X2=120.01 m=10 p < .001 p < .001 p < .001 rs=.267 rs=.304 r = 496 s Firm D W=.248 W=.307 W=.437 X2=89.12 X2=141.17 X2=192.28 m=20 p < .001 p < .001 p < .001 rs=.208 r =.271 s r =.407 s Average Spearman Correlation Coeffic-ient for al l possible pairwise correlations for partners from different firms. * Stat ist ical tests based on Ferguson, 1959, pp. 186-189, p. 309. ** In Firm A one partner did not complete the rankings at the fieldwork and evaluation stages and hence the coefficient of concordance computed at each of those stages is based on 16 sets of rankings. 122. APPENDIX VI Test of Hypothesis C2 Communality of Managers' Rank Assignments* Design Stage Fieldwork Stage Evaluation Stage Firm A W=.232 W=.329 W=.508 X2=62.69 X2=l13.44 X2=167.64 m=15 p < .001 p < .001 p < .001 r =.177 s r =.281 s r =.473 s Firm B** W=.281 W=.353 W=.494 X2=65.75 X2=97.43 X2=141.28 m = 13 p < .001 p «.001 p < .001 r =.221 s r =.294 s r =.452 s Firm C W=.400 W=.410 W=.412 X2=122.37 X2=160.31 X2=154.18 m=17 p < .001 p < .001 p < .001 r =.363 s r =.373 s r =.375 s Firm D W=.219 W=.275 W=.455 X2=78.76 X2=126.50 X2=200.11 m=20 p < .001 p < .001 p < .001 r =.178 s r =.237 s rs=.426 Average Spearman Correlation Coeffic-ient, for al l possible 246 pairwise correlations for partners from different firms. .296 .430 * Stat ist ical tests based on Ferguson, 1959, pp. 186-•189, p. 309. ** In Firm B one manager did not complete the rankings at the fieldwork stage and hence the.coefficient of concordance computed at that stage is based on 12 sets of rankings. 123. APPENDIX VII Test of Hypothesis C3 Rank Order Correlation Coefficients ( r s )* Correlation Between Templates Derived for Partner and For Manager Groups When Those Templates Describe What Managers Should Be Doing in a Given Problem Context Design Stage Fieldwork Stage Evaluation Stage Firm A r=-= .843 r = .738 r = .832 s s s p < .001 p < .001 p < .001 ' Firm B r s = .804 p < .001 r $ = .695 p < .001 r = .880 s p < .001 Firm C r s = .760 p < .001 r s = .803 p <- .001 r s = .856 p < .001 Firm D r s = .620 p < .01 r s = .527 p < .01 r = .851 s p < .001 * Stat ist ical tests are based on Ferguson, 1959, pp. 179-182, p. 308; Roscoe, 1975, pp. 106-110. Significance levels stated are al l based on two-tailed tests. 124. APPENDIX VII I Test of Hypothesis C4 Communality of Managers' Rank Assignments* Design Stage Fieldwork Stage Evaluation Stage Firm A** W=.233 W=.316 W=.512 X2=67.10 X2=116.29 X2=168.96 m=16 p < .001 p < .001 p < .001 r =.182 s rs=.270 r =.477 s Firm B W=.265 W=.410 W=.463 X2=62.01 X2=122.59 X2=132.42 m=13 p < .001 p < .001 p < .001 r =.204 s V . 3 6 1 r =.418 s Firm C W=.251 W=.368 W=.457 X2=76.77 X2=143.97 X2=170.96 m=17 p < .001 p < .001 p < .001 r =.204 s rs=.329 r =.423 s Firm D W=.244 W=.264 W=.444 X2=87.92 X2=121.30 X2=195.18 m=20 p «.001 p < .001 p < .001 r =.204 s r =.225 s r =.414 s Average Spear-man Correlation Coefficient for al l possible 2Q3 pairwise corre-lations, for managers from different f i rms. .303 .425 * Stat ist ical tests based on Ferguson, 1959, pp. 186-189, p. 309. In Firm A one manager did not complete the rankings at the evaluation stage and hence the coefficient of concordance computed at that stage is based on 1:5 sets of rankings. 125. APPENDIX IX Test of Hypothesis C5 Communality of Rank Assignments Across Partner Groups* DESIGN STAGE: W = .863 X2 = 62.15 p < .001 r = .817 s FIELDWORK STAGE: W = .839 X2 = 77.19 p < .001 r = .785 s EVALUATION STAGE: W = .918 X2 = 80.78 p < .001 r c = .891 s * Stat ist ical tests are based on Ferguson, 1959, pp. 186-189, p. 309. 126. APPENDIX X. Test of Hypothesis C6 Communality of Rank Assignments Across Manager Groups* DESIGN STAGE: W = .915 X2 = 65.88 p < .001 F = .887 s FIELDWORK STAGE: W = .868 X2 = 79.86 p < .001 r s = .824 EVALUATION STAGE: W = .988 X2 = 86.94 p < .001 r„ = .984 *Statist ical tests are based on Ferguson, 1959, pp. 186-189, p. 309. 127. APPENDIX XI Test of Hypothesis SI . The Manager as a Resource Allocator. Results by Role and Firm DESIGN STAGE: Number of Subjects FIRM Mean Responses* FIRM A B C D A B C D Partners 17 13 10 20 15 .35 15.08 15.90 15 .85 15 55 Managers 16 13 17 20 13 .69 13.54 . 14.00 14 .80 14 01 33 26 27 40 14 .52 14.31 19.95 15 .33 14 78 FIELDWORK STAGE: Number of Subjects FIRM Partners Managers A B C D 16 13 10 20 16 13 17 20 32 26 27 40 i Mean Responses* FIRM B C 13. 31 13. 54 15. 20 14. 85 14 23 13. 38 12. 92 14. 00 13. 40 13 43 13. 35 13. 23 14. 60 14. 13 13. 83 EVALUATION STAGE: Number of Subjects FIRM Partners Managers A B C D 16 13 10 20 15 13 17 20 31 26 27 40 A Mean Responses* FIRM B C 10.69 11.23 9.90 10. 60 9.13 9.46 9.41 8. 95 9.91 10.35 9.66 9. 78 10.61 9.24 9.93 * Values were assigned to ranks as follows: 1 to the highest ranked item, 2 to the next highest and so forth t i l l 7 was assigned to the lowest ranked item. 128. APPENDIX XII Test of Hypothesis SI Analysis of Variance of Q-sort Data*. DESIGN STAGE: Component Sum of Squares Degrees of Freedom Mean Square F Significance Between Groups: Role 70.96 1 70.96 6.17 p < .025 Firm 19.12 3 6.37 0.55 NS R x F 2.85 92.93 3 7 0.95 0.08 NS Within Groups: 1358.13 1451.06 118 125 11.51 FIELDWORK STAGE: Component Sum of Squares Degrees of Freedom Mean Square F Significance Between Groups: Role 19.05 1 19.05 2.34 p < .25 Firm 37.95 3 12.65 1.55 NS R X F 10.20 67.20 3. 7 3.40 0.42 NS Within Groups: 954.29 1021.49 117 124 8.15 EVALUATION STAGE: Component Sum of Squares Degrees of Freedom Mean Square F Significance Between Groups: Role 54.96 1 54.96 16.55 p < .0005 Firm 7.78 3 2.59 0.78 NS R x F 7.73 70.47 3 7 2.58 0.78 NS Within Groups: 385.48 455.95 116 123 3.32 * Stat ist ical tests are based on Lindman (1974, pp. 100-102) and make use of an harmonic mean approximation to correct for the unequal cell sizes. 129. APPENDIX XII I Test of Hypothesis S2 The Manager as a Negotiator. Results by Role and Firm FIELDWORK STAGE: Partners Managers Number of Subjects FIRM Mean Responses* FIRM A B C D A B C D 16 13 10 20 3.;44 3.62 3.:30 4.00 16 13 17 20 3.56 2.62 3.06 3.25 32 26 27 40 3.50 3.12 3.18 3.63 3.59 3.12 3.36 EVALUATION STAGE: Partners Managers Number of Subjects FIRM Mean Responses* FIRM A B C D A B C D 16 13 10 20 2.69 2.69 2.90 2.40 15 13 17 20 2.40 2.46 2.71 2.70 31 26 27 40 2.55 2.58 2.81 2.55 2.67 2.57 2.62 * Values were assigned to ranks as follows: 1 to the highest ranked item, 2 to the next highest and so forth t i l l 7 was assigned to the lowest ranked item. 130. APPENDIX XIV Test of Hypothesis S2 Analysis of Variance of Q-Sort Data*. The Manager as a Negotiator FIELDWORK STAGE: Component Between Groups: Role Firm R x F Within Groups: Sum of Squares 6.49 7.35 4.60 18.44 172.83 191.27 Degrees of Mean Freedom Square 1 3 3 7 117 124 6.49 2.45 1.53 1.48 f_ Significance 4.39 1.66 1.04 p < .05 p < .25 NS EVALUATION STAGE: Sum of Degrees of Mean Component Squares Freedom Square F Significance Between Groups: Role 0.09 1 0.09 0.11 . NS Firm 1.16 3 0.39 .46 NS R x F 1.85 3 0.62 .73 NS 3.10 7 Within Groups: 98.47 116 0.85 101.57 123 * Stat ist ical tests are based on' Lindman (1974, pp. 100-102) and make use of an harmonic mean approximation to correct for the unequal cell sizes. APPENDIX XV Test of Hypothesis S3 The Manager as an Entrepreneur. Results by Role and Firm DESIGN STAGE: Number of Subjects FIRM Mean Responses* FIRM A B C D A B C D Partners 17 13 10 20 2.06 1.69 1.80 2.35 Managers 16 13 17 20 3.31 3.31 • ; 2.71 2.85 33 26 27 40 2.69 2.50 2.26 2.60 1.98 3.04 2.51 FIELDWORK STAGE: Number of Subjects FIRM Mean Responses'* FIRM Partners Managers A B C D A B C D 16 13 10 20 9.00 8.23 7.70 8.40 16 13 17 20 8.69 9.15 8.76 9.65 36 26 27 40 8.85 8.69 8.23 9.02 8.33 9.06 8.70 EVALUATION STAGE: Number, of Subjects FIRM Mean Responses^ FIRM A B C D A B C D Partners. 16 13 10 20 18 .06 17. 23 17.70 18. 15 17 79 15 13 17 • 20 17 .73 18. 46 18.06 18. 45 18 17 31 26 27 40 17 .90 17. 85 17.88 18. 30 17 98 * Values were assigned to ranks as follows: 1 to the highest ranked\"item, 2 to the next highest and so forth t i l l 7 was assigned to the lowest ranked item. 132. APPENDIX XVI Test of Hypothesis S3 Analysis of Variance of Q-Sort Data*. The Manager as an Entrepreneur DESIGN STAGE: Component Sum of Squares Degrees of Freedom Mean Square F Significance Between Groups: Role 34.31 1 34.31 21.72 p < .0005 Fi rm 3.12 3 1.04 .66 NS R x F 5.12 42.55 3 7 1.71 1.08 NS Within Groups: 186.15 228.70 118 125 1.58 FIELDWORK STAGE: Component Sum of Squares Degrees of Freedom Mean Square F Significance Between Groups: Role 15.82 1 15.82 6.38 p < .025 Firm 10.31 3 3.44 1.39 p < .25 R x F 11.18 37.31 3 7 3.73 1 .50 p < .25 Within Groups 289.95 327.26 117 124 2.48 EVALUATION STAGE Component Sum of Squares Degrees of Freedom* Mean Square f. Significance Between Groups: Role 4.44 1 4.44 0.85 NS Firm 4.02 3 1.34 0.26 NS R x F 9.15 17.61 3 7 .3.05 0.58 NS Within Groups: 605.96 623.57 116 123 5.22 * Stat ist ical tests are based on.Lindman (1974, pp. 100-102) and make use uf an harmonic mean approximation to correct for unequal cell sizes. ,133. Chapter 5 FURTHERING THE DEVELOPMENT OF POSITIVE THEORY IN AUDITING: SOME CONCLUDING REMARKS This study has investigated the work of audit managers in large public accounting firms and has addressed two fundamental questions: (1) What do audit managers do? (2) Why do audit managers do what they do on the job? In Chapter 2, a checklist of c r i t i ca l task requirements is provided. This checklist details the c r i t i ca l aspects of an audit managers' task. Chapter 3 drew on that l i s t to examine partners' and managers' perceptions about the nature of managerial work in alternate decision contexts. For the purpose of describing what audit managers do, Flanagan's c r i t i ca l incident methodology (1954) was adopted. The application of that methodology in public accounting firms resulted in a f a i r l y complete description of managerial work in those firms. The description derived was validated by comparing the checklist with specifications about the nature of managerial work made available by the participating, firms. In addition, a small group of audit partners and managers reviewed the l i s t for completeness. For the purpose of examining why managers do what they do on the job, attention was directed to the characteristics of the incentive system con-fronting non-partners in public accounting firms and an assessment was made as to the way in which that system is l ike ly to influence the decision based behaviour of audit managers. I t was hypothesized that, relative to what their partners would want them to do, managers wi l l give a higher pr ior i ty to conformance to those organizational rules which can be readily monitored by their superiors. Conversely, managers wi l l give a lower 134. pr ior i ty to conformance to those rules which cannot be readily monitored or which can only be monitored at a signif icant, often prohibitive cost. Tests of this basic hypothesis made use of a Q-sort methodology within a f ie ld setting. The hypothesis was operationalized in terms of three decisional roles of the audit manager. Substantial support for this hypo-thesis was found amongst partners and managers from the participant firms. Field studies, such as th is , are strong in realism, significance, strength of variables and heuristic qual i t ies, but weak in stat is t ical control (Kerlinger, 1973, p. 406). Statements of relationships are necessarily weaker than they are in experimental research. The discrep-ancies observed between the responses of partners and managers have been attributed to the nature of the incentive system confronting non-partners in public accounting firms, however, such discrepancies may be a t t r i -butable to phenomena not related to the incentive function. Such pheno-mena may include age, experience, r isk taking propensity, and so for th. To the extent that such variables are related to role differences in public accounting firms, such variables might also be related to the results derived in this study. To provide some evidence in this regard, the number of years since each respondent in this study attained his professional qualifications: was correlated (using Spearman's r ) with his rank assignment with respect to resource al locative, negotiation and entrepreneurial cues in the Q-sort. The correlational analysis was disappointing, few s ign i f i -cant relationships were obtained and many were not in the direction that would be expected in the l ight of the basic hypothesis. A table of the relevant correlational data is presented in Appendix XVII. These results 135. are consistent with existing research in the audit judgment area which points to the lack of pervasive influence of psychological and demographic variables on decision making in auditing ( c f . , Weber, 1978, p. 386; Corless, 1972). Few claims of generalizabil ity from the results of the hypothesis testing stage of the study can be made due to restrict ions of the sampling plan. However, steps were taken to ensure that the study is internally valid and therefore provides a basis for further analysis. The Q-sort was developed so that i t is meaningful to the audit personnel who were involved in i ts completion. The cues were couched in the auditor's language and at each stage of the audit steps were taken to ensure that the Q-sort items were descriptive of the work of the.audit manager at that stage. Instructions for the completion of the research instrument were subjected to extensive pre-testing during p i lo t studies involving both academics and auditors to further ensure that any reactive effects would be minimized. To ensure that the Q-sorting procedure was clearly understood by subjects, an i l lus t ra t ive example from a f ie ld un-related to auditing was provided in the instructions. Care was taken to ensure that the i l lus t ra t ive example did not bias the Q-sorts provided by respondents. Respondents were requested not to discuss the research with others and were guaranteed anonymity. To investigate the r e l i a b i l i t y of the research instrument, test-retest stat is t ics were computed. These stat is t ics indicate that there was a reasonable degree of consistency in the Q-sorts provided by the same respondents over time. 136. The strength of the results derived in this study suggest that i t may be f r u i t f u l to consider the implications of the basic hypothesis posed above over a broader class of organizational actors than audit managers alone and over a broader range of problems than the situational contexts subjected to investigation in this study. Preliminary steps were taken to move the research in this direction. After respondents had completed the Q-sorting tasks, they were asked to respond to a few supplementary questions. These questions are detailed in Appendix IV to Chapter 3. Specif ically, three decision contexts were posed. One in which the manager was asked to act as an entrepreneur, one in which he was asked to act as a negotiator, and one in which he was asked to act as a resource allocator within his f i rm. Managers were asked to provide an action rating with respect to each problem posed. Partners were asked to indicate how they would want their managers to respond to the same problem i f the manager was acting in the best interests of the f i rm. These decision problems were designed to provide supplementary evidence in support of the substantive hypotheses. Once again, the rela-tionships proposed were amenable to analysis of variance test procedures. When the parametric F test used in analysis of variance was computed for each decision problem, no s ta t is t i ca l l y signif icant results were obtained with respect to role, firm or interaction effects. Nonetheless, in the case of the resource allocative and entrepreneurial decision, the mean responses of partners and managers were in the hypothesized direction. Respondents were c r i t i ca l of these supplementary questions in the comments that they provided about the research procedures at the end of the booklet. They f e l t that, in general, there was insuff icient information 137. to reasonably respond to them. Unlike the overall results derived with respect to the Q-sort analysis, there was a huge variation in the responses to each of the decisional problems posed. Such variation is indicative of the d i f f i cu l t y respondents seemed to have in dealing with these lat ter problems. In order to test the implications of the basic hypothesis described in Chapter 3 over a broader class of problems than the situational contexts used with respect to the Q-sorts, great care must be exercised in the development of detailed case analyses, so that variations amongst respond-ents cannot be attributed to a fai lure on their part to understand either the research problem or the research context with which they are required to deal. The systematic discrepancies between'partners1 and managers' responses in Chapter 4 are al l the more signif icant because the respondents demonstrated that they were in overall agreement as to how they would respond to the problem contexts under investigation. The results derived in this study suggest that relative to what their partners would want them to do, audit managers give a higher pr ior i ty to conformance to those organizational rules which can be readily monitored by their superiors. Conversely, they give a lower pr io r i ty to confor-mance to those rules which cannot be readily monitored or which can only be monitored at a prohibitive cost. Behaviour of this nature is entirely rational for a subordinate who is subjected to a promotion system which emphasizes up or out selection. The results derived in this study are important for those concerned with judgmental research in auditing and for organizational and decision theorists on a more general level. A focus upon economic incentives and 138. the way in which those incentives may influence the characteristics of the decisions effected in large organizations may account for the var iab i l i ty observed amongst the responses of audit personnel who participated in previous empirical research in this area. Furthermore, for those concerned with organizational decision making, the results derived suggest that understanding the characteristics of the decisions effected in organizations may be fac i l i ta ted by understanding the nature of the economic incentives confronting organizational par t i c i -pants and assessing the way in which those incentives influence their task related and decisional behaviour. 139. APPENDIX XVII Correlation Between the Respondents' Years of Experience and Rank Assignment with Respect to Specific Q-sort Items Resource Allocation: Firm A Firm B Firm C Firm D rs--= -.022 r s = -0.272 r$ = -0.10 r g = -0.10 n = 33 n = 25 n = 27 n =40 NS p < .1 NS NS Fieldwork Stage r g = -0.18 r g = -0.02 r g = -0.05 r g = -0.21 n = 32 n = 25 n = 27 n = 40 NS NS NS p < .1 Evaluation Stage r g = -0.52 r g = -0.26 r g = 0.15 r g = -0.41 n = 31 n = 25 n = 27 n = 40 p < .005 NS NS p < .005 Negotiation: Fieldwork Stage r g = .075 r g = -0.37 r g = -0.11 r $ = -0.39 n = 32 n = 25 n = 27 n = 40 NS p < .1 NS p < .01 Evaluation Stage r g = -0.19 r s = -0.11 r g = -0.30 r s = 0.16 n = 31 n = 25 n = 27 n = 40 NS NS p < .1 NS Entrepreneurial Design Stage r g = .48 r g = .54 r $ = .42 r g = 0.04 n = 33 n = 25 n = 27 n = 40 p < ^005 p < .005 p < .025 ' NS Fieldwork Stage r g = -0.03 r g = -0.02 r g = .23 r g = 0.27 n = 32 n = 25 n = 27 n = 40 NS NS NS p < .05 Evaluation Stage r's = -0.285 r g = .18 r s = -0.23 r g = -0.003 n = 31 n = 25 n = 27 n = 40 p < .1 - NS NS NS * Stat ist ical tests are based on Ferguson, 1959, pp. 179-182, p. 308, Roscoe, 1975, pp. 106-110. Significance levels stated are al l based on one tailed tests. 140. Bib!iography Alchian, A.A. and Demsetz, H., \"Production, Information Costs, and Economic Organizations,\".American Economic Review, Vol. LXII No. 5 (December 1972), pp. 777-795. Aly, H.F. and Duboff, J . I . , \"Stat ist ical vs. Judgmental Sampling: An Empirical Study of Auditing The Accounts Receivable of a Small Retail Store,\" The Accounting Review, Vol XLVI No. 1 (January 1971), pp. 119-128. American Inst i tute of Certif ied Public Accountants. Statement on Audit- ing Standards No. 1. New York: AICPA, 1973. Anderson,' R.J.', The External Audit. Volumes 1 and 2. Toronto: Pitman Publishing Co., 1977. Ashton, R.J., \"An Experimental Study of Internal Control Judgments,\" Journal of Accounting Research, Vol 12 No. 1 (Spring 1974), pp. 143-157. Baker, C.R., \"Management Strategy in a Large Accounting Firm,\" The Account- ing Review, Vol. LI I No. 3 (July 1977), pp. 576-586. , \"Management Strategy in a Large Accounting Firm: A Reply,\" The Accounting Review, Vol. LIV - No. 1 (January 1979), pp. 224-225. Bern, D.J. and Allen, A., \"On Predicting Some of the People Some of The Time: The Search for Cross-situational Consistencies in Behaviour,\" Psychological Review, Vol. 8V :,No. 6 (November 1974), pp. 506-520. Bern, D.J. and Funder, D.C, \"Predicting More of the People More, of the Time: Assessing the Personalityof-Situations,\" Psychological Review, Vol. 85 No. 6 (November 1978), pp. 485-501. Benveniste, G.B., Bureaucracy. San Francisco: Boyd and Fraser, 1977. Bernstein, L.A., \"The Concept of Material i ty,\" The Accounting Review, Vol. XLII No. 1 (January 1967), pp. 86-95. Block, J . , \"A Comparison of Forced and Unforced Q-Sorting Procedures,\" Educational and Psychological Measurement, Vol. XVI (1956), pp. 481-493. , The Q-Sort Method in Personality Assessment and Psychiatric Research, Springfield, 111.: Charles C. Thomas, 1961. Boatsman, J.R. and Robertson, J . C , \"Policy-capturing on Selected Materiality Judgments,\" The Accounting Review, Vol. XLIX No. 2 (April 1974), pp. 342-352. c 141. Bock, R.D. and Jones, L.V., The Measurement and Prediction of Judgment and Choice. San Francisco: Holden-Day, 1968. Bowers, K.S., \"Situationism in Psychology: An Analysis and a Critique,\" Psychological Review, Vol. 80> No. 5 (September 1973), pp. 307-336. Brooks, K. and Schroeder, R., \"Management Strategy in a Large Accounting Firm: A Comment,\" The Accounting Review, Vol. LIV No. 1 (January 1979), pp. 221-223. Brown, S.R., \"Bibliography on Q Technique and i ts Methodology,\" Perceptual and Motor Sk i l l s , Vol. 26 (1968), pp. 587-613. Campbell, J . , Dunnette, M.D., Lawler, E.E. and Weick, K.E., Managerial Behavior, Performance and Effectiveness. New York: McGraw Hi l l Book Co., 1970. Cascio, W.F., Applied Psychology in Personnel Management. Virginia: Reston Publishing Company, Inc., 1978. C.I.C.A., C.I.C.A. Handbook. Toronto: C.I.C.A., 1978. Committee on Auditing Procedures, A Case Study of the Extent of Audit Samples. New York: AIA, 1955. Corless, J . C , \"Assessing Prior Distributions for Applying Bayesian Statist ics in Auditing,\" The Accounting Review, Vol. 47- No. 3 (July 1972), pp. 556-566. Cronbach, L.J. and Gleser, G.C, Review of Stephenson, \"The Study of Behaviour: Q-technique and i t s Methodology,\" Psychometrika, Vol. 19 No. 4 (December 1954), pp. 327-330. Darlington, R.B., Radicals and Squares. Ithaca: Logan Hi l l Press, 1974. Dermer, J.D., \"Cognitive Characteristics and the Perceived Importance of Information,\" The Accounting Review, Vol. XLVIII No. 3 (July 1973), pp. 511-519. Dunnette, M.D. and Hoggatt, A . C , \"Deriving a Composite Score from Several Measures of the Same Attr ibute,\" Educational Psychology and Measurement, Vol. 17 (1957), pp. 425-434. Dyer, J .L. , \"Toward the Development of Objective Materiality Norms,\" The Arthur Andersen Chronicle, Vol. 35 (October 1975), pp. 38-49. Endler, N.S. and Magnusson, D., Interactional Psychology and Personality. Washington, D.C: Hemisphere Publishing, 1976. Felix, W.L., \"Alternative Means of Assessing Prior Probability Distribu-t ions,\" The Accounting Review, Vol. LI No. 4 (October 1976), pp. 800-807. 142. Ferguson, G.A., Stat ist ical Analysis in Psychology and Education. New York: McGraw-Hill Book Co., 1959. Flanagan, J . C , \"Defining the Requirements of the Executive's Job,\" Personnel, Vol. 28 No. 1 (July 1951), pp. 28-35. , \"The Crit ical Incident Technique,\" Psychological Bul let in, Vol. 51 No. 4 (July 1954), pp. 327-358. Gibbins, M., \"A Behavioral Approach to Auditing Research,\" in Symposium on Auditing Research 11. Urbana: The University of I l l i n o i s , 1977, pp. 141-186. Gibbins, M. and Wolf, F.M., \"Auditors' Perceptions of Their Decision Environment - The Case of a Normal External Audit,\" unpublished Work^ ing Paper, The University of Br i t ish Columbia, July 1979(a). , \"Audit Outcomes: An Exploration of What Auditors Perceive to Happen After the Audit has been Completed,\" unpublished Working Paper, The University of Br i t ish Columbia, September 1979(b). Hays, W.L., Statist ics for Psychologists. New York: Holt, Rinehart and Winston, 1963. Henderson, B.D. and Dearden, J . , \"New Systems for Divisional Control,\" Harvard Business Review, Vol. 44 No. 5 (Sept/October 1966), pp. 144-161. Jasinski, F.J., \"Use and Misuse of Efficiency Controls,\" Harvard Business.\" Review, Vol. 34 No. 4 (July/August 1956), pp. 105-112. Jensen, M.C, \"Reflections on the State of Accounting Research and the Regulation of Accounting,\" Stanford Lectures in Accounting, 1976. Jensen, M.C. and Meckling, W.H., \"Theory of the Firm: Managerial Behaviour, Agency Costs and Ownership Structure,\" Journal of Financial Economics, Vol. 3 No. 4 (October 1976), pp. 305-360. Jones, A., \"Distributions of Traits in Current Q-sort Methodology,\" Journal of Abnormal and Social Psychology, Vol. 53 (1956), pp. 90-95. Joyce, E.J., \"Expert Judgment in Audit Program Planning,\" Studies on Human Information Processing in Accounting, Supplement to Vol. 14 of Journal of Accounting Research (1976), pp. 29-60. Kaplan, R.S., \"The Roles for Research and Development in Auditing,\" in Symposium on Auditing Research 11. Urbana: The University of I l l i n o i s , 1977, pp. 3-9. Kay, B.R., \"Key Factors in Effective Foreman Behavior,\" Personnel, Vol. 36 No. 1 (Jan.-Feb. 1959), pp. 25-31. Kendall, M.G., Rank Correlation Methods. London: Gr i f f in and Co., L td. , 1970. 143. Kerlinger, F.N., \"Q Methodology in Behavioural Research\" in Brown, S. and Brenner, D. (eds.) Science, Psychology and Communication: Essays Honoring William Stephenson, New York: Teachers College Press, 1972. , Foundations of Behavioural Research, 2nd Edition. New York: Holt, Rinehart and Winston, Inc., 1973. Kinney, W.R. and Ri t ts , B.A., \"Auditors' Sampling Behaviour: An Empirical Study,\" Iowa Bureau of Business and Economics Research Working Paper Series, University of Iowa, Iowa City, 1973. Kogan, N. and Wallach, M.A., \"Risk Taking as a Function of the Situation, the Person and the Group,\" in Newcomb, T.H. (ed.) New Directions of Psychology I I I . New York: Holt, Rinehart and Winston, Inc., 1967. Lawler, E.E. and Rhode, J.G., Information and Control in Organizations. Santa Monica: Goodyear Publishing Co., 1976. Lewis, R. and Stewart, R., The Boss. London: Phoenix House, 1958. Lindman, H.R., Analysis of Variance in Complex Experimental Designs. San Francisco: W.H. Freeman and Company, 1974. Lindquist, E., Design and Analysis of Experiments. Boston: Houghton and M i f f l i n , 1953. Magnusson, D. and Endler, N.S. (eds.), Personality at the Crossroads: Current Issues in Interactional Psychology. Hil lsdale, N.J.: Erlbaum, 1977. Mautz, R.K., \"Discussant!s Response to 'Fads and Foibles in Auditing Research\"' in Symposium on Auditing Research 1, Urbana: The University of I l l i n o i s , 1976. Meckling, W.H., \"Values and the Choice of the Model of the Individual in Social Sciences,\" Revue Suisse d' Economie Politique et de Statisque, Vol. 112 No. 4 (December 1976), pp. 545-559. Merton, R.K., \"Bureaucratic Structure and Personality,\" (1957) in Etzioni, A. (ed.) Complex Organizations. A Sociological Reader. New York: Holt, Rinehart and Winston, Inc., 1962. Mintzberg, H., The Nature of Managerial Work. New York: Harper and Row, Publishers, 1973. Montagna, P.D., Certified Public Accounting: A Sociological View of a Profession in Change. Kansas: Scholars Book Co., 1974. Moriarity, S. and Barron, F.H., \"Modelling the Materiality Judgments of Audit Partners,\" Journal of Accounting Research, Vol. 14 No. 2 (Autumn 1976), pp. 320-341. Newton, L.K., \"The Risk Factor in Materiality Decisions,\" The Accounting Review, Vol. L11 No. 3 (July 1977), pp. 97-108. 144. Ng, D., \"Supply and Demand of Auditing Services and the Nature of Regula-tions in Auditing,\" forthcoming in Arthur Young Professors' Roundtable Conference Proceedings, 1978. Nunnally, J . C , Psychometric Theory. New York: McGraw-Hill Book Co., 1967. Oppenheim, A.N., Questionnare Design and Attitude Measurement. New York: Basic Books Inc., 1966. Pa t t i l l o , J.W., \"Material i ty: The (Formerly) Elusive Standard,\" The Financial Executive, Vol. 43 (August 1975), pp. 20-27. Pa t t i l l o , J.W. and Siebel, J.D., \"Factors Affecting the Materiality Judg-ment,\" CPA Journal, Vol. XLIV No. 7 (July 1974), pp. 39-44. Peat, Marwick, Mitchell and Co., Research Opportunities in Auditing. New York: Peat, Marwick, Mitchell and Co., 1976. Rhode, J.G. and Gr i f f i n , P.A., \"Professional Performance of Certified Public Accountants: Evidence on the Quality of Auditing Services,\" Research Paper No. 455, Graduate School of Business, Stanford University, 1978. Rhode, J.G., Sorensen, J.E. and Lawler, E.E., \"Source of Professional Staff Turnover in Public Accounting Firms Revealed by the Exit Inter-view,\" Accounting, Organizations and Society, Vol. 2 No. 2 09771, pp. 165-175. Ridgway, V.F., \"Dysfunctional Consequences of Performance Measurements,\" Administrative Science Quarterly, Vol. 1 • No. 2 (September 1956), pp. 240-247. Roscoe, J.T., Fundamental Research Stat ist ics for the Behavioural Sciences. New::York: Holt Rinehart and Winston Inc., 1975. Ross, S.A., \"The Economic Theory of Agency: The Principal's Problem,\" American Economic Review, Vol. LXII I No. 2 (May 1973), pp. 134-139. Siegel, S., Nonparametric Stat ist ics. New York: McGraw Hi l l Book Co., 1956. Sorensen, J.E. and Sorensen, T.L., \"The Conflict of Professionals in Bureaucratic Organizations,\" Administrative Science Quarterly, Vol. 19-No. 1 (June 1974), pp. 98-106. Spiegelman, M., Terwil l iger, C. and Fearing, F., \"The Rel iabi l i ty of Agreement in Content Analysis,\" The Journal of Social Psychology, Vol. 37. No. 2 (May 1953), pp. 175-187. Stephenson, W., The Study of Behaviour: Q-Technique and Its Methodology. Chicago: University of Chicago Press, 1953. 145. Terkel, S., Working. New York: Avon Books, 1975. Thompson, V.A., Bureaucracy and Innovation. Alabama: The University of Alabama Press, 1976. Ward, B.H., \"An Investigation of the Materiality Construct in Auditing,\" Journal of Accounting Research, Vol. 14 No. 1 (Spring 1976), pp.\"138-152. Warwick, D.P., A Theory of Public Bureaucracy: Pol i t ics , Personality and Organization in the State Department. Cambridge, Mass.: Harvard University Press, 1975. Watts, R.L. and Zimmerman, J .L . , \"Towards a Positive Theory of Determina-tion of Accounting Standards,\". The Accounting Review, LII No. 1 (January:,1978), pp. 112^134. , \"The Demand for and Supply of Accounting Theories: The Market for Excuses,\" The Accounting Review, LIV No. 2 (April 1979), pp. 273-305. Weber, R., \"Auditor Decision Making on Overall System Rel iabi l i ty : Accuracy, Consensus, and the Usefulness of a Simulation Decision Aid,\" Journal of Accounting Research, Vol. 16 No. 2 (Autumn 1978), pp. 368-387. Whitney, D.R., Elements of Mathematical Stat ist ics. New York: Henry Holt and Co., 1959. Willingham, J.J. and Carmichael, D.R., Auditing Concepts and Methods. New York: McGraw Hi l l Book Co., 1979. Wittenborn, J.R., \"Contributions and Current Status of Q-Methodology,\" Psychological Bul let in, Vol. 58 No. 2 (March 1961), pp. 131-142. Woolsey, S.M., \"Development of Criteria to Guide the Accountant in Judging Material i ty,\" Journal of Accountancy, Vol. 97 (February 1954), pp. 167-173. , \"Materiality Survey,\" Journal of Accountancy, Vol. 135 (September 1973), pp. 91-92. Zimmerman, J .L. , \"The Costs and Benefits of Cost Allocations,\" The Accounting Review, Vol. LIV No. 3 (July 1979), pp. 504-521. "@en ; edm:hasType "Thesis/Dissertation"@en ; edm:isShownAt "10.14288/1.0095404"@en ; dcterms:language "eng"@en ; ns0:degreeDiscipline "Business Administration"@en ; edm:provider "Vancouver : University of British Columbia Library"@en ; dcterms:publisher "University of British Columbia"@en ; dcterms:rights "For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use."@en ; ns0:scholarLevel "Graduate"@en ; dcterms:title "A positive theory of managers’ decisional behaviour in public accounting firms"@en ; dcterms:type "Text"@en ; ns0:identifierURI "http://hdl.handle.net/2429/22604"@en .