Error management in audit firms: Error climate, type and originator*

Anna Gold
Associate Professor
Faculty of Economics and Business Administration
VU University Amsterdam (The Netherlands)

Ulfert Gronewold
Professor
Chair of Accounting and Auditing in the Private and Public Sector
University of Potsdam (Germany)

Steven E. Salterio
Professor and PricewaterhouseCoopers/Thomas O‟Neill Faculty Fellow in Accounting
School Of Business
Queen‟s University (Canada)

A shorter version of this paper will appear in the January 2014 issue of The Accounting Review

August 20, 2013

* This paper was previously circulated under the title “The Impact of Error-Management
Climate, Error Type and Error Originator on Auditors‟ Reporting Errors Discovered on Audit
Working Papers.” We are grateful for comments provided by Christopher Agoglia, Jonathan
Grenier, Noel Harding, Jim Hunton, Hansrudi Lenz, Benjamin Luippold, Luc Quadackers,
Robert Ramsay, Martin Richter, Hun-Tong Tan, Ken Trotman, participants at Bern, HEC Paris,
KAIST, Korea University, Leuven, Maastricht, Münster, Tilburg, Umeå and University of Texas
at El Paso accounting workshops and participants at the AAA Audit Mid Year Meeting, the EAA
Congress, International Symposium on Audit Research, EARNet Symposium and AAA Annual
Meeting. We thank the partners and staff at the six German based audit firms for assisting us
with this research and to Ira Solomon for introductions with one firm and for a detailed
discussion of the paper at the EAA Congress. We thank Andy Grien, Myra Landsburg, Joan Luft
and Micheal Seadle for proofing and commenting on the English to German translation and to
Sarah Reuter for professionally retranslating the material.

Electronic copy available at: https://ssrn.com/abstract=1657117

 Error management in audit firms: Error climate, type and originator

ABSTRACT

Audit standard setters and regulators are increasingly focused on the management context within
the audit firm that conducts the audit for that context‟s effects on audit quality. We examine a
key, but little understood facet of organizational life in audit firms, namely how audit staff who
discover and report errors in audit files are routinely treated in response to reporting such errors.
This construct, denoted as audit office error-management climate, can be characterized on a
continuum from a relatively “blame-oriented” climate to a relatively more “open” climate. The
former is one where errors are not tolerated and those committing errors are punished. In
contrast, the latter characterizes error commitment as a normal, albeit unfortunate aspect of
organizational life, where the error commitment and discovery is used as an opportunity for the
organization to learn without sanctions on the originator if similar errors are not repeated by the
auditor. We examine audit office error-management climate in the context of audit-specific
factors that might affect the decision to report the discovered errors: audit error type (conceptual
or mechanical) and who committed the error (the individual who discovered it or a peer). We
find that an open error-management climate results in an increase in the reporting of mechanical
(but not conceptual) errors and an increase in reporting peer errors (bringing them to the level of
reporting of own errors) versus a blame climate. More directly our findings raise questions about
how the implementation of an open error-management climate might be enhanced so as to
promote greater reporting of own conceptual errors. Our post-hoc findings that auditors are
significantly more willing to report their own mechanical errors than their own conceptual errors
(while the difference between reporting willingness of peers‟ mechanical and conceptual errors is
insignificant) suggests that this obstacle for reporting conceptual errors particularly stems from
auditors‟ own impression management concerns. Our findings also suggest that audit firm
management needs to understand that what may appear to be innocuous differences in
management style at the individual audit office might affect audit quality at the client level
through varying the individual auditor perceptions of the office error management climate.
Further, our findings suggest that standard setters and regulators need to understand how their
standards and inspections may impact the audit firm error-management climates and how the
differences in such climates can affect audit quality by encouraging (or suppressing) the
reporting by staff of errors in working paper files.

KEYWORDS: error discovery, working paper review, error reporting, error management,
climate, audit quality

DATA AVAILABILITY: Experimental data are available from the second author subject to
data confidentiality restrictions placed on us by participating firms.

Electronic copy available at: https://ssrn.com/abstract=1657117

 Error management in audit firms: Error climate, type and originator

I. INTRODUCTION

We advance the concept of audit office error-management climate (hereafter denoted as

EMC) to help us understand how auditors act when they discover an audit error leading to

discovery of a potential error in the current year‟s financial statements. Organizational

researchers have defined EMC as the set of shared beliefs, norms, and common practices

regarding the management of discovered errors and mistakes in an organization (Van Dyck et al.

2005). This literature presents two archetypes of EMC, a relatively more “open” one where error

reports are used to enhance organizational learning and only repeated errors by individuals

attract sanctions, whereas a relatively more “blame” oriented climate that routinely sanctions the

individual who committed the error. The strong normative implications from this line of research

are that an open error climate results in the routine reporting of errors within the organization,

better organizational learning and overall better organizational performance. However, given the

need for a consistent application of a set of beliefs, norms and practices to support an open EMC,

it is likely that organizations can easily, and perhaps unintentionally, create blame EMCs in

practice, even though the organization‟s formal policies might be consistent with an open EMC

(Gillingham et al. 1997; Zhao and Olivera 2006). Hence, our primary research question is how

does the individual audit office EMC in public accounting firms affect the auditor reporting of

discovered errors within the audit firm?

Our focus on audit practice management issues stems from the observation that

professional standards (e.g., International Standard on Quality Control (ISQC) 1 (IAASB

2012b)) and audit regulators (e.g., PCAOB 2008a) are increasingly emphasizing that

management processes within audit firms need to be oriented to promote audit quality at the

Electronic copy available at: https://ssrn.com/abstract=1657117

individual audit level. Regulators are critically examining various aspects of the audit firm‟s

management from the “tone at the top” of the audit firm through to how the audit firm evaluates

both partners and staff globally and within offices to examine whether the audit firm‟s

managerial processes support high audit quality (PCAOB 2008a; FRC 2012, p. 13).

Furthermore, regulators are commenting on differences in the commitment to audit quality

between offices of the same firm (e.g., CPAB 2012, p. 17), which are consistent with findings in

the audit archival literature of such differences affecting audit quality (e.g. Ferguson et al. 2003;

Choi et al. 2010). Regulatory prescriptions to date range from suggesting that audit firms tie

audit quality to partner and staff performance evaluations and remuneration more explicitly

(FRC 2012, p. 13), to tying audit quality issues to the firms‟ disciplinary practices (e.g. PCAOB

2008b). To date, there has been limited research attention as to how these managerial practices

relate to audit quality performance by the individual auditor.

The classic definition of audit quality in the academic literature is the joint probability of

the auditor discovering a financial statement error during the audit and the auditor reporting the

error for correction by firm management or via disclosure through the auditor‟s report

(DeAngelo 1981). The detection of financial statement errors is dependent on the quality of audit

work carried out by audit staff and the audit firm‟s review process that is designed to ensure the

work is properly carried out (e.g. Asare and McDaniel 1996; Gibbins and Trotman 2002;

Owhoso et al. 2002). Given that discovered matters potentially requiring adjustment are brought

to the attention of the audit partners, then the decision about whether to report and negotiate

them with the client firm‟s management and the audit committee occurs (e.g., Antle and

Nalebuff 1991; Gibbins et al. 2001). Evidence from PCAOB and other jurisdictions‟ inspection

reports demonstrates that audit errors leading to the subsequent adjustment of issued financial

Electronic copy available at: https://ssrn.com/abstract=1657117

statements are not always discovered until the regulatory inspection process that occurs well

after the audit report is issued (e.g., PCAOB 2008a; FRC 2012). The PCAOB inspection reports

of all audit firms, including the Big 4 (e.g., PCAOB 2004a, 2004b, 2004c, 2004d, 2005), have

reported errors in the application of GAAP and GAAS in the USA (PCAOB 2008a), as have

other countries‟ regulators (e.g., the United Kingdom (FRC 2012) and Canada (CPAB 2012)).1

These post audit inspections have resulted in multiple clients having to restate their financial

statements and reveal many instances of potential non-application or misapplication of audit

procedures.

Recent regulatory reports have focused on audit firm management approaches at both a

firm-wide and office level that are beyond the original regulator emphasis on the individual audit

engagement quality (e.g., PCAOB 2012, pp. 8-9; FRC 2011). Regulators point to managerial

issues as being potential systematic contributors to the deficiencies identified in specific audit

engagements (CPAB 2012; FRC 2012). Prior research in other settings suggests that the EMC

can have a significant effect on the ability of organizations to discover errors in time to act on

them (e.g. Van Dyck et al. 2005). Further, there is survey research evidence that suggests these

EMC differences indirectly affect how auditors intend to deal with errors committed by clients

(Gronewold and Donle 2011).

We examine how differences in audit office EMC interact with audit-specific contextual

factors to affect the reporting of errors discovered by audit staff. In order to do so, we create a

case that controls for other factors (e.g., anticipation of review, audit firm reputation) so as to

focus on the effects of the EMC on the willingness of auditors to report discovered errors to

1
The PCAOB has recently warned audit committees that audit firm explanations of “differences in professional
judgment” to account for reported deficiencies in its public reports should be “viewed with skepticism” given that
“the inspection staff has considered any suggestions by the firm that a reasonable judgment could be made” and
“rejected any such suggestion” (PCAOB 2012, p. 6).

Electronic copy available at: https://ssrn.com/abstract=1657117

more senior audit staff in a timely fashion. Audit practice gives many potential reasons for

auditors to revisit the working papers post detailed file review but prior to the issuance of the

auditor‟s report and the financial statements. Examples include tidying up audit files for

archiving (paper or electronic), preparing files for in-house quality assurance (e.g., national

office peer review), preparing for next year‟s audit or, as used in our study, preparing the files for

second partner (also known as quality assurance partner) review, as mandated by ISA 230

(IAASB 2012a) and PCAOB AS No. 3 (PCAOB 2004e).

As the audit occurs in a unique professional context we examine EMC in light of two of

those contextual factors. First, we examine the types of audit errors (i.e., mechanical and

conceptual) frequently found in audit working papers during the audit review process (e.g.,

Ramsay 1994) that have implications for the financial statements. Specifically, prior research

suggests that error type is an important factor in whether the error is discovered by staff or

reviewers (Harding and Trotman 1999).2 Second, we consider that auditors work in teams (e.g.,

Rich et al. 1997) hence an auditor may discover an error that he/she created or one that a peer on

the audit team made. Research in other areas suggests that who makes the error is important to

the error reporting decision (e.g., Uribe et al. 2002).

Based on theory and the audit context, we hypothesize that the effect of EMC on

auditors‟ willingness to report discovered errors depends on the type of error and who committed

the error. We examine the interactions among these three factors via an experiment with 190

auditors who learn of an error in the working papers that is suggestive of a material misstatement

in the client‟s financial statements. We use the scenario of auditor-detected errors discovered

2
Note that regulators indicate that even simple matters such as “undetected clerical drafting errors in financial
statements” are not discovered by within audit firm review practices (FRC 2011, p. 29).

Electronic copy available at: https://ssrn.com/abstract=1657117

after all detailed file reviews have been completed but before the financial statements have been

released (see SQCS10 (AICPA 2009) and ISQC1 (IAASB 2012b)), so we can set aside issues of

external reputation effects from recalling or restating published financial statements and the

potential for litigation or regulator actions that are frequently associated with restatements (e.g.,

Palmrose et al. 2004; Palmrose and Scholz 2004). Hence, in our setting, the discovered errors, if

reported to superiors within the audit firm, could be dealt with by additional audit work to either

determine whether an adjustment is necessary (ISA 230.11 (IAASB 2012a)) or whether

negotiation need be entered into with the client firm to adjust the financial statements prior to

their release (Gibbins et al. 2001).3

Our results show that an open versus blame EMC increases an auditor‟s willingness to

report mechanical (but not conceptual) errors and brings likelihood of reporting peers‟ errors to

the same level as self-committed errors. Hence, it appears that promoting an open EMC results in

a greater willingness to report errors in all but the self-discovered own conceptual errors setting

versus a blame EMC. We discuss how an open EMC might be modified to support obtain greater

reporting of conceptual errors especially in light of our post hoc finding that the unwillingness to

report conceptual errors is focused on self-made conceptual errors. We also discuss how audit

firm management may wish to consider the amount of latitude it gives various individual audit

offices‟ management in how they interpret general firm policies that might affect how the

individual auditor views the office EMC. Finally, we discuss how these findings relate to recent

regulatory focus on audit firm managerial practices including considerations of how linking audit

3
In the extreme, the audit firm could issue a modified audit opinion if the client firm declines additional auditing or
correction of the financial statements (ISA 705 (IAASB 2012c)). Antle and Nalebuff (1991) model these
possibilities.

Electronic copy available at: https://ssrn.com/abstract=1657117

quality to various firm management practices (e.g., disciplinary) can affect the EMC in audit

firms.

The remainder of this paper is organized as follows: the next section provides background

information about audit quality, audit firm management practices and error-management climate,

leading to the development of three hypotheses. We then present the research design for an

experiment to test these hypotheses. Results including manipulation checks and robustness

checks are then reported. The paper concludes with a discussion of results and their implications

for audit practice and future research in the final section.

II. BACKGROUND AND HYPOTHESES DEVELOPMENT

Audit firm management practices and audit quality

After several years of audit engagement inspection experience by regulators in Canada,

the UK and the USA, regulators are beginning to focus on audit firm management practices that

might cause systematic audit quality problems rather than viewing each audit engagement in

isolation (CPAB 2012; FRC 2012; PCAOB 2012). This is not to say that regulators have not

been interested in audit firm management issues previously but rather that the explicit linkage to

engagement level quality concerns has until recently remained in the background. For example,

regulators suggest, “Firms must ensure that the performance of partners and staff is appropriately

evaluated against specific audit quality objectives and that there is a direct and proportionate

impact on remuneration arising from adverse audit quality assessments.” (FRC 2012, p. 13).

Further, regulators increasingly have begun to attribute audit quality issues to “matters such as

how the firm‟s management structure and processes, including the tone at the top, affect audit

quality; how the firm‟s partner management practices, including evaluation, compensation,

admission, and disciplinary practices, affect audit quality” (PCAOB 2012, pp. 8-9). The office-

Electronic copy available at: https://ssrn.com/abstract=1657117

specific level of these concerns has also been noted by regulators as “CPAB also noted audit

quality issues arose more often in certain offices of the firms” and “that firms have been asked to

develop office-specific action plans” (CPAB 2012, p. 18) to remediate deficiencies. Further, in a

rare insight into the detailed quality control aspects of the PCAOB‟s private report on Deloitte‟s

US audits in 2007 we find the call for “The enforcement, through appropriate monitoring and

disciplinary activities, of compliance with the Firm‟s policies and procedures.” (PCAOB 2008b).

As will be discussed, these factors (nature of monitoring, linking of quality to remuneration,

performance appraisal and disciplinary processes) can all influence the EMC of a firm and can

also be differentially interpreted by firm management at the office level.

In an audit firm, the EMC will likely depend on the local office or even team culture;

potentially following the local managing partner‟s or head of audit practice‟s leadership style

(see Covaleski et al. 1998; Dirsmith and Covaleski 1985; Otley and Pierce 1995, 1996). Prior

audit research has shown that organizational differences affect auditors‟ behavior (e.g., Bowrin

1998; Hermanson 1997; Hyatt and Prawitt 2001; Otley and Pierce 1995, 1996). Relevant to our

study, Peytcheva and Gillett (2012) report that auditors‟ willingness to suppress evidence of

errors made in internal control evaluations varies across offices of the same firm. These between-

office differences support our contention that there can exist different EMCs between offices of

the same firm. Gronewold and Donle (2011), employing a survey research method, demonstrate

that auditors perceive differences in EMCs across a variety of audit organizations (e.g., Big-4

and smaller audit firms, internal audit departments, government auditing, etc.) further supporting

cross-sectional differences in EMC among audit organizations. Recent archival audit research

also demonstrates differences in audit quality between audit firm offices (e.g., Choi et al. 2010;

Ferguson et al. 2003; Francis et al. 1999; Reynolds and Francis 2000). All of these findings are

Electronic copy available at: https://ssrn.com/abstract=1657117

consistent with our theoretical premise that if audit offices have different EMCs they would

exhibit differences in performance.

The academic definition of audit quality focuses on both the discovery of the error and

the reporting decision (DeAngelo 1981). Applying this definition to our context, once auditors

discover previously committed errors, they then face the question of whether to report such

errors. Extant research in psychology, healthcare management and organizational behavior

suggests that employees often do not disclose their own errors (e.g., Edmondson 1996; Tax and

Brown 1998; Tucker and Edmondson 2003; Uribe et al. 2002), because error reporting may

involve serious potential costs (such as damaged reputations, additional work effort, poor peer

relationships, and financial costs) and negative emotions (e.g., fear, embarrassment, and guilt)

(Zhao and Olivera 2006). In the audit context, error reporting may be crucial as it allows for the

timely correction of potentially material errors in the audit files or the client financial statements

(Stefaniak and Robertson 2010). Extant research shows individuals appear to differentiate among

multiple referents during considerations about reporting errors, such as their own interests, the

interests of the organization or team they work for, and the effects on potential victims (Zhao and

Olivera 2006; Coram et al. 2008; Donnelly et al. 2003).

More directly, Gronewold and Donle‟s (2011) survey research documents associations

between auditors‟ perceptions of the EMC of the entity they work for, how they handle their own

errors and how they would consequently handle client-committed errors. However, Gronewold

and Donle do not find a direct link between EMC and handling of client errors. Stefaniak and

Robertston (2010) provide initial evidence that auditors expect that their peers will be less likely

to admit to an error about their incorrect performance of an audit procedure if a compensating

audit procedure was also carried out but only in the case where there is a prior history of

Electronic copy available at: https://ssrn.com/abstract=1657117

managers reacting to mistake admissions by subjecting the auditor to a “negative, humiliating

experience” (p. 48).4 Interestingly, they report weak to insignificant results when auditors report

what they would do personally in these circumstances (p. 50).

Error-Management Climate

Organizational error-management climate comprises the set of shared beliefs, norms, and

common practices regarding the management of errors in the organization (Van Dyck et al.

2005). Research has shown that organizations differ in their climates, which can be characterized

by different degrees of “error management” (Van Dyck 1997, 2009). In what we call an open

climate (denoted in psychology research as “high error-management”), the underlying

management philosophy is based on the premise that errors are likely to occur whenever fallible

humans carry out tasks. Hence, organizations should accept this reality and plan to take

advantage of the error discovery to improve work processes and outcomes. Therefore, an open

climate is characterized by ready discussion of errors with others, upper management being

positive toward the communication about errors (i.e. “tone at the top”), obtaining others‟ help

after discovering errors, a thorough analysis of errors and their potential causes, a concern with

getting the errors corrected, and preventing individuals from being punished for having

committed such errors (Edmondson 1996; Van Dyck et al. 2005; Van Dyck 2009). Van Dyck et

al. (2005), among others, conclude that such a climate can be expected to be beneficial to the

4
Stefaniak and Robertson (2010) do not inform their participants whether an error was discovered in the account
where the audit procedure was carried out incorrectly. They tell participants that drawing of the sample from the
wrong population could lead to a failure to detect material misstatements in the account, however if the
compensating audit procedure was done there would be a relatively higher chance of detecting a material
misstatement, if one was present. However, participants are informed that no material misstatement was discovered
due to the erroneous audit procedure. Further, Stefaniak and Robertson place participants at a stage in the audit
where it is likely that the workpapers will still be subject to detailed review as audit fieldwork was not yet completed
(p. 47) potentially confounding anticipation of review with error reporting climate.

Electronic copy available at: https://ssrn.com/abstract=1657117

overall quality and performance of the organization. 5 However, it is expected that the person who

committed the error (as well as other organizational members) will learn from the mistake and

not repeat the same error again. If the same types of errors are repeatedly committed by

individuals then organizational sanctions will be employed, even in an open climate. On the

other hand, a blame EMC creates an organization where employees are better at hiding errors

they have discovered but does not reduce the number of errors initially committed (e.g.,

Edmondson 1996; Itoh et al. 2009; Reason 2000). For example, Edmondson (1996) did a

detailed analysis of eight hospital units from two large teaching hospitals and concluded that

differences in reported error rates among the units were largely due to willingness to report

factors that were affected by EMC rather than underlying error rate differences across different

EMCs.

All audit firms like to be characterized as “learning organizations” (e.g., Peecher et al.

2007; Andrews and Freeman 2001) and most espouse formal policies that are consistent with an

open EMC (e.g., Epps and Messier 2007). However, in what we call the blame climate (denoted

in the psychology literature as “low error management” or “error aversion” climate; see Van

Dyck 1997, 2009; Van Dyck et al. 2005) while these positive learning-oriented practices may be

part of the firm‟s formal policies they are frequently only paid “lip service” to by local office

management.6 Hence, the actual office EMC is one where any reported error has the strong

likelihood to lead to sanctions (e.g., disciplinary, performance appraisal, remuneration), no

matter how small the error, with larger errors leading to larger sanctions (e.g., Rybowiak et al.

1999; Van Dyck 2009; Van Dyck et al. 2005). In other words in this climate, “it must be right

5
Homsma et al. (2009) also refer to the positive effects of error discussion on organizational learning.
6
See Emett et al. (2012) on the problem of differences between upper-management views (“tone at the top”) and
(team or office-based) lower-level superiors‟ views (which they call “tone at the bottom”), and the possible
undesired outcomes of such discrepancies.

10

Electronic copy available at: https://ssrn.com/abstract=1657117

the first time” without considering the fact that fallible error-prone humans are likely to err even

when attempting to perform well in their jobs. 7

Researchers point out that it is the overall consistency in the set of error-management

responses (i.e., from “tone at the top” through to not invoking sanctions except for repeated

errors) that determines the climate and that one or two features in isolation do not effectively

create a climate (e.g., Van Dyck 2009; Edmondson 1996). Hence, it is relatively easier to create

a blame climate in an audit office than it is to promote an open climate given the need for

consistent application of shared beliefs, norms and practices in the office that go far beyond

having formal audit firm level policies that are consistent with an open EMC (Gillingham et al.

1997; Zhao and Olivera 2006). Gronewold et al. (2013) provide initial experimental evidence

that audit office level EMC affects auditors‟ perceptions of whether others would report an error

by predicting and finding a main effect for greater reporting by those other auditors under an

“open” EMC.8 Hence, we posit the following main effect for organizational EMC, ceteris

paribus:

7
Note that this formal positive-sounding emphasis in a blame climate of “getting things right the first time” actually
means that errors are not tolerated, which reflects a negative view about errors and is thus characteristic of a blame
climate. Using positively framed views about errors in a blame climate (though with negative meanings and
implications) is consistent with the notion that blame climates may pay “lip-service” to norms associated with an
open EMC. That is, negative views about errors may to some degree be “camouflaged” by positively sounding
representations. Finally, note that “getting things right the first time” could never be a leading motto of an open
climate. An open climate is characterized by accepting that errors may occur whenever humans act, which would be
in conflict with a general claim to get everything right the first time.
8
Gronewold et al. (2013) experiment examines what participants believe another auditor will do when he/she
discovers his/her own error. However, Gronewold et al. do not address what the auditor would do him or herself
when discovering his/her own error, nor do they examine his/her reaction to discovering a peer‟s error under
different EMCs. Stefaniak and Robertson (2010), who also use a “what would other auditors do” dependent variable,
report that auditors expect that the reporting of other auditors‟ errors in performing audit procedures (errors that
have ambiguous effects on the client financial statements) are affected by the historical reaction of audit managers to
other auditors reporting such errors (“negative and humiliating” versus “positive and supporting”). While in
hindsight, Stefaniak and Robertson‟s (2010) supervisor history manipulation may touch upon one of the dimensions
of EMC; their manipulation is neither grounded in theory nor does it include the various components that
characterize an EMC. Our EMC construct is explicitly derived from theory and features the typical dimensions that
make up an EMC according to error-management theory, which is beyond the scope of the study by Stefaniak and
Robertson (2010) and which enables us to carry out testing of error-management theory.

11

Electronic copy available at: https://ssrn.com/abstract=1657117

 H1: Auditors are more willing to report post-review discovered errors in an open error-
management climate than in a blame climate.

Error-Management Climate and Error Type

Prior audit research (e.g., Ramsay 1994) distinguishes between two error types:

mechanical and conceptual. Mechanical errors relate to deviations from “technical accuracy and

completeness” in the working papers (Ramsay 1994, p. 129). An example of a mechanical error

would be an erroneous calculation (e.g., an auditor performing an analytical procedure with an

arithmetic error in it leading to an error in the financial statements).9 In contrast, conceptual

errors relate to “complex, subjective, or significant matters and the adequacy of the audit work as

a whole” (Ramsay 1994, p. 129). For example, a conceptual error could be illustrated by not

employing the appropriate method (e.g., using sampling as a technique but not projecting the

errors to the population).10 Other researchers (e.g. Owhoso et al. (2002)) conclude that Ramsay‟s

(1994) “categories are quite robust” (p. 885) to alternative definitions and operationalization.11

Underlying psychology research suggests that both types of errors can result in organizational

learning, albeit there is a tendency to underestimate what can be learned from errors analogous to

our environment‟s mechanical errors but when actually followed up, these errors lead to

significant learning (Homsma et al 2009).

9
FRC (2011) reports “Weaknesses were also identified in audit finalisation procedures, including undetected
clerical drafting errors in financial statements” (p. 29). If such errors are found in the financial statements after
review it is not a stretch to suggest that they could occur in the working papers.
10
PCAOB (2004d) reports an example of this “The engagement team did not, however, project a known
misstatement to the untested portion of the materials inventory balance. Such projection is necessary to evaluate
whether the aggregate misstatement, which is the known misstatement plus the projected misstatement, exceeded the
SUD de minimus posting threshold.” (p. 20).
11
Owhoso et al. (2002) employed a different definition: “Mechanical errors are errors that require little or no
subjective judgment on the part of the auditor. Judgmental (i.e., conceptual) errors, on the other hand, require
significant subjective judgment on the part of the auditor.” (p. 893). They concluded that, contrary to their
expectations, the exact definition of mechanical versus conceptual errors did not matter to their results.

12

Electronic copy available at: https://ssrn.com/abstract=1657117

 Prior research has primarily examined hierarchical differences in the ability to detect

errors, given this error type distinction (e.g., Ramsay 1994; Bamber and Ramsay 1997; Harding

and Trotman 1999; Tan and Trotman 2003). However, our setting holds the hierarchical level

relatively constant and examines whether the type of detected error influences the auditor‟s

willingness to report the error to a superior across different EMCs. As such, we predict that any

error type effect is driven by the expectation that the auditor holds with respect to the EMC and

the consequent reactions of his/her superior. First, in a blame climate, we expect auditors to be

relatively unwilling to report errors, regardless of the error type, due to the expected likely

occurrence in such an EMC of attribution by superiors of all mistakes to personal incompetence

of the error originator leading to sanctions being imposed. In other words, a culture of “getting it

right the first time” does not allow for errors so that auditors will always fear being punished,

leading to great reluctance to report any error, regardless of type.

By contrast, in an open climate we expect variations in report willingness across error

types. First, mechanical errors (e.g., arithmetical) can be considered “careless” work but can be

easily rationalized away because such errors can happen even to a competent auditor and can be

explained by engagement-related pressures prevailing during most audits (e.g., time pressure

(Coram et al. 2004)). Hence, in an open climate, in isolated instances of mechanical errors, the

superior is unlikely to invoke sanctions or otherwise attribute the error to inherent auditor

incompetence. However, in the conceptual error setting within an open climate, the superior,

while being supportive and learning in orientation, might assign the auditor easier tasks until the

auditor has gained more experience or encourage the auditor to undertake additional training.

While the superior is not deliberately attempting to “punish” the auditor but rather to promote

learning from the error for both the auditor and the organization, the auditor may perceive the

13

Electronic copy available at: https://ssrn.com/abstract=1657117

supports designed in an open climate for this type of error being “punishment-like.” Therefore,

even in an open climate, where the response is supportive such that learning results from the

mistake, the auditor would be worried about the repercussions to him or herself of reporting

discovered conceptual errors. This analysis leads to the following interaction hypothesis (see

Figure 1, Panel A):

H2: The difference in auditors‟ willingness to report post-review discovered errors in an

open versus a blame organizational error-management climate will be greater for a
mechanical error than for a conceptual error.

<<<Insert Figure 1 about here>>>

Error-Management Climate and Error Originator

Research on error originator reporting in medical settings suggests that individuals feel

more comfortable reporting their own errors than reporting errors made by others (Rathert and

May 2007).12 Specifically, peers‟ errors in such contexts are reported by a small percentage of

discoverers and only when the reporter does not like the individual who committed the error

(Wild and Bradley 2005).13 Further, social norms against “tattle tales” and “snitches” (e.g.,

Pershing 2003) are well-embedded in Western culture and suggest that auditors would be

generally hesitant to reporting peers‟ errors. Hence, before considering the effects of the EMC,

auditors are likely to be more willing to report their own errors than those of peers. An open

EMC should provide further encouragement to report one‟s own errors versus a blame climate,

12
The literature defines “peer reporting” as occurring “when group members go outside their group to report a
member‟s misconduct” (Trevino and Victor 1992, p. 39). While most peer-reporting research concerns intentional
wrongdoings (similar to whistleblowing); unintentional misconduct is also encompassed by this definition as is
reporting to superiors if one defines the group as peers only (King 2001).
13
Other support includes: Miller and Thomas (2005) found management students‟ willingness to report an observed
unethical act to the next highest level in the organization to be the lowest when it was committed by a fellow team
member than any other relationship to self. The peer-reporting and medical error-reporting literatures document that
people may not feel responsible for reporting errors committed by others to higher levels within the organization
(Trevino and Victor 1992; Uribe et al. 2002; Victor et al. 1993). Since the audit environment features several levels
of review (e.g., Harding and Trotman 1999) that process may “shift away” responsibility for reporting others‟ errors
to either the originators themselves or the formally assigned reviewers (Trevino and Victor 1992; Victor et al. 1993).

14

Electronic copy available at: https://ssrn.com/abstract=1657117

however, given the documented high rate of reporting of own errors in other contexts the effects

of EMC are likely to be small. In other words, we do not expect a significant effect of open EMC

on own error reporting.

The key issue is whether in the peer error originator setting, does the type of EMC affect

the willingness to report peers‟ errors? An open EMC shares many of the characteristics of

positive organizational climates that prior research shows encourages relatively more reporting

of peer errors to higher levels of management. According to Vadera et al. (2009, p. 563),

whistleblowing “[…] research shows that individuals in organizations with team or friendship

climates, strong ethical climates or democratic climates are more likely to engage in

whistleblowing when they observe a wrongdoing.”14 Kaptein (2011) explicitly links these

positive organizational climates to a willingness to report peer wrongdoings to higher levels of

management. Seifert et al. (2010) show that management accountants need to be in a strong

justice (i.e., fairness) environment to increase the likelihood of reporting unethical acts within

their organization to higher management.15 Thus, following this literature, we expect that the

willingness to report on peer errors is relatively higher in an open EMC than in a blame EMC as

14
The whistleblowing literature characterizes “peer reporting” to more senior levels of management as “a form of
whistleblowing” (e.g., Barnett et al. 1996, p. 1162). While our setting does not address unethical acts or
wrongdoings in terms of the error committed the very decision to report the error is an ethical act (Barnett et al
1996). This ethical dimension is equally relevant to the general whistleblowing setting as to our specific error
reporting setting. Whether an observed peer error is unintentional or due to intentional misconduct will not always
be clear from the perspective of the person reporting the error. Hence, the reporting of unintentional and intentional
peer errors cannot be separated, at the time of the decision to report, in theory nor in practice. Hence, structurally
similar dynamics in the reporting decision between the whistleblowing setting and the unintentional error setting
appear plausible and have been studied as such in prior research (e.g., Barnett et al. 1996). To the extent that
findings from the whistleblowing literature are dependent on perceived issue seriousness and intentionality of the
misconduct (e.g. fraud vs. errors) or on the perceived “guilt” of the wrongdoer, then our analogy from this research
to the unintentional error reporting setting would not be applicable and hence one would expect no effect from the
EMC manipulation if this analogy is erroneous.
15
While the organizational climates studied in the whistleblowing literature stress this “justice” or “fairness” notion,
an EMC shares the structural similarity that the typical accepting and constructive view and handling of errors in an
open EMC will likely be conceived as “fair.” In contrast, a blame climate with its strict non-acceptance and
sanctioning of errors will be much more likely perceived as an “unfair” environment, because it does not tolerate the
“normal” and very human weakness of making errors from time to time.

15

Electronic copy available at: https://ssrn.com/abstract=1657117

the blame EMC would reinforce the auditor‟s initial reluctance to report peer errors, especially

given the persistent social norms against “tattling” and the reluctance to report team members.

Given the lower level of peer reporting to start out with and with the blame EMC reinforcing that

lower willingness to report, we expect that an open EMC will lead to a significant increase in

willingness to report peer committed errors. Based on this analysis, we predict the following

interaction, ceteris paribus:

H3: The difference in auditors‟ willingness to report post-review discovered errors in an

open versus a blame organizational error-management climate will be greater for
peer errors than for own errors.

See Figure 2, Panel A for an illustration of the predicted interaction. 16

<<<Insert Figure 2 about here>>>

III. RESEARCH METHOD

Our overall goal in designing this experiment is to create a setting where we can isolate

the effects of EMC and its interactions with the contextual features of error type and error

originator without dealing with extraneous concerns (e.g. anticipation of detailed review,

litigation due to withdrawal of previous audit reports etc.). Hence, we take advantage of the

power of the experimental method to develop a plausible scenario that isolates the key factors

that we want to examine from those extraneous concerns so we could clearly test our causal

reasoning with strong internal validity. We discuss the effects of these design choices on external

validity in our discussion section that follows the results.

16
Consistent with our focus on EMC‟s contextual effects and due to lack of theory to guide predictions, we do not
make predictions about interactions between error type and error originator and consequently make no predictions
about the three-way interaction between climate, type and originator. However, in our main analysis we control for
these interactions and report the results in the supplemental analysis.

16

Electronic copy available at: https://ssrn.com/abstract=1657117

Design, Task and Procedure

We designed and administered our experiment by using a two (EMC: blame versus open)

by two (error type: mechanical versus conceptual) by two (error originator: own versus peer)

full-factorial design. Participants were randomly assigned to one of the eight treatment

conditions. Materials were provided to the participants in a paper-based format.17

The case used in the current study is an adaptation of the Gronewold et al. (2013) case

that was originally based on the Burgstahler et al. (2000) case (with permission of both author

teams). We employed a PCAOB report (2004d) about sampling errors made in one of the Big-4

firms‟ audits to guide us in making these modifications. The PCAOB case report disclosed that

the audit firm did not project errors found in an inventory sample to the population as required

for any sampling application. Hence, while the PCAOB report was about a conceptual error, it is

relatively easy to modify a sample-related error to be either mechanical (i.e., the correct

arithmetic projection was not done) or conceptual (i.e., no projection at all was done). This

allowed us to hold the context constant rather than using different errors in different contexts to

instantiate error type, resulting in greater internal validity.

Each participant received a packet of case materials, which included a brief introduction

to the research project. The materials included (1) background information about the

hypothetical audit engagement and client, including the description of the EMC manipulation

(open versus blame); (2) a series of working paper extracts including the error type manipulation

17
All experimental instruments were in German and administered in Germany by a native German-speaking author.
The US-based case was adapted as necessary for the German audit environment and then translated into German by
one of the authors. A professional translator independently translated the case materials back from German to
English and the two reconciled any differences, which focused on technical accounting and auditing terms. Further,
four German-speaking business professionals read the case for general understandability and clarity. Next, the audit
firms that agreed to participate read and commented on the case prior to a pre-test. The case was then pre-tested on
15 auditors (none of whom took part in the experiment) from the first audit firm that agreed to take part in our
experiment and modifications were made based on that pre-test to further enhance understandability of the case
materials.

17

Electronic copy available at: https://ssrn.com/abstract=1657117

(mechanical versus conceptual); (3) the dependent variables; (4) manipulation checks; and (5) a

debriefing questionnaire.

Participants learned of an audit carried out by an international audit firm. The background

information described the audit of a medium-sized privately held manufacturing firm with debt

agreements that require audited financial statements. 18 In the planning phase of the audit,

inherent risk, control risk and risk of material misstatement were assessed as moderate and

materiality was set at €1,000,000 (equal to 5% of earnings before taxes). Participants were

informed that the audit plan relied on substantive testing and that all proposed or potential

adjustments based on errors in excess of €50,000 are posted to the “Summary of Non-Booked

Audit Differences (SNBAD).” The materials also stated that in prior years, management had

readily made required adjustments. However, this year, without prior consultation with the audit

firm, the client‟s president has provided shareholders and creditors with preliminary unaudited

earnings information and strongly prefers to minimize adjustments to the financial statements

because he believes that “such adjustments will unduly reduce shareholder and creditor

confidence in management.” Hence, the president has asked the audit firm to propose no

adjustments unless absolutely necessary. Consequently, the office-managing partner of the audit

firm has informed the audit team that only essential adjustments should be proposed to client

management.

Following the EMC description, participants were shown three working papers including

the SNBAD that had been detached from the audit file during the initial review process. The

SNBAD summarized five adjustments leading to a total proposed adjustment that had been

waived as it was well below materiality. The two other working papers were extracts from the

18
There is also a legal requirement in Germany that companies have what is referred to as a statutory audit in the
Anglo-Saxon world.

18

Electronic copy available at: https://ssrn.com/abstract=1657117

lead sheets, providing the details for the adjustment and its calculations. The participants were

told they had been assigned the task to “tidy up” the file for final review by the quality assurance

partner before the audit report would be signed and the financial statements released. It was

during this hypothetical tidying up of the working paper file that participants were informed they

discovered an error (manipulated as either conceptual or mechanical) that had been made during

working paper preparation.19 The participants were told that the error they had found resulted in

discovering additional income-reducing adjustments that needed to be added to the SNBAD. The

combined total of the previous and the newly discovered overstatement of income totaled

€998,000, which is slightly below materiality of €1,000,000.20 The participants learned that this

error had not been discovered by previous reviews and would be unlikely to be detected in the

upcoming quality review.21 There was no indication in the working papers that these errors were

19
Professional standards explicitly recognize that some tidying up of working papers must occur, even after the
completion of the audit, and hence documentation standards allow a specified period after the audit report is signed
before the files are “frozen” (ISA 230 (IAASB 2012a) – 60 days and PCAOB AS No. 3 (PCAOB (2004e) – 45
days). Payne et al. (2010) document that auditors are sensitive to remaining errors in audit work papers despite the
stage of review, hence, providing support that it would be plausible that auditors would detect such errors.
Psychology research, including neuroscience evidence, demonstrates that a basic level of self-monitoring occurs
automatically and consistently to control one‟s own behavior (e.g., Ganushchak and Schiller 2006, 2008a, 2008b;
Rabbitt 1990). Such monitoring occurs even in situations where people are explicitly instructed to ignore any
possible errors that they may have committed (Rabbitt 1969, 1990).
20
If participants had automatically considered a reasonable allowance for sampling risk, the total misstatement
would have to be over €1 million given that the misstatement without sampling risk is just €2,000 less than the
materiality of €1 million. Our computations, employing the information provided in the case, suggest misstatements
would be nearing €1.3 million if sampling risk were to be incorporated in the calculation of misstatement.
21
While the working paper review process results in detecting errors in working papers, previous research has
demonstrated that overall error detection rates rarely exceed 50 percent of errors seeded in the working papers (e.g.,
Asare and McDaniel 1996; Bamber and Ramsay 1997; Tan and Trotman 2003) in experimental studies. This
experimental evidence is reinforced by PCAOB inspection reports (e.g., PCAOB 2004a, 2004b, 2004c, 2004d, 2005;
Hermanson et al. 2007) findings as well as other countries‟ regulator‟s reports (e.g., CPAB 2004, 2005; FRC 2011,
2012). Over three hundred firms with such errors were identified in the largest six auditors‟ PCAOB public reports
(e.g., see PCAOB 2008a for a partial summary) in the USA between 2004 and 2009. Similar findings are reported in
the UK in the Audit Inspection Unit annual reports (see FRC 2011, 2012).

19

Electronic copy available at: https://ssrn.com/abstract=1657117

investigated and deemed to be “isolated” or “contained,” hence these errors needed to be

projected to the population (Hermanson 1997).22

After examining the background information and the working papers, participants were

asked to indicate their willingness to report the additional overstatements they discovered to an

appropriate responsible person for the audit engagement. In addition, participants responded to a

number of manipulation checks, covariate measurements, demographic questions, and questions

related to the understanding of the experimental case.

Independent variables

Error-Management Climate

Following the client description and under the heading “Office climate at your firm,” the

EMC was manipulated as either „blame‟ or „open.‟ See Table 1 for the wording of the EMC

manipulation. The items about EMC are drawn from the set of factors employed by the

underlying psychology and limited accounting research to measure EMC (e.g., Van Dyck et al.

2005; Gronewold and Donle 2011) as specialized to our case setting and are in accordance with

the EMC descriptions in Gronewold et al. (2013). Note that we manipulate multiple EMC items

due to the theoretical nature of the construct as the EMC theory indicates a climate is created by

a consistent set of beliefs, norms and practices being exhibited, especially to support an open

EMC.

<<<Insert Table 1 about here>>>

22
Hermanson (1997) in her literature review and experiment finds that the frequency of occurrence of an error is the
strongest factor in predicting the decision by auditors to project discovered errors to the population rather than
attempt to isolate the error. Our scenario has multiple errors of the same nature with no indications in the working
papers of an attempt to investigate and determine the cause of the error. Thus, there is no basis provided for the
auditor in our scenario to believe that error projection is not required.

20

Electronic copy available at: https://ssrn.com/abstract=1657117

Error Type

All “mechanical error” participants were told of the incorrect arithmetic in the projection

to the population and provided with the correct computation, whereas in the “conceptual error”

condition participants were told about the absence of the projection to the population and given

the correct one.23 The design choice to inform participants about the error discovery rather than

allowing them to self-discover the error was done to ensure we were considering the reporting

decision, not a joint test of discovery ability and reporting. Given that prior research (e.g.,

Harding and Trotman 1999; Tan and Trotman 2003) strongly shows that different error types are

detected at different rates with different amounts of auditor experience we did not want to

confound error discovery with the error reporting judgment.

Error originator

The error originator manipulation featured language that focused on either the participant

learning of his/her own errors (e.g., “you,” “your,” etc.) or those of a named peer. Error

origination was introduced during the description of the audit engagement and consistently

employed throughout the experimental materials. For example, participants in the “own error”

treatment were informed that “you and your audit team” had completed all planned fieldwork

23
To the extent that participants consider sampling risk, we bias against finding our expected results as we expect
that more important errors are more likely to be reported and as noted above adding in sampling risk takes the likely
errors to well above materiality. In the second part of this experiment, after eliciting the dependent variables
reported in the text, we carefully followed Burgstahler et al.‟s (2000, pp. 88-90) within-subjects design, to alert
participants of the need to consider sampling risk in making error projections. Following Burgstahler et al. we add to
the previous part‟s scenario by explicating in the second part that a projection should also include sampling risk
(quantified in the second part as the computed upper confidence limit of €300,000), rendering revised total potential
misstatement to €1.3 million. Participants then responded to the same question set as in the first part reported in this
paper. Overall, the means moved towards a greater willingness to report (F(1, 181)=12.18, p<0.001) consistent with
the argument that more important errors are more likely to be reported. However, while each cell‟s mean moved
towards a greater willingness to report the discovered error the only significant effect was in the peer‟s mechanical
error in the blame EMC. This increase in report willingness brought the mechanical error treatment mean closer to
the position predicted in Figure 1 Panel A (the solid black line) from the first part‟s result in Panel B. Further, this
added within-subjects manipulation rules out the potential confound in Gronewold et al.‟s (2013) research design
where implicitly both the mechanical error and conceptual error treatments in that paper had a further conceptual
error embedded in both of them, confounding the error type manipulation.

21

Electronic copy available at: https://ssrn.com/abstract=1657117

and that ”you yourself” had prepared all working papers that were referred to in the materials. In

the “peer error” condition, another auditor working at the same local office and member of the

audit team (identified as “Andreas Meyer”) was described as having prepared the working papers

in which the errors were detected.24 By design, and to enhance the internal validity of our

research design (enabling us to randomly assign participants to treatments), we chose to have the

participants in the own error discovery setting learn that they made the error (i.e., we could not

ensure such errors would be created as unintentional errors by their very nature occur randomly).

While this design choice potentially weakens this manipulation, we made strong efforts in our

instrument to ensure that participants in the own error condition recognize and invoke the key

circumstances of the situation of becoming aware that some of their own work apparently

contains an error. This manipulation is consistent with underlying error management theory in

that it is becoming aware of a presumed own error that triggers the responses about how to deal

with an own error rather than the original creation of the error (Zhao and Olivera 2006).

Dependent Variable

Participants responded on a scale from 0 (Very Unlikely) to 10 (Very Likely) to the

question: “How likely are you as an auditor at Good and Better WPG to report the problem you

found in the finished goods working paper to an appropriate responsible person for this audit

engagement at Good and Better (e.g., audit team leader, engagement manager, responsible

partner, or quality reviewer)?” 25 This willingness to report measure reflects the emphasis on the

individual auditor‟s responsibility to report and was the same in both own and peer error

conditions.

24
The use of a first name (instead of only the surname) in the German language supports the notion that it is a
relationship with a familiar and friendly peer.
25
Good and Better WPG is the name of the fictitious audit firm in our experiment.

22

Electronic copy available at: https://ssrn.com/abstract=1657117

Sample

198 German auditors from a Big-4 audit firm participated in the experiment administered

during firm training. 26 Eight participants were eliminated: four with missing values for the

dependent variable and four extreme outliers (i.e., the residuals more than 4 standard deviations

from the ANOVA‟s residual mean).27 Two sessions were administered in two different German

locations. At least one of the researchers was present at both experimental sessions to ensure that

the participants followed the experimental instructions and worked quietly and independently.

According to their firms, participants had sufficient audit experience (µ=10.47 months served on

audit engagements) to carry out the task.28 Further, as desired, the participants had limited

experience with reviewing activities (µ=1.2 months). Self-reported experience with reviewer

activities (measured in months) varies significantly across experimental conditions (p<0.05) and

is correlated with the dependent variable; hence, we include it as a control variable. 29 See Table 2

for more demographic details about the participants. 30

<<<Insert Table 2 about here>>>

26
Note that in total, six different audit firms participated in various versions of this research.
27
Further, we examined the responses of the four statistically identified outliers and found their response pattern to
lack internal consistency across questions providing a substantive reason to consider these participants as outliers.
Including the four outliers in the statistical analyses that follows results in substantially similar results with only the
interaction of error climate and error originator becoming insignificant at the p<0.13 level two-sided. All mean
patterns are the same as reported in the paper.
28
Stefaniak and Robertson (2010) report no significant interactions between experience and the variables they
manipulated with participants that varied from a mean of 7.6 months to 39.5 months but there was a main effect for
experience (audit interns versus all others).
29
Ten participants had missing values for review experience. For these missing values we entered the variable‟s
median/mode of 0 months as well as the mean of 1.2 months. Choice of missing value substitution makes no
difference in our analyses.
30
To verify whether randomization across experimental treatments was successful, we examined whether the
following demographic factors vary across treatment conditions: age, gender, auditor rank, academic degree, audit
engagement experience, and experience with similar errors like those in the case. None of these items are
significantly different across treatment conditions (all p‟s>0.20). No other demographic variables are correlated with
the dependent variable.

23

Electronic copy available at: https://ssrn.com/abstract=1657117

 Respondents demonstrated they understood the case by recalling correctly that (1) in prior

year audits, the client had always readily booked audit proposed adjustments (µ=2.90, all scales

were from -5 (strongly disagree) through +5 (strongly agree)), and (2) management is currently

strongly committed to reporting the net income figure as is, unless the audit firm threatens to

qualify the report (µ=2.40). Participants also recalled that the potential adjustments recorded in

the SNBAD prior to the current error being discovered were well below the materiality threshold

(µ=-2.55 reverse-scored). Overall, these responses indicate that participants understood the case.

IV. RESULTS

Manipulation Checks

We employ seven items (e.g., “At this office of Good and Better WPG, there is a climate

characterized by „being open for improvement‟ in the case of errors that have occurred in

carrying out audit procedures.”) based on previous error management research to check the

effectiveness of our manipulation of EMC. Participants indicated their agreement with each item

on a scale from -5 (“Strongly Disagree”) to +5 (“Strongly Agree”). As the seven items loaded

onto a single factor that captures over 50% of the variance of the items we computed an

unweighted average index of the seven.31 The mean of the index for participants in the „open

climate‟ treatment (1.83) is significantly greater than the index mean of the „blame climate‟

(-1.37, p<0.001).32 Thus, we conclude that the relative manipulation of the EMC was successful

without being extreme.

31
The seven items were taken from the “organizational error culture” scale developed by Van Dyck et al. (2005),
which has already been successfully tested in an auditing setting (Gronewold and Donle 2011). In addition to the
exploratory factor analysis reported in the text, a reliability analysis reveals a high Cronbach‟s alpha (0.84),
indicating that the single identified factor measures the underlying construct with a high degree of consistency.
32
All seven items were highly significantly different (p<0.001) in the correct direction when analyzed individually.

24

Electronic copy available at: https://ssrn.com/abstract=1657117

 To test for understanding of the error type manipulation, using again the “Strongly

Disagree” (-5) to “Strongly Agree” (+5) scale, participants indicated their agreement with the

statement “The detected error in the work papers was more like a „mechanical‟ error (e.g.

calculation error, typing error, writing error, or the like, in audit or evaluation steps carried out)

than a „conceptual‟ error (e.g. missing of a complete audit step or of a crucial evaluation step, a

fundamental thinking error, or the like).” The level of agreement for the mechanical error

condition (µ=1.82) was significantly greater (p<0.001) than for the conceptual error (µ=-1.42).

This suggests an effective manipulation of error type.

To test for understanding of the error originator manipulation (using the same -5/+5 scale

as before) participants indicated if they themselves had prepared the working papers on which

the errors were detected. The mean response was 4.02 for the „own error‟ treatment and -3.34 for

the „peer‟s error‟ treatment. The two means are significantly different (p<0.001), suggesting a

successful manipulation of error originator.33

Beyond recognizing and recalling that they were in the own versus peer experimental

condition, two other findings suggest that the participants internalized the own error attribution.

First, error management theory suggests that if people recognize their own work as containing an

error, they engage in coping behavior in response to this discovery (see Rybowiak et al. 1999,

pp. 528-529, with reference to the general coping concept by Lazarus and Folkman 1984; Van

Dyck et al. 2005). One of these coping behaviors is discussion with colleagues about their error

as a way to “normalize” their behavior. Hence, if a self-made error were internalized one means

for the auditor to “normalize” the discovery of this self-made error would be to discuss his/her

33
A second question to check the manipulation of error origination asked participants whether the work papers had
been prepared by their colleague, which provided similarly strong results (-3.59 versus 3.31) in the opposite
direction in accordance with the opposite framing of the question (p<0.001).

25

Electronic copy available at: https://ssrn.com/abstract=1657117

making this error with a colleague; a behavior that would be much less likely to occur than if it

was a peer‟s error discovered as such discussion would merely increase the chance that the

auditor would be labelled as a “snitch” or “tattletale.” Hence, we elicited, on a scale ranging

from “Very Unlikely” (0) to “Very Likely” (10), whether the auditor would discuss the discovery

of the error with his/her colleagues. We found that participants in the own error condition were

likely to discuss the discovered own error with colleagues (Own µ=6.8 which is significantly

above the scale midpoint (p<0.001)) and marginally significantly more likely than in the peer

error condition (Peer µ=6.1; Own-Peer µ difference=0.7, p<.08, one-tailed).

Second, participants had fairly high overall assessment of realism of the case (overall

µ=2.05 on a scale ranging from “Strongly Disagree” (-5) to “Strongly Agree” (+5)) and the own

error treatment was consistent with that assessment (Own µ=2.07). Further, a similar assessment

of case difficulty was elicited (grand µ= 2.31 on the same scale as before) and again the

assessment in the own error condition (µ=2.28) was consistent with the overall assessment.

These findings suggest that the participants saw no difference in the realism or relative difficulty

of the two settings hence they appear to have internalized the manipulations equally at least from

these perspectives.

Taken together, these manipulation checks and the added internalization analyses provide

support that participants in the own error treatment both recognized and recalled that they were

in the own error condition instead of being in the peer error condition. This combination of

results provides support for the assertion that the auditors‟ “internalized” the error as their own,

at least to the same extent that they internalized the peer error discovery and attribution.

26

Electronic copy available at: https://ssrn.com/abstract=1657117

Hypothesis Testing

Table 3 reports the overall analysis for the experiment. The dependent variable is how

likely the participants themselves would be to report the error to relevant superiors.34 Table 3

Panel A‟s ANCOVA shows the overall model is highly significant at conventional levels

(F(1,181)= 3.30, p<0.002).

<<<Insert Table 3 about here>>>

Error-management climate

H1 predicts that the open EMC leads to more report willingness than the blame EMC

across all conditions. The observed means suggest that the open EMC results in a greater

propensity to report errors in some but not all conditions. Hence, while the main effect of EMC

is significant (µ1-µ2=8.78-8.33=0.45, t=1.86, p<0.04, one-tailed) it is qualified by two significant

interactions: EMC with error type (F(1,181)=4.38, p<0.04) and EMC with error originator

(F(1,181)=6.26, p<0.015).

Error-management climate and error type

We find, as predicted in H2 (Figure 1, Panel A), the willingness to report mechanical

errors in the open EMC is greater than in the blame EMC (µ1-µ2=9.09-8.14=0.95, t=2.78,

p<0.01, one-tailed, see Figure 1, Panel B and Table 3, Panel C). Further, while we predicted a

similar directional, albeit weaker effect for conceptual errors being reported, there is no

difference in the reporting likelihood for conceptual errors between the blame and open EMC

34
To check for potential social desirability issues about reporting errors (e.g., Clement and Krueger 2000;
Mikulincer and Horesh 1999; Ruvolo and Fabin 1999; Smith 1997) we also asked participants how their peers
would respond to the same question. There was a significant main effect for own report willingness (µ=8.55 (0.13))
versus predictions about peers‟ report willingness (µ=7.68 (0.15)) (F(1,181)=51.22, p<0.001), that is the auditors
reported they would be more likely to report errors in the working papers than their peers would. Other than this
overall level difference there was no difference between the two responses formats for EMC or its interactions with
error type and error originator.

27

Electronic copy available at: https://ssrn.com/abstract=1657117

conditions (µ1-µ2=8.48-8.53=-0.05, p>0.85). Consistent with H2, we find that there is a greater

increase in willingness to report mechanical errors in the open EMC whereas there is no

difference due to EMC for conceptual errors. This finding qualifies H1 such that an open EMC

increases the willingness to report errors only when the errors in question are mechanical, albeit

all errors are reported at a relatively high level (i.e., greater than 8 out of 10).

Error-management climate interaction with error originator

Error originator interacts significantly with EMC (F(1,181)=6.26, p<0.02). H3 predicts

that the effect of EMC will be greater for peer-committed errors than for self-committed errors.

An examination of the simple effects (see Table 3, Panel D) of EMC in the two error originator

conditions (see Figure 2, Panel B) reveals that auditors are more likely to report peer-committed

errors in an open EMC, as compared to a blame EMC (µ1-µ2=8.91-7.86=1.05, t=3.04, p<0.01),

but when they discover a self-committed error, the effect of EMC is not significant (µ1-µ2=8.65-

8.80=-0.15, p>0.65). Further, in the open EMC the mean report willingness of peer-committed

errors is not significantly different from either own error condition (Open µ=8.91 versus 8.65,

p>0.40; Blame µ=8.91 versus 8.80, p>0.75). Therefore, the effect of open EMC is to increase the

likelihood of reporting peer errors to the same level as reporting own errors (there is no statistical

difference between the three cells, all p-values greater than 0.40), thus supporting the prediction

made in H3.

Other effects

Table 3 Panel A shows that neither of the two main effects (error type: F(1,181)=0.20,

p>0.65; error originator: F(1,181)=2.02, p>0.15) nor the three-way interaction (EMC, error type

and error originator) are significant (F(1,181)= 0.98, p>0.30), the latter being consistent with the

lack of theory to predict a three way interaction. However, while we had no theory that clearly

28

Electronic copy available at: https://ssrn.com/abstract=1657117

supported an interaction prediction for error originator and error type, we find a significant

interaction (F(1,181)= 6.36, p<0.02) (see Table 3, Panel A). Examining the simple effects (see

Table 3, Panel E) we find in the own error condition that mechanical errors are more likely to be

reported than conceptual errors (µ1-µ2=9.09-8.37=0.72, t=2.13, p<0.05). In the peer error

condition the difference appears to reverse (µ1-µ2=8.14-8.64=-0.50), however, the effect is not

significant (t=-1.43, p>0.15). Hence, our post-hoc analysis of the results suggests that auditors‟

willingness to report errors is greater for mechanical errors than for conceptual errors when they

themselves commit the error, but there is no difference in willingness to report the two different

types of errors when they discover a peer committed the error.35 Drilling down further into the

own conceptual errors cell we find that the relatively low mean in that cell is driven by the open

climate (µ=7.92 see Table 3 Panel B) not as might be expected by the blame climate (µ=8.82).

This more nuanced comparison suggests that the relative unwillingness to report conceptual

errors by auditors is driven by auditor‟s own impression management concerns in that auditors

may believe that reporting self-made conceptual errors is more problematic for their image than

reporting their own mechanical errors or peer errors of any type to appropriate parties.

Robustness of experimental findings

All reported tests of mean differences are for least square adjusted means (i.e., adjusted

for the covariate months of review experience). Results do not change when raw mean tests are

carried out (see Table 4, Panels A, B and C for raw means and standard deviations for the three

significant interactions), and significance levels are generally higher. Further, we buffered for

multiple comparisons, where appropriate, in analyzing the simple effects.

35
This finding also suggests that internalization of the own error manipulation was successful as otherwise there
should be no difference in reporting rates for different types (i.e. mechanical versus conceptual) of their own errors.

29

Electronic copy available at: https://ssrn.com/abstract=1657117

 During post-experimental debriefing participants responded on a scale from -5 to +5 to

the following question: “Has your audit experience been characterized as tending to be

predominately one of „getting it right the first time‟ or „being open for improvement‟ when errors

have been found in carrying out audit procedures?” anchoring on each descriptor plus a small

explanation at each end of the scale. We included this continuous variable in an ANCOVA to

examine how overall results would be affected when controlling for participants‟ own error-

climate assessment. The ANCOVA main effect of EMC moves from p<0.06 to p<0.04. The three

interaction terms also become somewhat more significant, but the overall results remain

equivalent to reported results. As further evidence of successful random assignment, if we

replace the manipulated climate variable with the participants‟ own EMC experience (we

dichotomized the scale responses based on the median own EMC experience response), we

found no significant main effect for own EMC or interactions with our other two variables.

Hence, these robustness checks provide additional evidence that participants accepted the error

climate manipulation and that their own current experience with EMC did not affect their

experimental judgments.

The grand (adjusted) mean for the dependent variable in the experiment was µ=8.57 with

roughly 33% of respondents at the top value of the scale “10.” This may indicate the presence of

“ceiling effects” (i.e., too high of a percentage of the data is clustered at the top of the scale).

However, research on ceiling effects finds analysis problems only with a much higher percentage

of the respondents clustered at one end of the scale (e.g., Uttl 2005). Further, any ceiling effect

biases against our finding significant effects in our hypotheses tests.

30

Electronic copy available at: https://ssrn.com/abstract=1657117

 V. CONCLUSION

In this study we examine the effects of audit office error-management climate differences

on an auditor‟s willingness to report errors focusing on interactions with error type and error

originator. We employ the advantages of the experimental method to control for factors like

review process and auditor reputation effects so as to focus on the effects of EMC on error

reporting. We provide evidence that an open EMC causes a significant increase in reporting

propensity in several, but not all of the contexts, which we examined. The open versus blame

EMC results in a greater willingness to report errors when the error is mechanical, bringing the

reporting of mechanical errors up to being statistically indistinguishable from the high level of

report willingness for conceptual errors. Moreover, an open EMC increases the likelihood of

reporting peer errors versus a blame EMC. Overall, the evidence is compelling that an open

EMC affects the willingness of auditors to report post-initial review working paper errors under

most circumstances and with the exception of self-discovered own conceptual errors in an open

climate does no worse than a blame climate.

Limitations to this study include the following. First, our manipulation of the EMC was

possibly too strong and obvious, resulting in potential demand effects. However, a demand effect

suggests a lower likelihood of discovering interactions as there would be no difference in the

demand effect across conditions, yet the two hypothesized two-way interactions are significant.

Second, we manipulated error type (mechanical versus conceptual) only one way and other

operationalizations might lead to different results. However, our approach is consistent with the

approach in previous research (e.g., Ramsay 1994), drawn from a published “real world”

example (PCAOB 2004d) and previous research finds that operationalization differences, even

based on redefining the construct do not matter (Owhoso et al 2002). Third, it is possible that

participants (being staff auditors) who were denoted in the case as being “audit seniors” might

31

Electronic copy available at: https://ssrn.com/abstract=1657117

not have responded as they would have given their current role. Only further research with more

experienced auditors can resolve this issue but we note that experience is reported as having only

a main effect in the closest analogous research (Stefaniak and Robertson 2010). Fourth, in order

to exert strong experimental control we told participants in the own error condition about having

made the error instead of allowing them to potentially make the error themselves, which by its

very nature can only provide an approximation to the real-world situation of a truly self-made

error. However, we made strong efforts in our instrument to ensure that participants recognized

and invoked the key circumstances of the typical situation of becoming aware of an own error. A

number of checks based on error management theory and our empirical findings provide support

for participants‟ “internalization” of the error as their own, so we believe they behaved

accordingly in the experiment.

Fifth, we chose a setting where the audit error was one that directly affected the financial

statement (i.e., overstatement of income). Auditors may deal differently with the discovery of

omitted audit procedures or results that indicate the previous auditor made incorrect decisions

(see Peytcheva and Gillett 2012; Stefaniak and Robertson 2010). Sixth, it may be argued that a

blame EMC might lead auditors to commit fewer errors in the first place because it imposes

higher sanctions and hence motivates auditors to avoid errors. This is ultimately an empirical

question to be addressed by future research. Research in other domains finds that a blame EMC

creates an organization where employees are better at hiding errors they have discovered but not

reducing the number of errors initially committed (e.g., Edmondson 1996; Itoh et al. 2009;

Reason 2000). Seventh, it is possible that in some of our tests we experienced ceiling effects,

which might obscure effects that are present but not significant due to end of scale. However, our

finding of the two hypothesized interactions as significant suggests this is not a large concern.

32

Electronic copy available at: https://ssrn.com/abstract=1657117

Finally, we employed a paper-based case whereas audit firms are increasingly employing

electronic working paper systems. However, prior research suggests that post-initial review

working paper consultation is likely to be more common where review occurs real-time rather

than being carried out in a strict sequential manner after all field work by preparers is finalized

(e.g., Barrett et al. 2005; Brazel et al. 2004; Winograd et al. 2000). This finding suggests that

changes to the review environment towards real time review might lead to even more

opportunities for auditors to discover errors.

Our findings have several implications. We find that an open EMC enhances the

reporting of mechanical errors as well as peers‟ errors that auditors come across in the expected

fashion, and with the exception of self-discovered own conceptual errors, which results in no less

likelihood of reporting other types of errors than a blame EMC does. This finding suggests that

audit firms‟ efforts to create an open EMC that is an espoused goal found in all of the major

firms‟ websites has positive benefits if actually implemented in the individual audit office.

While audit firms espouse formal policies that are consistent with an open EMC they may not be

as vigilant about individual audit office management interpretation of those policies if those

managers achieve their audit office profit goals (e.g., see Ettredge et al. 2011). Further, there is a

strong rhetorical value associated with a “getting it right the first time” culture that might appeal

to audit firm office management, especially in light of the regulatory pressures in the current

audit environment that arise from third party inspection of audit files. Given that audit firms can

directly control the EMC in their practice offices via selection, socialization, promotion and

evaluation/compensation decisions our results suggest that activities to support the

implementation of the open EMC have value to the firm beyond personnel retention and extend

to substantive improvements in audit quality, at least at the margins. Hence, our results provide

33

Electronic copy available at: https://ssrn.com/abstract=1657117

support for firm level management that supports more consistent implementation of open EMC

norms in all practice offices, irrespective of whether there is strong shorter term office

profitability or not.

Moreover, our findings suggest that the reporting of conceptual errors is unaffected by an

open EMC. In light of our expectations from the underlying EMC literature, we find this

puzzling. Given that conceptual errors may be more important than mechanical errors to audit

quality, further research might be undertaken to confirm this finding as well as to attempt to see

what aspects of the EMC that might have to be strengthened in the audit context to enhance the

willingness of auditors to report conceptual errors. Our post-hoc finding that auditors are

significantly more willing to report their own mechanical errors than their own conceptual errors

(while the difference between reporting willingness of peers‟ mechanical and conceptual errors is

insignificant) suggests that this obstacle for reporting conceptual errors stems from auditors‟ own

impression management concerns. This suggestion is reinforced by our deeper probing into the

own conceptual errors where we found that this cell‟s relatively low mean was driven by the

open EMC not, as might be expected, by the blame EMC. This may provide one line of inquiry,

that is, to help determine what aspects of the open EMC could be emphasized in an audit firm to

obtain, what we consider, the desired outcome of increased reporting of conceptual errors.

Another possible way to proceed might be to further adapt or enhance the open EMC so

that it is rendered successful in relieving auditors‟ concerns about such negative competence

attributions associated with impression management. The challenge for such an open EMC

would be to incorporate notions that make it believable for (and ultimately believed by) auditors

that reporting errors with potential “negative competence signals” will not entail any negative

competence attributions or other “disadvantages,” but to the contrary would be positively noticed

34

Electronic copy available at: https://ssrn.com/abstract=1657117

by superiors and ultimately assessed as a strength. It is conceivable, though, that very strong

signals in this regard would be necessary to achieve this (increasing the difficulty of putting

these signals into everyday practice consistently which is crucial to establishing and maintaining

an open EMC). One way might be to promote the notion that reporting an own conceptual error

that allegedly might be indicative of a lack of professional competence actually reveals that the

reporter does have the competence, because otherwise he or she could not have determined that

this was an error and now report it.

The generally high willingness to report own errors in our experiment suggests that a key

question to make use of this additional potential to uncover errors (that were not detected in the

review process) is how to enhance the likelihood that auditors will indeed self-detect these

errors. A number of arguments plausibly suggest that self-detection may occur automatically to a

certain extent given humans‟ self-monitoring processes and institutional features of the audit

process such as typically referring back to finished steps when, e.g., auditing the notes or other

steps that are usually performed at rather late stages of an audit. However, it would of course be

desirable to systematically increase this probability of detection. Promoting auditors‟ self-

assessment (or “self-reflection”) of their performed work as a formally required task might be a

possible means.

Finally, while research regarding a potentially effective open EMC that enhances the

reporting of conceptual errors is still required, the quality-enhancing functionality of an open

EMC with respect to peers‟ errors and mechanical errors that was demonstrated in our study is

also subject to certain challenges for audit firms and offices and for audit regulators. Given EMC

is a set of shared beliefs, norms, and common practices regarding the management of errors in

the organization (Van Dyck et al. 2005), consistency of application of these beliefs, norms and

35

Electronic copy available at: https://ssrn.com/abstract=1657117

practices is especially important in creating and maintaining an open climate. There is an

increased focus by audit standard setters and regulators on management processes in audit firms

as a potential systematic cause for audit quality concerns. To date, regulators have suggested that

audit firms tie audit quality to partner and staff performance evaluations and remuneration more

explicitly (FRC 2012, p. 13) as well as to the firm‟s disciplinary practices (e.g., PCAOB 2008b).

Furthermore, unless audit firms follow regulator advice on these matters, additional disclosures

from the regulators‟ private reports to firms are made public (e.g., PCAOB 2008b on Deloitte),

potentially causing reputational damages to the audit firms affected.

If an open EMC has the benefits that we suggest it has in this study, a key question is

whether a regulator would recognize the value of an open EMC. An open EMC would have

individual auditors‟ errors (i.e., potential audit quality indicators) go unpunished if they were

learned from and not repeated by the auditor. Further, the audit firm would have to ensure that

others in the firm would learn from the individuals‟ errors potentially enabling regulators to find

“audit quality” issues that they could report on publicly. Hence, if ways were found to encourage

auditors to report more conceptual errors rather than the mechanical errors our current study

documents as occurring in an open EMC, would the audit firm not just be delivering to

regulators an easy means to further sanction audit firms for low audit quality? This could occur

even though the enhanced self-identification and correction of such errors would contribute to

better information being made available to markets. Finally, regulators‟ suggestions about

linking performance evaluation and remuneration of individual auditors to audit quality would

require regulators to be able to distinguish individual errors that are learned from, from repeated

errors that might indicate a lack of concern for audit quality. This may be a very demanding task

for regulators although regulators in other industries (e.g. known as high reliability organizations

36

Electronic copy available at: https://ssrn.com/abstract=1657117

such as nuclear power and medicine; see Provera et al. 2008 for a review) have shown such

abilities. It remains an open question if audit regulators can do the same.

In the broader accounting context, the generalizability of the functionality of an open

EMC to the reporting of peers‟ errors as documented in our study suggests that such a climate

may also be functional in other team work contexts where people may come across errors

committed by their peers, as well as in related fields such as corporations‟ compliance

management activities and systems. In other words, firms that exhibit an open EMC might have

a stronger implementation of the espoused “tone at the top” thus strengthening the internal

control system through a greater willingness of employees to report errors that they have

encountered with the opposite occurring at firm‟s where a blame EMC predominates. In other

wordsfact, if supported by further research, client firm EMC might well be a factor to take into

consideration when evaluating internal controls. Indeed, at the broader level, compliance

management across managerial functions shares a number of structural characteristics with the

context that we studied, as it might well affect the performance of firms if the organizations‟

members were able and willing to report both own and other organization members‟

noncompliant behaviors that they become aware of. Hence, EMC might well have applications

in the broader study of management control systems in firms as well as in studying decisions by

financial statement preparers in larger firms where many individuals are involved in financial

statement preparation often across multiple locations and indeed countries. Further, implications

could be found for making decisions about internal audit activities as different EMCs among

units of a company might well affect the strength of the control systems use in different

locations.

37

Electronic copy available at: https://ssrn.com/abstract=1657117

 In conclusion there appear to be benefits for audit quality from audit offices with an open

EMC under certain circumstances. However, there are significant forces at work from audit

regulators who want to see audit firms link audit quality to performance evaluation, remuneration

and disciplinary practices (e.g., PCAOB 2008b; FRC 2011) that might unintentionally work

against creating such an open climate. Further, within audit firm pressures for profitable audit

practices might value local office managers who meet profit targets (see Ettredge et al. 2011) but

who exhibit beliefs, norms and practices inconsistent with an open EMC. As we noted, while

audit firms tend to have formal policies that are consistent with an open EMC, such policies by

themselves cannot create an open EMC. Hence, audit firm management, regulators and

researchers need to be aware of the beliefs, norms and practices at the local office level when

considering the effectiveness of a firm‟s EMC. Finally, as we noted, the open EMC is not a

panacea for all error reporting problems unless means can be found so that an open EMC results

in an auditor‟s willingness to report own conceptual errors to be similar to the high level of

reporting their own mechanical errors and all types of peer errors.

38

Electronic copy available at: https://ssrn.com/abstract=1657117

 REFERENCES

American Institute of Certified Public Accountants (AICPA). 2009. Statement on Quality

Control Standard (SQCS) 10: A Firm’s System of Quality Control. New York, NY:
AICPA.
Andreoli, N., and J. Lefkowitz. 2009. Individual and organizational antecedents of misconduct in
organizations. Journal of Business Ethics 85 (3): 309–332.
Andrews, A., and S. Freeman. 2001. E-lessons learned. CA Magazine. September.
Antle, R., and B. Nalebuff. 1991. Conservatism and auditor-client-negotiations. Journal of
Accounting Research 29 (Supplement): 31–54.
Asare, S. K., and L. S. McDaniel. 1996. The effects of familiarity with the preparer and task
complexity on the effectiveness of the audit review process. The Accounting Review 71 (2):
139–159.
Bamber, E. M., and R. J. Ramsay. 1997. An investigation of the effects of specialization in audit
workpaper review. Contemporary Accounting Research 14 (3): 501–513.
Barnett, T., K. Bass, and G. Brown. 1996. Religiosity, ethical ideology, and intentions to report a
peer‟s wrongdoing. Journal of Business Ethics 15 (11): 1161–1174.
Barrett, M., D. J. Cooper, and K. Jamal. 2005. Globalization and the coordinating of work in
multinational audits. Accounting, Organizations and Society 30 (1): 1–24.
Bowrin, A. R. 1998. Review and synthesis of audit structure literature. Journal of Accounting
Literature 17: 40–71.
Brazel, J. F., C. P. Agoglia, and R. C. Hatfield. 2004. Electronic versus face-to-face review: The
effects of alternative forms of review on auditors‟ performance. The Accounting Review 79
(4): 949–966.
Burgstahler, D., S. M .Glover, and J. Jiambalvo. 2000. Error projection and uncertainty in the
evaluation of aggregate error. Auditing: A Journal of Practice & Theory 19 (1): 79–99.
Canadian Public Accountability Board (CPAB). 2004. Public Report on Initial Quality
Inspections of the Four Largest Accounting Firms. October 6. Available at:
http://www.cpab-ccrc.ca/EN/content/1st_Public_Report_OCt_6_2004.pdf. Toronto, CA:
CPAB.
–––––. 2005. Second Public Report on Quality Inspections of Public Accounting Firms. August
18. Available at: http://www.cpab-
ccrc.ca/EN/content/2nd_Public_Report_Aug_18_2005_eng.pdf. Toronto, CA: CPAB.
–––––. 2012. Meeting the Challenge “A Call to Action” Annual Report 2011. Available at
http://www.cpab-ccrc.ca/EN/content/2011Public_Report_EN.pdf. Toronto, CA: CPAB.
Choi, J.-H., F. Kim, J.-B. Kim, and Y. S. Zang. 2010. Audit office size, audit quality and audit
pricing. Auditing: A Journal of Practice & Theory 29 (1): 73–97.
Clement, R. W., and J. Krueger. 2000. The primacy of self-referent information in perceptions of
social consensus. British Journal of Social Psychology 39 (2): 279–299.

39

Electronic copy available at: https://ssrn.com/abstract=1657117

–––––, A. Glavovic, J. Ng, and D. R. Woodliff. 2008. The moral intensity of reduced audit
quality acts. Auditing: A Journal of Practice & Theory 27 (1): 127–149.
Coram, P., J. Ng, and D. R. Woodliff. 2004. The effect of risk of misstatement on the propensity
to commit reduced audit quality acts under time budget pressure. Auditing: A Journal of
Practice & Theory 23 (2): 159–167.
Covaleski, M. A., M. W. Dirsmith, J. B. Heian, and S. Samuel. 1998. The calculated and the
avowed: Techniques of discipline and struggles over identity in Big Six public accounting
firms. Administrative Science Quarterly 43 (2): 293–328.
DeAngelo, L. E. 1981. Auditor size and audit quality. Journal of Accounting and Economics 3
(3): 183–199.
Dirsmith, W., and M. A. Covaleski. 1985. Informal communications, nonformal communications
and mentoring in public accounting firms. Accounting, Organizations and Society 10 (2):
149–170.
Donnelly, D. P., J. J. Quirin, and D. O‟Bryan. 2003. Auditor acceptance of dysfunctional audit
behavior: An explanatory model using auditors‟ personal characteristics. Behavioral
Research in Accounting 15 (1): 87–110.
Edmondson, A. C. 1996. Learning from mistakes is easier said than done: Group and
organizational influences on the detection and correction of human error. Journal of
Applied Behavioral Science 32 (1): 5–28.
Emett, S., J. Pickerd, S. L. Summers, and D. A. Wood. 2012. Which bad apple spoils the batch?
An examination of how entry-level employees respond to tone at the top vis-à-vis tone at
the bottom. Cornell University working paper. Available at SSRN:
http://ssrn.com/abstract=2008932.
Epps, K. K., and W. F. Messier, Jr. 2007. Engagement quality reviews: A comparison of audit
firm practices. Auditing: A Journal of Practice & Theory 26 (2): 167–181.
Ettredge, M., C. Li, and E. Emeigh. 2011. Fee pressure and audit quality. University of Kansas
working paper. Available at SSRN: http://ssrn.com/abstract=1903860 or
http://dx.doi.org/10.2139/ssrn.1903860.
Financial Reporting Council (FRC). 2011. Audit Inspection Unit Annual Report 2010/11.
Available at http://www.frc.org.uk/Our-Work/Conduct/Audit-Quality-Review/Audit-
Quality-Review-annual-reports.aspx. London: UK FRC.
–––––. 2012. Audit Quality Inspections Annual Report 2011/12. Available at
http://www.frc.org.uk/Our-Work/Conduct/Audit-Quality-Review/Audit-Quality-Review-
annual-reports.aspx. London: UK FRC.
Ferguson, A., J. Francis, and D. Stokes. 2003. The effects of firm-wide and office-level industry
expertise on audit pricing. The Accounting Review 78 (2): 429–448.
Francis, J., L. E. Maydew, and H. C. Sparks. 1999. The role of big 6 auditors in the credible
reporting of accruals. Auditing: A Journal of Practice & Theory 18 (2): 17–34.
Ganushchak, L. Y., and N. O. Schiller. 2006. Effects of time pressure on verbal self-monitoring:
An ERP study. Brain Research 1125: 104–115.

40

Electronic copy available at: https://ssrn.com/abstract=1657117

–––––, and –––––. 2008a. Motivation and semantic context affect brain error-monitoring
activity: An event-related brain potentials study. NeuroImage 39 (1): 395–405.
–––––, and –––––. 2008b. Brain error-monitoring activity is affected by semantic relatedness:
An event-related brain potentials study. Journal of Cognitive Neuroscience 20 (5): 927–
940.
Gibbins, M., S. Salterio, and A. Webb. 2001. Evidence about auditor–client management
negotiation concerning client‟s financial reporting. Journal of Accounting Research 39 (3):
535–563.
–––––, and K. T. Trotman. 2002. Audit review: Managers‟ interpersonal expectations and
conduct of the review. Contemporary Accounting Research 19 (3): 411–444.
Gillingham, D. W., J. Blanco, and J. H. Lewko. 1997. An integrated model of error management.
Disaster Prevention and Management 6 (3): 186–190.
Gronewold, U., and M. Donle. 2011. Organizational error climate and auditors‟ predispositions
toward handling errors. Behavioral Research in Accounting. 23 (2): 69–92.
–––––, A. Gold, and S. Salterio. 2013. Reporting self-made errors: The impact of organizational
error-management climate and error type. Journal of Business Ethics. Forthcoming.
Harding, N., and K. T. Trotman. 1999. Hierarchical differences in audit workpaper review
performance. Contemporary Accounting Research 16 (4): 671–684.
Hermanson, D. R., R. W. Houston, and J. C. Rice. 2007. PCAOB inspections of smaller CPA
firms: Initial evidence from inspection reports. Accounting Horizons 21 (2): 137–152.
Hermanson, H. M. 1997. The effects of audit structure and experience on auditors‟ decisions to
isolate errors. Behavioral Research in Accounting 9 (Supplement): 76–93.
Homsma, G. J., C. Van Dyck, D. De Gilder, P. L. Koopman, and T. Elfring. 2009. Learning from
error: The influence of error incident characteristics. Journal of Business Research 62 (1):
115–122.
Hyatt, T. A., and D. F. Prawitt. 2001. Does congruence between audit structure and auditors‟
locus of control affect job performance? The Accounting Review 76 (2): 263–274.
International Auditing and Assurance Standards Board (IAASB). 2012a. International Standard
on Auditing (ISA) 230: Audit Documentation. Available at:
http://www.ifac.org/sites/default/files/publications/files/A012 2012 IAASB Handbook ISA
230.pdf. New York, NY: International Federation of Accountants (IFAC).
–––––. 2012b. International Standard on Quality Control (ISQC) 1: Quality control for firms
that perform audits and reviews of financial statements, and other assurance and related
services engagements. Available at:
http://www.ifac.org/sites/default/files/publications/files/A008 2012 IAASB Handbook
ISQC 1.pdf. New York, NY: IFAC.
–––––. 2012c. International Standard on Auditing (ISA) 705: Modifications to the opinion in the
Independent Auditor’s Report. Available at:
http://www.ifac.org/sites/default/files/publications/files/A038 2012 IAASB Handbook ISA
705.pdf. New York, NY: IFAC.

41

Electronic copy available at: https://ssrn.com/abstract=1657117

Itoh, K., N. Omata, and H. B. Andersen. 2009. A human error taxonomy for analysing healthcare
incident reports: Assessing reporting culture and its effects on safety performance. Journal
of Risk Research 12 (3–4): 485–511.
Kaptein, M. 2011. From inaction to external whistleblowing: The influence of the ethical culture
of organizations on employee responses to observed wrongdoing. Journal of Business
Ethics 98 (3): 513–530.
King, G. 2001. Perceptions of Intentional Wrongdoing and Peer Reporting Behavior Among
Registered Nurses. Journal of Business Ethics 34 (1): 1–13.
Lazarus, R. S., and S. Folkman. 1984. Stress, appraisal, and coping. New York, NY: Springer.
Mikulincer, M., and N. Horesh. 1999. Adult attachment style and the perception of others: The
role of projective mechanisms. Journal of Personality and Social Psychology 76 (6): 1022–
1034.
Miller, D. L., and S. Thomas. 2005. The impact of relative position and relational closeness on
the reporting of unethical acts. Journal of Business Ethics 61 (4): 315–328.
Otley, D., and B. Pierce. 1995. The control problem in public accounting firms: an empirical
study of the impact of leadership style. Accounting, Organizations and Society 20 (8): 405–
20.
–––––, and –––––. 1996. The operation of control systems in large audit firms. Auditing: A
Journal of Practice & Theory 15 (2): 65–84.
Owhoso, V. E., W. F. Messier, Jr., and J. G. Lynch, Jr. 2002. Error detection by industry-
specialized teams during sequential audit review. Journal of Accounting Research 40 (3):
883–900.
Palmrose, Z.-V., V. J. Richardson, and S. Scholz. 2004. Determinants of market reactions to
restatement announcements. Journal of Accounting & Economics 37 (1): 59–89.
–––––, and S. Scholz. 2004. The circumstances and legal consequences of non-GAAP reporting:
Evidence from restatements. Contemporary Accounting Research 21 (1): 139–180.
Payne, E. A., R. J. Ramsay, and E. M. Bamber. 2010. The effect of alternative types of review on
auditors‟ procedures and performance. Auditing: A Journal of Practice & Theory 29 (1):
207–220.
Peecher, M., R. Schwartz, and I. Solomon. 2007. It‟s all about audit quality: Perspectives on
strategic-systems auditing. Accounting Organizations and Society 32 (6): 463–485.
Pershing, J. L. 2003. To snitch or not to snitch? Applying the concept of neutralization
techniques to the enforcement of occupational misconduct. Sociological Perspectives 46
(2): 149–178.
Peytcheva, M., and P. R. Gillett. 2012. Auditor perceptions of prior involvement and reputation
threats as antecedents of quality threatening audit behavior. Managerial Auditing Journal.
27 (9): 796–820.
Provera, B., A. Montefusco, and A. Canatow. 2008. A „no blame‟ approach to organizational
learning. British Journal of Management 21 (4): 1057–1074.

42

Electronic copy available at: https://ssrn.com/abstract=1657117

Public Company Accounting Oversight Board (PCAOB). 2004a. Report on the 2003 Limited
Inspection of Deloitte & Touche LLP. August 26. Available at: http://pcaobus.org/
Inspections/Reports/Documents/2003_Deloitte_and_Touche.pdf. Washington, D.C.:
PCAOB.
–––––. 2004b. Report on the 2003 Limited Inspection of Ernst & Young LLP. August 26.
Available at: http://pcaobus.org/Inspections/Reports/Documents/
2003_Ernst_and_Young.pdf. Washington, D.C.: PCAOB.
–––––. 2004c. Report on the 2003 Limited Inspection of KPMG LLP. August 26. Available at:
http://pcaobus.org/Inspections/Reports/Documents/2003_KPMG.pdf. Washington, D.C.:
PCAOB.
–––––. 2004d. Report on the 2003 Limited Inspection of PricewaterhouseCoopers LLP. August
26. Available at: http://pcaobus.org/Inspections/Reports/Documents/
2003_Pricewaterhouse_Coopers.pdf. Washington, D.C.: PCAOB.
–––––. 2004e. Auditing Standard (AS) No. 3: Audit Documentation. June 9. Available at:
http://pcaobus.org/Standards/Auditing/Pages/Auditing_Standard_3.aspx. Washington,
D.C.: PCAOB.
–––––. 2005. Report on the 2004 Inspection of Ernst & Young LLP. November 17. Available at:
http://pcaobus.org/Inspections/Reports/Documents/2005_Ernst_and_Young.pdf.
Washington, D.C.: PCAOB.
–––––. 2008a. Report on the PCAOB’s 2004, 2005, 2006, and 2007 Inspections of Domestic
Annually Inspected Firms. PCAOB Release No. 2008-008, December 5. Available at:
http://pcaobus.org/Inspections/Documents/2008_12-05_Release_2008-008.pdf.
Washington, D.C.: PCAOB.
–––––. 2008b. Report on 2007 Inspection of Deloitte & Touche LLP (including the Part II
quality control criticisms that were originally redacted) May 18. Available at
http://pcaobus.org/Inspections/Reports/Documents/2008_Deloitte.pdf. Washington, D.C.:
PCAOB.
–––––. 2012. Information For Audit Committees About The PCAOB Inspection Process Release
2012-003. Available at
http://pcaobus.org/Inspections/Documents/Inspection_Information_for_Audit_Committees.
pdf. Washington, D.C., PCAOB.
Rabbitt, P. 1969. Psychological refractory delay and response-stimulus interval duration in serial,
choice-response tasks. Acta Psychologica 30: 195–219.
–––––. 1990. Age, IQ and awareness, and recall of errors. Ergonomics 33 (10/11): 1291–1305.
Ramsay, R. J. 1994. Senior/manager differences in audit workpaper review performance. Journal
of Accounting Research 32 (1): 127–135.
Rathert, C., and D. R. May. 2007. Health care work environments, employee satisfaction, and
patient safety: Care providers‟ perspectives. Health Care Management Review 31 (1): 2–
11.
Reason, J. T. 2000. Human error: Models and management. British Medical Journal 320 (7237):
768–770.

43

Electronic copy available at: https://ssrn.com/abstract=1657117

Reynolds, J. K., and J. Francis. 2000. Does size matter? The influence of large clients on office-
level auditor reporting decisions. Journal of Accounting and Economics 30 (3): 375–400.
Rich, J. S., I. Solomon, and K. T. Trotman. 1997. The audit review process: A characterization
from the persuasion perspective. Accounting, Organizations and Society 22 (5): 481–505.
Ruvolo, A. P., and L. A. Fabin. 1999. Two of a kind: Perceptions of own and partner's
attachment characteristics. Personal Relationships 6 (1):57–79.
Rybowiak, V., H. Garst, M. Frese, and B. Batinic. 1999. Error orientation questionnaire (EOQ):
Reliability, validity, and different language equivalence. Journal of Organizational
Behavior 20 (4): 527–547.
Seifert, D. L., J. T. Sweeney, J. Joireman, and J. M. Thornton. 2010. The influence of
organizational justice on accountant whistleblowing. Accounting, Organizations and
Society 35 (7): 707–717.
Smith, E. 1997. Private selves and shared meanings: Or forgive us for our projections as we
forgive those who project into us. Psychodynamic Counseling 3 (2): 117–131.
Stefaniak, C., and J. C. Robertson. 2010. When auditors err: How mistake significance and
superiors‟ historical reactions influence auditors‟ likelihood to admit a mistake.
International Journal of Auditing 14 (1): 41–55.
Tan, H.-T., and K. T. Trotman. 2003. Reviewers‟ responses to anticipated stylization attempts by
preparers of audit workpapers. The Accounting Review 78 (2): 581–604.
Tax, S. S., and S. W. Brown. 1998. Recovering and learning from service failure. Sloan
Management Review 40 (1): 75–88.
Trevino, L. K., and B. Victor. 1992. Peer reporting of unethical behavior: A social context
perspective. Academy of Management Journal 35 (1): 38–64.
Tucker, A. L., and A. Edmondson. 2003. Why hospitals don‟t learn from failures: Organizational
and psychological dynamics that inhibit system change. California Management Review 45
(2): 55–72.
Uribe, C. L., S. B. Schweikhart, D. S. Pathak, and G. B. March. 2002. Perceived barriers to
medical-error reporting: An exploratory investigation. Journal of Healthcare Management
47 (4): 264–279.
Uttl, B. 2005. Measurements of individual differences: Lessons from memory assessment in
research and clinical practice. Psychological Science 16 (6): 460–467.
Vadera, A.K., R. V. Aguilera, and B. B. Caza. 2009. Making sense of whistle-blowing‟s
antecedents: Learning from research on identity and ethics programs. Business Ethics
Quarterly 19 (4): 553–586.
Van Dyck, C. 1997. Error climate and management issues: Its links to organizational
performance. In Managerial Behavior and Practices: European Research Issues, edited by
R. Pepermans, A. Buelens, C. J. Vinkenburg, and P. G. W. Jensen, 107–116. Amersfoort,
NL: Acco.
–––––. 2009. The tragic 1996 Everest expedition: A tale of error culture. Netherlands Journal of
Psychology 65 (1): 22–34.

44

Electronic copy available at: https://ssrn.com/abstract=1657117

–––––, M. Frese, M. Baer, and S. Sonnentag. 2005. Organizational error management culture
and its impact on performance: A two-study replication. Journal of Applied Psychology 90
(6): 1228–1240.
Victor, B., L. K. Trevino, and D. L. Shapiro. 1993. Peer reporting of unethical behavior: The
influence of justice evaluations and social context factors. Journal of Business Ethics 12
(4): 253–263.
Wild, D., and E. H. Bradley. 2005. The gap between nurses and residents in a community
hospital‟s error-reporting system. Journal on Quality and Patient Safety 31 (1): 13–20.
Winograd, B. N., J. S. Gerson, and B. L. Berlin. 2000. Audit practices of
PricewaterhouseCoopers. Auditing: A Journal of Practice & Theory 19 (2): 175–182.
Zhao, B., and F. Olivera. 2006. Error reporting in organizations. Academy of Management
Review 31 (4): 1012–1030.

45

Electronic copy available at: https://ssrn.com/abstract=1657117

 Table 1
Error-Management Climate Manipulation

Panel A: First Introduction of Office Error-Management Climate Manipulation

(differences between versions are underlined)

Blame Open
The overall climate of this office of your firm is The overall climate of this office of your firm is
noted for a “getting it right the first time” noted for an “open for improvement”
mentality that reflects the office managing mentality that reflects the office managing
partner‟s own beliefs and actions. partner‟s own beliefs and actions.
Errors in carrying out audit procedures are seen Errors in carrying out audit procedures are seen
as signs of incompetence and as a natural part of learning and
reviewers who miss errors are reviewers who miss errors are
at risk of at least demotion back to being coached by more experienced reviewers so they
preparers. do not miss similar problems in the future.
You have seen top seniors coming from review You have seen top seniors coming from review
meetings with managers and partners – after meetings with managers and partners – after
discussions about errors in files – discussions about errors in files –
nearly reduced to tears. with a renewed determination to work harder
next time.
Performance evaluations Performance evaluations do not
document such problems, document such problems
if the person does not repeat them or is making
progress in addressing them,
complicating possibilities for future promotion allowing possibilities for future promotion
within the firm. within the firm.

Panel B: Error-Management Climate Manipulation is Reinforced Later in Case

Blame Open

In part you are relieved when you think that In part you are relieved when you think that
the error cannot be easily detected; the error cannot be easily detected
because you are not certain you want to go as it will result in a great learning opportunity
through a detailed debriefing about how you when you sit down with the reviewers to
could have been “so stupid” as to miss determine how to ensure that this type of error
something this obvious – despite is not missed again – especially given
the fact that two levels of review had also the fact that two levels of review had also
missed it. One thing is for certain: if you report missed it. One thing is for certain: if you report
this, your annual performance evaluation is this, your annual performance evaluation is
going to become a much more negative one. not going to be affected unless you continue to
make the same error, which certainly will not
occur.

46

Electronic copy available at: https://ssrn.com/abstract=1657117

 Table 2
Sample Demographics

Variable N=190*
Gender
Male 109 (57.4%)
Female 70 (36.8%)
Missing 11 (5.8%)
Rank
Assistant auditor 176 (92.6%)
Senior-associate 2 (1.1%)
Senior 3 (1.6%)
Other 1 (0.5%)
Missing 8 (4.2%)
Academic degree
Bachelor 23 (12.1%)
Master 6 (3.2%)
German Diploma, Master, or State Examination 151 (79.5%)
PhD 1 (0.5%)
Missing 9 (4.7%)

Variable Mean Standard N=190*

Deviation
Age 27.58 2.15 181
Audit engagement experience (months) 10.47 4.66 183
Reviewer experience (months) 1.20 3.14 180
Statistical audit sampling experience (0=Never; 10=All 6.14 3.05 187
clients)

* Data is missing where N is less than 190.

47

Electronic copy available at: https://ssrn.com/abstract=1657117

 Table 3
Willingness to Report Error to an Appropriate Person a

Panel A: Results of a 2x2x2 ANCOVA of Report Willingness per Error-Management

Climateb, Error Typec, and Error Originatord Controlling for Reviewer
Experiencee

Type III Sum Mean

Source of Squares df Square F-value p-value
Model 72.3 8 9.04 3.30 0.002***
Error-Management Climate (EMC) 9.51 1 9.51 3.47 0.065*
Error Type (ET) 0.54 1 0.54 0.20 0.657
Error Originator (EO) 5.54 1 5.54 2.02 0.156
EMC x ET 12.00 1 12.00 4.38 0.038**
EMC x EO 17.16 1 17.16 6.26 0.014**
ET x EO 17.42 1 17.42 6.36 0.013**
EMC x ET x EO 2.69 1 2.69 0.98 0.323
Reviewer Experience 7.94 1 7.94 2.90 0.091*
Error 496.09 181 2.74

Panel B: Adjusted Means and Standard Errors of Error-Management Climate, Error

Originator, and Error Type on Report Willingness Controlling for Reviewer
Experience
Error-
Management Error
Climate Error Type Originator Mean SE n
Conceptual Own 8.82 0.33 25
Peer 8.24 0.35 23
Blame
Mechanical Own 8.79 0.34 24
Peer 7.48 0.35 22
Conceptual Own 7.92 0.33 25
Peer 9.03 0.35 23
Open
Mechanical Own 9.38 0.34 24
Peer 8.79 0.34 24

48

Electronic copy available at: https://ssrn.com/abstract=1657117

Panel C: Adjusted Means and Standard Errors of Report Willingness by Error-
Management Climate and Error Type Controlling for Reviewer Experience
Error-
Management
Climate Error Type Mean SE n
Conceptual 8.53 0.24 48
Blame
Mechanical 8.14 0.25 46
Conceptual 8.48 0.24 48
Open
Mechanical 9.09 0.24 48

Panel D: Adjusted Means and Standard Errors of Report Willingness by Error-

Management Climate and Error Originator Controlling for Reviewer Experience
Error-
Management Error
Climate Originator Mean SE n
Own 8.81 0.24 49
Blame
Peer 7.86 0.25 45
Own 8.65 0.24 49
Open
Peer 8.91 0.24 47

Panel E: Adjusted Means and Standard Errors of Report Willingness by Error Originator
and Error Type Controlling for Reviewer Experience
Error
Error Type Originator Mean SE N
Own 8.37 0.23 50
Conceptual
Peer 8.64 0.25 46
Own 9.09 0.24 48
Mechanical
Peer 8.14 0.25 46

Notes:
*** p<0.01, ** p<0.05, * p<0.10
a
Report Willingness: Response to the likelihood you would report the problem found in the finished
goods working paper to an appropriate responsible person for this audit engagement at Good and
Better (e.g., audit team leader, engagement manager, responsible partner, or quality reviewer).
(0=Very unlikely; 10=Very likely).
b
Error-Management Climate: Client office is described as one with an “open” learning and supportive
error-management climate versus a “blame” climate.
c
Error Type: The type of the error is “mechanical” (e.g., an arithmetical error in a computation) versus
“conceptual” (e.g., not performing the necessary computation).
d
Error Originator: The originator of the error is the auditor (own error) or the auditor‟s peer (peer error).
e
Reviewer Experience: Self-reported experience with reviewer activities measured in months.

49

Electronic copy available at: https://ssrn.com/abstract=1657117

 Table 4
Willingness to Report Error to an Appropriate Person Raw Means

Panel A: Raw Means and Standard Deviations of Report Willingness by Error-

Management Climate and Error Type
Error-Management
Climate Error Type Mean SD N
Conceptual 8.53 1.50 48
Blame Mechanical 8.21 2.27 46
Overall 8.37 1.91 94
Conceptual 8.41 1.74 48
Open Mechanical 9.10 1.20 48
Overall 8.75 1.52 96

Panel B: Raw Means and Standard Deviations of Report Willingness by Error-

Management Climate and Error Originator
Error-Management Error
Climate Originator Mean SD N
Own 8.85 1.27 49
Blame Peer 7.86 2.33 45
Overall 8.37 1.91 94
Own 8.63 1.66 49
Open Peer 8.89 1.37 47
Overall 8.75 1.52 96

Panel C: Raw Means and Standard Deviations of Report Willingness by Error Type and
Error Originator
Error
Error Type Originator Mean SD n
Own 8.37 1.63 50
Conceptual Peer 8.58 1.61 46
Overall 8.47 1.61 96
Own 9.12 1.20 48
Mechanical Peer 8.19 2.27 46
Overall 8.66 1.85 94

Notes:
For definitions of variables see Table 3.

50

Electronic copy available at: https://ssrn.com/abstract=1657117

 Figure 1
Error-Management Climate and Error Type

Panel A: Predicted Interaction Effect of Error-Management Climate x Error Type on

Report Willingness (H2)

Panel B: Observed Interaction Plot: Error-Management Climate by Error Type on Report

Willingness (Adjusted Means)

51

Electronic copy available at: https://ssrn.com/abstract=1657117

 Figure 2
Error-Management Climate and Error Originator

Panel A: Predicted Interaction Effect of Error-Management Climate x Error Originator

on Report Willingness (H3)

Panel B: Observed Interaction Plot: Error-Management Climate by Error Originator

Report Willingness (Adjusted Means)

52

Electronic copy available at: https://ssrn.com/abstract=1657117