Printer Friendly

The effect of relationship and reward on reports of wrongdoing.

SUMMARY

Failure to detect fraud results in serious consequences for external auditors. Increasingly, professional guidance and evidence point to the importance that communication plays in detecting fraud. The research reported in this paper investigated several auditing experts' propositions about communication's role in detecting fraud: (1) communication from client personnel is important in detecting fraud; (2) the likelihood of receiving sensitive communications from client personnel is heavily dependent on the strength of the relationship between the auditor and the person knowledgeable of any wrongdoing; and (3) the willingness to communicate is influenced by the knowledgeable person's understanding of who might be benefited by the wrongful act--the perpetrator, the organization to which the perpetrator belongs, or both. This last matter is relevant to reporting on management fraud because the fraud often involves (at least) short-term benefit to the perpetrator, the company, and the knowledgeable observer.

The experts' suggestions and previous research related to communication of wrongdoing resulted in our securing input from 20 highly experienced partners. Five partners served as expert consultants prior to execution of a controlled experiment which used auditing students. The other 15 participated in telephone interviews to establish a check on our experts' perceptions. The experiment examined two propositions: (1) strength of relationship between an observer of wrongdoing and a potential report recipient affects the likelihood that the wrongdoing will be reported, and (2) the likelihood of a report on the wrongdoing is impacted by who will benefit from the misdeed. The appropriateness of using a student setting is discussed.

The experimental results, which came exclusively from students, indicated that stronger relationships between the observer of wrongdoing and the potential report recipient resulted in more reports. The results showed no relationship between the beneficiary of the wrongdoing and the frequency of reports. Partners' experiences and perceptions clearly support the experimental variables' relevance to the external audit environment and the importance of communication. In particular, their collective experiences and perceptions support the proposition that the stronger the relationship between the auditor and client personnel, the greater the likelihood of receiving a report about sensitive matters related to the financial statements.

Key Words: Fraud detection, Communication.

Data Availability: Instruments are available from the authors. Due to confidentiality agreements made with the university Institutional Review Board, the data are not available to others.

"This is a relationship business."

--Richard Measelle, former managing partner,

Arthur Andersen (Green and Barrett

1994)

Voluntary communication about a wrongful act by a witness to that act is gaining increasing attention as an important source of fraud detection. Two KPMG Peat Marwick (1993, 1995) fraud surveys indicate that communications of this nature, sometimes as anonymous reports, are a significant source for discovering organizational wrongdoing. An Association of Certified Fraud Examiners (1995) survey shows that over one-third of financial statement frauds are detected as a result of suspicions of individuals. Ponemon (1994) points out that several notorious frauds have been uncovered by "tips" from client personnel. Hooks et al. (1994) cite the relationship between auditor and client as a potential instigator of sensitive communications and encourage research exploring operational effects of these relationships.

The study reported here, which deals with the likelihood that observers will report a wrongdoing, is relevant and important to the external auditing environment. Expert external auditors believe that auditor-client relationships affect the likelihood of discovering fraud. (1) This is consistent with the evidence cited above that shows communications from others to be the most common way to discover fraudulent financial reporting (FFR). Further, improved communications is efficient and effective vis a vis alternatives. The purposes of the study were to provide empirical data on variables affecting reports of wrongdoing and to test behavioral theories suggested in the literature that may be relevant to auditing.

The need for information about organizational wrongdoing in auditing contexts is well recognized. There is little debate about the importance of auditors' failure to detect management fraud (e.g., Palmrose 1987; Cook et al. 1992); high costs have highlighted its importance. Changes in professional guidance within the last decade emphasize assessments of client characteristics that involve significant subjectivity. Two examples are required assessments of management integrity (SAS No. 53, AICPA 1988a) and the control environment (SAS No. 55, AICPA 1988b). (2) Research findings suggest that these two factors may play the most important roles in differentiating between audits where management fraud is present or absent (Bell et al. 1993). Further, COSO (1992) points to client communications as a major component of the control environment.

The value of facilitating communications can reasonably be extended to the topic of fraud detection by external auditors--especially given the inefficiency of methods in existing guidance and other prospective approaches (Albrecht and Willingham 1993; SAS No. 53, AICPA 1988a, section 17). If the strength of an auditor-client relationship affects the auditor's likelihood of receiving communications about problems and wrongful acts, then a strengthened relationship may provide a low-cost enhancement to reducing audit risk.

Research findings suggest several variables that influence the occurrence of sensitive communications. However, studies conducted in varied settings with different methodologies are inconclusive. (3) This study investigates two of these variables. Specifically, this inquiry investigates how two variables affect communication of wrongdoing: (1) the strength of relationship between the person with knowledge of the wrongdoing and the potential report recipient of that information, and (2) the knowledgeable person's understanding of who might be benefited from the wrongful act. In an audit setting, the audit partner parallels the potential report recipient in this study.

The two variables investigated were proposed as most important and worthy of study by expert partners and are consistent with constructs recommended for further research (Miceli and Near 1985). Although communication experts (e.g., Graham 1986; Miceli and Near 1992) treat cost-benefit perceptions as a major element in their theoretical models underlying the propensity to communicate wrongdoing, there is little research on how the strength of the relationship between one knowledgeable of a wrongdoing and an interested party affects the knowledgeable party's willingness to communicate. Research findings about the influence of personally benefiting from fraud committed by another on the tendency to disclose the fraud are mixed (Trevino and Victor 1992). As Miceli and Near (1992, 172) point out, in situations involving management fraud, the fraud often benefits both perpetrators and witnesses.

IMPETUS FOR THE STUDY AND PRACTITIONER ASSISTANCE

The primary motivation for this study came from a discussion with a Big 6 partner in charge of risk containment for his region and from a comment by Steven J. Aldersley in his role as discussant at the University of Southern California Audit Judgment Symposium (see Hooks et al. 1994, 112). Four other Big 6 "experts" played key roles in motivating the study. Two had formal designation as risk containment experts and the other two had relevant experience. One of the partners designated as a risk containment expert was in charge of his firm's national forensic practice. The other was recently retired, having divided his final 20 years between serving as his firm's litigation expert and as director of quality control. One of the two without "expert" designation served extensively in assessing damages at financial institutions subject to Resolution Trust Corporation action. The other served as partner in charge of a Big 6 office. Three of the Big 6 firms were represented. Five partners provided ex ante input to the variables to be studied. Several reviewed the instrument or provided counsel about matters in the instrument. (4)

After executing the experimental portion of the study, we conducted phone interviews with 15 more partners to gauge the representativeness of our input from the five consulting partners and to determine if our beliefs about the environment were reasonable. In part because of the low rate of FFR, we chose partners, rather than lower ranking auditors. Their greater experience provided a longer period over which they may have experienced sensitive communications or been a member of an audit team where others had experienced such communication. Taken together, the 20 partners' experience levels ranged from 14 to 35 years; more than half had held the rank for 20 years or more. Except for two tax partners, all were audit partners; similarly, all but two were male. (5)

Before conducting the interviews, we contacted each partner by telephone or email to alert them to the study's general purpose and to arrange a convenient time. We waited until near the end of each interview to make direct inquiries about the factors studied in our experiment to avoid inducing demand effects and interfering with the partners' recall. Using a structured interview instrument, we began by indicating our interest in ways auditors learn about needed financial statement adjustments--particularly those resulting from intentional acts by management to distort financial statements. We then asked the partners to relate any personal experiences regarding intentional management distortions. We pursued the seriousness of the partners' experiences. Near the end of the interview, we asked them to assess the importance of the auditor-client relationship and related factors. (6)

Although the five consulting partners were more adamant about the role of relationships and beneficiary of the wrongdoing in encouraging sensitive communication, the input from the 15 partners subsequently interviewed was similar. Some believed that "private" communication from clients is so important to the audit process that some material issues (e.g., obscure, hidden, or unrecorded items) would likely be missed if the auditor were not told. Other partners expressed complete confidence that they would eventually uncover everything that is important. However, virtually all of the interviewed partners expressed the conviction that discovery would be delayed and resolution more difficult were they not tipped. Some believed the delay would be several years. Since it is common for major frauds to begin as minor transgressions to overcome "temporary" problems, this belief, if correct, justifies more concern for understanding facilitation of sensitive communications. When asked what they believed to be the reasons underlying client personnel decisions about reporting a wrongdoing, many of the partners expressed uncertainty. Some stated that client personnel communicated with auditors when they trusted them and believed that was the best course of action to avoid illegal activity. The interview responses supported the notion that FFR and reports of FFR occur infrequently. The next section focuses on prior research underlying the two variables selected for experimental study and contains other partner observations.

LITERATURE REVIEW AND HYPOTHESES

Hooks et al. (1994) develop a communication model directed at fraud prevention and detection. Although their model is directed toward reporting fraud, it is tailored from a general model proposed by Miceli and Near (1992) and can be applied to wrongful acts that are less severe. The model relates four classes of variables to the necessary intermediate judgments that the observer of a wrongful act must make prior to reporting the act. The stages include becoming aware of the act, judging the act to be wrong, assessing seriousness, assessing reporting responsibility, and determining viable potential actions and their related costs and benefits. Within accounting, Schultz et al. (1993) show that the perceived net cost of communicating a wrongdoing plays a significant role in judgments on whether to communicate. The study reported here most directly addresses the costs and benefits of the reporting decision.

As noted, the strength of relationship variable was chosen based on the expressed beliefs of a group of expert auditing partners. The strength of the relationship between the auditor and individuals working in key client positions was believed crucial to the trust required for sensitive communication. Theory and research reinforce the validity of the partners' beliefs in the variable's expected importance (Hooks et al. 1994). The experts' input also provided support for including "beneficiary of the wrongdoing" as a variable. As discussed below, literature provided mixed support for this variable, but showed it as an important consideration regarding a decision to report management fraud.

Strength of Relationship

An important element in arriving at costs and benefits associated with the decision to report involves the prospective benefit of effective action. This belief depends on the observer's reading of existing evidence about the receiving individual's or organization's responsiveness to reports. If the observer believes that the receiver's response will be effective, evidence suggests that reporting tendency will increase (Dozier and Miceli 1985; Miceli and Near 1992). COSO (1992, 60) comes to this same conclusion.

Expected reporting costs often involve prospective retaliation. Farrell and Petersen (1982), Trevino (1986), Dozier (1988), Arnold and Ponemon (1991) and Schultz et al. (1993) report that threatened retaliation lessens the likelihood of reporting, while others (e.g., Miceli and Near 1992, 176) suggest it has little effect. A system which allows anonymous reporting seems to reduce expected costs since anonymity is associated with increased reporting frequency (Miceli and Near 1992, 176).

As the rank of the manager committing the wrongdoing increases, expected reporting costs increase and reporting frequency decreases (Dozier and Miceli 1985). With high-ranking perpetrators, external reports are favored over internal reports (Miceli and Near 1992). These conditions seem directly applicable to the environment often surrounding FFR cases. Perpetrators of FFR are typically high-ranking managers, frequency of reports is low, and the likelihood of an external (vs. an internal) report can be expected to be reasonably high. These arguments bolster the applicability of the behavioral parallel between this study and the external audit environment.

For highly sensitive communications, a trustworthy prospective report recipient can be expected to decrease reporting cost. For example, the relationship in religion between the one confessing and the one hearing the confession involves privileged communication. Similarly, the concept of privileged communication is accepted between defendants and their attorneys to encourage defendants to share sensitive information necessary for their defense. A potential reporter's perception of this trustworthiness is likely to depend considerably on the strength of his or her relationship with the prospective recipient. A stronger relationship generally creates a greater level of trust and a greater likelihood of a report.

Research to date has addressed this relationship-trust-reporting link through exploring the use of internal and external reporting channels. Most reporting occurs internally--presumably due to familiarity with, and confidence in, the report recipient. Reports to internal channels occur in supportive environments (Miceli and Near 1985) in which trustworthiness would be expected to be high. Within the internal setting and in a communication context not involving wrongdoing, research suggests that ambitious, lower ranking organizational members who trust their superiors communicate more frequently about problems (Gaines 1980). In a similar setting, Jones et al. (1975) show that confidence and trust in superiors is positively related to subordinates' perceptions of their superiors' success in upward interaction. Other work (e.g., Blackburn 1988) also supports trust as embedded in the strength of relationship construct.

Both the consulting partners and the partner interviews supported strength of relationship and trust as being important to encouraging sensitive communication. For example, one partner wrote:

regarding the trust and mutual respect that

must be developed between management

personnel and the partner ... I have had two

experiences.... One [client employee] actually

told me the details of a scam being

conducted jointly by the CEO and CFO of

a major division of a large company, and

the other merely suggested that we pay

very close attention to certain transactions.

(emphasis added)

Another stated that he considered strength of relationship with client personnel to be second in importance only to the auditor's understanding of the client's business and industry for discovering serious problems.

Another example involved a report of a major problem that came only after a client source had tipped the partner to a minor transgression. The partner carefully protected his source as he pursued audit testing of the minor transgression, avoiding any cost to the reporter. Only after passing the first "test" of protecting the reporter did the partner receive the tip related to the far more material matter. The partner believed that the reporter used the minor problem to test his trustworthiness. It appears that the reporter was attempting to reduce his personal cost by preserving his anonymity. Journalists often employ this protection to generate tips about newsworthy items and to instill trust.

These views exemplified the general sentiment that communication about sensitive matters would come only from client personnel with whom auditors have a strong relationship. The literature and partners' suggestions led to the following operational hypothesis:

H1: Observers' perceptions of the strength of their relationship with the potential report recipient will be positively related to reporting frequency.

Beneficiary

Management wrongdoing may provide a benefit to the perpetrator and to the organization. Often the benefit is short term. The benefit to the organization affects a group member's cost-benefit calculus in deciding whether to report. When organizational survival depends on the wrongdoing, group pressure is likely to reduce the chances of reporting (Greenberger et al. 1987; Near and Miceli 1985). Research results for situations involving less consequences of impropriety are mixed (Schultz et al. 1993). Trevino and Victor (1992) support the importance of group pressure in suppressing reporting tendencies in less severe cases.

Tangible personal benefits (e.g., rewards, promotions) may contribute favorably to the decision to report. Research findings are mixed. Trevino and Youngblood (1990) show that rewards to reporters favorably influence observers' reporting tendencies. Miceli and Near (1992, 1985) and Keenan (1988), find that personal monetary rewards for reporting have no significant effect. Anecdotal evidence (e.g., Naj 1992) indicates that if rewards are large, reporting frequency will be positively affected.

A number of partners supported exploring beneficiary of the wrongdoing as an independent variable. One example involved a report received from a CFO and the ensuing discussion about the CFO's decision to notify the partner. The CFO was reluctant to disclose the matter because he was concerned about betraying the CEO and the organization. Yet, if the CFO did not act to stop the deception, he feared that he would become mired in the fraud. Even though the CEO and the organization might have benefited from the misreporting in the shortrun, the CFO concluded that supporting the action through his silence was too costly, particularly since the CEO was the chief beneficiary. Several partners echoed this motivation. They asserted that they received reports because of the reporter's sense of self-preservation, despite costs associated with not supporting the company and management personnel.

These accounts from the external audit environment support the notion that potential reporters weigh the consequences of the personal costs prior to contacting an auditor. Part of this process includes balancing the cost and benefits that accrue to different parties. Previous research and the partners' contributions led to the following operational hypothesis:

H2: When the benefit of wrongdoing accrues primarily to the individual committing the wrongdoing (as opposed to his or her group), reporting frequency will increase.

EXPERIMENTAL METHOD

Overview of the Experiment

Senior auditing students were assigned to groups consisting of three individuals. In each group, one student was secretly selected and trained by the researcher to fill the role of wrongdoer. The wrongdoing was cheating on an assignment; using a calculator when it was prohibited. The other two students were constrained by assigned roles and physical seating arrangements so that only one student could see the cheating, and only the other student (who could not see it) could report it to the professor. Variables manipulated were the closeness of the relationship between the student who could see and the one who could report, and the impact of cheating on grades. For some groups, cheating would benefit only the individual grade of the cheater. For other groups the group grade would benefit. The intent was to explore the likelihood of the cheating being reported to the professor under the various alternatives. Only the perpetrator knew the actual purpose of the experiment; other participants were led to believe that the purpose of the study was to determine the effect of curriculum change.

As described above, the student experiment used a design that relied on between-subjects manipulation of relationship strength at two levels ("know well;" "not know well") and of reward structure at two levels (group; individual). The dependent variable was either a report or nonreport of wrongdoing. The participants were undergraduate auditing students nearing the end of the course. Thus, they were of similar education levels, tenure within the curriculum and had been exposed to similar group norms and conduct standards, all of which might affect the recognition and acknowledgment of a wrongdoing. Students were used as participants because of the homogeneity of these factors and because of the researchers' ability to control the variables within a natural, classroom environment--not because of the students' ability to submerge themselves into a practical business setting.

In this experiment, the communication behavior of the "observer of the wrongdoing" is being studied. The intent is to surrogate a client employee's behavior, not that of an auditor. Since the experiment focuses on employee behavior, it differs from many auditing studies which must consider the professionalism and expertise that auditors possess. The experimental setting, including the penalties and incentives, roughly parallel those found in a business environment. Additional rationale for using students rather than business people is deferred to near the end of this paper because an understanding of the experimental context is useful for that discussion.

Procedure

Figure 1 portrays the experimental process. Prior to stage 1, participants were assigned to one of four treatment levels of the two-by-two design. Each group consisted of a potential report recipient, an observer and a perpetrator. Prior to stage 1, all roles and seating assignments were determined by the researchers. Students playing the perpetrator role were trained. In Stage 1, each group member read instructions, met other group members, and was notified of the reward structure. The process was explained in the instructions so that observers would be aware of opportunities for reporting. The instructions informed participants of all group members' ostensive roles. The potential report recipient, who was designated group coordinator and charged with handling and turn-in of all materials, was specifically named. This made the coordinator the sole direct link to the professor. The emphasis on a designated report recipient, the communication process and direct link to the professor was intended to establish a viable alternative to silence.

[FIGURE 1 OMITTED]

The instructions were complex. Therefore, each member was required to sign a statement that he or she understood the process. These features were intended to increase salience and effort (e.g., Cohen 1991; Cressey 1980). The instructions emphasized the reward structure (individual or group) and that calculators were not allowed. The stated justification for the calculator ban was that results were to be compared to a group that completed the task several years earlier, also without calculators.

Stage 2 required each member to read an audit planning case, compute four commonly used analytical procedures ratios, and note areas needing investigation. The task was routine for the participants. Contrary to the announced instructions, the perpetrator cheated by using a calculator for the computations. Cheating was selected as an appropriate wrongful act because it was believed that upper-level accounting students' norms should uniformly classify it as unacceptable behavior and because it had been used effectively in other academic settings (Trevino and Victor 1992). Using the calculator surreptitiously is similar to a wrongful act in a business setting that is intended to go unseen, but such use is natural in a student context. Using a calculator may be seen as a minor issue, which is often the case with frauds that start with small transgressions and grow into material fraud. To make calculator use more consequential, time pressure was introduced. Results from pilot studies guided the time constraints. Computations were also required to be correct to two decimal places to receive any credit.

As indicated in stage 2 of figure 1, the participants' seating arrangements allowed only the observer to see the wrongdoing. This arrangement was intentional. It was to provide the observer unambiguous evidence that wrongdoing had occurred. Further, the arrangement should have encouraged the observer to accept responsibility for action since it was highly likely that no one else could see or, consequently, report the wrongdoing (Miceli and Near 1992). By reporting the wrongdoing, the observer could forestall the effect of the wrongdoing, much as a reporter would do in a business setting.

Stage 3 (see figure 1) involved group consideration and solution of the same case ratios that were previously solved individually--a typical procedure in these classes. It presented a first opportunity for the observer to report the wrongdoing. To do so, however, involved confronting the perpetrator. At the conclusion of this stage the perpetrator left the classroom. Once outside, he or she completed a questionnaire dealing with what transpired in stages 1, 2 and 3. He or she also received a written debriefing statement which explained the experiment's behavioral purpose, which appears in the appendix. (7) The appendix also contains key questions and requests addressed to each group member. The other appendix questions were included to enhance the stated purpose of the exercise.

Stage 4 emphasized an affective assessment without the perpetrator present. A portion of these issues dealt with the two participants' working relationship with the perpetrator and the fairness of the exercise (appendix items 7, 8 and 10) and were designed to increase the likelihood of reporting. The two continuing participants were students from the same section. The perpetrator came from a different section. This stage presented a second, less confrontational forum for reporting. Although the reporter would not have anonymity, which is inversely related to personal cost (Miceli and Near 1992), he or she would be afforded considerable confidentiality if the decision were made to report.

At the completion of stage 4, the observer left the classroom and completed a questionnaire outside the testing site about stages 1 through 4 and received a written debriefing statement. The questionnaire's relevant items appear in the appendix. Stage 5 was used to establish with the potential report recipient that he or she was the sole direct link to the professor. Emphasis on the link to the professor was intended to reinforce the concept that a report would likely be effective. After the observer left, the group coordinator completed an individual assessment, turned in the completed experimental package, and received debriefing materials.

Envelopes were used to manage the experimental procedures and documents. Doctoral students proctored the sessions to avoid contamination by differential professor-student relationships and to avoid demand effects from one professor-researcher. Each session was conducted in a large classroom with only six groups to minimize the chance of observation between groups. Manipulation checks on the independent variables and cross-checks on different participants' reports were employed (see appendix).

Participants

All students in each of four sections of the undergraduate auditing course participated. The students were aware that the academic unit was undergoing significant curriculum change that increased the emphasis on affective (e.g., communication, interpersonal relations) and independent learning skills. All had worked extensively in fixed-member groups of four or five throughout the semester and were accustomed to having individual- and group-grade components. This study was included as part of the announced grading structure and was represented as a benchmarking procedure to compare to past and future students' performances. As discussed later, this obvious deception was thought to be necessary in order to elicit reliable measures. Each section provided all the observers and potential report recipients for one of the four treatment cells. Each group consisted of an observer and a potential report recipient from the "home room" section and a "transient"-perpetrator student from another section. Since research indicates a low to moderate reporting tendency (e.g., Arnold and Ponemon 1991; Trevino and Victor 1992; Schultz et al. 1993), this arrangement was adopted to encourage reporting. (8) It also provided for maximum cell size. Class sizes limited the number of groups to 12 per section. Perpetrators were paid a small amount for their inconvenience of undertaking both individual training and group testing. (9)

Strength of Relationship Treatment

Effecting the strength of relationship treatment depended on the assignment of the potential report recipient and the observer to the appropriate treatment cell. As noted, all potential report recipients and observers came from the same auditing section. For the know-well treatment, students from the same fixed-member teams were assigned these two roles. For the not-know-well condition, students from different fixed-member teams were assigned these roles. The experimental sessions were conducted in the last two weeks of class. The extensive interaction that had occurred throughout the semester between students from the same teams was expected to generate a much stronger relationship than among students who were merely in the same section. The experimenters believed that this setting offered the ability to manipulate the strength of relationship that could not be matched in a business environment. This stronger relationship was expected to positively influence reporting behavior. To guard against students from the same fixed-member team having an antagonistic relationship and being categorized as "know well" for the experiment, the researcher from the experimenting university reviewed two sets of routine peer evaluations. In one case, a student was shifted to a perpetrator role to avoid a conflict-laden situation.

Beneficiary of Wrongdoing Treatment

This treatment involved the use of two reward structures. The benefit accrued primarily to either the individual perpetrator or group. For students in the individual (group) reward structure treatment, the instructions stated that 40 points out of 1,000 would be awarded for perfectly correct answers based on the individual (group) score. In addition, a bonus of one point would be awarded for each correct group (individual) answer. As evidenced in figure 1, the individual (group) score was earned in stage 2 (stage 3). Using rewards that primarily emphasize individual or group benefits paralleled fraud in a business setting which can be beneficial to only the perpetrator (e.g., in individual bonus plans) or can benefit both the perpetrator and the organization (e.g., overstatement of assets to forestall immediate insolvency).

Final grades in the course are assigned on the basis of relative rank; the grades are curved. Thus, an observer in the individual reward setting would be penalized if one member used a calculator unfairly. Conversely, an observer in the group reward setting would benefit by failing to report calculator use within his or her own group. The group members would gain points relative to those in other groups. Given students' competitiveness, this arrangement was expected to promote an elevated tendency for those students in the individual reward condition to report. The competition was expected to elevate this tendency even if students did not use the anticipated value structure which categorizes cheating as wrong. This treatment was suggested by Trevino and Victor (1992) who used curved grading vs. absolute scoring to effect a similar concept.

Data Coding

Two graduate students unfamiliar with the study's hypothesized relationships coded the data. To preserve student anonymity, final conflicts were resolved with the author from the nonexperimenting university. The emphasis on realism led to the loss of considerable data. The reasons for data loss are summarized in table 1.
TABLE 1
Data Reduction (a)

Total number of possible observations (12 groups per cell)    48
Less deleted potential observations:
Observer reported not seeing calculator use and perpetrator
  was unsure of whether he or she was observed                12
Observer or coordinator also used a calculator                 2
Observer and perpetrator were close friends                    2
Perpetrator missed session                                     1
Observer stopped perpetrator before calculator could be
  used                                                         1
Perpetrator decided not to use calculator
Total number of unusable occurrences                          19
Net number of useful observations                             29

See appendix for key questions and issues addressed to
participants that allowed for these conclusions.


The most important loss resulted from the observer's failure to see the perpetrator using the calculator. This occurred because the perpetrators were instructed to use considerable discretion to mimic reality and avoid observation by members of the five other groups in the room. (10) Any failure to observe calculator use was detected in debriefing questions answered by both the observer and perpetrator from the same group (see appendix). The observer in one group and the coordinator in another used calculators, invalidating these results in two groups (no reports were made). In two other cases, random allocation of the perpetrators resulted in pairing perpetrators and observers with strong pre-existing relationships (no reports were made). In three other cases other problems arose (see table 1).

DATA ANALYSIS AND RESULTS

The usable data are portrayed in table 2. Panel A shows the data by each level of experimental variable, reporting behavior and results from a CATMOD analysis which relies on a LOGIT procedure for determining statistical probabilities (SAS No. 1992).
TABLE 2
Report and Nonreport by Experimental Treatment

Panel A: Overall Report Distribution

                      Strength of Relationship

                 Know-Well            Not-Know-Well

            Individual    Group    Individual    Group
             Benefit     Benefit    Benefit     Benefit

Report          8           7          2           3
No Report       0           2          4           3

                   Categorical Modeling Results

Source             DF   Chi-Square   Probability

Intercept                  3.14         .076
Relationship       1       6.11         .014
Beneficiary        1       0.17         .684

Likelihood Ratio   1       2.96         .086

Panel B: Report Distribution by Strength of Relationship

            Strength of Relationship

            Know-Well   Not-Know-Well

Report         15             5
No Report       2             7

Panel C: Report Distribution by Beneficiary

               Beneficiary

            Individual   Group

Report          10        10
No Report        4         5


Manipulation checks indicate the experimental design and manipulations were successful For example, students assigned to the know-well observer roles averaged 6.0 (range 4 to 7) on a 7-point scale (with 7 indicating the closest relationship) when asked to assess the strength of their relationship with the potential report recipient. By contrast, their counterparts in the not-know-well roles averaged 2.4 (range 1 to 5), which constituted a significant difference (p = .01). Comparable data for the beneficiary treatment were 5.9 (range 3 to 7) for individual and 3.3 (1 to 5) for group, indicating a significant difference (p =.10). Because the observer-perpetrator relationship may result in differential reporting behavior, tests were also run to determine any patterns in this relationship. Using the same scale as above, observers who made a report indicated a mean of 2.95 for their relationship with the perpetrator; those observers who did not make a report indicated a mean of 2.00 (p =. 18). Perpetrators whose misdeed was reported indicated a 2.6 relationship with the observer; those whose actions were not reported indicated a 2.0 relationship (p =.45). Thus, the strength of the observer-perpetrator relationship had no influence on reporting behavior. No students indicated prior knowledge that some students might use calculators. This can be interpreted to signal that students accepted the experiment's value and did not inform students in later experimental sessions. The last section's administration registered only a 50 percent report rate compared to a 74 percent rate for the other three sections. This also supports the conclusion that later sessions were not contaminated by leakage of the experimental activity.

Chi-square tests indicate that the strength of relationship (H1) had a significant effect on reporting behavior (p = .014) and that the beneficiary of the wrongdoing (H2) did not (p = .684). When the data are pooled (panels B and C), the positive effect on reporting attributable to strength of relationship is readily observable. Conversely, the weak association of beneficiary is obvious.

Panel A also reveals that more students chose to report (20) than not report (9), which was different from an evenly split distribution (percentage reporting = 69; p = .077). This was greater than expected based on some other studies, which reported only a moderate propensity toward disclosure (e.g., Arnold and Ponemon 1991; Schultz et al. 1993). The frequency obtained in this study indicates that the procedures designed to encourage reporting behavior were successful.

DISCUSSION AND CONCLUSIONS

This section discusses the choice of methodologies and offers some conclusions from the study.

Methodology Considerations

Any audit-relevant study with student participants merits discussion of the use of students. Clearly, students are not client employees who might witness a wrongdoing or auditors who might be told about a wrongful act. For example, in this experiment the worst prospective penalty was a zero on the assignment if caught cheating (including collaboration) or serving as a witness if reporting. In practice, an actual reporter might be fired. Also, since the experiment was near the end of the semester, most of the students would have little interaction with their team members in the future. Reporting on a fellow student does not involve the same consequences as reporting on a perpetrator who, in practice, is often superior to the observer. Admittedly, there are many differences between the environments of the student in our experiment and of potential reporters in an actual business environment.

While recognizing the limitations of using student participants, we explain our rationale to use a student experiment by drawing heavily on two methodological discussions from Auditing: A Journal of Practice & Theory. One deals with experimental economics by Smith et al. (1987). The other is a forum that deals with accountability and deception (Lord 1992; Dopuch 1992; Gibbins 1992). The seminal methodological piece by Swieringa and Wieck (1982) also influenced our choice. (11) In choosing to use students, an issue of critical importance was that the experimental setting was designed to parallel variables in the students' environment that are similar to variables in the business and audit world. The student participants were not intended to cope with the forms those variables take for auditors or experienced business people.

The following discussion proceeds by reviewing fundamental scientific concerns and relating them specifically to our study's context.

Scientific research has two interrelated objectives:

to develop and test theories that

help us to understand, explain, and predict

naturally occurring phenomena; and to

provide a basis for evaluating alternative

strategies for improving practice.

Neither of these objectives can be

achieved through a priori reasoning alone.

Another necessary component of scientific

research is the systematic collection

and analysis of empirical data. (Smith et al.

1987, 71-72)

Smith et al. (1987, 72) conclude that a major reason limiting dialogue between auditing theorists and empiricists is the difficulty faced by "empiricists trying to collect and interpret field data; i.e., observations of naturally occurring events." They cite two advantages for experimental economics that relate directly to our decision to use student participants in a controlled setting. First, the experimental approach provides a way to test theoretical predictions under rigorous, replicable conditions that are high in internal validity. Second, it provides a practical and low cost means for producing evidence on environmental variables.

Campbell and Stanley (1963, 5) state:

Internal validity is the basic minimum without

which any experiment is

uninterpretable ... External validity asks the

question of generalizability.... While internal

validity is the sine qua non, and while

the question of external validity, like the

question of inductive inference, is never

completely answerable, the selection of designs

strong in both types of validity is obviously

our ideal. (emphasis in the original)

A key argument supporting the external validity of any study is the similarity of the forces within the experimental setting to those relevant forces occurring naturally. Factors that mirror the environment, but are not relevant theoretically, are thought of as enhancing mundane realism. Ordinarily, experiments using student participants should not be high in audit environment reality as students have very limited familiarity with that environment (i.e., this artificiality would reduce validity). Consequently, when students are used it is essential that the setting have high experimental realism to them and that the important variables within the experimental setting mirror forces present in their natural setting.

Experimental realism relates to how seriously the participants treat the laboratory events. This contributes to a high level of internal validity. For example, Gibbins (1992) has argued that placing auditors in an experimental setting raises serious questions about experimental realism, and Lord (1992) has argued and empirically shown that motivation in a contrived setting is a substantive issue affecting audit practitioners. Mirroring key variables in the environments is obviously essential if any reasonable generalization is to be achieved.

As stated from the outset, our study's goal was to provide empirical data for practice consideration and to contribute to theory testing. Much of the research and theory on which this study is based comes from areas other than business. That research typically uses methods (e.g., surveys, field studies) that generally possess better external than internal validity. The practitioner interview and consultation employed in our study parallel these techniques. However, we believed that it was unwise to involve practitioners in a more structured experimental study at this stage of the research. First, there are few studies that directly support the specific propositions tested. Second, some aspects of the existing studies suggest limited internal validity. Third, there was no efficient and effective approach for such a study. Therefore, given the severe difficulty and cost of a study involving practitioners, the first step was to test the behavioral propositions with appropriate controls and high validity in an environment with parallel variables.

Even if auditors and selected client personnel were willing to participate in a study of this nature, many concerns would persist. Observation in a natural field setting is virtually impossible. The frequency of material wrongdoings is unknown, but quite low. The occurrence rate of reports about such wrongdoings is also unknown, but necessarily low. As observers, researchers would be obtrusive and would likely alter behavior. Observers who did report would likely have limited insight into the motives underlying their actions. Identifying and understanding what restrained knowledgeable parties from reporting the wrongdoing is questionable. Some who did not report might not even acknowledge that they knew. Even if we could overcome these concerns, the time required for observation would be prohibitive. A natural field setting is not feasible. (12)

Collecting archival/historical data on fraud and nonfraud audits is another possible approach. (13) However, the feasibility of this approach is doubtful. Where fraud was discovered, but only after several audits, persons who allegedly knew about the fraud would need to be asked why they chose to remain silent. Problems of identifying knowledgeable parties, overcoming their hindsight bias, and assuring valid recall and insight represent serious challenges. Perhaps even more serious problems would arise when auditors had received reports that stopped wrongdoing before it became material. Identifying these situations would entail considerable error. If this could be overcome, then the problem of identifying reporters and nonreporters would arise. If this were overcome, then the problems of recall, bias and insight persist. This approach is not practical.

Using practicing auditors and businesspersons in a contrived experimental setting that used actual behavior seemed fraught with problems. Placing them in such a setting would likely produce behavior quite different from that in practice (e.g., Gibbins and Newton 1994; Gibbins 1992; Lord 1992; Gibbins and Emby 1984). Constructing a setting that would produce experimental realism in terms of strength of relationship (which is often developed over time) and beneficiary of the wrongdoing seemed an insurmountable problem (cf., Smith et al. 1987; Swieringa and Wieck 1982).

We also considered a more traditional pencil-and-paper exercise asking participants to describe what reporting choice they or a third party observer might make, but we believed that demand effects constituted significant obstacles to validity. To shore up our belief about demand effects, we presented the current scenario to other similarly situated students at another university in the form of a case exercise. Forty-one of the 43 participants indicated that a report would be made. This contrasts with only 20 of 29 that actually reported, which constitutes a significant difference in frequency (p<.01). Also, the paper-and-pencil exercise resulted in such a large majority who said that a report would be forthcoming that insufficient variability existed to study the two experimental variables.

Militating against the use of deception is the possible harm that might accrue to the participants. The relationship between teacher and student is very special; one that demands trust. In this experiment, nearly all the participants were students of the experimenter in a senior-level auditing class. A portion of that class involved the study of professional ethics; consequently, the concern for any harm was heightened. Fisher and Fryberg (1994) note that invasion of privacy, stress and discomfort, loss of self esteem and negative reactions to being induced to perform inappropriate acts are all potential concerns when using deception. However, their study indicates no serious consequences accruing to students who were participants in experiments involving deception. Nevertheless, the procedures to protect the students in this study reflected the best practices for this purpose. (14) As noted earlier, these steps exceeded those required by the Institutional Review Board at the experimenter's university.

After weighing the issues and consulting with our expert practitioners, we concluded that we could surrogate the factors in the external environment adequately to contribute evidence about the contentions of Aldersley (Hooks et al. 1994) and the regional risk management partner. The use of a rigorously controlled experimental setting allowed us to stabilize such issues as known frequency of wrongdoing and reports/ nonreports, participant randomization, ambiguity associated with whether a wrongdoing had occurred, obtrusiveness, and a natural setting for the selected participants. Coupled with the use of student participants, it also allowed us to manipulate strength of relationship and beneficiary of the wrongdoing in a realistic manner that would have been difficult, if not impossible, with auditors and potential reporters. As noted, several experts counseled us on the motivational aspects used in the study. (15) Clearly limitations resulting from the use of student participants persist, but parallels to the audit-client setting are also present. Table 3 presents a succinct consideration of those parallels.
TABLE 3
Similarity Between FFR and Experimental Task and Motives

Fraudulent Financial Reporting   Experimental Task

Within business context, FFR     Within student environment,
  evolves from concealed           wrongful act occurs from
  wrongful act                     concealed calculator usage

Normally starts small            Usage not overtly a large
                                   matter

Can result in serious            Can change course grade
  consequences                     (gain/loss) to observer and
                                   perpetrator (especially near
                                   term's end where grades are
                                   curved and standing is
                                   known)

Reporting FFR can                Reporting cheating can
  stop/remediate consequences      stop/remediate damage

Expert auditors say that         Theory strongly supports
  strength of relationship         potential reporter's
  affects receipt of report        benefit/cost analysis;
                                   relationship and
                                   benefit/cost positively
                                   related

FFR can involve benefit to       Reward structure manipulated
  perpetrator and organization     to benefit primarily
  (including potential             perpetrator or perpetrator's
  reporter)                        group. Theoretically tied to
                                   benefit/cost analysis to
                                   reporting decision.


Research design may have affected the beneficiary-variable portion of the experiment. Libby and Lipe (1992) note that attention to the task can be raised by increasing the task's difficulty and its perceived importance (Humphreys and Revelle 1984). Such actions can increase intrinsic incentives to a sufficiently high level that an extrinsic manipulation (e.g., the beneficiary treatment in this study) may appear to have no significant effect. Inasmuch as the participants in this experiment clearly had a difficult, time-pressured task related to their school's curriculum change progress, it is possible that the beneficiary treatment's behavioral impact was marginal. That is, the full four percent of the final grade was not at stake due to the awarding of partial credit. The manipulation was effective, as evidenced by the students' recall of the beneficiary treatment, although it was not significant to reporting behavior. As suggested above, perceived task importance may have overpowered the manipulation.

Practical Significance

This study's results show that the participants were more inclined to report a wrongdoing when they had a close relationship with the report recipient than when they did not. To the extent that this finding carries over to the relationship between an audit partner and key client personnel, one could expect a close relationship between the two parties to yield increased communication about sensitive matters. The importance of the strength of auditor-client relationships is of consequence because of independence concerns and because some conditions in practice may jeopardize these relationships. Competition places increased emphasis on efficiency (Cook et al. 1992) and directs attention to revenue enhancement. Achieving maximum audit efficiency and spending time to build relationships with clients are not necessarily compatible activities.

Increased partner turnover also threatens the probability of strong relationships with client top management. Significant turnover among auditing personnel has posed a concern for over 20 years (Sorensen et al. 1973), but has recently extended into partner ranks (Berton 1994). Dalton et al. (1994) find that pressure to raise new revenue plays a role in managers' and partners' decisions to leave their firms. SEC Practice Section guidelines, which mandate engagement partner rotation (AICPA 1978), also result in partner reassignments and truncated relationships.

Client turnover impedes building and maintaining strong relationships. About one of 20 public clients change auditors annually (Francis and Wilson 1988), and this rate may be increasing or higher for all clients (O'Keefe et hi., 1994). (16) First year audits tend to constitute serious problems (e.g., St. Pierre and Anderson 1984; Bell et al. 1993 (17)). Based on opinions voiced by partners and the experimental results, it seems reasonable to expect that weak relationships may be detrimental to first-year audit effectiveness.

If auditors strengthen their relationships with selected client personnel to encourage trustworthiness and communication, will independence and objectivity be necessarily impaired? Requirements such as SEC engagement partner rotation as well as numerous ethical warnings implicitly rely on the danger from auditors' loss of healthy skepticism. Clearly, this study does not advocate a relationship of this nature. Rather, it suggests that auditors develop close professional relationships that engender trust that any sensitive communication will be treated judiciously and that any action will be pursued diplomatically but effectively. Independence and objectivity remain the key ingredients that allow auditors to lend credibility to financial statements. The prospect of increased client communication about wrongdoings also enhances the credibility of those same statements. As noted, communication about wrongdoing seemingly is more effective in discovering fraud than documentary methods (Association of Certified Fraud Examiners 1995).

This study's findings pertinent to beneficiary of the wrongdoing provide only clues that may affect auditors. More research dealing with beneficiary impact is needed. One possible explanation for the lack of significance in this study is that the manipulation (points available toward a final grade) may not have been sufficiently extreme to affect behavior. However, other explanations for the lack of significance may also be plausible. An alternative explanation may be that the group or organization must desperately need the benefit of the wrongdoing--as in a case where organizational survival is at risk--for the observer's reporting decision to be affected (Near and Miceli 1985; Greenberger et al. 1987). This conjecture can only be resolved by further research.

Another explanation for the nonsignificant experimental results may be that some observers decide against reporting because they adapt to and tolerate the wrongdoer's behavior. If "honest" witnesses ignore what seem to be small or moderate wrongdoings, they run the risk of becoming "accomplices." After studying differences between actual reporters and nonreporters, Glazer and Glazer (1987) concluded that behavior, but not moral, adaptation is a common justification used by nonreporters. However, as implications from the wrongdoing worsen, the nonreporter witnesses find their position compromised as they have, in effect, conspired with the wrongdoers by failing to report in a timely manner.

This theory is of special concern to auditors as many serious frauds begin with small amounts that increase over time. Should subsequent research findings support this "adapt-and-tolerate-behavior theory," then internal controls and control assessment procedures may need to increase their emphasis on reporting all transgressions; not just those that are obviously serious. Control procedures could also be drafted to protect observers who report after a significant time delay. (18)

More investigation and analysis is needed before auditors modify their practices. However, even without further research it seems prudent that they attend to the strength of relationships. There are a few indications that some existing practices may facilitate this action. For example, when partner rotation is occurring, some practitioners indicated that an effort was made to continue the existing manager on the engagement. However, the stated reasons for retaining the managers were for coordination purposes. Relationships might also be managed in ways that are not a typical part of current practice. For example, when a new audit client is accepted, efforts could be made to deal extensively with "sensitively" positioned client personnel, i.e., persons capable of observing significant wrongdoing can be targeted for special attention. Our current level of understanding is limited, but the theory reviewed and the results in this study indicate that the impact of relationships on audit-client communication may be important.

APPENDIX

General Debriefing (19)

Explanation of Auditing Study

First, thank you for participating in this study. Perhaps the most important thing to reiterate is that your fellow students who used calculators during the exercise were acting magnanimously to increase our understanding of some very important matters. They played the crucial role in making this study possible.

A second major point is that no one here at (this university) will ever be able to associate your name with your responses as the data will be coded at another university by a graduate assistant. Upon completion of that coding, the paper sources that allow tracking to your number will be shredded.

You will receive full credit for participating in the exercise.

As you probably know from discussions and readings from Auditing, employee fraud and theft have reached record levels in the United States. Financial statement fraud by management has led to added billions in loss to investors. Often, auditors are held accountable for these matters. One approach is to help auditors and honest managers to understand what leads to reporting of wrongdoing when people see it.

While this may at first seem elementary, the fact is that there are often very significant personal costs associated with the reporting of wrongdoing. This is particularly the case when the wrongdoer is a close business associate or a classmate. Conflicting loyalties arise: that is, loyalty to the associate versus loyalty of fair play. Personal values also may make it difficult for an observer of wrongdoing to "rat" or "rink" on a friend. When one reports on his or her boss, who often is the party committing serious fraud, there are also significant potential economic costs.

This study was designed in part to help you experience, and, thus appreciate, some of this pressure.

John Doe (name changed) is the person responsible for the part of the study being carried out here at (this university). He will be happy to talk about these matters either in person or by phone (number deleted). He will be in his office on Thursday, May 5, from 9:00 till 1:00 and from 3:00 to 6:00 and on Friday, May 6, from 8:45 to 11:45 and from 2:00 to 4:45. Feel free to drop by or to call if you would like to discuss anything dealing with the exercise.

Key Information Elicited (20)

Perpetrator Debriefing (Stage 4 (21))

1. 1 knew the student sitting beside me in my group well. --

2. I knew the student sitting in front of me in my group well. --

3. The student sitting beside me in my group saw me using my calculator. --

5. My use of the calculator was addressed during group deliberations. YES NO

Potential Report Recipient (Coordinator) and Observer Assessment (Stage 4)

3. Generally, working in groups is beneficial to learning, improving interactive skills, and improving communication skills. --

7. The scoring procedure (including point allocation) was fair. --

8. The exercise represented a fair test of my knowledge of APs. --

10. Compared to students in my own class, working with the student from the other class was more difficult. --

11. The seating arrangement made interaction very difficult. --

Observer Debriefing (Stage 5)

1. I knew the group member from my own auditing class well. --

2. I knew the other group member well. --

3. I saw the student sitting beside me use a calculator. YES NO

4. All three of us discussed his/her use of a calculator during the group solution. YES NO

5. After the student sitting beside me left, the other student in my group and I discussed his/her use of a calculator. YES NO

8. Instructions indicated that my individual performance would be stressed much more than my group's performance. --

11. Prior to starting this exercise, had you heard that students might use calculators? YES NO

Potential Report Recipient (Coordinator) Assessment (Stage 5)

1. I knew the group member from my own auditing class well. --

2. I knew the other group member well. --

3. Instructions indicated that my individual performance would be stressed much more than my group's performance. --

4. After the group member from the other auditing class left, the other student in my group and I discussed the way the exercise would be evaluated. YES NO

Potential Report Recipient (Coordinator) Debriefing (Stage 6)

1. During the exercise, I saw the group member from the other auditing class use a calculator. YES NO

2. All three of us discussed his/her use of a calculator during the group solution. YES NO

3. After the group member left, the other student in my group and I discussed his/her use of a calculator. YES NO

6. Prior to starting this exercise, had you heard that students might use calculators? YES NO

We appreciate the assistance or counsel of Steve Kaplan, Jeffrey Wilson, Susan Ayers, Ann Byk, Elizabeth Dreike, Charlene Henderson, Kay Newberry, Bill Stout, Juliette Webb, Arnie Wright, the Boston College accounting workshop, the Texas Tech University accounting workshop, and the University ' North Texas accounting workshop. This research was supported, in part, by a grant from the Research Support Grunt Program at Arizona State University, College of Business, and by Price Waterhouse LLP.

REFERENCES

Albrecht, W. S., and J. J. Willingham. 1993. An evaluation of SAS No. 53, the auditor's responsibility to detect and report errors and irregularities. In The Expectation Gap Standards. Jersey City, NJ: AICPA.

American Institute of Certified Public Accountants. 1978. Report of Progress: The Institute Acts on Recommendations for Improvements in the Profession. New York, NY: AICPA

--. 1988a. The Auditor's Responsibility to Detect and Report Errors and Irregularities. Statement on Auditing Standards No. 53. New York, NY: AICPA.

--. 1988b. Consideration of the Internal Control Structure in a Financial Statement Audit. Statement on Auditing Standards No. 55. New York, NY: AICPA.

--. 1990. Audit Guide: Consideration of the Internal Control Structure in a Financial Statement Audit. New York, NY: AICPA.

--. 1996. Considerations of Internal Control in a Financial Statement Audit. Statement on Auditing Standards No. 78. New York, NY: AICPA.

--. 1997. Consideration of Fraud in a Financial Statement Audit. Statement on Auditing Standards No. 82. New York, NY: AICPA.

Arnold, D. F., and L. Ponemon. 1991. Internal auditors' perceptions of whistle-blowing and the influence of moral reasoning: An experiment. Auditing: A Journal of Practice & Theory (Fall):1-15.

Association of Certified Fraud Examiners. 1995. Characteristics of fraud schemes--Some empirical evidence. Working paper, Austin, Texas.

Bell, T. B., S. Szykowny, and J. J. Willingham. 1993. Assessing the likelihood of fraudulent financial reporting: A cascaded logit approach. Working paper, KPMG Peat Marwick.

Berton, L. 1994. KPMG Peat Marwick is in a) great shape b) turmoil. Wall Street Journal (July 28).

Blackburn, M. S. 1988. Employee dissent: The choice of voice versus silence. Unpublished doctoral dissertation, University of Tennessee, Knoxville.

Boatsman, J. R., L. P. Grasso, M. B. Ormiston, and J. H. Reneau. 1992. A perspective on the use of laboratory market experimentation in auditing research. The Accounting Review (January): 148-156.

Campbell, D., and J. Stanley. 1963. Experimental and Quasi-Experimental Designs for Research. New York, NY: Rand McNally.

Cohen, E. G. 1991. Teaching in multiculturally heterogeneous classrooms: Findings from a model program. McGill Journal of Education 26: 7-23.

Committee of Sponsoring Organizations of the Treadway Commission (COSO). 1992. Internal Control: Integrated Framework. Harborside, NJ: AICPA.

Cook, J. M., E. M. Freedman, R. J. Groves, J. C. Madonna, S. F. O'Malley, and L. A. Weinbach. 1992. The liability crisis in the United States: Impact on the accounting profession. Deloitte & Touche Review (November 2).

Cressey, D. R. 1980. Management fraud, accounting controls, and criminal logical theory. In Management Fraud: Detection and Deterrence, edited by R. K. Elliott, and J. J. Willingham. Princeton, NJ: Petrocelli Books.

Dalton, D. R., J. W. Hill, and R. J. Ramsay. 1994. The big chill. Journal of Accountancy (November): 53-56.

DeJong, D. V., and R. Forsythe. 1992. A perspective on the use of laboratory market experimentation in auditing research. The Accounting Review (January): 157-170.

Dozier, J. B, and M. P. Miceli. 1985. Potential predictors of whistle-blowing: A prosocial behavior perspective. Academy of Management Review 10: 823-836.

--. 1988. Is whistle-blowing helping behavior? A laboratory study of team members' reporting of an unethical team leader. Unpublished dissertation, Ohio State University, Columbus, OH.

Dopuch, N. 1992. Another perspective on the use of deception in auditing experiments. Auditing: A Journal of Practice & Theory (Fall): 109-112.

Farrell, D., and J. C. Peterson, 1982. Patterns of political behavior in organizations. Academy of Management Review 7:403-412.

Fisher, C. B., and D. Fyrberg. 1994. Participant partners: College students weight the costs and benefits of deceptive research. American Psychologist (May): 417-427.

Francis, J., and E. Wilson. 1988. Auditor changes: A test of theories relating to agency cost and auditor differentiation. The Accounting Review (October): 663-682.

Gaines, J. 1980. Upward communication in industry: An experiment. Human Relations 33: 929-942.

Gibbins, M., and C. Emby. 1984. Evidence on the nature of professional judgment in public accounting. Papers from the Auditing Research Symposium. Champaign, IL: College of Commerce and Business Administration.

--. 1992. Deception: A tricky issue for behavioral research in accounting and auditing. Auditing: A Journal of Practice & Theory (Fall): 113-126.

--, and J. D. Newton. 1994. An empirical exploration of accountability in public accounting. Journal of Accounting Research 32 (2): 165-186.

Glazer, M. P., and P. M. Glazer. 1987. Pathways to resistance: An ethical odyssey in government and industry. In Research in Social Problems and Public Policy, Vol. 4, edited by M. Lewis, and J. L. Miller. Greenwich, CT: JAI Press.

Graham, J. W. 1986. Principled organizations dissent: A theoretical essay. In Research in Organizational Behavior, vol. 8, edited by L. L. Cummings, and B. M. Staw. Greenwich, CT: JAI Press.

Green, R., and K. Barrett. 1994. Auditing the accounting firms. Financial World (September): 30-55.

Greenberger, D. B., M. P. Miceli, and D. Cohen. 1987. Oppositionists and group norms: The reciprocal influence of whistle-blowers and co-workers. Journal of Business Ethics 7: 527-542.

Hooks, K. L, S. E. Kaplan, and J. J. Schultz, Jr. 1994. Enhancing communication to assist in fraud prevention and detection. Auditing: A Journal of Practice & Theory (Fall): 86-117.

--, and J. J. Schultz, Jr. 1996. The use of deception in behavioral accounting research. Behavioral Research in Accounting: 25-47.

Humphreys, M., and W. Revelle. 1984. Personality, motivation and performance: A theory of relationship between individual differences and information processing. Psychological Review (April): 153-184.

Jones, A. P., L. R. James, and J. R. Brunni. 1975. Perceived leadership behavior and confidence in the leader as moderated by job involvement. Journal of Applied Psychology 60: 146-149.

Keenan, J. P. 1988. Communication climate, whistle-blowing, and the first-level manager: A preliminary study. Academy of Management Best Paper Proceedings 40:247-251.

KPMG Peat Marwick. 1993. Fraud Survey Results 1993. New York, NY: KPMG Peat Marwick.

--. 1995. Fraud Survey Results 1995. New York, NY: KPMG Peat Marwick.

Kreutzfeldt, R. W., and W. A. Wallace. 1986. Error characteristics in audit populations: Their profile and relationship to environmental factors. Auditing: A Journal of Practice & Theory (Fall): 20-43.

Libby, R., and M. G. Lipe. 1992. Incentives, effort and the cognitive processes involved in accounting related judgments. Journal of Accounting Research (Autumn): 249-273.

Lord, A. T. 1992. Pressure: A methodological consideration for behavioral research in auditing. Auditing: A Journal of Practice & Theory (Fall): 89-108.

Miceli, M. P., and J. P. Near. 1985. Characteristics of organizational climate and perceived wrongdoing associated with whistle-blowing decisions. Personnel Psychology 38: 525-544.

--, and --. 1992. Blowing the Whistle: The Organizational and Legal Implications for Companies and Employees. New York, NY: Lexington Books.

Naj, A. K. 1992. Federal judge awards ex-GE staffer record amount in whistle-blower case. Wall Street Journal (December 7).

Near, J. P., and M. P. Miceli. 1985. Organizational dissidence: The case of whistle-blowing. Journal of Business Ethics 4:1-16.

O'Keefe, T. B., D. A. Simunic, and M. T. Stein. 1994. The production of audit services: Evidence from a major accounting firm. Journal of Accounting Research (Autumn): 241-261.

Palmrose, Z. 1987. Litigation and independent auditors: The role of business failures and management fraud. Auditing: A Journal of Practice & Theory (Spring): 90-103.

Ponemon, L. A. 1994. Whistle-blowing as an internal control mechanism: Individual and organizational considerations. Auditing: A Journal of Practice & Theory (Fall): 118-130.

SAS User's Guide: Statistics. 1992. Version 6.03. Cary, NC: SAS Institute.

Schultz, J. J., Jr., D. A. Johnson, D. Morris, and S. Dyrnes. 1993. An investigation of the reporting of questionable acts in an international setting. Journal of Accounting Research. (Supplement): 75-103.

Smith, V. L., J. Schatzberg, and W. S. Waller. 1987. Experimental economics and auditing. Auditing: A Journal of Practice & Theory (Fall): 71-93.

Sorensen, J. E., J. G. Rhode, and E. L. Lawler III. 1973. The generation gap in public accounting. Journal of Accountancy 136: 42-50.

St. Pierre, K., and J. A. Anderson. 1984. An analysis of the factors associated with lawsuits against public accountants. The Accounting Review (April): 242-263.

Swieringa, R., and K. Wieck. 1982. An assessment of laboratory experiments in accounting. Journal of Accounting Research. (Supplement): 56-117.

Trevino, L. K. 1986. Ethical decision making in organizations: A person-situation interactionist model. Academy of Management Review 11: 601--617.

--, and S. A. Youngblood. 1990. Bad apples in bad barrels: A causal analysis of ethical deci sion-making behavior. Journal of Applied Psychology 75: 378-385.

--, and B. Victor. 1992. Peer reporting of unethical behavior: A social context perspective. Academy of Management Journal XXXV: 38-64.

(1) The strength of relationship may apply to internal auditors as well.

(2) These SASs were effective during the design and execution of this study. SAS No. 78 (AICPA 1996) replaced SAS No. 55 and became applicable for all audits for periods beginning on or after January 1, 1997. SAS No. 82 (AICPA 1988a) replaced SAS No. 53 (AICPA 1997) for the audit of financial statements ending on or after December 15. 1997. These new standards do not affect this study's core objectives.

(3) For conciseness, reference to the studies addressing these variables are not reported. See Hooks, et al. (1994) for a discussion of the research.

(4) For clarity, we refer to the five partners who provided input prior to the experiment as consulting partners. Several faculty colleagues and accounting doctoral students also assisted in factor selection, design, and instrument refinement. The instruments were also pilot tested and subsequently revised.

(5) Industries served by these partners varied and were described as: manufacturing, service, high tech, financial services, real estate and health care. Partner descriptions of the size and ownership structures of various clients were: emerging business, long-time public (blue chip), changing from private to public, and closely held Tax partners were included for an additional perspective since they may have knowledge regarding aggressive tax avoidance behavior with financial statement impact.

(6) The interview protocol, including the structured interview instrument, is available from the authors.

(7) The study met and exceeded requirements of the experimenting university's Institutional Review Board. which is charged with review of human subjects' issues. For example, the IRB required debriefing only after all experimentation was completed, but participants were debriefed immediately upon completing their portion of the experiment. This timing prevented some perpetrator-confederates from having to spend about one week as "cheaters." See Lord (1992), Dopuch (1992) and Gibbins (1992).

(8) Using a nonclass member as the perpetrator was expected to increase the chances of reporting over what it would have been had a "home room classmate" been the perpetrator.

(9) The individual training sessions conveyed what to expect and how to behave in the group sessions. These sessions revealed the behavioral purpose of the experiment and secured students' signed acknowledgments. In recruiting volunteers, the sessions were described as individual testing sessions.

(10) Each section (12 groups) was split into two separate large classrooms, leaving space between groups. It was expected that time pressure during the individual task stage would occupy students' attention. However, in the first administration, a coordinator from one group noticed a perpetrator in another group using her calculator. Perpetrators in future administrations were cautioned to be very careful. Data were run with and without these six groups and no results changed. An alternative explanation is that the perpetrators used the calculators in such a secretive manner to avoid any temporary embarrassment of being cast as cheaters. The authors doubt this explanation because of thorough explanations at the training sessions which consisted only of perpetrators.

(11) Perhaps the most common use of student subjects in auditing research occurs within experimental economics. Also, refer to Boatsman et al. (1992) and DeJong and Forsythe (1992) for the use of experimental economic studies in auditing.

(12) Many partners we spoke with noted one or more of these concerns in expressing their opinions. This was especially true for assessing the motives that drove people who knew about wrongdoing and who did or did not communicate with them.

(13) Bell et al. (1993) represents one example of such an approach. The purpose here is not to criticize that study, but to point out some of the concerns with the approach. Indeed. Bell et al. (1993) acknowledge some of the methodological limitations discussed here.

(14) For a summary discussion of these steps, see Hooks and Schultz (1996).

(15) These experts represented three different Big 6 firms; since the paper was written, a member of another Big 6 firm's executive office has also provided supportive remarks for the surrogation.

(16) O'Keefe at al. (1994) report a stratified random sample of 1,000 1989 audit clients for one firm yielded eight percent first-year clients (see their table 1).

(17) See their table 1 panel C, which indicates a disproportionate rate of undiscovered fraud arose in first year engagements.

(18) Policies to protect those who filed delayed reports may be beneficial for other reasons, as well. Observers often need time to gather less ambiguous evidence of the wrongdoing. For example, in a multimillion dollar case involving General Electric, the court upheld the observer's justification for delaying his report. That is, he remained silent for several years until he could demonstrate that a wrongdoing had occurred (Naj 1992).

(19) In addition to this general debriefing, a clear statement about the study appeared on each individually completed debriefing form. See examples of items from individual forms below. Extraneous matter and names have been deleted.

(20) All items, except those followed by a "YES NO," called for responses on a seven-point scale, which ranged from 1 = Strongly Disagree to 7 = Strongly Agree.

(21) Refer to figure 1 for stage at which the elicitation occurred.

Joseph J. Schultz, Jr., is a Professor at Arizona State University and Karen L. Hooks is a Professor at Florida Atlantic University.
COPYRIGHT 1998 American Accounting Association
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1998 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Schultz, Joseph J., Jr.; Hooks, Karen L.
Publication:Auditing: A Journal of Practice & Theory
Date:Sep 22, 1998
Words:11837
Previous Article:The effects of organizational structure on communication within audit teams.
Next Article:The error detection of structural analytical procedures: a simulation study.
Topics:

Terms of use | Privacy policy | Copyright © 2022 Farlex, Inc. | Feedback | For webmasters |