Copyright American Accounting Association May 2008| [Headnote] |
| SUMMARY: |
| In this paper, we test whether experience limits auditors' refiance on management-provided information when that information is more favorable than an objective benchmark. In an experiment, we find that experience as an audit senior and the favorableness of management information interact to influence auditor assessments of the reliability of internal controls. While auditors with low levels of experience at the senior rank give assessments that are more favorable, when management's assessment is favorable, high-experience senior auditors' judgments are not influenced by management's assessment. We conclude that as auditors gain experience, they also gain persuasion knowledge, which allows them to deflect management's persuasion attempts. |
INTRODUCTION
One of the most pervasive sources of information auditors receive during the conduct of an audit engagement is management. Auditors' consideration and evaluation of information from management presents a dilemma for the auditor. On one hand, management should be knowledgeable of the firm's business strategies and risks, as well as details of accounting transactions and related internal controls. Given limitations on the scope of the audit process, auditors have incentives to utilize information from management (Haynes 1999), and professional standards recognize that ignoring management-provided information would be costly and impractical (PCAOB 2004, para. 109).
On the other hand, management is not an objective information source and may have incentives that are not consistent with auditor objectives. For example, management has incentives to manage earnings (Anderson et al. 2004; Graham et al. 2005). Thus, auditors should be particularly skeptical when evaluating information obtained from management, and professional standards direct auditors to maintain an attitude of professional skepticism when they integrate information provided by management into their auditing judgments (for example, see PCAOB 2004, para. 106). Given this conflict, it is not surprising that auditing researchers (Andersen et al. 1994; Anderson et al. 2004; Goodwin 1999; Hirst 1994a; Haynes 1999; Joyce and Biddle 1981; Rebele et al. 1988) have a longstanding interest in examining how attributes of the source of audit evidence, and attributes of the auditors who evaluate that information, can affect the extent to which information obtained from management influences auditor judgment.
Reflecting the complexity and importance of management-provided information, the Public Company Accounting Oversight Board (PCAOB) has recently proposed revisions to professional standards that provide additional guidance on using evidence provided by management to accomplish audit objectives (PCAOB 2006, 21-25). As auditing policy makers continue to revisit this issue to promote effective and efficient audit practices, it is important to understand the factors that influence auditors' consideration of information provided by management.
Psychology researchers recognize that the persuasiveness of information is a complex process likely to be affected by attributes of (1) the source, (2) the message, and (3) the recipient (Pornpitakpan 2004). Auditing researchers have primarily focused on examining how attributes of the source (e.g., competence, trustworthiness, reliability, objectivity) influence the persuasiveness of management information. A limited number of studies have examined how attributes of the recipient (e.g., auditor experience) alter the extent to which attributes of the source influence auditor judgment. However, audit research has not examined the joint influence that attributes of the message and attributes of the recipient have on auditor judgment. We believe that further examination of auditors' consideration of information provided by management will be informative to practicing auditors, standard setters, and academic auditors because little is known about whether attributes of the message and attributes of the recipient interact to influence the persuasiveness of information from management.
As auditors gain experience (an attribute of the recipient), they are expected to develop and accumulate persuasion knowledge. Friestad and Wright (1994, 1999) describe persuasion knowledge as what laypeople believe about persuasion-that it "is a socially constructed set of beliefs created over the years from the pooling of private perceptions and social communications about persuasion" (Friestad and Wright 1999, 187). Persuasion knowledge is activated when another person's behavior raises suspicion (Campbell and Kirmani 2000). In the context of auditing, persuasion knowledge represents an auditor's beliefs about how, when, and why managers try to influence auditors, and consequently, help auditors respond to these persuasion attempts (Anderson et al. 2004). Also in the context of auditing, a persuasion attempt represents auditors' perception about managerial behaviors designed to influence their beliefs, decisions, and/or actions (Friestad and Wright 1994).
We contend that as auditors gain experience (an attribute of the recipient), they will find information provided by management that is congruent with management's self-interest (an attribute of the message) to be increasingly less persuasive. Auditors are accessing their persuasion knowledge when they recognize that information from management is a persuasion attempt, congruent with management's self-interest.
Thus, experience is expected to influence auditors' judgments about managementprovided information when it is congruent with management's self-interest. Alternatively, auditors receiving information from management that is not congruent with management's self-interest will view that information with less suspicion, and consequently, auditors will not access their persuasion knowledge. Thus, experience is not expected to influence auditors' judgments about management-provided information when it is not congruent with management's self-interest.
This study examines the effect that information about management's assessment of internal control reliability (the message) has on internal control reliability judgments that auditors (the recipient) develop when they integrate information from management with information gathered by other members of the audit team. We selected this setting for several reasons. An overall judgment about the reliability of an internal control system, whether by management or the auditor, is an unstructured and complex task. Thus, professional judgment is involved and there is no single, "right" answer (Mock and Turner 1981). Internal control evaluation is also a setting in which the auditor is informationalIy dependent on management. That is, although auditors acquire some information about control procedures through observation and by performing walkthroughs, considerable information about internal controls comes from management. Further, rigorous management self-assessment of internal control has long been advocated within the professional literature (Baker and Graham 1996; Figg 1999; Green 2004; Hawkins and Huckaby 1998).
Self-assessment of the reliability of internal controls helps satisfy management's responsibility for monitoring internal controls, one of the five components of an effective system of controls (Committee of Sponsoring Organizations [COSO] 1992). In addition, the Sarbanes-Oxley Act of 2002 (SOX) requires that the managements of publicly traded companies assess and report on the effectiveness of their system of internal controls over financial reporting. Recently, Earley et al. (2006) also examined auditor judgment in this setting. Lastly, internal control evaluation is a setting in which management has reporting preferences. In applying their professional judgment as part of a self-assessment process, management should prefer reporting that the internal control system is stronger rather than weaker. In this regard, Ashbaugh-Skaife et al. (2007) find that idiosyncratic risk, betas, and cost of equity is higher for firms reporting an internal control deficiency compared to firms not reporting an internal control deficiency. Thus, self-reports of a reliable internal control system are congruent with management self-interest, but self-reports of an unreliable internal control system are incongruent with management self-interest.
We examine the relationship between the nature of manage ment-provided information (e.g., whether the information is congruent or incongruent with management self-interest) and auditor experience with data from two laboratory experiments. Audit seniors were asked to rate the overall reliability of internal controls for an e-commerce sales system based on: (1) management's reliability self-assessment and (2) reliability ratings based on tests performed by other members of the audit team. Results from tests of controls performed by other members of the audit team were held constant in both experiments, but management's reliability assessment was manipulated bet ween-participants.
The participants in our first experiment possessed a broad range of experience at the senior rank. Participants received information about management's rating that was either more favorable or less favorable than the rating implied by the results from tests of controls performed by other members of the audit team. Our second experiment, which provided follow-up evidence for corroborating and interpreting results from the first experiment, included participants with limited amounts of senior auditor experience. Participants in the second experiment either received information that management had rated control reliability as more favorable than the rating implied by the results from tests performed by other members of the audit team, or did not receive information about management's rating.
Findings from Experiment 1 and corroborating evidence provided by Experiment 2 supported our predictions about associations among the congruency of management's message with management's self-interest, auditor experience, and the extent to which information from management influenced auditor judgment. The remainder of this study is organized into four sections. The second section, we review related literature, explain our theoretical framework, and present hypotheses. The third section describes the experiments we used to gather data for evaluating our hypotheses. The fourth section presents the results of our hypothesis tests. The fifth section discusses our findings and provides suggestions for future research.
LITERATURE REVIEW AND HYPOTHESES
This section develops predictions about the effect of auditor experience on the extent to which information acquired from management influences auditor judgment. To support these predictions, we review relevant research on source credibility, management incentives, and auditor experience.
Source Credibility
Hirst (1994a, 113) contends that source credibility is important "because the inferential value of evidence must always be considered in light of its source." Audit research has generally examined the construct of source credibility along two dimensions, including (1) objectivity-a source maintaining an impartial, unbiased attitude, and (2) competencepossessing the ability to provide accurate and reliable information (Reimers and Fennema 1999). Research has demonstrated that source credibility influences auditors' consideration of information during judgment tasks that involve account fluctuations (Anderson et al. 1994; Hirst 1994a), asset impairment (Rebele et al. 1988; Goodwin 1999; Haynes 1999), contingent liabilities (Goodwin 1999), and internal controls (Bamber 1983). These studies operationalized source credibility as technical competence of subordinates (Bamber 1983), expertise of client personnel (Rebele et al. 1988; Anderson et al. 1994), integrity of independent counsel and integrity of management (Goodwin 1999), or client management versus firm personnel (Hirst I994a; Haynes 1999). However, other studies (Joyce and Biddle 1981; Reimers and Fennema 1999) have found that auditor judgment did not differ when information came from an external source (e.g., an independent agency) versus an internal source (e.g., client personnel).
Research has also shown that audit experience influences consideration of source credibility. Haynes (1999) examined the interaction between auditor experience and the source credibility of evidence. In her study, the high-experience group was comprised of governmental auditors with an average of more than seven years of audit experience whereas the low-experience group was comprised of MBA students with no audit experience. She manipulated management credibility by changing the probability that management would truthfully report certain information. Findings suggest that the persuasiveness of information obtained from management was jointly influenced by auditor experience and source credibility. The credibility of management had a stronger influence on judgments of the highexperience group compared to judgments of the low-experience group.
Management Incentives
Research has examined the extent to which auditor judgment is influenced by whether information from management is congruent with explicit management reporting incentives. This research is grounded in psychology linking source credibility and whether information is (in)congruent with the source's self-interest (Sternthal et al. 1978; Pornpitakpan 2004). Hirst (1994b) contends that, when considering explanations from management for significant unexpected differences in audit results, auditors will be persuaded less when an explanation is congruent with management self-interests than when an explanation is incongruent with management self-interests. The intuition is that auditors should be aware that management is not an objective source and typically has incentives to report in a particular fashion. When management provides information that is congruent with their incentives, auditors are expected to discount the information because the source lacks credibility and the information conforms to the source's bias. When management provides information that is incongruent with their incentives, auditors are expected to be persuaded more by the information because one is less able to attribute the information to the source's bias. The results were mixed, with the results from Study 1 providing support. In Study I, the CFO's self-interest was manipulated by indicating that the CFO was part of a management group taking the company private (e.g., incentive to understate) or that no major changes in the business had taken place (e.g., incentive to overstate). However, the results from Study 2 showed that auditors' judgments were not influenced by explicit incentives (e.g., whether current period unaudited earnings were either above or below the upper bound for a performance based bonus).
More recently, Anderson et al. (2004) contend that that in considering explanations from management for significant unexpected differences in audit results, the persuasiveness of the explanation will be jointly determined by whether or not the explanation is quantified and whether or not management has strong, explicit incentives to manage earnings. Specifically, Anderson et al. (2004) predict that auditors' judgments will be influenced by whether the explanation is quantified only when management's incentives to manage earnings are low. The results of the study, however, show that quantification did not matter. However, auditors were more persuaded when management's incentives to manage earnings were low compared to when management's incentives were high.
Two auditing studies (Ayers and Kaplan 2003; DeZoort et al. 2003) examine the influence of reporting incentives from sources other than management. Ayers and Kaplan (2003) examine how review audit partners' judgments about a prospective audit client are influenced by the judgments and implicit incentives of contact audit partners. The study is premised on the belief that contact audit partners are rewarded for obtaining new clients, and consequently, have incentives to bias their judgments towards an acceptance decision. In the study, the contact partner's engagement risk assessment was either more favorable than the assessment implied by the underlying client background information or consistent with the assessment implied by the underlying client background information. The results of the study indicated that review partners discounted a contact partner's overly favorable engagement risk assessment when forming their acceptance and audit fee judgments.
DeZoort et al. (2003) examine whether audit committee members' support for an auditor-proposed adjustment is stronger when unaudited earnings are above analysts' forecast than when unaudited earnings are below analysts' forecast. Management has incentives to report earnings above rather than below analysts' forecast. The results show that audit committee members' support for an auditor-proposed adjustment was not influenced by the relationship between unaudited earnings and forecasted earnings.
The results from studies reviewed above suggest a mixed picture. That is, the results indicate that professionals (auditors or audit committee members) sometimes discounted information that was congruent with the incentives of the source, and sometimes did not discount information that was congruent with the incentives of the source.
Auditor Experience
When auditors develop an evaluative judgment by integrating evidence from management with evidence obtained from an independent source, their professional experience will likely influence their evaluative judgment. Field experience provides the opportunity to accumulate knowledge that helps professionals develop effective mental models for interpreting and integrating evidence during evaluative judgment tasks (Bonner and Lewis 1990; Libby and Luft 1993; Patel and Groen 1986). Knowledge gained from experience improves people's ability to recognize potential problems by providing them with the capacity to distinguish and integrate diagnostic information more effectively (Chi et al. 1982). Experience provides knowledge that helps auditors assign appropriate decision weights to the evidence they acquire (Bonner 1990) because experience helps them develop more comprehensive knowledge structures, and improves their ability to use that knowledge more effectively (Tubbs 1992). For example, Bhattacharjee and Moreno (2002) report that affective information, which should be irrelevant to the task, significantly influenced inexperienced auditors' inventory obsolescence risk assessments but did not influence experienced auditors' assessments.
Additionally, audit experience is expected to foster the development and use of persuasion knowledge (Campbell and Kirmani 2000; Friestad and Wright 1994). Friestad and Wright (1994, 6) describe persuasion knowledge as a "set of interrelated beliefs about (1) the psychological events that are instrumental to persuasion, (2) the causes and effects of those events, (3) the importance of the events, (4) the extent to which people can control their psychological responses, (5) the temporal course of the persuasion process, and (6) the effectiveness and appropriateness of particular persuasion tactics."
In the context of auditing, persuasion knowledge helps auditors identify how, when, and why managers try to influence them, and consequently, helps auditors respond to these persuasion attempts (Anderson et al. 2004). Persuasion knowledge includes an auditor's understanding of a manager's persuasion motives and persuasion tactics and, for auditors, is expected to develop over time, as they interact with managers and other client personnel as well as through their discussions with other auditors.
Persuasion knowledge is activated when another person's behavior raises suspicion (Campbell and Kirmani 2000). Campbell and Kirmani (2000) found that consumers' perceptions of a salesperson were significantly influenced by the accessibility of persuasion knowledge. In their study, accessibility of persuasion knowledge was manipulated through a salesperson making a flattering remark to the customer either before (raising suspicion) or after (less suspicious) a purchase. As expected, consumers' judgments of a salesperson's sincerity were more favorable when the motive for flattery was less obvious because the consumer knew that the salesperson made the flattering remark after the purchase.
Research Hypotheses
In the context of evaluating information from management, we contend that as auditors gain experience, they will develop persuasion knowledge about the motives and tactics of managers. Auditors who, through experience, have developed greater persuasion knowledge will be more informed about management's motives and tactics, and better able to recognize and respond to situations in which management may be attempting to influence auditor judgment.
We believe that management has incentives to encourage a favorable auditor assessment of control reliability. Managers are likely to realize that decreases in auditors' reliance on controls negatively impact the manager and the company by exposing the company to an adverse opinion on the effectiveness of its internal controls. Furthermore, managers also are likely to know that auditors who rely less on controls increase the scope of substantive testing, which may result in higher audit fees, and could also result in communications with the audit committee that do not reflect favorably on management effectiveness.
We hypothesize that the extent to which management-provided information about control reliability influences auditor judgment about control reliability will depend jointly on (1) the favorable versus unfavorable nature of information provided by management, and (2) the extent to which auditors have had the opportunity to develop persuasion knowledge through field experience. Specifically, we predict that attributes of the message (favorable versus unfavorable management information) will interact with attributes of the recipient (amount of experience) to influence auditors' consideration of information from management.
Consider a situation in which management's internal control assessment appears less favorable than an assessment based solely on independent evidence gathered by the audit team. Regardless of auditor experience, this situation is unlikely to arouse suspicion. That is, auditors are unlikely to treat such a management assessment as an attempt at persuasion. In this situation, auditors will tend to interpret management's information at face value (e.g., as a reflection of their specific knowledge about control reliability). This discussion leads to our first hypothesis, stated in alternative form as follows:
H1: When management assessments of control reliability are less favorable than reliability assessments implied by independent evidence, auditors' experience will not be associated with their judgments about control reliability.
Next, consider a situation in which management's internal control assessment appears more favorable than an assessment based solely on independent evidence gathered by the audit team. In this situation, management's assessment is congruent with their self-interest, which should be suspicious, and auditors will access their persuasion knowledge. Auditors possessing persuasion knowledge that is more developed are more likely to identify management's assessment as a persuasion attempt, and consequently, respond to this attempt by discounting management's assessment. That is, as experience increases, auditors will weight management's assessment less. As discussed above, auditors are expected to acquire and develop their persuasion knowledge over time. This discussion leads to our second hypothesis, stated in alternative form as follows:
H2: When management assessments of control reliability are more favorable than reliability assessments implied by independent evidence, increases in auditors' experience will be negatively associated with their judgments about control reliability.
METHOD
Hypotheses were tested with data gathered from two laboratory experiments where audit seniors were asked to evaluate the overall reliability of a client's internal controls over a computerized e-commerce sales system. Experiment 1 compared the overall reliability judgments for auditors receiving information from management that was either more or less favorable than the overall reliability assessment implied by independent evidence. Participants in Experiment 1 possessed a wide range of senior auditor experience. Experiment 2 compared the overall reliability judgments for auditors either receiving information from management that was more favorable than the overall reliability assessment implied by independent evidence, or did not receive information from management. Participants in Experiment 2 possessed a limited range of senior auditor experience. Performing two experiments allowed us to examine the effect of different levels of reliability information from management as well as the effect of the availability of reliability information from management on auditor judgment.
Auditors at the rank of senior were selected for the research because they are generally responsible for conducting internal control evaluations (Abdolmohammadi 1999). Participants were employed by a single Big-4 audit firm and completed the experiment during a national training session under the supervision of a research proctor. Participants completed the internal control task for Experiments 1 and 2 immediately after they had completed an unrelated auditing task that involved performing analytical procedures on a case that did not include evidence about internal controls.
Materials
Participants were asked to rate the overall reliability of internal controls over the sales processing system based on: (1) the reliability ratings for various control objectives from their firm's computer audit specialists and (2) management's self-assessment of overall control reliability. (As noted above, Experiment 2 included a group that did not receive information about management's rating.) Ratings from computer specialists provided participants with independent, objective evidence gathered by other members of the audit team. The rating by management provided information about control reliability that auditors may encounter when they evaluate control effectiveness.
Reliability ratings from computer specialists were he!d constant across all experimental conditions. case materials indicated that the audit firm's computer specialists had tested control procedures for six control objectives and rated the reliability of controls that addressed each objective on a nine-point scale where 1 represented low reliability and 9 represented high reliability. Control objectives were taken from the CobiT framework (ISACA 2003) for computerized transaction processing. Objectives and reported reliability ratings (in parentheses) included, processing integrity (7), system security (6), service levels (4), customer assistance (6), monitoring problems (4), and system availability (3). The average reliability rating across these six objectives is 5. In Experiment 1, management's rating was manipulated at two levels, high and low. In Experiment 2, management's rating was manipulated at two levels, high and not provided.
In the low-management-rating condition, management had rated overall control reliability for the sales system at 3 on a scale where I represented low reliability and 9 represented high reliability. In this situation, management's self-assessment is lower than the implied rating of 5 based on evidence available from the computer specialists. In contrast, the high-management-rating condition stated that management had rated control reliability at 7 on the same nine-point scale. In this situation, management's self-assessment is greater than the implied rating of 5 based on evidence available from the computer specialists.
Our intention was to provide evidence gathered by other members of the audit team that would support a reliability rating of 5 if auditor ratings were not influenced by management ratings. We asked members of the committee charged with revising the CobiT audit guidelines to rate overall control reliability based on the information provided in the case, but without any information about management assessment. Seven of the 17 members responded, including three audit partners at Big-4 firms, two IT managers from Fortune 500 companies, and two academics. Four respondents rated reliability at 5, and the other respondents rated reliability at 3, 4, or 7. The mean rating of these seven members is approximately 4.9, and is consistent with our intention. These results are also consistent with our task being unstructured, and involving auditor judgment.
Task for Experiment 1
The experimental task involved three components. First, after reading background information about their client and the control evaluation task that they were about to perform, participants were told that, pursuant to the requirements imposed by SOX, management had evaluated controls over the sales system and rated overall reliability at 3 (7). A management rating of three represented a low management rating as it is below the implied rating from other, more independent sources of audit evidence. Alternatively, a management rating of 7 represented a high management rating as it is above the implied rating from other, more independent sources of audit evidence.
Second, participants were told that their firm's computer specialists had identified six critical control objectives for the sales system, and participants were asked to rate the importance of each objective as high, medium, or low. Participants were also asked to document their judgment about the importance of each objective to provide a metric for evaluating the influence that opinions about the relative importance of different controls had on judgment about control reliability. Next, participants were provided with reliability ratings from the computer specialists for each of the six objectives. To control for presentation order effects, approximately half of the participants received information about the computer specialists' reliability ratings first, and the rest received information about management's rating first.
Third, participants were asked to rate the overall reliability of internal controls over the e-commerce sales system, and subsequently, provide information about their auditing experience. The experimental design included two manipulated factors: (1) favorable and unfavorable management rating conditions, and (2) presenting information about management's rating either before or after information about ratings developed by other members of the audit team. These two factors were manipulated in a sample of auditors with a range of O to 36 months of experience as an audit senior, our measure of experience. The instrument for Experiment 1 is presented as Appendix A.
Task for Experiment 2
Experiment 2 provides evidence on whether audit seniors receiving a favorable management control rating evaluated controls more favorably than a comparison group not receiving a management rating. Based on convenience, this task was completed by participants in the lower range of experience among auditors at the rank of senior. However, we expect audit seniors at this level of experience to rely on a favorable management control rating, thus, an important group. case materials for Experiment 2 were identical to the case materials for Experiment I with two exceptions. First, to ensure that attention was focused exclusively on integrating management information with other, independent evidence, participants were not asked to rate the importance of the control objectives tested by the firm computer specialists. second, Experiment 2 manipulated management's control rating at two levels, a favorable management rating (similar to Experiment I), and a comparison condition that did not include a management rating. Similar to Experiment 1, the presentation order for management's rating and computer specialist ratings was counter-balanced for participants in the favorable management rating condition. Participants' rating of the overall reliability of internal controls over the e-commerce sales system was again used as the dependent variable. The instrument for Experiment 2 is presented as Appendix B.
RESULTS
A total of 78 seniors participated in Experiment 1. We had expected 88 participants and, because of the way our experimental materials were sorted for distribution, the ten unused instruments all included cases with a high management rating. As a result, 44 of the completed instruments were for cases with a low management rating and 34 were for cases with a high management rating. Three instruments were removed from each group for missing data, and our final sample included 41 observations for the low-managementrating condition and 31 observations for the high-management-rating condition. Our sample of 72 participants had experience at the senior rank that ranged from one to 48 months, with an average of 19.8 months and a standard deviation of 10.7 months.
Validity Checks
Before testing our hypotheses, we examined two threats to internal validity. First, we tested whether perceptions about the relative importance of different control objectives were associated with experience as an audit senior. If the association was significant, the influence of experience on participants' reliabiiity ratings could be attributable to differences in perceptions about the relative importance of different control objectives rather than hypothesized differences in the persuasiveness of management's reliability rating.
Second, we tested whether the presentation order for information about management's rating and ratings developed by other members of the audit team affected the influence that management's rating had on participants' reliability ratings. If order influenced participants' reliability ratings, then our findings could be attributable to either primacy or recency effects, which are not envisioned in our hypotheses.
To determine whether experience influenced participants' perceptions about the relative importance of control objectives, we constructed a metric that estimates the rating participants would have developed if each participant had computed an average of control reliability ratings provided by computer specialists that was weighted by the importance rating the participant attributed to each objective. Recall that reliability ratings from computer specialists differed across the six control objectives. To calculate the weighted reliability rating, we multiplied the reliability ratings provided by computer specialists for each of the six control objectives times a weight of 1.25 when the importance of the objective was rated high, 1.00 when the importance rating was medium, and 0.75 when the importance rating was low. We summed these products and divided the total by six to calculate the weighted reliability rating. Next, we calculated Spearman correlations between months at the senior rank and the weighted reliability rating for each of the six control objectives. None of the correlations is significant at the p < .10 level. These results suggest that experience-related differences in the relative importance of controls does not represent a potential confound.
To examine whether participants were influenced by the order in which they received information from management versus evidence provided by other members of the audit team, we performed analysis of variance using participants' reliability ratings as the dependent variable. The independent variables are management's rating at two levels (high versus low); presentation order for management's rating at two levels (before versus after ratings developed by other members of the audit team), and an interaction term based on the two independent variables. Participants1 reliability ratings were significantly greater (at the p < .01 level) in the high-management-rating group compared with the low-managementrating group. Neither order of presentation nor the interaction term was significant (at the p < .10 level). Because we found no evidence of an order effect, we concluded that neither primacy nor recency effects represents a potential confound.
Hypothesis Tests
Hypothesis 1 predicts no association between participants' reliability ratings and experience in the low-management-rating condition. Hypotheses 2 predicts that in the high-management-rating condition there will be a negative association between participants' reliability ratings and experience. Taken together, these hypotheses predict an interaction between the level of management's rating and experience as an audit senior with respect to participants' reliability ratings. We used ordinary least-squares regression to conduct a joint test of these predictions. The model included management's rating (a high rating is coded 1, and a low rating is coded 0), months of experience as an audit senior, and the interaction term between management's rating and months of senior experience. The dependent measure is participants' overall reliability rating of internal controls. Results are presented in Panel A of Table 1.
When the management rating is low (HMR = 0), months of experience at the senior level do not affect the participant reliability rating, as evidenced by the insignificant (p = 0.1945) parameter estimate for months at the senior rank. These findings support H1. However, when the management rating is high (HMR = 1), on average, the participants' reliability ratings significantly lower (p = 0.0059) for more experienced audit seniors. As shown, the parameter estimate is -0.08 for the interaction between high management rating and months at the senior rank. Overall, these results indicate that, on average, participants' reliability ratings decline by 0.06 for each additional month of audit senior experience when the management rating is high. These findings support H2.
| TABLE 1 |
| Influence of Management Reliability Ratings and Audit Experience on Participants' Reliability Ratings |
We performed additional analysis by splitting our sample into two groups of low- and high-experience participants at the sample median of 18 months, then comparing cell means for participant1 reliability ratings across the four resulting groups. Results are presented in Panel B of Table 1. Consistent with H1, mean reliability ratings of 3.91 for participants in the low-management-rating-low-experience group are not significantly different from mean ratings of 4.52 participants in the low-management-rating-high-experience group (t = 1.56; p > .10). Consistent with H2, mean reliability ratings of 5.93 for participants in the high-management-rating-low-experience group are significantly greater from mean ratings of 4.93 participants in the high-management-rating-high-experience group (t = 2.24; p < .05).
Results from Experiment 1 demonstrate that audit seniors with less experience are influenced to a greater extent by a management rating that is congruent with management's self interest than are audit seniors with more experience. Because Experiment 1 does not include a comparison group excluding management's control rating, we do not have evidence of the ratings auditors would have made in the absence of management's control rating. Experiment 2 was designed to provide evidence that included a comparison group that did not receive management's control rating and the high-management-ratings condition used in our first experiment. Experiment 2 was similar to the first experiment, except that (1) all participants had less than 19 months of experience at the senior rank, and (2) participants did not rate the importance of individual control objectives.
A total of 58 auditors with one to 18 months of experience at the senior rank (mean = 5.22; standard deviation = 4.72) participated in Experiment 2. Eighteen participants completed the no-management-rating condition. Forty participants completed the highmanagement-rating condition. In the high-management-rating condition, 21 participants received information about management's rating before ratings from firm computer specialists, and 19 participants received information about management's rating after ratings from firm computer specialists. Consistent with findings from Experiment 1, we found no evidence in Experiment 2 of an order effect. That is. participants' reliability ratings did not differ significantly across alternative presentation orders, and presentation order did not interact with the management rating condition.
The mean participant reliability rating in Experiment 2 is 5.25 for participants who received a favorable management rating in contrast to a mean reliability rating of 4.66 for participants who did not receive a management rating (t = 2.12; p < .05). This finding provides additional evidence that less-experienced senior auditors are persuaded by a favorable control assessment by management.
DISCUSSION
Auditors commonly receive information from management throughout the engagement. Evaluating information from management, however, presents a dilemma for auditors. While management should generally be knowledgeable and competent when it comes to assessing performance, in the context of an audit, management is not an objective information source. Thus, the extent to which auditors incorporate information from management into their judgments represents a significant issue for audit practice, especially when information from other, more objective sources is also available. Not incorporating information from management into their judgments would be very costly to auditors, making audits inefficient and impractical for auditors and their firm. Not recognizing management's lack of objectivity, however, could result in inappropriate reliance on information from management, potentially making audits ineffective and exposing their firm to audit failures and large litigation costs.
Building on persuasion research (Pornpitakpan 2004), we contend that the persuasiveness of information from management will be influenced by attributes of the source, the message, and the recipient. While prior audit research has primarily focused on source attributes, our research focuses on the attributes of the message and the recipient and the interaction between these two in influencing auditor judgment. We believe that our focus extends auditor judgment research in an important way and represents a substantial contribution to this literature. In this regard, our study, along with Anderson et al. (2004), suggests that under certain conditions auditors' persuasion knowledge is likely to play a key role in auditors' interactions with and judgments about information from management.
Our research focused on auditors' consideration of information from management about internal control reliability in reaching their own independent judgment about internal control reliability. This judgment is particularly important as it is intrinsically linked to the scope of the audit. Specifically, our research examined the congruency between management's self-assessment of internal control reliability and management self-interests, and whether auditor experience moderates the effect that this congruency has on auditor judgment. In the context of our research, information that the internal control system is highly reliable is congruent with management's self-interest, whereas information of that the internal control system is relatively unreliable is incongruent with management's self-interest.
We predicted that experience would not be related to auditors' reliability judgments when management rated internal controls at a low level of reliability. Alternatively, we predicted that experience would be related to auditors' reliability judgments when management rated internal controls at a high level of reliability. Both of these predictions are based on models of persuasion knowledge (Campbell and Kirmani 2000; Friestad and Wright 1994, 1999), which hold that persuasion knowledge will not be accessed when information from management is not suspicious (e.g., management rating internal controls low). Consequently, in this situation, audit experience is not expected to influence auditor judgment. Auditors, however, are expected to be suspicious of a report from management with a high rating of internal controls, which should trigger auditors to access their persuasion knowledge. Because persuasion knowledge develops gradually over time (Friestad and Wright 1994; Wright et al. 2005), auditors with greater experience should respond by increasingly discounting management's high rating report.
The results from two studies provided support for our hypotheses. First, the results indicate that when management's self-assessment is not congruent with management's selfinterest, auditors consider management's assessment when reaching their own judgment of control system reliability. This tendency was not related to auditor experience, which indicates that when management provides information that is not congruent with management's self-interest, auditors will use this information when forming their own judgments, and its use is not associated with auditor experience.
Second, the results indicate that experience reduces auditors' tendency to rely on information from management when it is congruent with management's self-interest. Our finding that less experienced senior auditors relied on information from management that was congruent with management self-interest to a greater extent than more experienced senior auditors is somewhat troubling. To the extent that such relatively inexperienced audit seniors have responsibility for assessing overall internal control reliability, it suggests that internal controls will be assessed more favorably than may be warranted (e.g., in comparison to assessments developed by more experienced audit seniors) in light of other, independent evidence. This inflated assessment, in turn, could lead to an underestimation of misstatement risk and jeopardize the auditor's ability to collect sufficient, competent evidence to appropriately support the audit opinion.
Our findings inform previous auditing research showing a mixed picture of whether professionals discount information that is congruent with the source's self-interests. Our results suggest that it is not enough for suspicions to arise, but auditors will increasingly discount information that is congruent with management self-interest as experience grows. This pattern is consistent with auditors slowly acquiring persuasion knowledge over time. The relative presence of persuasion knowledge may explain the contrasting results of Ayers and Kaplan (2003) and DeZoort et al. (2003). Participants in Ayers and Kaplan (2003) were highly experienced audit partners, who presumably possess extensive persuasion knowledge about the incentives of contact audit partners. Consequently, these participants discounted information from contact audit partners consistent with their self-interest. Alternatively, participants in DeZoort et al. (2003) were audit committee members. Relative to the task, it is less clear whether they would have sufficiently developed persuasion knowledge. Only 18 percent of the participants were CPAs or equivalent and less than 40 percent served on more than one audit committee. Further, data was collected in mid-2000, before a series of financial scandals and the increased scrutiny that was placed on audit committee members. The relative lack of experience may explain why audit committee members support for a proposed adjustment was not sensitive to management's self-interest.
Further research regarding auditors' consideration of information from management is encouraged. Our research focused on attributes of the message and the recipient. This research could be expanded to provide insight into whether auditors' consideration of a message that is congruent or incongruent with management's self-interest depends on attributes about management (e.g., management has high versus low integrity). Alternatively, research could examine whether auditors' consideration of information from management depends on the credibility of the source providing "other information." In our research, other information was provided by members of the audit firm, presumably viewed as highly credible. Whether auditors' consideration of the congruency between information provided by management and management's self-interest also depends on the credibility of the source providing the other information remains an interesting avenue for further research.
| [Reference] » View reference page with links |
| REFERENCES |
| Abdoimohammadi, M. J. 1999. A comprehensive taxonomy of audit task structure, professional rank and decision aids for behavioral research. Behavioral Research in Accounting 11: 51-92. |
| Anderson, U., L. Koonce, and G. Marchant. 1994. The effects of source-competence information and its timing on auditors' performance of analytical procedures. Auditing: A Journal of Practice & Theory 13 (Spring): 137-148. |
| ______, K. Kadous, and L. Koonce. 2004. The role of incentives to manage earnings and quantification in auditors' evaluations of management-provided information. Auditing: A Journal of Practice & Theory 23 (March): 11-27. |
| Ashbaugh-Skaife, H., D. W. Collins, W. R. Kinney, Jr., and R. LaFond. 2007. The effect of SOX internal control deficiencies on firm risk and cost of equity. Working paper, University of Iowa. |
| Ayers, S., and S. E. Kaplan. 2003. Review partners' reactions to contact partner risk judgments of prospective clients. Auditing: A Journal of Practice & Theory 22 (March): 29-45. |
| Baker, L. L., and R. D. Graham. 1996. Control self-assessment. The Internal Auditor (April): 52-57. |
| Bamber, M. 1983. Expert judgment in the audit team: A source reliability approach. Journal of Accounting Research 21 (Autumn): 396-412. |
| Bhattacharjee, S., and K. Moreno. 2002. The impact of affective information on the professional judgments of more experienced and less experienced auditors. Journal of Behavioral Decision Making 15: 361-377. |
| Bonner, S. 1990. Experience effects in auditing: The role of task-specific knowledge. The Accounting Review 65 (January): 72-92. |
| ______, and B. Lewis. 1990. Determinants of auditor expertise. Journal of Accounting Research 28 (Supplement): 1-20. |
| Campbell, M. D., and A. Kirmani. 2000. Consumers' use of persuasion knowledge: The effects of accessibility and cognitive capacity on perceptions of an influence agent. Journal of Consumer Research 21 (June): 69-83. |
| Chi, M. T. H., R. Glaser, and E. Rees. 1982. Expertise in problem solving. In Advances in the psychology of human intelligence, edited by R. Sternberg, 7-75. Hillsdale, NJ: Erlbaum. |
| DeZoort, F. T., D. R. Hermanson, and R. H. Houston. 2003. Audit committee member support for proposed audit adjustments: A source credibility perspective. Auditing: A Journal of Practice & Theory 22 (September): 189-205. |
| Barley, C., J. Joe, and V. Hoffman. 2006. The effect of management's classification of control deficiencies on auditors' judgments under PCAOB Auditing Standard Number 2. Working paper, Bentley College. |
| Figg, J. 1999. The power of CSA. The Internal Auditor (August): 28-35. |
| Friestad, M., and P. Wright. 1994. The persuasion knowledge model: How people cope with persuasion attempts. Journal of Consumer Research 21 (June): 1-31. |
| _____, and_____. 1999. Everyday persuasion knowledge. Psychology and Marketing 16 (March): 185-194. |
| Goodwin, J. The effects of source integrity and consistency of evidence on auditors' judgments. Auditing: A Journal of Practice & Theory 18 (Fall): 1-16. |
| Graham, J. R., C. R. Harvey, and S. Rajgopal. 2005. The economic implications of corporate financial reporting. Journal of Accounting and Economics 40: 3-73. |
| Green, S. 2004. Internal controls: Take seven key actions before you certify. The Journal of Corporate Accounting & Finance (May/June): 21-27. |
| Hawkins, K. W., and B. Huckaby. Using CSA to implement COSO. The Internal Auditor (June): 2835. |
| Haynes, C. 1999. Auditors' evaluation of evidence obtained through management inquiry: A cascaded inference approach. Auditing: A Journal of Practice & Theory 18 (Fall): 87-104. |
| Hirst, D. E. 1994a. Auditors sensitivity to source reliability. Journal of Accounting Research 32 (Spring): 113-126. |
| _____. 1994b. Auditors sensitivity to earnings management. Contemporary Accounting Research 11 (Fall): 405-422. |
| International Systems Audit and Control Association (ISACA). 2003. Control Objectives for Business Information Technology (CobiT). Rolling Meadows, IL: ISACA. |
| Joyce, E., and G. Biddle. 1981. Anchoring and adjustment in probabilistic inference in auditing. Journal of Accounting Research 19 (Autumn): 120-145. |
| Libby, R., and J. Luft. 1993. Determinants of judgment performance in accounting settings: Ability, knowledge, motivation, and environment. Accounting, Organizations and Society 18: 425-450. |
| Mock, T, and J. Turner. 1981. Internal Accounting Control Evaluation and Auditor Judgment. New York, NY: American Institute of Certified Public Accountants |
| Patel, V., and G. Groen. 1986. Knowledge based solution strategies in medical reasoning. Cognitive Science 10: 91-116. |
| Pornpitakpan, C. 2004. The persuasiveness of source credibility: A critical review of five decades' evidence. Journal of Applied Social Psychology 34: 243-281. |
| Public Company Auditing Oversight Board (PCAOB). 2004. An Audit of Internal Control over Financial Reporting Performed in Conjunction with an Audit of Financial Statements. Auditing Standard No. 2 (March 9). Washington D.C.: PCAOB. |
| _____. 2005. Auditing internal control over financial reporting 2005. Staff Questions and Answers (May 16). |
| _____. 2006. Release No. 2006-07-Proposed Auditing Standard (December 19). |
| Rebele, J. E., J. A. Heintz, and G. E. Briden. 1988. Independent auditor sensitivity to evidence reliability. Auditing: A Journal of Practice & Theory 8.1 (Fall): 43-52. |
| Reimers, J., and M. Fennema. 1999. The audit review process and sensitivity to information source objectivity. Auditing: A Journal of Practice & Theory 18 (Spring): 117-123. |
| Sternthal, B., L. W. Phillips, and R. Dholakia. 1978. The persuasive effect of source credibility: A situational analysis. The Public Opinion Quarterly 42 (Autumn): 285-314. |
| Tubbs, R. 1992. The effect of experience on the auditor's organization and amount of knowledge. The Accounting Review 67 (October): 783-801. |
| Wright, P., M. Friestad, and D. M. Boush. 2005. The development of marketplace persuasion knowledge in children, adolescents, and young adults. Journal of Public Policv & Marketing 24 (Fall): 222-233. |
| [Author Affiliation] |
| Steven E. Kaplan is a Professor at Arizona State University, Edward R O'Donnell is an Assistant Professor at The University of Kansas, and Barbara M. Arel is an Assistant Professor at the University of Vermont. |
| We appreciate feedback provided by Stan Biggs, Marianne Bradford, Michelle Diaz. Rich Houston, two anonymous reviewers, Kathryn Kadous, the Associate Editor, and participants in research workshops at the University of Arkansas, University of Connecticut, University of Kansas. Louisiana State University, North Carolina State University, the University of Texas at San Antonio, Texas A&M University, and the University of Waterloo. |
| Submitted: February 2006 |
| Accepted: August 2007 |