Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

A positive theory of managers’ decisional behaviour in public accounting firms Wolf, Frank Michael 1980

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

Item Metadata

Download

Media
831-UBC_1980_A1 W64.pdf [ 6.8MB ]
Metadata
JSON: 831-1.0095404.json
JSON-LD: 831-1.0095404-ld.json
RDF/XML (Pretty): 831-1.0095404-rdf.xml
RDF/JSON: 831-1.0095404-rdf.json
Turtle: 831-1.0095404-turtle.txt
N-Triples: 831-1.0095404-rdf-ntriples.txt
Original Record: 831-1.0095404-source.json
Full Text
831-1.0095404-fulltext.txt
Citation
831-1.0095404.ris

Full Text

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 fulfilment of the requirements for an advanced degree at the University of British 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 British 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 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? 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 c a 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 c a l to the effective or ineffective completion of their work. An extensive series of interviews was conducted in five public accounting firms in a large c i t y in Western Canada for the purpose of implementing the c r i t i c a l incident methodology.  Partners, managers and  senior f i e l d staff participated in the study and provided a total of 582 c r i t i c a l incidents.  These incidents were summarized and abstracted to  yield 88 unique c r i t i c a 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 w i 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 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 w i l l give a higher p r i o r i t y to conformance to those rules which can be readily monitored by their superiors.  Conversely, managers w i l l give a lower p r i o r i t y 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 t a t i s t i c a 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 stimuli. In the context of this study, a Q-sort was developed which characterizes the work of the audit manager in terms of behaviours c r i t i c a 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 p a r 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 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 p r i o r i t y to conformance to organizational rules which can be readily monitored within the firm.  Conversely, managers give a lower p r i o r i t y 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 ii  Abstract List of Appendices  vii  List of Figures  ix  List of Exhibits  x  List of Tables  xi  Acknowledgments-  xii  CHAPTER 1:  INTRODUCTION  1  CHAPTER 2:  THE NATURE OF MANAGERIAL WORK: AN INVESTIGATION OF THE WORK OF THE AUDIT MANAGER Introduction The Work of Audit Personnel:  6 °.  6  A Review of the  Extant Literature  7  The Critical Incident Methodology  9  The Application of the Critical Incident Methodology  11  Data Collection  12  Interpreting the Critical Requirements of  CHAPTER 3:  Managerial Work  16  Discussion  22  Validating the List of Critical Requirements  24  Conclusions  28  Appendices  31  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  vi. Paje  CHAPTER 4:  CHAPTER 5:  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  THE COLLECTION AND ANALYSIS OF THE Q-SORT DATA . . . .  101  Data Collection  101  Statistical Analysis  101  Discussion of the Contextual Hypotheses  Ill  tests of the Substantive Hypotheses  112  Discussion of the Substantive Hypotheses  119  Appendices  121  FURTHERING THE DEVELOPMENT OF POSITIVE THEORY IN AUDITING: SOME CONCLUDING REMARKS Appendix  133 139  Bibliography  140  vi i .  List of Appendices Page Appendix I  Appendix I I  Applying the Critical Incident Methodology: An I l l u s t r a t i o n of the Structured Interview Procedure  31  Checklist of Critical Requirements of Managerial Work . . .  38  Appendix I I I  The Critical 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 Test of Hypothesis C2 Communal i t y of Managers' Rank Assignments  Appendix VI  Appendix VII : Test of Hypothesis C3 Rank Order Correlation.Coefficients. . Appendix V I I I Appendix IX  Appendix X  Appendix XI  Appendix XII  Appendix X I I I  Appendix XIV  Appendix XV  121 122 123  Test of Hypothesis C4 Communality of Managers' Rank Assignments  124  Test of Hypothesis C5 Communality of Rank Assignments Across Partner Groups  125  Test of Hypothesis C6 Communality of Rank Assignments Across Manager Groups  126  Test of Hypothesis SI The Manager as a Resource Allocator. Role and Firm  Results by  127  Test of Hypothesis SI Analysis of Variance of Q-Sort Data. as a Resource Allocator  The Manager  Test of Hypothesis S2 The Manager as a Negotiator. and Firm  Results by Role  Test of Hypothesis S2 Analysis of Variance of Q-Sort Data. as a Negotiator Test of Hypothesis S3 The Manager as. an Entrepreneur. and Firm  The Manager  Results by Role  128  129  130  131  viii.  List of Appendices (continued) Appendix XVI  Appendix XVII  Page  Test of Hypothesis S3 Analysis of Variance of Q-Sort Data. The Manager as an Entrepreneur  132  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 l u s t r a t i v e Flowchart of the Work Undertaken on an Audit  57  X.  List of Exhibits Exhibit 1  Page A Description of Mintzberg's Managerial Roles. . . . 20  Exhibit 2  I l l u s t r a t i o n of the Q-Sort Methodology  62  xi..  List of Tables Page Table I  Table I I Table I I I  Participants in the Interview Program Designed for the Implementation of the Critical Incident Methodology  13  Frequency with Which Critical Task Requirements Pertained to Each Managerial Role  23  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 Partner-Manager Differences in the Rank Assignment of the Negotiation Cue  116  Table VI Table VII  Partner-Manager Differences in the Rank Assignment of the Entrepreneurial Cues  117 118  xii.  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 i k e 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  w i l l i n g l y co-operated with this research through i t s various phases. A great deal of assistance and guidance in the planning and execution of this research was provided by my dissertation committee at The University of British "Columbia, Vancouver.  I would like 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 British Columbia are gratefully 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 continuously 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 i e n t ' s state of affairs so that a professional opinion may be issued.  Ultimately, such an opinion w i 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, particularly in the United States, but also elsewhere, have continued to extend the audit o r ' 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 c s 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 significant resources towards the development 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 t e 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 process 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 i e l d staff and juniors or staff accountants.  This  study is specifically 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 process (Peat, Marwick & Mitchell and Co., 1976, p. 8), and for the evaluation of positive and normative aspects of the audit.  Such an understanding  w i 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 c a 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 c a l to the effective completion of their task.  The c r i t i c a 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 s u f f i c i e n t l y complete in i t s e l f to permit inferences and predictions to be made about the person performing the act.  A critical inci-  dent is one which occurs in a situation where the purpose or intent of the act seems clear to the observer and where i t s consequences are such that there is l i t t l e doubt as to i t s effects. A detailed specification of the work of audit managers"is presented in Chapter 2.  In that chapter the c r i t i c a 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 c a 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 w i 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 firm. Jensen (1976, p. 13) indicated that positive theory building is a precondition for addressing normative questions which may be of interest to us. Traditionally, research in accounting and auditing has been normative and d e f i n i t i o n a l .  L i t t l e attention has been directed towards the  explanation of observable phenomenon or to the verification 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. ing phenomena.  Such theories do not set out to explain exist-  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.  It  attempts to describe phenomena which are observable and hence a positive theory might address questions of the following form: (1) (2) (3) (4)  Why do firms change auditors? Why do different audit firms charge different rates for comparable services? How has accounting research influenced actual practice? 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 i o n , the question 'how much auditing is enough', cannot be satisfactorily 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 evaluation 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 limitations of the research are discussed. tions for further work in this area are alluded to.  Possible direc-  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. to change from one setting to another:  A manager's job is subject  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 executive suites in Unilever House " (Lewis and Stewart, 1958, p. 17). More recently Mintzberg indicated: "Although an enormous amount of material has been published on the manager's job, we continue to know very l i t t l e about i t . Much of the l i t e r a t u r e is of l i t t l e use, being merely endless repetition 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 managing," an academic wrote recently Descriptions like 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 f u 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 w i l l form a descriptive foundation for the remainder of the research. This chapter w i l l proceed as follows.  F i r s t , a review of the extant  research which describes the nature of work in public accounting firms w i l l be provided.  Certain problems with that research w i l l be alluded to  and the need for alternate methodologies for the description of managerial work within the public accounting environment w i 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 w i l l be discussed. collection w i l l be delineated.  The method of data  A detailed checklist of c r i t i c a l require-  ments of managerial work w i 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 w i 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 profession, devoted one chapter of his book to the career pattern of a largefirm 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 firm.  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 informative 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 considerations which seriously mitigate the value of the description actually provided (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 s a t e l l i t e office 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 s a t e l l i t e office as his  research site was problematic i f a general description of managerial strategy was to be derived as s a t e l l i t e 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 i k e l y to be contingent upon the time period during which the observations were undertaken unless a longitudinal or a c t i v i t y 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 interactions between the audit personnel and their clients would be severely 1imi ted.  C.  The Critical 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 w i 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 description of managerial work (Mintzberg, 1973, pp. 221-229).  In this study,  a methodology originally proposed by Flanagan (1954) which assesses task related behaviours in terms of incidents c r i t i c a l to the completion of those tasks is adopted.  Campbell et al (1970) see the c r i t i c a 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 c a 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 c a l incidents of a job as those behaviours which are crucial in making a difference between doing the job effectively and doing i t ineffectively.  Critical incidents are simply reports by  qualified 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 lies 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 interpretation 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 c a l events may need to be collected and abstracted.  Cascio  (1978) indicated that the c r i t i c a l incident technique w i l l typically 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 c a 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 c a l incident methodology is that one can never be sure that important aspects of the job are not missing from the description.  He suggests  n.  that there may be a tendency to ignore a c t i v i t i e s which are routine, compl or sensitive. Tb..ensure that the l i s t of c r i t i c a l requirements derived using the methodology outlined above is comprehensive, two validation procedures are adopted.  F i r s t , the l i s t of c r i t i c a l requirements identified is com-  pared to the requirements identified 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 c a l task requirements and asked to comment on the completeness of the list.  These procedures were designed to ensure that the specification  of managerial work derived using the c r i t i c a l incident methodology is compl ete.',  D.  The Application of the Critical Incident Methodology The c r i t i c a l incident technique does not consist of a single set of  rules governing the data collection procedure and should be though of as a flexible 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 a c t i v i t y subject to description.  A prerequisite for the formulation of a functional  description of an a c t i v i t y is a basic orientation in terms of the general aims of that a c t i v i t y .  In i t s simplest form, the functional description  of an a c t i v i t y requires the specification of what must be done and what must not be done i f participation in that a c t i v i t y is to be judged success ful.  12.  In this study, incidents c r i t i c a 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 qualified 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 firm.  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 firm. 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 c a l incidents, five of the 'big 14'  public accounting firms in a large Western Canadian c i t y were selected at random.  The participant firms were asked to provide twenty subjects:  five audit partners, ten audit managers and five senior f i e l d s t a f f , so that a broad range of incidents could be collected.  There was no firm  that f u l l y satisfied these subject requirements and in many firms a l l available staff in each of the categories specified participated in the study.  Critical incidents were collected using a structured interview  procedure and often repeated v i s i t s to a particular firm were required to ensure that all 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 promotion decision.  Conversely, respondents were requested to provide five  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  19  12  9  69  13  * In Firm B six senior staff were made available for the interview program and all 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 significant in terms of the manager's promotability within the firm. By identifying just what the manager did that was c r i t i c a l to his promotion decision, i t was possible to identify the c r i t i c a l requirements of  14.  his work.  To aid respondents an example of the c r i t i c a 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 t o t a l , 582 incidents were collected. provided 8.4 incidents.  On average, each respondent  Each incident was summarized and abstracted so  that the c r i t i c a l requirements of managerial work could be identified. From the set of 582 incidents reported, 88 unique c r i t i c a l task requirements were identified 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 c a 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 c a l requirements.  The  i n i t i a l collection of 582 incidents satisfied that criterion 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 all 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 reflect.the criticalness of the i n d i 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 s e 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 i n c 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 . client.  The partner.asked the manager to raise the issues with the  When the partner returned a week later nothing had been done.  The client 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 i n a b i l i t y  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. back  The senior f e l t that feed-  should have been provided to him.  - A manager did not get along with one of his clients particularly well, and was nervous in his dealings with them. to make the client happy.  For this purpose the manager made rash  promises about delivery to the c l i e n t . entirely inappropriate.  As a result he wanted  However, this strategy was  When the commitments to the client were not  met (and they simply could not be met) the manager had an unhappy client -  16.  and this led, in turn, to an unhappy partner.  This incident impaired  the promotability of the manager in the firm. - A senior identifies a potential scope limitation 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 client has done a poor job in completing the audit f i l e s so that the partner to whom he reported directly 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 Critical Requirements of Managerial Work The collection of 88 c r i t i c a l requirements of  managerial work in an  audit firm 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 qualified, 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 directly 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 c a l requirements of managerial  work were categorized in terms of the ten managerial roles identified 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 five chief executives, each for a period of one week. However, Mintzberg indicated that there exists considerable empirical evidence to support the generality of his role prescriptions. are seen to form a gestalt - an integrated whole.  The ten roles  The delineation of the  roles is simply a categorizing process - a somewhat arbitrary partition of the characteristics of managerial work into a f f i n 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  a c t i v i t i e s are divided into three groups - those concerned with interpersonal relationships, those concerned with the transfer of information and those concerned with the decision making a c t i v i t i e s of the manager. The ten managerial roles are divided into three groups - three interpersonal roles, three informational roles and four decisional roles. Each of the 88 c r i t i c a l requirements of managerial work was assigned to the managerial role which best encapsulated i t .  The assignment of  c r i t i c a l requirements to role descriptors was based on an extensive v e r i fication 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 c a l requirement to that role description which he f e l t was most appropriate.  The c r i t i c a l requirements were pre-  sented to judges on a randomly arranged l i s t and in order to f a c i l i a t e the assignment process, each judge was provided with a summary l i s t describing the intent of each of the ten managerial roles.  The l i s t of role  descriptions is presented in Exhibit 1. Critical 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 c a 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 c u l t i e s associated with their task due to the interrelationships which exist in the audit process and the possibility that the c r i t i c a 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 c u l t i e s associated with any taxonomizing procedure, the degree of category r e l i a b 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, Terwilliger 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 all 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). rank order obtained was 5.95.  The mean  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 fifteen patterns of agreement can be illustrated in the Table below: PATTERN OF AGREEMENT 7 6 5 5  1  (complete agreement) .1 2 1 1  1 1 1 1 1 1 (no agreement)  RANK ORDER 1 2 3 4  15  20. Exhibit 1:  A Description of Mintzberg's Managerial Roles  INTERPERSONAL ROLES The FIGUREHEAD is the most basic and most simple of all 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 interpersonal a c t i v i t i e s but none involve significant information processing or decision making. As LEADER the manager is responsible for the motivation and activation of subordinates; their s t a f f i n g , training and associated a c t i v i t i e s . The LIAISON role deals with the significant web of relationships that the manager maintains with individuals and groups who provide information and favors. (This definition 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 t s environment. He seeks information in order to detect changes, ident i f y problems and opportunities and in order to be informed when information 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 organization and upward within the organization. He may be a spokesman for the organization or for members within the organization. He transmits i n formation about the organization's actions, plans, policies or products. DECISIONAL.ROLES As an ENTREPRENEUR, the manager searches the organization and its.environment for opportunities and problems so that he can act as an i n i t i a t o r and designer of changes that w i l l further the objectives of the organization. As a DISTURBANCE HANDLER, the manager is responsible for i n i t i a t i n g corrective 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 t s broadest context. Among organizational resources are money, time, material, equipment, manpower and reputation. As NEGOTIATOR, the manager takes charge when his organization must engage in important negotiation a c t i v i t i e s . He represents the interests of the organization in important conflict situations.  21.  More than 40% of the c r i t i c a l task requirements subject to c l a s s i f i cation had at least five 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 c a 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 r e l i a b i l i t y for a study of this nature. Associated with each of the 582 c r i t i c a l incidents is a c r i t i c a l task requirement.  The distribution of c r i t i c a l task requirements among mana-  gerial roles was sensitive to the characteristics of the voting rule applied.  When a l 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 distribution was significantly 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 i n t e r p r e t a b i l i t y 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 c a 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 voting 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 a c t i v i t i e s of the audit manager.  I t may well be that such a c t i v i t i e s can be better described  in terms of other managerial roles. In Appendix I I to this chapter, a detailed l i s t i n g of the c r i t i c a 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 c a l task requirement was reported is presented. A summary of the frequency data presented in Appendix I I is reproduced in Table I I .  G.  Discussion Table I I summarizes the frequency with which c r i t i c a l task require-  ments reported by partners, managers and senior f i e l d staff were assigned to each of the ten managerial roles originally identified by Mintzberg (1973). Most frequently c r i t i c a l task requirements were reported which pertained to the decisional roles of the audit manager.  In t o t a l , 55% of  a l l the requirements reported described some aspect of the manager's decisional behaviour.  Critical requirements pertaining to the manager's  interpersonal and informational responsibilities together accounted for 45% of a l l requirements reported. When a l l the c r i t i c a l incidents and the task requirements associated with them are considered, significant 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 ROLE  Over Task Requirements Reported By Over All Senior Managers Requirements Partners Field Staff n=l 78 n=216 n=582 n=188 !  Figurehead Leader** Liaison* All Interpersonal Roles  29.9%  Monitor Disseminator** Spokesman*** All Informational Roles Entrepreneur*** Disturbance Handler Resource Allocator Negotiator All Decisional Roles  All Roles  7.7% 11.7% 10.5%  •6.3% 9.6% 7.4%  .  9.2% 7.9% 8.9%.  7.3% 18.1% 15.4%  23.3%  26.0%  40.8%  4.4% 2.1% 8.6% 15.1%  5.1% 0.6% 14.5% 20.2%  3.3% 0.9% 7.8% 12.0%  5.3% 4.8% 3.8% 13.9%  28.5% 2.1% 21.3% 3.1%  36.6% 1 .4% 20.8% 3.2%  55.0%  28.0% 2.8% 24.0% 1.7% 56.5%  62.0%  19.7% 2.1% 19.2% 4.3% 45.3%  100.0%  100.0%  100.0%  100.0%  *  Differences in the frequency with which partners, managers\ahd f i e l d staff reported task requirements pertaining to a given managerial role significant at p = 0.05.  '*  Differences in the frequency with which partners, managers and f i e l d staff reported task requirements pertaining to a given managerial role significant at p = 0.01. Differences in the frequency with which partners, managers and f i e l d staff reported task requirements pertaining to a given managerial role significant at p = 0.001.  24.  pattern with which requirements assigned to specific managerial roles were reported by partners, managers and senior f i e l d staff (X p < 0.001; Ferguson, 1959, pp. 165-169, p. 309).  = 56.06, d.f. = 18,  The position of an  individual reporting a c r i t i c a 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 significant differences were observed in the frequency of the incidents reported by partners, managers and senior f i e l d staff.  In p a r t i -  cular, significant differences in the frequency with which different groups of auditors reported task requirements of a given characteristic were found to exist with respect to the leadership, l i a i s o n a l , 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 i e l d staff perceived incidents c r i t i c a l to the effective or ineffective completion of managerial work d i f f e r e n t i a l l y .  At  best, we are able to assert that there were significant differences in the pattern with which those incidents, and the task requirements associated with them, were reported by different groups of individuals within the firm.  Such differences may be suggestive of perceptual differences. In subsequent chapters of this study attention w i l l specifically be  directed to the way in which partners and managers within public accounting firms perceive the importance of a number of the c r i t i c a l requirements of managerial work in alternate decision contexts. H.  Validating the List of Critical Requirements In the preceding discussion, two procedures were outlined for the  validation of the l i s t of c r i t i c a l requirements identified from the 582  25.  c r i t i c a l incidents reported during the interview program.  F i r s t , a review  of the secondary sources made available to the researcher by the p a r 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 consistent with firm job descriptions, appraisal forms, and so f o r t h . Second, a small group of auditors were asked to verify the completeness of the l i s t of c r i t i c a l task requirements. In general, the l i s t of c r i t i c a l requirements derived using Flanagan's (1954) c r i t i c a 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 particularly c r i t i c a l to his promotion decision within the firm.  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 c a 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 c a 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 firm.  However, i t was l e f t to a  small group of auditors to make the final 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 c a l requirements.  Two auditors  were drawn from one of the five 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 c a 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 c a 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 like 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 i e n t ' s affairs and keeps superiors informed about the work of subordinates.  This l a t t e r 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 i n g of subordinates, the discussion of their remuneration, and their overall management.  Activities of this sort may be broadly  27.  described as part of the managers' leadership a c t i v i t i e s and whilst such a c t i v i t i e s 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 i e n t .  As  such, the manager often communicates problems to the c l i e n t , expresses concerns to them, and may even be responsible for dismissing a client on behalf of the firm. Two further points of c l a r i f i c a t i o n were called for by the validating auditors.  Two of:the auditors questioned thejiaisonal a c t i v i t y '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 responsibility 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 c a l requirements of managerial work derived using the c r i t i c a l incident methodology was both insightful and reassuring.  Basically, the l i s t of 88 c r i t i c a 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 i e n t , and his responsibility to deal effectively with the client 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 identification of these factors was not consistent with that used for the identification of other dimensions 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. Conversely, 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 c a l incident methodology.  Eighty-eight  c r i t i c a l requirements of managerial work were identified from the 582 c r i t i c a l incidents that were reported by the 19 audit partners, 25 audit managers and 25 senior f i e l d staff who participated in this study. Critical requirements were assigned to one of ten managerial roles originally identified 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 c a l requirements were reported by different groups of auditors was undertaken.  Significant differences were found to exist in  the pattern with which partners, managers, and senior f i e l d staff reported c r i t i c a 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 f u 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 c a 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 responsibilities of the audit manager were absent.from the description presented in Appendix I I . The c r i t i c a 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 i e l d staff reported specific requirements 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 i e l d staff  may share identical perceptions about what constitutes effective behaviour within the audit firm but may have simply provided illustrations of that behaviour in significantly different ways. In Chapter 3, partners' and managers' perceptions about the importance of a number of c r i t i c a l requirements of managerial work are investigated in alternate problem contexts.  Systematic and predictable discrep-  ancies between the way in which partners and managers perceive various decision related requirements w i l l be hypothesized.  A positive theory  w i l l be developed which w i l l explain the nature of these discrepancies. The theory w i l l be tested empirically by drawing extensively upon the descriptive foundation developed in this chapter.  31.  Appendix I Applying the Critical Incident Methodology: An •Illustration of the Structured Interview Procedure 1.  The oral script used to motivate each interview session. First, l e t me thank you for agreeing to participate in this study.  This study w i l l form a portion of the research I am undertaking for my Ph.D. dissertation at the University of British Columbia. the work into two stages:  I have divided  the information I w i l l be asking you to pro-  vide today w i 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 c a 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 c a l incident method. The c r i t i c a 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 l u s t r a t e 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 carefully, 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 i g h t , or a bus driver having unusual d i f f i c u l t y replacing a displaced t r o l l e y 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 illustrations 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  particular, I want to focus on incidents in which the audit manager was involved and in which he did something that was either favourable or unfavourable 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 i n c i dents may lead to a delay in promotion. Before we proceed, l e t me assure you that your responses w i l l be s t r i c t l y confidential.  In no way w i 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)  2.  33.  The Specific Interview Questionnaire*  QUESTION 1: For this question I would like you to indicate a number of incidents you have participated in or observed in which the audit manager did something 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 c a 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 c a 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 c a 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 c a l to his promotion 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 c a l to his promotion decision?  QUESTION 2: For this question I would like 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 c a 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 c a 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 c a 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 c a 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 c a l to his promotion decision?  QUESTION 3: I would l i k e 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 firm 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)  QUESTION 6: What other rewards/penalty influence your behaviour on the job?  QUESTION 7: How is the promotion decision determined?  37.  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  1.1  0.0  0.0  0.6 0.0 0.0 0.6 0.0  0.9 0.5 0.9 0.9 0.0  0.0 0.5 1.6 0.5 1.1  0.6  0.0  0.0  0.6 0.0  0.5 0.0  0.5 0.5  1.6  1.8  0.5  0.0 0.0 0.6 0.6  0.9 1.8 0.5 0.5  0.5 0.0 0.5 1 .1  6.3  9.2  7.3  Leader: Accepts responsibility for work of subordinates  0.0  0.9  0.0  Creates meaningful learning experiences for subordinates  3.4  1.8  1 .1  1.7  0.9  1.1  1.1  1 .4  1.6  INTERPERSONAL ROLES OF THE AUDIT MANAGER 1. Figurehead: Involves himself in the professional environment Develops a profile in community Appears to be well organized Develops a profile in the office Conforms to Civil Standards Abides by professional code Keeps abreast of professional developments Maintains professional appearances Maintains client confidentiality Adheres to organizational procedure Meets personal commitments Displays loyalty to the firm Develops a profile with clients Maintains independence  Provides supervision for subordinates Recognizes capabilities of subordinates  Appendix I I (continued)  39.  Provides technical assistance to subordinates  0.0  0.0  1.6  Assists subordinates with any problems that arise  0.0  0.5  3.2  1.7  0.9  3.7  0.0  0.5  2.7  1.1 0.0  0.5 0.0  0.5 0.5  Manages subordinates  0.6 0.0  0.0 0.5  0.0 0.5  Recognizes and accepts individual differences in subordinates  0.0  0.0  1.6  9.6  7.9  18.1  0.6 0.6  0.5 0.5  3.7 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 0.6 0.6  0.9 0.9 1.4  0.5 0.0 0.0  0.0  0.0  1.6  0.0  0.5  0.5  7.4  8.9  15.4  Communicates effectively with subordinates Imposes r e a l i s t i c requirements on subordinates Provides instruction to subordinates Adopts an outgoing position Motivates subordinates  3. Liaison: Fosters c l i e n t - s t a f f relationships Develops a rapport with subordinates Interacts with subordinates in the f i e l d Maintains contact with client through year Works as a member of a team Develops good relationships with client Maintains existing client good will Develops team relationships Adapts to superiors Subordinates personal interests Relates effectively to staff in front of client Plays the 'organizational p o l i t i c s game'  Appendix I I (continued)  40.  INFORMATIONAL ROLES OF THE AUDIT MANAGER 1. Monitor: Ensures the completeness of audit f i l e  '•  _ ~  n  A  n  n  Reviews audit f i l e s promptly  0.0  0.0  3.2  Ensures that the financial statements are error-free Involves the partners only in c r i t i c a l problems Develops a f a m i l i a r i t y with audit f i l e  0.6  1.4  0.5  Ensures that the financial statements meet professional requirements  n  fi  n u  '  -  n  u  n  n  n  ,  n n u , u  fi 1 l b  0.6  0.0  0.0  5.1  3.3  5.3  0.0  0.9  3.7  0.6  0.0  1.1  0.6  0.9  4.8  y  2 2  00  0.6  0.5  0.0  1.1  0.0  1.1  Disseminator: Provides feedback .to subordinates Keeps subordinates informed about developments of clients during the year  3.  Spokesman: Advises superiors about potent i a l problems Completes the management l e t t e r Informs the client about the execution of the audit Provides j u s t i f i c a t i o n s for actions taken Communicates effectively with superiors Communicates effectively with client Makes r e a l i s t i c commitments to clients Stands up for his own beliefs and opinions Maintains the reputation of the firm  6  Q  Q  Q  2 2 0 2  -  0 8  0  n n u > u  14.5.  Q  18 K  •, ,  4  1  0.5 0  n u  5  n  -  7.8  J  0.0 n  u  Q  05  , „  1.1 0  g  n  , , '•' 3.8  Appendix I I (continued) DECISIONAL ROLES OF THE AUDIT MANAGER 1.  Entrepreneur: Displays imagination on the audit Sells additional .services to V. existing clients Is geographically mobile  0 .6  0 .0  0 .5  2 .2  1 .4  2 .7  0 .0  2 .3 2 .8  Acts independently of his partner Assumes the responsibility of partner  3.3 1 .1  0 .0 2 .7  1 .4  0 .5  2 .8  3.2  1 .6  Assumes additional responsibilities  2 .8  1 .4  2 .1  Pursues challenging assignments Provides management type advice and service to the client  0 .6  0 .5  0.0  5 .6  4 .6  2 .1  Displays an a b i l i t y to identify problems Brings in new clients Develops technical s k i l l s in a specialist area Adapts to the individual needs of clients  6 .7  6 .5  1 .6  0 .6  5.1  3.7  1.1  5 .1  1 .1  0 .6  2 .3  1 .1  28 .0  36 .6  19 .7  2 .8  1 .4  2 .1  2 .8  1 .4  2 .1  0 .6 2 .2  0 .9 0 .5  0 .5 0 .0  1.1  0.0  0 .0  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  Displays technical competence  Disturbance Handler: Proposes solutions for problems identified  Resource Allocator: Meets c r i t i c a l deadlines Meets commitments to partners Makes r e a l i s t i c commitments to partners Prepares an appropriate audit package for client  Appendix I I (continued)  42.  Makes a time commitment to the firm.(puts in his time)  2.8  0.5  2.7  Meets commitments made to clients  1.7  1.4  0.0  ^ 7 ''  t\~i  Completes the job in a timely manner Completes the job cost effectively Accommodates the needs of others in the firm Effectively u t i l i z e s available personnel  c n  5 , u  n 2  4.  Negotiator: Reconciles fee differences with client Develops an audit plan with client Reconciles disagreements about the scope and function of the audit with client Reconciles technical differences with client  0 2  '  n u  -  b  , , '•'  ,  '  5  0.0  Delegates authority to subordinates Prepares a coherent audit package  6  ,  6  0.5 fi  n Q u , y  n  n  °-  c  '-  6  1.6 n  ,-  u , b  ,  9  c  '-  6  0.9  0.0  24.0  20.8  19.2  0.0  1.4  1.6  n  fi  n u  '  n u  0.6 , , 1  J  1.7  n n u > u  n  K  '  u  0.0 n K  0 8  3.2  b  1.1 , 1  J  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. Traditionally, research in accounting and auditing has been normative and definitional and as a result there is a dearth of theory or evidence which bears on positive issues ( c f . , Jensen, 1976). ature focuses,  The available l i t e r -  in the main, on questions of a 'what ought to be nature'  and has not tended to examine positive questions concerned with the explanation of observable phenomena. In this chapter a positive theory of decision making in auditing w i l l be developed by focusing on the decision based behaviour of audit managers and by specifically examining why the decisions that they take on the job are what they are.  Such a theory w i 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 firm. 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 v a r i a b i l i t y 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.  It will  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 w i l l be examined and the implications of that function on the nature of the decisions effected by the audit manager w i l l be considered. Specific hypotheses which w i l l be subjected to investigation w i l l be detailed and the research design w i l l be elaborated.  The principal results  from this section of the analysis w i 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  liter-  ature in auditing w i l l be provided. The American Institute of Accountants 'Committee on Auditing Pro1  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 s t a t i s t i c a l and judgmental samples as applied to the audit of accounts receivable balances for a hypothetical r e t a i l store.  They  reported wide variations in the size of the judgment samples recommended by the 158 I l l i n o 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 f u l l investigation. Corless (1972), in a third study in this area, examined the a b 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 n g information pertinent to the construction of prior d i s t r i butions.  Corless (1972) observed considerable v a r i a b i l i t y among prior  distributions assessed by different auditors in each of the two audit cases which he presented to them.  He suggested that because an auditor  "may typically consider a relatively small error rate as being material, the variation among prior distributions appears to be great" (Corless, 1972, p. 560).  Corless also observed considerable v a r i a b i l i t y 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 provided with a detailed case description of a plumbing supply firm. Respondents were asked to assess the inventory sampling plan required for the audit of this firm.  A preliminary analysis of the data indicated  that there were considerable differences in the plans proposed by respondents. 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 personal background of the participants in his study.  However, he found no  46.  significant relationships between the characteristic values of the prior distributions and either the level of audit experience or the s t a t i s t i c a l background of participants in the study. A major exception to the above studies reporting substantial i n d i 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 r m , given the pattern of responses generated.  Each auditor was required to evaluate 32 different  stimulus combinations. Ashton reported a f a i r l y high level of consistency in the internal control evaluations of different auditors given the same stimulus combinations.  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 earlier 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. audit  Joyce focused upon the consistency of  program plans assessed to be necessary in the audit of a t i r e whole-  saler.  Thirty-five 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' a b i l i t y to assess prior probability distributions from a case description of an orderreceiving, shipping and b i l l i n g system.  His study is closely related to  the earlier work of Corless (1972) reported above. participated in the study.  Ten audit personnel  Felix (1976, p. 805) found considerable agree-  ment amongst the subject-auditors' responses both in terms of the location 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 sensitivity.  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 literature 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 significantly on what are the same or similar sets of facts.' Studies by Dyer (1975), P a t t i l l o (1975) and Woolsey (.1954) reveal that different materiality thresholds are used by different auditors and the survey work of P a t t i l l o 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 i n d i -  cated that audit personnel scaled the contributing variables differently  49.  (Moriarity and Barron, 1976,.p. 338). The extant research points to substantial v a r i a b i l i t y among auditors' responses to decisional problems in alternate settings.  This finding is  particularly significant 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 j u d g -  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 confronting those individuals.  Systematic discrepancies are hypothesized to  exist between the decisional responses that partners would want their managers to adopt and those that would typically be adopted by the managers.  Such discrepancies are predictable and verifiable.  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 i k e l y to be a major determinant of task related behaviour (Lawler and Rhode, 1976), turnover (Rhode, Sorensen, and Lawler, 1977) and interorganizational conflict (Sorensen and Sorensen, 1974). Lawler and Rhode (1976), in their observations of the audit environment, 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 w i l l be adopted. There is a growing literature in economics concerned with the derivation of optimal contracts between principals and agents under conditions of uncertainty.  This l i t e r a t u r e 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 w i l l be with the employer-employee relationship which exists amongst members of the audit firm.  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 all individuals are resourceful, evaluative, maximizing 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 responsib 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 w i l l typically install 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 w i l l reduce the principal's welfare or which w i 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 monitoring and bonding a c t i v i t i e s by the principal and the agent, there w i 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 w i l l be developed to explain the manner in which the actions of audit manager d i f f e r 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 sanction 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 well, 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 w i l l be reduced since the only permanent positions in public accounting firms are those of partners.  Staff accountants  and managers have temporary positions; they w i l l either become partners or leave the firm.  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 firm's partners would want them to adopt in a specific problem context.  Rhode and G r i f f i n (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 i n -  complete procedures was more prevalent amongst junior vis a vis senior audit personnel.  Presumably, the evaluation of junior staff is more sensi-  tive to deviations from budgeted time allocations than is the evaluation of more senior personnel. There is a significant body of literature 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 G r i f f i n (1978) is i l l u s t r a t i v e of much of that l i t e r a t u r e . 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 typically 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 i k e l y to induce, on the part of a rational manager, a s t r i c t conformity to organizational rules which influence his promotability within the.organization.  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 internally, and may use the accounting firm only as a vehicle for the enhancement of their external marketability, conformance to organizational rules and regulations enhances 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 w i l l be subject to organizational attention.  In addition, rules form a basis  for the definition of appropriate behaviour within the organization.  How-  ever, within any organization, superiors are not usually able to monitor the performance o f - t h e i r subordinates with respect to each organizational rule, and performance with respect to some rules can only be monitored at a significant (often prohibitive) cost.  In this study, attention  w i l l be directed towards a broad range of managerial behaviours and an assessment w i 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 w i 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 failure  to conform to these l a t t e r 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 organizational rule w i l l be detected, the size of that probability and the nature of the penalty imposed, i f detected, w i l l be influential in determining which rules a rational subordinate w i 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  w i 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 prescriptions of the behaviour that their partners would want them to adopt in a specific problem context. Attention is specifically directed to the decisional a c t i v i t i e s of the audit manager i,n order to investigate whether his decisions and judgments in a given problem context d i f f e r 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 w i 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 w i l l give a higher p r i o r i t y to conformance to those rules which can be readily monitored by their superiors.  Conversely, managers w i l l give a lower p r i o r i t y 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 i n d i cate how they would typically 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 l d 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. sentative of the flow of work on the audit.  This flowchart is repre-  I t is based on an inter-.,  national firm's audit manual, oh the o f f i c i a 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. hypotheses that are examined are elaborated below.  The specific  Figure 2:  An I l l u s t r a t i v e 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 all 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 all 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 n g • Writing of management l e t t e r to client  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. theses are of supplementary interest in this study.  The contextual hypoThey f a c i l i t a t e  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 v a l i d i t y of the measurement procedure. In general, the contextual hypotheses posit a high degree of agreement between and among partner and managers groups as to the behaviour managers w i l l typically 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 accreditation procedure operative within the audit environment and the careful system of monitoring and evaluation which exists at all levels within audit firms.  Support for such hypotheses w i l l enhance the interpreta-  b i l i t y of the substantive hypotheses, which propose systematic discrepancies in the decisional responsesof partners and managers.  These dis-  crepancies are hypothesized to exist even though the intent of the incentive system and the nature of the environmental context is understood by the partners and managers themselves. Whilst the basic hypothesis w i l l be tested across accounting firms,  59.  using the firm as a control device, the contextual hypotheses are examined within each firm.  The six contextual hypotheses are elaborated directly  below. The f i r s t three hypotheses below focus on the explication of the i n 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 firm. HYPOTHESIS Cl:  There w i l l be agreement amongst partners within a given firm as to the behaviour that their managers should take at a givenstage of the audit i f the managers' behaviour is to be in accord with firm policy.  HYPOTHESIS C2:  There w i 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 w i 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 w i 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 w i l l be agreement amongst managers in a given firm as to the behaviour that .they would typically 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 similarity of':the incentive system in different firms. HYPOTHESIS C5:  There w i 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 w i 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 interaction of the person and the situation that supplies most of the psychologically interesting variance in behaviour.  In this study, the question  examined is a question of person-situation interactions.  In particular,  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). described as the Q-sort methodology.  That language is generally  61.  A Q-sort methodology is the general name used by Stephenson (1953) to characterize a set of philosophical, s t a t i s t i c a l , and psychometric ideas oriented towards research on individuals.  Recent reviews of the Q  methodology point to i t s extensive use (Brown, 1968; Kerlinger, 1972). The Q procedure may be viewed as essentially a kind of rating procedure for rank ordering stimuli.  I t has grown out of a general methodology for  the study of verbalized attitudes, individual preferences and self descriptions.  This methodology has been widely used to study a broad range of  issues (Wittehborn, 1961).  In the accounting and auditing l i t e r a t u r e , the  Q procedure has been applied to the investigation of the information processing characteristics of sales managers and supervisors (Dermer, 1973) and for the purpose of investigating the materiality construct in auditing (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 c r i t e r i o n .  For the purpose of s t a t i s t i c a l  analysis, the sorter is instructed to put varying numbers of cards into each p i l e , the whole making up a predetermined distribution which is t y p 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 c a 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 f a c i l i t a t e data collection. Subjects received randomly ordered sets of c r i t i c a 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 l u s t r a t i o n is presented in  Exhibit 2.  ILLUSTRATION OF THE Q-SORT METHODOLOGY Task Descriptors (randomly ordered) 1.  Manage subordinates  2.  Inform client about the execution of the audit  3.  Provide instruction to subordinates  4.  Develop a f a m i l i a r i t y with audit f i l e  5.  Try to meet c r i t i c a l deadlines  6.  Provide technical assistance for subordinates Chart to Aid Ranking  Highest Rank (Most Important Item)  4  Next most important i terns  5 ' 1  Lowest Rank (Least Important Item)  3  Exhibit 2  i  2 6  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 l u s t r a t e the Q-sorting procedure, I have ordered  the six task requirements in terms of my perceptions about their importance at the design stage of the audit. Nunnally (1967, p.. 555) sees the chief advantage of the Q-sort procedure in i t s a b i l i t y to e l i c i t relatively precise comparative responses amongst a large number of stimuli in a relatively 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 caref u l l y chosen for their "known" or presumed possession of some significant characteristic or characteristics. One explores unknown and unfamiliar areas and variables for their identity, their interrelation and their functioning. 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 v a l i d , the cues must be unambiguous to those who w i 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 c a l task requirements identified using the c r i t i c a l incident methodology. The task requirements are recorded using the phrases provided by interviewees and should be unambiguous to those who w i 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 implemented.  These procedures enhance t h e w a l i d i t y 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 w i l l form the test of the substantive hypotheses to follow. In Chapter 2, each of the 88 c r i t i c a 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 v e , 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 requirement.  To ensure that the task requirements were descriptively valid at  each stage of the audit under investigation, i t was necessary to t a i l o 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 v a l i d i t y of the cues to be used. The research instrument was thus tailored to each stage of the audit subject 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 i g h t l y .  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 presented.  The number of items placed in each stack was specified so that  a bell shaped distribution resulted from the sort of the Q-sort items at each stage of the audit.  This form of distribution 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 typically respond to that context.  Specifically,  i t is hypothesized that relative to what their partners would want their managers to do, managers w i l l give a higher p r i o r i t y to conformance to those rules which can be readily monitored by their superiors.  Con-  versely, managers w i l l give a lower p r i o r i t y 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 firm's a b i l i t y to monitor subordinates' performance on those dimensions.  I f the firm can  monitor the performance of their staff 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 v a l i d i t y of the Q-sort items were applied, no c r i t i c a l task requirements were assigned to the disturbance handler a c t i v i t i e s 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 listed above. The Q-sort items which describe the manager's resource allocative a c t i v i t i e s 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 a c t i v i t i e s w i l l determine the extent to which he is able to meet c r i t i c a l deadlines, satisfy budgetary constraints and meet timing-considerations.  Resource allocative a c t i v i t i e s 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 performance with respect to the manager's resource allocative a c t i v i t i e s was the easiest to monitor within the organization.  I t is therefore hypothe-  sized that: HYPOTHESIS SI:  Managers w i l l give a higher p r i o r i t y to conformance to resource allocative a c t i v i t i e s than their partners would want them to give.  As a negotiator, the manager represents the firm in important confrontations with their c l i e n t .  One c r i t i c a l task requirement which was  assigned to the negotiator role of the manager appears amongst the Q-sort items.  Specifically, as a negotiator the manager reconciles technical  differences with their c l i e n t .  The four audit managers generally f e l t  that the performance of personnel can be readily monitored in terms of their a b 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 allocative decisions.  I t is hypothesized that:  HYPOTHESIS S2:  Managers w i l l give a higher p r i o r i t y to conformance to negotiation a c t i v i t i e s than their partners would want them to give.  68.  F i n a l l y , a t t e n t i o n is directed to the r o l e of the manager as an entrepreneur w i t h i n the organization.  The Q-sort items (see Appendix  III)  indicate that as an entrepreneur the manager'is concerned with the maintenance and furtherance of the e x i s t i n g practice (by providing management type advice and services to e x i s t i n g c l i e n t s and s e l l i n g additional services to those c l i e n t s ) and with the acceptance of r e s p o n s i b i l i t i e s associated with his p o s i t i o n and status w i t h i n the organization (by d i s playing an a b i l i t y to i d e n t i f y problems and by developing technical  skills  in a s p e c i a 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 t h e r the negotiation or resource a l l o c a t i o n 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 extending the practice 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 d e n t i f y problems or develop technical s k i l l s in a s p e c i a l i s t area. Whilst objective measures are available which r e l a t e to the manager's a b i l i t y to maintain and develop the e x i s t i n g p r a c t i c e , performance evaluat i o n in terms of these dimensions is d i f f i c u l t to monitor.  A failure  to provide management type advice and services to e x i s t i n g c l i e n t s or a f a i l u r e to s e l l additional services to them may not be i n d i c a t i v e of poor entrepreneurial performance.  Furthermore, the f a c t that such services  are provided may not indicate that the auditor concerned displayed entrepreneurial 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  69.  norms were available upon which to base such evaluations.  I t is hypothe-  Q  sized that: HYPOTHESIS S3:  Managers w i l l give a lower p r i o r i t y to conformance to entrepreneurial a c t i v i t i e s 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 p r i o r i t y assigned to specific managerial roles by partners and managers, are based on the ease with which partners are able to monitor 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 specifically 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 all hypotheses  w i l l be tested at each stage of the audit because specific task requirements assigned to a particular managerial role were not seen to be descriptive 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 identified.  70.  TABLE I I I TABLE OF THE SUBSTANTIVE HYPOTHESES TO BE TESTED AT EACH AUDIT STAGE Design Stage  H.  Resource Allocation Hypothesis (Hypothesis SI) Negotiation Hypothesis (Hypothesis S2)  /  Entrepreneurial Hypothesis (Hypothesis S3)  /  Fieldwork Stage  Evaluation Stage  /  /  /  /  /  /  The Research Material Each participant in the study received a booklet containing the  research material. managers.  Separate booklets were prepared for partners and  Copies of the research material are reproduced in Appendix IV.  The Q-sort material was preceded by an introductory l e t t e r explaining in very broad terms the nature of the study and e l i c i t i n g the co-operation of each participant.  Respondents were guaranteed anonymity and were  asked not to discuss the research with others.  To enhance the v a l i d i t y  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 l u s t r a t i v e 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 accounting 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 s t a t i s t i c s  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 satisfactory for a study of this nature. Each subject participating in the study was asked to sort the c r i t i c a 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 listed  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 listed 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 listed 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 subjects in each firm. Once subjects have rank ordered the stimuli, 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 c e l l s , three is assigned.to the items in the third set of c e l l s , and so on until seven is assigned to the item in the bottom most c e l l .  This procedure  for the assignment of numerals to Q-sort data is the conventional one (Block, 1961, pp. 77-82) and typically 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 firm. 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 prediction 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 managerial strategy within audit firms. In the following chapter, the hypothesis tests w i l l be outlined and the test results presented.  74.  APPENDIX I I I The Critical Task Requirements in the Q-Sort (Items asterisked appear in the Q-sort at the stage of the audit indicated) DESIGN STAGE Resource Allocation Decisions Try to u t i l i z e available personnel effectively Try to complete the job in a timely manner Try to meet c r i t i c a l deadlines Try to complete the job cost effectively Negotiation Decisions Reconcile technical differences with the client Entrepreneurial Decisions Display an a b i l i t y to identify problems Develop technical s k i l l s in a specialist area Provide management type advice and services Sell additional services to existing clients Other Managerial Activities Impose r e a l i s t i c requirements on subordinates Develop team relationships Maintain professional appearances Create meaningful learning experience for subordinates Inform the client about the execution of the audit  FIELDWORK STAGE  EVALUATION STAGE  Appendix I I I (continued)  75.  DESIGN STAGE  FIELDWORK STAGE  EVALUATION STAGE  Provide instruction to subordinates  *  *  ^  Provide technical assistance to subordinates  *  *  ^  Communicate effectively with clients  *  *  *  Manage subordinates  *  *  *  Provide supervision for subordinates  *  *  *  Communicate effectively with superiors  *  *  ^  Motivate subordinates Complete the management l e t t e r  *  *  * *  Accept responsibility for the work of others Develop a f a m i l i a r i t y 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.  f i r s t stage was completed in Vancouver during the summer.  The  Frank has already  provided feedback to each of the participant firms and has indicated that he w i 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 c a l incidents pertaining to the work of audit managers were e l i c i t e d . of c r i t i c a l task requirements.  These were analysed to yield a number  Some s t a t i s t i c a l tests were carried out.  During this stage you w i l l be asked to rank various subsets of these c r i t i c a l task requirements at a number of stages of the audit.  The rank-  ing task has been simplified by calling for only a partial rank ordering of each set of requirements. I have discussed the research material and the entire project with Frank and f u l l y endorse i t .  I feel that the results which he hopes to  derive w i l l be worth the investment of our time in i t s 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 descriptions 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 j u r i s d i c t i o n . The company has been your firm's client for several years and the audit manager is assumed to have been responsible for the audit of this client last year.  The audit is a normal statutory audit - there are no special  reporting requirements or major international implications.  Appendix IV (continued)  (2)  79.  THE FLOW OF WORK ON THE AUDIT  The flowchart presented outlines the various stages of the audit. J_n this study attention w i 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 all of above * Design of this year's audit  FIELDWORK STAGE FOCUS OF THIS STUDY  * * * *  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 * * * [i *  Issuance of audit report Preparation of f i l e notes for next year Preparation of fee b i l l i n g Writing of management l e t t e r to client  Appendix IV (continued)  80.  COMPLETING THE QUESTIONNAIRE The completion of the questionnaire w i l l involve ranking a randomly ordered subset of c r i t i c a l task requirements at each stage of the audit. A s l i g h t l y 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. You w i l l be asked to provide three sets of rankings; one at each stage of the audit.  Each set of rankings w i 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. an example is provided of the ranking procedure. of a bus driver.  Below  I t relates to the work  I have ranked the six task requirements listed 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 c a 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:  Chart to aid ranking:  a) b) c) d) e) f)  Highest Rank Next Highest  Drive within speed limits Try-to meet timetables Distribute transfers quickly Display courtesy to passengers Drive safely Announce street names  Lowest Rank NOTE:  e b  a  d  c  f  The Bus Co..always wants safe drivers: hence item 'e' is in the top most box. The next 2 items are b' and ' a ' . They are ranked equally. Of least concern is item ' f . 1  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 w i l l have the same rank assigned to them. pleting the questionnaire:  (1)  Please use a lead pencil.  This w i l l f a c i l i t a t e  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.  In com-  Appendix IV (continued)  82.  THE WORK OF THE AUDIT MANAGER AT THE DESIGN STAGE Rankings Based on What the Firm Would Want Managers to Comply With at the Decision Stage..of the Audit Please rank the items listed 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. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.  Try to u t i l i z e available personnel effectively. Inform the client about the execution of the audit Try to complete the job cost effectively. Develop a f a m i l i a r i t y with the audit f i l e . Display loyalty to the firm. Communicate effectively with superiors. Provide motivation for subordinates. Communicate effectively with clients. Provide supervision for subordinates. Try to meet c r i t i c a l deadlines. Try to complete the job in a timely manner. Manage subordinates. Develop team relationships. Display an a b i l i t y to identify problems. Provide instruction for subordinates. Create meaningful learning experiences for subordinates. Accept responsibility for the work of others. Impose r e a l i s t i c requirements on subordinates. Provide technical assistance to subordinates.  Highest Rank Next Highest Identification # of items the firm would want their managers to comply with at the DESIGN STAGE of the audit. Lowest Rank  Appendix IV (continued)  83.  THE WORK OF THE AUDIT MANAGER AT THE FIELDWORK STAGE Rankings Based on What the Firm Would Want Managers to Comply With at the Fieldwork Stage of the Audit Please rank the items listed 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. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.  Provide technical assistance to subordinates. Try to meet c r i t i c a l deadlines. Maintain professional appearances. Provide instruction for subordinates. Develop technical s k i l l s in specialist area. Display loyalty to the firm. Reconcile technical differences with c l i e n t . Develop a f a m i l i a r i t y with the audit f i l e . Try to complete the job cost effectively. Ensure the financial statements are error free. Create meaningful learning experiences for subordinates. Display an a b i l i t y to identify problems. Manage subordinates. Try to u t i l i z e available personnel effectively. Try to complete the job in..timely manner. Communicate effectively with clients. Abide by the professional code of ethics. Provide supervision for subordinates. Inform the client about the execution of the audit. Communicate effectively with superiors. Provide motivation for subordinates. Accept responsibility for the work of others. Impose r e a l i s t i c requirements on subordinates. Develop team relationships.  Identification # of items the firm would want their managers to comply with at the FIELDWORK STAGE of the audit.  Highest Rank Next Highest  Lowest Rank  Appendix IV (continued)  84.  THE WORK OF THE AUDIT MANAGER AT THE EVALUATION STAGE Rankings Based on What the Firm Would Want Managers to Comply With at the Evaluation Stage of the Audit Please rank the items listed 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. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.  Accept the responsibility for the work of others. Try to complete the job cost effectively. Display loyalty to the firm. Develop technical s k i l l s in a specialist area. Provide management type advice and services. Display an a b i l i t y to identify problems; Provide supervision for subordinates. Complete the management l e t t e r . Sell additional services to existing clients. Provide motivation for subordinates. Try to meet c r i t i c a l deadlines. Communicate effectively with clients. Adapt to superiors. Provide technical assistance to subordinates. Provide instruction for.subordinates. Communicate effectively with superiors. Develop a f a m i l i a r i t y with the audit f i l e . Try to complete the job in a timely manner. Create meaningful learning experiences for subordinates. Abide by professional code of ethics. Manage subordinates. Ensure the .'financial statements are error free. Reconcile technical differences with c l i e n t .  Identification # of items the firm would want their managers to comply with at the EVALUATION STAGE of the audit.  Highest Rank Next Highest [ " [ ' ] '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 i e n t ' s state of affairs but cannot pinpoint any specific problems. The audit has gone well - all tests have checked out, a l l deadlines have been met, and the audit is well within budget. The client 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 i n e . Commit Resources  Do Not Commit 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 integr i t y of QWER's management is subject to question. Just the other day QWER management had a quarrel with their auditors who were apparently not ' f l e x i b l e ' 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? 1  Recommend QWER  Do Not Recommend QWER  3. The senior on the job has come into conflict with the c l i e n t ' 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.if any problems arise. The senior has told the accountant that certain preparatory work must be completed by the c l i e n t . 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 l o t of work involved but the senior does not want to establish an undesirable precedent;. The audit manager has been called-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 Firm Policy  Negotiate a 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 f a i r coin, how much would you be prepared 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 descriptions 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 j u r i s d i c t i o n . The company has been your firm's client for several years and the audit manager is assumed to have been responsible for the audit of this client last year.  The audit is a normal statutory audit - there are no special  reporting requirements or major international implications.  Appendix IV (continued)  (2)  89.  THE FLOW QF WORK ON THE AUDIT  The flowchart presented outlines the various stages of the audit. this study, attention w i 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  I DESIGN STAGE If. Evaluation of a l l of above * Design of this year's audit  I FIELDWORK STAGE * * * *  FOCUS OF THIS STUDY  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 n g Writing of management l e t t e r to client  In  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 c a l task requirements at each stage of the audit.  A s l i g h t l y 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 w i l l be asked to provide w i 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 like 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. an example is provided of the ranking procedure. of a bus driver.  Below  I t relates to the work  I have ranked the six task requirements listed 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 c a 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 limits 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  NOTE: The Bus. Co. always wants safe drivers: hence item 'e' is 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 w i 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  facilitate  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)  92.  THE WORK OF THE AUDIT MANAGER AT THE DESIGN STAGE 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 presumably the firm would not be concerned i f they were NOT ALWAYS COMPLIED WITH at this stage. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.  Try to u t i l i z e available personnel effectively. Inform the client about the execution of the audit. Try to complete the job cost effectively. Develop a f a m i l i a r i t y with the.audit f i l e . Display loyalty to the firm. Communicate effectively with superiors. Provide motivation for subordinates. Communicate effectively with clients. Provide supervision for subordinates. Try to meet c r i t i c a l deadlines. Try to complete the job in a timely manner. Manage.:siibordinates. Develop team relationships. Display an a b i l i t y to identify problems :; Provide instruction for subordinates. Create meaningful learning experiences for subordinates. Accept responsibility for the work of others. Impose r e a l i s t i c requirements on subordinates. Provide technical assistance to subordinates. 1  Highest Rank Next Highest Identification # of items the firm would want their managers to comply with at the DESIGN STAGE of the audit.  Lowest Rank  Appendix IV (continued)  93.  THE WORK OF THE AUDIT MANAGER AT THE FIELDWORK STAGE Rankings Based on What the Firm Would Want Managers to Comply with at the Fieldwork Stage of the Audit Please rank the items listed 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. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.  Provide technical assistance to subordinates. Try to meet c r i t i c a l deadlines. Maintain professional appearances. Provide instruction for subordinates. Develop technical s k i l l s in specialist area. Display loyalty to the firm. Reconcile technical differences with c l i e n t . Develop a f a m i l i a r i t y with the audit f i l e . Try to complete the job cost effectively. Ensure the financial statements are error free. Create meaningful learning experiences for subordinates. Display an a b i l i t y to identify problems. Manage subordinates. Try to u t i l i z e available personnel effectively. Try to complete the job in a timely manner. Communicate effectively with clients. Abide by the professional code of ethics. Provide supervision for subordinates. Inform the client about the execution of the audit. Communicate effectively with superiors. Provide motivation for subordinates. Accept responsibility for the work of others. Impose r e a l i s t i c requirements on subordinates. Develop team relationships.  Identification # of items the firm would want their managers to comply with at the FIELDWORK STAGE of the audit.  Highest Rank Next Highest  Lowest Rank  Appendix IV (continued)  94.  THE WORK OF THE AUDIT MANAGER AT THE EVALUATION STAGE Rankings Based.on What the Firm Would Want Managers to Comply with at the Evaluation Stage of the Audit Please rank the items listed 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. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.  Accept the responsibility for the work of others. Try to complete the job cost effectively. Display loyalty to the firm. Develop technical s k i l l s in a specialist area. Provide management type advice and services. Display an a b i l i t y to identify problems. Provide supervision for subordinates. Complete the management l e t t e r . Sell additional services to existing clients. Provide motivation for subordinates. Try to meet c r i t i c a l deadlines. Communicate effectively with clients. Adapt to superiors. Provide technical assistance to subordinates. Provide instruction for subordinates. Communicate effectively with superiors. Develop a f a m i l i a r i t y with the audit f i l e . Try to complete the job in a timely manner. Create meaningful learning experiences for subordinates. Abide by professional code of ethics. Manage subordinates. Ensure the financial statements are error free. Reconcile technical differences with c l i e n t .  Identification # of items the firm would want their managers to comply with at the EVALUATION STAGE of the audit.  Highest Rank Next Highest  Lowest Rank  Appendix IV (continued)  95.  COMPLETING THE REST OF THE QUESTIONNAIRE The next three rankings you w i l l be asked to provide w i 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 like 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 listed 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. Chart to aid Ranking:  Task Requirements a) b) c) d) e) f)  Drive within speed limits Try to meet timetables Distribute transfers quickly Display courtesy to passengers Drive safely Announce street names  Highest Rank Next Highest  Lowest Rank  NOTE:  I have found that bus drivers always drive within speed limits hence item 'a' is in the top most box. The next 2 items are 'c and ' f . They are ranked equally. Of the 6 items l i s t e d , I find that bus drivers t r y to meet timetables least frequently; hence item 'b' is in the bottom most box. 1  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 listed 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. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.  Try to u t i l i z e available personnel effectively. Inform the client about the execution of the audit. Try to complete the job cost effectively. Develop a f a m i l i a r i t y with' the audit f i l e . Display loyalty to the firm. Communicate effectively with superiors. Provide motivation for subordinates. Communicate effectively • with clients. Provide supervision for subordinates. Try to meet c r i t i c a l deadlines. Try to complete the job in a timely manner. Manage subordinates. Develop team relationships. Display an a b i l i t y to identify problems. Provide instruction for subordinates. Create meaningful learning experiences for subordinates. Accept responsibility for the work of others. Impose r e a l i s t i c requirements on subordinates. Provide technical assistance to subordinates.  Highest Rank Next Highest Identification # 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 listed 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. 2. 3. ;. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. ' 18. 19. 20. 21. 22. 23. 24.  Provide technical assistance to subordinates. Try to meet c r i t i c a l deadlines. Maintain professional appearances. Provide instruction for subordinates. Develop technical s k i l l s in specialist area. Display loyalty to the firm. Reconcile technical differences with c l i e n t . Develop a f a m i l i a r i t y with the audit f i l e . Try to complete the job cost effectively. Ensure the financial statements are error free. Create meaningful learning experiences for subordinates. Display an a b i l i t y to identify problems. Manage subordinates. Try to u t i l i z e available personnel effectively. Try to complete the job in a timely manner. Communicate effectively with clients. Abide by ;the professional code of ethics. Provide supervision for subordinates. Inform the client about the execution of the audit. Communicate effectively with superiors. Provide motivation for subordinates. Accept responsibility for the work of others. Impose r e a l i s t i c requirements on subordinates. Develop team relationships. Highest Rank Next Highest  Identification # 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 listed 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. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.  Accept the responsibility for the work of others. Try to complete the job cost effectively. Display loyalty to the firm. Develop technical s k i l l s in a specialist area. Provide management type advice and services. Display an a b i l i t y toi identify problems. Provide supervision for subordinates. Complete the management l e t t e r . Sell additional services to existing clients. Provide motivation for subordinates. Try to meet c r i t i c a l deadlines. Communicate effectively with clients. Adapt to superiors. Provide technical assistance to subordinates. Provide instruction for subordinates. Communicate effectively with superiors. Develop a f a m i l i a r i t y with the audit f i l e . Try to complete the job in a timely manner. Create meaningful learning experiences for subordinates. Abide by professional code of ethics. Manage subordinates. Ensure the financial statements are error free. Reconcile technical differences with c l i e n t .  Identification # of items managers actually do at the EVALUATION STAGE of the audit.  Highest Rank Next Highest  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 i e n t ' s state of affairs but cannot pinpoint any specific problems. The audit has gone well - all tests have checked out, a l l deadlines have been met, and the audit is well within budget. The client 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 line. Commit Resources  Do Not Commit 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 profession. He has often told you that the integrity 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 l e x i b l e ' enough. You are advised by your friend that QWER is looking for a change of auditors 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 commodities. PLEASE INDICATE HOW INCLINED YOU ARE TO PUT THIS CLIENT UP FOR CONSIDERATION BY YOUR FIRM? Recommend QWER  Do Not Recommend QWER  3. The senior on the job has come into conflict with the c l i e n t ' 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 i e n t . 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 l o t 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 Firm Policy  Negotiate a 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 firm:  4.  In five years do you expect to * Be working for this firm * Be working for another CA firm * Be working for a non CA firm (Please tick the most appropriate response)  5.  I f you were offered the opportunity of winning $500 on the toss of a f a i r coin, how much would you be prepared to pay to p a r t i c 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 firm. buted.  A total of 160 questionnaires were d i s t r i -  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 distributed with the research material. were returned.  Overall, 127 usable questionnaires  This represents a response rate of over 79% which is  high for studies of this nature.  The distribution and response rate from  each firm are shown in Table IV below.  B.  Statistical Analysis Hypotheses Cl and C2 focus on the degree of agreement which exists  within partner and manager groups in a given firm.  Specifically, 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 w i 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 communa 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.  If  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 s t a t i s t i c 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 calculating the Spearman-rank order correlation coefficients (designated r 1 among all possible pairs of rankings and finding the average value (designated r ) .  This l a t t e r value is linearly 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 r m ,  the rankings provided at each stage of the audit were not independent. 1  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 significant 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 directly. value of r  g  In order to f a c i l i t a t e interpretation, the average  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 r  g  g  ranged from .183 to .276, at the fieldwork stage  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 coefficient for all 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 assignment 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 identified. 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 c a l deadlines'which is used for the  test of the resource allocative hypothesis at that stage.  As a result,  particular care w i 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 f a m i l i a r i t y 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 client 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 was noted at both the fieldwork 1  and evaluation stages as a cue with a high variation in the rank assignment 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 i v e l y , variations in the ranks assigned to a given cue may indicate that the p r i o r i t y 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 certain questions at the end of the research material.  The cue 'Ensure the  financial statements are error free was questioned by a number of respond1  ents who sought c l a r i f i c a t i o n 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 conformity 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 v a l i d i t y 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 significant W s t a t i s t i c s 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. significant at the .001 level.  Each W coefficient was  Once again, the average Spearman rank  order correlation coefficient associated with each W coefficient is presented.  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 all their responses.  Cues in which there  was a high degree of variation in the rank assignments were noted.  These  l a t t e r 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 client about the execution of the a u d i t ' , 'Develop a f a m i l i a r i t y with the audit f i l e  1  and  'Try to meet c r i t i c a l deadlines' at the design stage and the cues 'Abide by the professional code of e t h i c s ' , 'Ensure the financial statements are error free' and 'Develop a f a m i l i a r i t y with the audit f i l e ' at the f i e l d 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 criterion above.  However, those with the highest variance are the same  as those which were identified 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 t a t i s t i c a l l y significant degree of agreement as to what managers should do at a given stage of the audit i f their behaviour is to be consistent with f i r m 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 incentive 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 all partners and another to summarize the responses of a l l managers within a given firm.  The v i a b 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 Qsort 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 reliable.  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 coefficient (.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 typically 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 firm. Hypothesis C4 focuses upon the degree of agreement between managers within a given firm as to the sort of behaviour that they typically 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 typically adopt at each stage of the audit.  The coefficients of concordance pre-  sented in Appendix V I I I are each s t a t i s t i c a l l y significant 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 managers participating in this study. The Spearman coefficient associated with each W s t a t i s t i c in Appendix V I I I was computed.  This data is also presented in that appendix.  the design stage the average r  g  At  over.all possible pairs of rank orders in  each firm varied from .182 to .204, at the fieldwork stage the average r •varied from .225 to .361:and at the evaluation stage the average r  g  varied  g  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  all 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 criterion identified 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 client about the execution of the audit and 1  'Develop a f a m i l i a r i t y 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 f a m i l i a r i t y 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 r e e ' . the same as those identified previously.  These cues are  The only cue which had a high  degree of variation in the rank assigned to i t by managers which w i l l be subject to further analysis is the cue 'Try to meet c r i t i c a l deadlines' and in the evaluation of the hypothesis to which that cue relates care w i 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-  c a l l y , 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 earlier. results are presented in Appendix IX.  The specific  The coefficient of concordance  computed at each stage of the audit was uniformly high, and each was significant 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 s t a t i s t i c 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 variable for the analysis of the data derived in this study. Hypothesis C6 examines whether manager groups from different firms  Ill.  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 earlier. 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 all 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 i o n , at each stage of the audit there was significant agreement amongst the managerial group as to the sort of behaviour they would typically 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 c a l deadlines' did have a high degree of variation associated with i t s rank assignment and this cue w i l l be used in subsequent analysis.  Particular care w i 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 discrepancies 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 failure on the part of respondents in this study to understand the environmental 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 t a t i s t i c a 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 necessary for an F test are ever' f u 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 distributions depart considerably from normal, provided that the number of observations in each sample is relatively large and Lindquist (1953, pp. 7886) has indicated that the F distribution is "amazingly insensitive" to the form of the distribution,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 distribution of F which is independent of the v a l i d i t y of the null hypothesis i t s e 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  t o t a l , there were eight applications of the analysis of variance test procedure. .For each test the x coefficient was computed, to one.  x was always close  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 t h i s , the F computed using  analysis of variance procedures is l i k e l y to be s l i g h t l y 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 independent 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 i s .  While  i t is clear that a Q-sort with 20 items does not have 19 degrees of freedom associated with i t , and thus the analysis of variance procedure is v i t i a t e d , the number of potential ranking combinations is enormous. A rank order scale with only ten items has 10! or over 3 million 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 l o s t , and thus the independence assumption violated, the possibilities 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 f a c i l i t a t e 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 s t a t i s t i c a l control. Hypothesis SI focuses on the resource allocative a c t i v i t i e s of the audit manager.  Because such a c t i v i t i e s can be readily monitored within  audit firms, i t is hypothesized that managers w i l l give a higher p r i o r i t y to conformance to such a c t i v i t i e s 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 significantly higher than that assigned to those cues by their partners.  In Appendix X I , 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.  s t a t i s t i c s 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 t a 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 p r i o r i t y to conformance to resource allocative a c t i v i t i e s 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 c a l deadlines'. Such variations were identified in rank assignments at the design stage of the audit.  Test of Hypothesis SI above includes that  high variance cue in the definition of the resource allocative construct at the design stage.  The analysis of variance test procedure was  conducted with the high variance cue specifically excluded when the test of Hypothesis SI was conducted at the design stage.  Significant d i f f e r -  117. ences in partners' and managers' rank assignments were s t i l l observed, however, the significance of the test s t a t i s t i c was s l i g h t l y reduced (F = 3.69 p < 0.10). Hypothesis S2 focuses on the role of a manager as a negotiator within the firm.  Like resource allocative a c t i v i t i e s , the negotiation a c t i v i t i e s  of the manager are relatively easy to monitor within the organization and as a result i t is hypothesized that managers w i l l give a higher p r i o r i t y to conformance to those a c t i v i t i e s than their partners would want them to.  The rank assigned to the cue which describes the manager's negotia-  tion a c t i v i t i e s w i l l , therefore, be higher when that assignment is by managers than by partners. In Appendix X I 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 s t a t i s t i c s 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 significant 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 a c t i v i t i e s of the audit manager which relate to his role as a negotiator within the audit firm become so c r i t i c a 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 a c t i v i t i e s of the manager are  more d i f f i c u l t to monitor than either the negotiation or resource allocative a c t i v i t i e s of the manager and hence i t is hypothesized that managers w i l l give a lower p r i o r i t y to conformance to entrepreneurial responsibili t i e s than their partners would want them to. at the three stages of the audit.  This hypothesis was tested  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 s t a t i s t i c s is set out below. TABLE VII Partner-Manager Differences in the Rank Assignment of the Entrepreneurial Cues Design Stage  Fieldwork Stage  F = 21.72  F = 6.38  F = 0.85  p < .0005  p < .025  p < .5:'  Evaluation Stage  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 a c t i v i ties 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 a c t i v i t i e s 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 typically 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 significant 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 f e r ences were not s t a t i s t i c a l l y significant.  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 p r i o r i t y to conformance to those organizational rules which can be readily monitored by their superiors. Conversely, managers give a lower p r i o r i t y 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.  121.  APPENDIX V Test of Hypothesis Cl Communality of Partners' Rank Assignments* Design Stage Firm A**  W=.323  W=.503  X =70.54  X =l18.86  X =177.06  p < .001  p < .001  p < .001  r =.183 s  r =.278 s  r =.470 s  W=.332  W=.289  W=.381  X =77.69  X =86.41  X =108.97  p < .001  p < .001  p < .001  r =.276  r =.230  r =.329 s  W=.340  W=.374  X =61.11  X =86.00  W=.546 X =120.01  p < .001  p < .001  p < .001  r =.267  r =.304  r = 496 s  W=.248  W=.307  W=.437  X =89.12  X =141.17  X =192.28  p < .001 r =.208  p < .001  p < .001 r =.407 s  m=17  2  m=13  s  Firm C  Evaluation Stage  W=.231 2  Firm B  Fieldwork Stage  2  m=10  s  Firm D  2  m=20  s  2  2  s  2  s  2  r =.271 s  2  2  2  2  Average Spearman Correlation Coefficient for all possible pairwise correlations for partners from different firms. * Statistical 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 Firm A  Fieldwork Stage  Evaluation Stage  W=.232  W=.329  X =62.69  X =l13.44  W=.508 X =167.64  m=15  p < .001 r =.177 s  p < .001 r =.281 s  p < .001 r =.473 s  Firm B**  W=.281  W=.353  X =65.75  X =97.43  W=.494 X =141.28  p < .001  p «.001 r =.294 s  p < .001 r =.452 s W=.412 X =154.18  W=.455  2  2  m = 13  Firm C  r =.221 s W=.400 X =122.37  2  2  m=17  p < .001 r =.363 s  W=.410 X =160.31 p < .001 r s=.373  Firm D  W=.219 X =78.76 p < .001 r s =.178  W=.275 X =126.50 p < .001 r s=.237  46  .296  2  2  m=20 Average Spearman Correlation Coefficient, for all possible pairwise correlations for partners from different firms.  2  2  2  2  2  2  p < .001 r =.375 s X =200.11 p < .001 r =.426 2  s  .430  * Statistical 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 ( s )* r  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 Firm A  ' Firm B  Firm C  Firm D  Fieldwork Stage  r=-= .843 s  r  p < .001  p < .001  p < .001  r  r  r s = .880  s  = .804  s  $  = .738  Evaluation Stage  = .695  r  s  = .832  p < .001  p < .001  p < .001  r  r  r  s  = .760  s  = .803  s  = .856  p < .001  p <- .001  p < .001  r  r  r  s  = .620  p < .01  s  = .527  p < .01  = .851 s p < .001  * Statistical tests are based on Ferguson, 1959, pp. 179-182, p. 308; Roscoe, 1975, pp. 106-110. Significance levels stated are all based on two-tailed tests.  124.  APPENDIX V I I I Test of Hypothesis C4 Communality of Managers' Rank Assignments* Design Stage Firm A**  Evaluation Stage  W=.233  W=.316  W=.512  X =67.10  X =116.29 p < .001  X =168.96  2  m=16  Fieldwork Stage  p < .001 r =.182 s  2  2  r =.270  p < .001 r =.477 s  W=.265 X =62.01  W=.410 X =122.59  W=.463 X =132.42  p < .001 r =.204 s  p < .001  p < .001  V.361  r =.418 s  Firm C  W=.251  W=.368  W=.457  m=17  X =76.77 p < .001 r s=.204  X =143.97 p < .001 r =.329  X =170.96 p < .001 r s=.423  W=.244 X =87.92  W=.264  W=.444 X =195.18 p < .001 r =.414 s  Firm B  2  m=13  2  Firm D  2  m=20  p «.001 r =.204 s  Average Spearman Correlation Coefficient for all possible Q3 pairwise correlations, for managers from different f i rms. 2  s  2  2  s  X =121.30 p < .001 r =.225 s 2  .303  2  2  2  .425  * Statistical 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  X = 62.15 2  p < .001 r  s  = .817  FIELDWORK STAGE: W = .839 X = 77.19 2  p < .001 r  s = .785  EVALUATION STAGE: W = .918 X = 80.78 2  p < .001 r  c  = .891  s  * Statistical 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  X = 65.88 2  p < .001 Fs = .887 FIELDWORK STAGE: W = .868 X = 79.86 2  p < .001 r  = .824  s  EVALUATION STAGE: W = .988 X = 86.94 2  p < .001 r„ = .984  *Statistical 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  Partners  A 17  B 13  C 10  D 20  A 15 .35  B 15.08  C 15.90  D 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  15 .33  14 78  19.95  FIELDWORK STAGE:  Partners  Number of Subjects FIRM A B C D 16 13 10 20  Mean Responses* FIRM B C 13. 31 13. 54 15. 20  14. 85  14 23  Managers  16  13  17  20  13. 38  12. 92  14. 00  13. 40  13 43  32  26  27  40 i  13. 35  13. 23  14. 60  14. 13  13. 83  10. 60  10.61 9.24 9.93  EVALUATION STAGE:  Partners  Number of Subjects FIRM A B C D 16 13 10 20  Managers  15  13  17  20  9.13  9.46  9.41  8. 95  31  26  27  40  9.91  10.35  9.66  9. 78  A 10.69  Mean Responses* FIRM B C 11.23 9.90  * 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  1  70.96  2.85  3 3  6.37 0.95  92.93 1358.13  7 118  11.51  1451.06  125  F  Significance  Between Groups: Role Firm RxF Within Groups: FIELDWORK STAGE: Component Between Groups: Role Firm R XF Within Groups:  70.96 19.12  Sum of Squares  Degrees of Freedom  Mean Square  19.05 37.95  1 3  19.05 12.65  10.20 67.20 954.29 1021.49  3. 7 117 124  3.40  6.17 0.55 0.08  F 2.34 1.55 0.42  p < .025 NS NS  Significance p < .25 NS NS  8.15  EVALUATION STAGE: Component Between Groups: Role Firm RxF Within Groups:  Sum of Squares 54.96 7.78 7.73  Degrees of Freedom  70.47 385.48  1 3 3 7 116  455.95  123  Mean Square  F  54.96 2.59 2.58  16.55 0.78 0.78  Significance p < .0005 NS NS  3.32  * Statistical 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 X I I I Test of Hypothesis S2 The Manager as a Negotiator. Results by Role and Firm FIELDWORK STAGE: Mean Responses* FIRM  Number of Subjects FIRM Partners  A 16  B 13  C 10  D 20  A 3.;44  B 3.62  C 3.:30  D 4.00  3.59  Managers  16  13  17  20  3.56  2.62  3.06  3.25  3.12  32  26  27  40  3.50  3.12  3.18  3.63  3.36  EVALUATION STAGE: Number of Subjects FIRM  Mean Responses* FIRM  Partners  A 16  B 13  C 10  D 20  A 2.69  B 2.69  C 2.90  D 2.40  2.67  Managers  15  13  17  20  2.40  2.46  2.71  2.70  2.57  31  26  27  40  2.55  2.58  2.81  2.55  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 RxF Within Groups:  Sum of Squares  Degrees of Freedom  Mean Square  6.49  1 3 3 7  6.49  117 124  1.48  7.35 4.60 18.44 172.83 191.27  2.45 1.53  f_  Significance  4.39 1.66 1.04  p < .05 p < .25 NS  EVALUATION STAGE: Component Between Groups: Role Firm RxF Within Groups:  Sum of Squares  Degrees of Freedom  Mean Square  F  Significance  0.09 1.16 1.85 3.10  1  0.09 0.39 0.62  0.11 . .46 .73  NS NS NS  98.47 101.57  116 123  3 3 7  0.85  * Statistical 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  Partners  A 17  B 13  C 10  D 20  A 2.06  B 1.69  C 1.80  D 2.35  Managers  16  13  17  20  3.31  3.31 • ; 2.71  2.85  33  26  27  40  2.69  2.50  2.60  3.04 2.51  8.33  2.26  1.98  FIELDWORK STAGE: Number of Subjects  Mean Responses'*  FIRM  FIRM  Partners  A 16  B 13  C 10  D 20  A 9.00  B 8.23  C 7.70  D 8.40  Managers  16  13  17  20  8.69  9.15  8.76  9.65  36  26  27  40  8.85  8.69  8.23  9.02  9.06 8.70  EVALUATION STAGE:  Partners.  Number, of Subjects FIRM A B C D 13 10 20 16  Mean Responses^ FIRM A B C 18 .06 17. 23 17.70  15  13  17 • 20  17 .73  18. 46  31  26  27  17 .90  17. 85  40  D 18. 15  17 79  18.06  18. 45  18 17  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:  Sum of Squares  Degrees of Freedom  Mean Square  F  Role  34.31  1  34.31  21.72  Fi rm RxF  3.12 5.12  3  1.04  .66  NS  1.71  1.08  NS  186.15  3 7 118  228.70  125 F  Significance  Component Between Groups:  42.55 Within Groups: FIELDWORK STAGE: Component Between Groups: Role Firm RxF Within Groups  Sum of Squares 15.82 10.31 11.18 37.31 289.95 327.26  Degrees of Freedom 1 3 3 7 117 124  Significance p < .0005  1.58  Mean Square 15.82 3.44 3.73  6.38 1.39 1 .50  p < .025 p < .25 p < .25  2.48  EVALUATION STAGE Sum of Component Squares Between Groups: 4.44 Role 4.02 Firm RxF 9.15 Within Groups:  Degrees of Freedom*  Mean Square  f.  Significance  1 3  4.44 1.34 .3.05  0.85 0.26 0.58  NS NS NS  3  17.61 605.96  116  623.57  123  7  5.22  * Statistical 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 c a l task requirements is provided. This checklist details the c r i t i c a 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 c a 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 confronting non-partners in public accounting firms and an assessment was made as to the way in which that system is l i k e l y 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 w i l l give a higher p r i o r i t y to conformance to those organizational rules which can be readily monitored by their superiors.  Conversely, managers w i l l give a lower  134.  p r i o r i t y to conformance to those rules which cannot be readily monitored or which can only be monitored at a significant, often prohibitive cost. Tests of this basic hypothesis made use of a Q-sort methodology within a f i e l d 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 t h i s , are strong in realism, significance, strength of variables and heuristic qualities, but weak in s t a t i s t i c a l 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, risk taking propensity, and so f o r t h . 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 allocative, negotiation and entrepreneurial cues in the Q-sort.  The correlational analysis was disappointing, few s i g n i f i -  cant relationships were obtained and many were not in the direction that would be expected in the l i g h t 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 generalizability from the results of the hypothesis testing stage of the study can be made due to restrictions 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 t s 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 l o 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 l u s t r a t i v e example from a f i e l d unrelated to auditing was provided in the instructions.  Care was taken to  ensure that the i l l u s t r a t i v e 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 s t a t i s t i c s were computed.  These s t a t i s t i c s 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. in Appendix IV to Chapter 3. posed.  These questions are detailed  Specifically, three decision contexts were  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 firm.  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 firm.  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 t a t i s t i c a l l y significant 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 c a 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 insufficient 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 c u l t y respondents seemed to have in dealing with these l a t t e r 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 respondents cannot be attributed to a failure on their part to understand either the research problem or the research context with which they are required to deal.  The systematic discrepancies between'partners and managers' 1  responses in Chapter 4 are all the more significant 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 p r i o r i t y to conformance to those organizational rules which can be readily monitored by their superiors.  Conversely, they give a lower p r i o r i t y 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 v a r i a b i l i t y 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 f a c i l i t a t e d by understanding the nature of the economic incentives confronting organizational p a r 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 Firm A Resource Allocation:  Evaluation Stage  Negotiation: Fieldwork Stage  Evaluation Stage  Entrepreneurial Design Stage  Fieldwork Stage  Evaluation Stage  r  = -0.10  Firm D  r  n = 33 NS r = -0.18  n = 25 p < .1 r = -0.02  n = 27 NS r = -0.05  n =40 NS r = -0.21  n = 32  n = 25  n = 27  n = 40  NS r = -0.52  NS r = -0.26  NS r = 0.15  p < .1 r = -0.41  n = 31 p < .005  n = 25 NS  n = 27 NS  n = 40 p < .005  r = .075 n = 32 NS r = -0.19  r n p r  r = -0.11 n = 27 NS r = -0.30  r n p r  n = 31 NS  n = 25 NS  n = 27 p < .1  n = 40 NS  r = .48 n = 33 p < ^005 r = -0.03 n = 32 NS r' = -0.285 n = 31 p < .1 -  r = .54 n = 25 p < .005 r = -0.02 n = 25 NS r = .18 n = 25 NS  r = .42 r = 0.04 n = 27 n = 40 p < .025 ' NS r = .23 r = 0.27 n = 27 n = 40 NS p < .05 r = -0.23 r = -0.003 n = 27 n = 40 NS NS  g  g  g  g  g  g  s  s  = -0.272  Firm C  r --= -.022 s  Fieldwork Stage  Firm B  g  g  g  s  g  g  g  = -0.37 = 25 < .1 = -0.11  $  g  g  g  g  r  g  = -0.10  g  g  $  s  $  g  g  g  s  g  = -0.39 = 40 < .01 = 0.16  * Statistical tests are based on Ferguson, 1959, pp. 179-182, p. 308, Roscoe, 1975, pp. 106-110. Significance levels stated are all 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 . , "Statistical 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 Institute of Certified Public Accountants. ing Standards No. 1. New York: AICPA, 1973. Anderson,' R.J.', The External Audit. Publishing Co., 1977.  Volumes 1 and 2.  Statement on AuditToronto:  Pitman  Ashton, R.J., "An Experimental Study of Internal Control Judgments," Journal of Accounting Research, Vol 12 No. 1 (Spring 1974), pp. 143157. Baker, C.R., "Management Strategy in a Large Accounting Firm," The Accounting 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 Materiality," 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. 481493. , 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 t s Methodology," Perceptual and Motor S k 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 Hill Book Co., 1970. Cascio, W.F., Applied Psychology in Personnel Management. Reston Publishing Company, Inc., 1978. C.I.C.A., C.I.C.A. Handbook.  Toronto:  Virginia:  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 Statistics 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 Hill 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 Attribute," 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 Distributions," The Accounting Review, Vol. LI No. 4 (October 1976), pp. 800807.  142. Ferguson, G.A., Statistical 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 Critical Incident Technique," Psychological B u l l e t i n , 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 British 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 British Columbia, September 1979(b). Hays, W.L., Statistics for Psychologists. Winston, 1963.  New York:  Holt, Rinehart and  Henderson, B.D. and Dearden, J . , "New Systems for Divisional Control," Harvard Business Review, Vol. 44 No. 5 (Sept/October 1966), pp. 144161. 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. 1970.  London:  G r i f f i n and Co., L t d . ,  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. Holt, Rinehart and Winston, Inc., 1973.  New York:  Kinney, W.R. and R i t t s , 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. Francisco: W.H. Freeman and Company, 1974. Lindquist, E., Design and Analysis of Experiments. M i f f l i n , 1953.  Boston:  San  Houghton and  Magnusson, D. and Endler, N.S. (eds.), Personality at the Crossroads: Current Issues in Interactional Psychology. Hillsdale, 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. Publishers, 1973.  New York:  Harper and Row,  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 Regulations 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. Basic Books Inc., 1966.  New York:  P a t t i l l o , J.W., "Materiality: The (Formerly) Elusive Standard," The Financial Executive, Vol. 43 (August 1975), pp. 20-27. P a t t i l l o , J.W. and Siebel, J.D., "Factors Affecting the Materiality Judgment," CPA Journal, Vol. XLIV No. 7 (July 1974), pp. 39-44. Peat, Marwick, Mitchell and Co., Research Opportunities in Auditing. York: Peat, Marwick, Mitchell and Co., 1976.  New  Rhode, J.G. and G r 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 Interview," 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 Statistics 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. LXIII No. 2 (May 1973), pp. 134-139. Siegel, S., Nonparametric Statistics. 1956.  New York:  McGraw Hill Book Co.,  Sorensen, J.E. and Sorensen, T.L., "The Conflict of Professionals in Bureaucratic Organizations," Administrative Science Quarterly, Vol. 19No. 1 (June 1974), pp. 98-106. Spiegelman, M., Terwilliger, C. and Fearing, F., "The Reliability 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 Press, 1976.  Alabama:  The University of  Ward, B.H., "An Investigation of the Materiality Construct in Auditing," Journal of Accounting Research, Vol. 14 No. 1 (Spring 1976), pp."138152. Warwick, D.P., A Theory of Public Bureaucracy: P o l i t i c s , 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 Determination 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 R e l i a b i l i t y : Accuracy, Consensus, and the Usefulness of a Simulation Decision Aid," Journal of Accounting Research, Vol. 16 No. 2 (Autumn 1978), pp. 368387. Whitney, D.R., Elements of Mathematical Statistics. Holt and Co., 1959.  New York:  Henry  Willingham, J.J. and Carmichael, D.R., Auditing Concepts and Methods. New York: McGraw Hill Book Co., 1979. Wittenborn, J.R., "Contributions and Current Status of Q-Methodology," Psychological B u l l e t i n , Vol. 58 No. 2 (March 1961), pp. 131-142. Woolsey, S.M., "Development of Criteria to Guide the Accountant in Judging Materiality," 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.  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0095404/manifest

Comment

Related Items