Journal of Business Ethics (2009) 85:109–136
DOI 10.1007/s10551-008-9752-x
The Effects of Satisfaction with a Client’s
Management During a Prior Audit
Engagement, Trust, and Moral Reasoning
on Auditors’ Perceived Risk
of Management Fraud
ABSTRACT. The recent accounting scandals have
raised concerns regarding the closeness of auditor–client
relationships. Critics argue that as the relationship
lengthens a bond develops and auditors’ professional
skepticism may be replaced with trust. However, State-
ment on Auditing Standards No. 99 states that auditors
‘‘should conduct the engagement with a mindset that
recognizes the possibility that a material misstatement due
to fraud could be present, regardless of any past experi-
ence with the entity and regardless of the auditor’s belief
about management’s honesty and integrity’’
(AICPA 2002, Statement on Auditing Standards No. 99,
paragraph 13, p. 10). The purpose of this study is to
investigate whether auditors develop trust in a client’s
management and whether this trust affects auditors’
decisions. Specifi- cally, this study examines whether
auditors’ satisfaction with a client’s management during
a prior audit engage- ment affects auditors’ self-reported
trust in that client’s management and whether that trust
affects their fraud risk assessment. The decision to trust a
client’s management should be an ethical decision
because excessive trust may impair auditors’ skepticism,
which auditors are required to maintain by their
professional responsibilities. We therefore also investigate
whether auditors’ trust is affected by their moral
reasoning. An experimental case was completed by 89
professional auditors, all with experience assessing the risk
of fraud. The results suggest auditors’ satisfaction with
the client affects their trust in the client (higher satisfaction
associated with higher trust and lower satisfaction
associated with lower trust). Fur- ther, after an overall
unsatisfying experience, auditors’ trust affects their fraud
risk assessments. However, after an overall satisfying
experience, their trust does not affect their fraud risk
assessments. The results indicate auditors are able to
maintain their professional skepticism after
© Springer 2008
William A. Kerler III
Larry N. Killough
satisfying past experiences with the client regardless of
their beliefs about the honesty and trustworthiness of the
client’s management. Lastly, auditors’ moral reasoning
was not related to their trust in the client’s management.
KEY WORDS: auditor trust, fraud risk, moral reason-
ing, professional skepticism
ABBREVIATIONS: AICPA: American Institute of
Certified Public Accountants; CFI: comparative fit
index; CPA: Certified Public Accountants; DIT:
Defining Issues Test; PCAOB: Public Company
Accounting Oversight Board; RMSEA: root mean
square error of approxima- tion; SAS: Statement on
Auditing Standards
Introduction
Recent accounting scandals have ignited an
increased interest in fraudulent financial reporting
among professionals, regulators, and academics. In
1997, the American Institute of Certified Public
Accountants (AICPA) issued Statement on Auditing
Standards (SAS) No. 82, Consideration of Fraud in a
Financial Statement Audit. This standard requires
auditors to specifically assess the risk of material
misstatement due to fraud and provides guidance on
the response to the results of the assessment. 1 Fur-
thermore, in October 2002, the AICPA superseded
SAS No. 82 with SAS No. 99, also titled Consid-
eration of Fraud in a Financial Statement Audit. This
new standard provides auditors additional guidance
in fulfilling their responsibilities relating to detecting
110 William A. Kerler III and Larry N.
Killough
fraud in a financial statement audit. A key provision
in SAS No. 99 is an increased emphasis on profes-
sional skepticism, which is defined as ‘‘an attitude
that includes a questioning mind and a critical
assessment of audit evidence’’ (AICPA, 2002, para-
graph 13, p. 10).
While the purpose of an audit is to provide an
objective review of a company’s financial
statements (Messsier et al., 2006), the value of an
audit depends on stakeholders’ (e.g. investors,
potential investors, creditors, etc.) perceptions of
auditors’ objectivity, independence, and professional
skepticism. Although auditors are required by SAS
No. 99 to maintain their professional skepticism, it
has been argued that as the auditor–client
relationship lengthens, an auditor’s skepticism can
become impaired. In a Wall Street Journal article
published shortly after the Enron fiasco, Herrick and
Barrionuevo (2002) raise the question of whether
Arthur Andersen’s close relationship with Enron
affected Andersen’s ability to objectively review
Enron’s financial statements. Johnstone et al. (2001)
suggest an interpersonal relationship between auditor
and auditee may ‘‘affect the auditor’s ability to
exercise an appropriate level of professional skepti-
cism’’ (p. 5). One potential threat that may impair
auditors’ professional skepticism is if auditors
develop trust in a client’s management. Trust is
defined in this study as auditors’ belief in
management’s intentions to compile the company’s
financial statements following applicable laws and
standards and to provide auditors with all relevant
information they need to audit the financial
statements. Most auditor–client relationships involve
year-round communication via face-to-face meetings,
telephone conversations, etc. Research in other arenas
suggests that trust develops, at least in part, as a
function of successful/satisfying prior experience with
an individual or group (e.g. Doney and Cannon, 1997;
Flores and Solomon, 1998; Ganesan, 1994; Ring and
Van De Ven, 1992). Further, some accounting
researchers claim that ‘‘a behavioral bond develops
between auditor and auditee as they become more
familiar with each other and mutual trust replaces the
auditor’s necessary professional skepticism’’
(Latham et al., 1998, p. 168). If auditors’ trust in
client’s management impairs their professional
skepticism, it can affect their audit deci- sions and
ultimately the quality of the audit. For example,
auditors who develop trust in a client’s management
may be willing to accept a client’s
explanation of an unusual account fluctuation
without collecting evidence to verify the client’s
explanation.2
SAS No. 99 specifically addresses the potential of
auditors’ professional skepticism being impaired due
to previous interactions with an audit client. The
standard states that auditors ‘‘should conduct
the engagement with a mindset that recognizes the
possibility that a material misstatement due to fraud
could be present, regardless of any past experience
with the entity and regardless of the auditor’s belief
about management’s honesty and integrity’’
(AICPA, 2002, paragraph 13, p. 10). This study
examines whether auditors develop trust in a client’s
management and whether this trust affects their audit
decisions. Specifically, we test whether auditors’
satisfaction working with a client during a previous
audit engagement affects auditors’ trust in that cli-
ent’s management. We also explore whether audi-
tors’ trust in the client’s management affects their
fraud risk assessment.
While in many business contexts the decision to
trust someone is not likely an ethical decision, the
auditing environment is unique because trusting a
client’s management may impair auditors’ profes-
sional skepticism and hence violates the requirements
of SAS No. 99. Any decision that goes against
auditing standards and may lead to an impairment of
skepticism should be an ethical decision. One way
individuals make ethical judgments is through the use
of justice- based concepts. Kohlberg (1969) developed
a justice- based model of moral reasoning, which has
been used for research in a variety of disciplines (e.g.
education, marketing, psychology). The research
demonstrates that individuals’ moral reasoning
affects their deci- sions and behaviors (see Rest et al.,
1999a). In accounting settings, researchers find
accountants with higher levels of moral reasoning
generally make more ethical decisions and behave in a
more ethical manner (e.g. Ponemon, 1992; Tsui and
Gul, 1996). Because trusting a client’s management
may impair auditors’ skepticism, and thus may
compromise auditors’ pro- fessional responsibilities
(e.g. SAS No. 99), this study examines whether the
extent to which auditors trust a client’s management
is affected by auditors’ moral reasoning.
This paper reports the results of an experimental
case completed by 89 professional auditors, all with
experience assessing fraud risk. The experimental
 Auditors’
case utilized was identical for all subjects with the
exception of the description of the results for the
previous year’s audit. Auditors received one of two
experimental cases. In the ‘‘satisfying’’
manipulation, the results of the prior year’s audit
engagement were presented to create an overall
satisfying experience working with the client’s
management, while the ‘‘unsatisfying’’
manipulation was designed to create an overall
unsatisfying experience working with the client’s
management. The results of path analyses indicate
that auditors’ satisfaction with the client’s
management during a prior audit engagement is
positively related to their trust in the client’s man-
agement (i.e. higher satisfaction associated with
higher trust and lower satisfaction associated with
lower trust). We also find that after a satisfying
previous experience, this trust does not affect their
level of perceived risk of management fraud. Thus,
although auditors’ trust in a client’s management
may increase due to a satisfying previous experience,
auditors are able to maintain their professional
skepticism regardless of their beliefs about the hon-
esty and trustworthiness of the client’s management.
However, after an unsatisfying previous experience
with the client’s management, auditors’ trust is
negatively related to their perceived risk of man-
agement fraud (i.e. lower trust associated with higher
risk). This suggests, as we would expect, as auditors’
trust in a client’s management decreases due to an
unsatisfying prior experience, auditors become more
skeptical of that management. Finally, contrary to
our expectations, auditors’ moral reasoning was not
related to their trust in a client’s management.
Implications for financial statement users, standard
setters, and the profession will be discussed in the
concluding section of this paper.
This study differs from previous research and
contributes to the literature in three ways: (1) to our
knowledge, this is the only study to explicitly
measure and investigate whether auditors develop
trust in a client’s management and to examine
whether trust affects their decisions; (2) we develop a
measure of auditors’ trust in a client’s
management that may prove useful for future auditing
judgment and decision-making research; and, (3) we
examine whether moral reasoning affects auditors’
trust in the client’s management. The concluding
section of this paper will elaborate on these
contributions as well as provide suggestions for future
research.
1
The remainder of this study is organized into
four sections. We first present the relevant literature
and develop our hypotheses and research question.
The second section discusses the research methods
and the third presents the results. We conclude with
a summary of the main research findings and a dis-
cussion of the implications, contributions, and lim-
itations of the study.
Prior literature
Satisfaction with a client’s management during
a previous audit engagement and trust
We utilize the framework of trust from Nooteboom
(1996) because of its adaptability to an auditing
context. In this framework, an individual’s trust
in another person or group consists of two
dimensions. The first, competence trust, refers to
the trust an individual may have in another’s
ability to perform; while the second, goodwill trust,
refers to the trust in another’s intentions to
perform (Nooteboom, 1996,
p. 990).3 We control for competence trust, as it is
not of primary interest in this study. Goodwill trust
(hereafter, trust) is the focus in this study and is
defined in an auditor–client context as auditors’
belief in management’s intentions to accurately
compile the company’s financial statements and to
provide auditors with all relevant information audi-
tors needs to evaluate the fairness of the client’s
financial statements. Auditors’ trust in management
would include beliefs regarding management’s
honesty, dependability, helpfulness, etc. In essence,
trusting management indicates auditors believe:
management means well, management is represented
by good people, and management will do the right
thing.
Although not the object of extensive empirical
testing, previous research suggests that trust
develops (deteriorates), at least in part, as a
function of suc- cessful/satisfying
(negative/unsatisfying) prior expe- riences with an
individual or group (e.g. Doney and Cannon, 1997;
Flores and Solomon, 1998; Ganesan, 1994; Ring
and Van De Ven, 1992). We found two studies in
the field of auditing that identify and test this
relationship. Shaub (1996) finds that auditors who
receive information about inaccuracies in the
previous year’s inventory evaluate both the
likelihood
112 William A. Kerler III and Larry N.
Killough
of the current year’s inventory as being overstated
and the likelihood of requiring more audit proce-
dures as higher compared to auditors who do not
receive information regarding inaccuracies. These
results support that auditors’ trust in the client’s
management decreases when inaccuracies are revealed
in the previous year’s audit results. Although Shaub
(1996) provides preliminary evidence regarding the
effect on trust of satisfaction with a client during a
previous experience, the inaccuracies represent an
unsatisfying prior audit experience with the client’s
competence and therefore the study addresses
competence trust. King (2002) conducted an
experiment using student subjects performing a task
devoid of any auditing context (i.e. the exper- imental
task made no mention of auditing, auditors, or fraud).
King’s (2002) results suggest that the
communication of a client’s management can induce
auditors to develop unwarranted trust in the man-
agement; however, this effect is mitigated when
auditors belong to groups, which create social
pressures. King’s study provides evidence that audi-
tors may develop trust in client’s management as a
result of interactions with management.
This study attempts to explicitly measure and test
the effect that auditors’ satisfaction working with
the client’s management during a prior audit engage-
ment has on auditors’ trust in that client. Based on
the above research, we expect higher levels of
auditors’ satisfaction working with the client’s
management to be related to higher levels of trust in
that client’s management. Stated formally:
H1: Auditors’ satisfaction working with the
client’s management during a prior audit
engagement will be positively related to
auditors’ trust in that management.
Trust and perceived risk of management fraud
SAS No. 99 distinguishes two types of fraud auditors
should be cognizant of during a financial statement
audit: misstatements arising from fraudulent financial
reporting and misstatements arising from misappro-
priation of assets. The first involves intentionally
falsifying the financial statements to deceive users;
while the second involves the theft of the
company’s assets. For the purposes of this study,
management
fraud will refer to the intentional misstatement of
the financial statements
To our knowledge, no previous auditing research
has explicitly measured and examined the effect of
trust on auditors’ decisions. Further, the connection
between trust and perceived risk of management
fraud has not been examined. One stream of
research, however, has investigated the effect that
auditors’ beliefs regarding management’s integrity
has on auditors’ decisions and suggests that such
beliefs may affect auditors’ trust in that
management. This research stream finds the client’s
integrity influences auditors’ business and combined
risk assessments (Beaulieu, 2001), acceptance of client
explanations (Peecher, 1996), and likelihood assess-
ments of account overstatement and required write-
downs (Goodwin, 1999). For example, Beaulieu
(2001) finds that the auditors’ perceived level of
client integrity is negatively related to their assess-
ments of the client’s business and combined risk
(i.e. lower integrity related to higher risk). Although
none of these studies specifically addresses the issue
of measuring trust, it is possible that the manipula-
tions of integrity affect auditors’ trust. Based on this
possibility, we would expect a negative relationship
between auditors’ trust in client’s management and
auditors’ perceived risk of management fraud.
However, this potential relationship between
auditors’ trust and their audit decisions provides a
unique environment where expectations from prior
research conflict with professional requirements.
SAS No. 99 states that auditors ‘‘should conduct
the engagement with a mindset that recognizes the
possibility that a material misstatement due to fraud
could be present, regardless of any past experience
with the entity and regardless of the auditor’s
belief about management’s honesty and integrity’’
(AICPA, 2002, paragraph 13, p. 10). This implies
auditors’ decisions should not be affected by in-
creased levels of trust in the client’s management.
Therefore, auditors should maintain their profes-
sional skepticism despite a satisfying experience
working with the client’s management during a
prior audit engagement and despite any trust they
may have in the client’s management. Based on
this auditing standard, we would expect no rela-
tionship between auditors’ trust in the client’s
management and auditors’ perceived risk of man-
agement fraud.
 Auditors’
Because expectations based on previous research
conflict with expectations based on professional
standards, we are unable to make a formal
hypothesis regarding the relationship between
auditors’ trust and their perceived risk of
management fraud. Instead, we will explore the
following research question:
RQ1: Is auditors’ trust in the client’s management
related to their fraud risk assessments?
Moral reasoning and trust
Although auditors are required by SAS No. 99 to
maintain their professional skepticism, some
believe that mutual trust may develop between
auditors and their clients as they become more
familiar with each other and this trust may replace
auditors’ necessary skepticism (e.g. Latham et
al., 1998). Johnstone et al.’s (2001) framework
for independence risk posits that an interpersonal
relationship between auditors and their clients is an
incentive for impaired independence. Presumably
then, the trust that may develop from the
relationship would also be a threat to auditors’
independence. Furthermore, shortly after the Enron
accounting scandal, Herrick and
1
Barrionuevo (2002) questioned whether the close
relationship between Enron and their auditors
affected the auditors’ ability to objectively review
Enron’s financial statements. These beliefs indicate
that auditors’ decision to trust a client’s
management may impair their skepticism and may
compromise auditors’ professional responsibilities
(e.g. SAS No. 99); thus, the extent to which to trust a
client’s management should be an ethical decision.
At the heart of moral reasoning research is the
question of how and why individuals make
decisions when an ethical dilemma presents itself. 4
Based on the work of the child psychologist Piaget,
Lawrence Kohlberg (1969) developed a justice-
based model of moral reasoning that consists of a
series of three cognitive levels, with each level
containing two stages (see Table I). According to
Kohlberg, indi- viduals develop sequentially one
stage at a time. An individual’s ethical reasoning
process in the first level, preconventional level,
revolves around pun- ishment and self-interest. In
the conventional level, reasoning is influenced by
wanting to please others, such as peers or the law.
In the final level, post- conventional level, an
individual’s reasoning involves individual rights,
self-chosen principles, and belief in ideals, such as
equality and justice.

TABLE I
Kohlberg’s six-stage model of moral reasoning

Preconventional level
Stage 1 Obedience and punishment orientation
At this stage punishment or harm determines what is right or wrong.
Stage 2 Naively egoistic orientation
At this stage an individual acts to serve his/her own interests.
Conventional level
Stage 3 Good-Boy (or Good-Girl) orientation
At this stage one strives to be a good person in the eyes of others. Good behavior
is that which pleases others
Stage 4 Authority and social-order maintaining orientation
At this stage there is an orientation toward obeying authority and maintaining the social order.
Postconventional level
Stage 5 Contractual legalistic orientation
At this stage the right action tends to be defined in terms of general individual rights,
not necessarily what is agreed upon by society.
Stage 6 Conscience or Principle Orientation
At this stage one follows self-chosen ethical principles. There is a belief in ideals such
as justice and equality.

Adapted from Kohlberg (1969).

114 William A. Kerler III and Larry N.
Killough
Research involving the ethical reasoning of
auditors shows that for a wide variety of judgments
and behaviors, higher moral reasoning auditors
generally make more ethical decisions and act in a
more ethical manner (e.g. Bernardi, 1994; Pone-
mon, 1992; Ponemon and Gabhart, 1993; Auditors’
Tsui Satisfaction
and
Gul, 1996; Windsor and Ashkanasy, 1995). For
example, Ponemon (1992) finds that auditors with
low moral reasoning underreported audit time
more severely than auditors with high moral rea-
soning. Windsor and Ashkanasy (1995) show that
auditors with the highest moral reasoning were
better able to resist the client’s economic power,
compared to groups of auditors with lower moral
reasoning. Because trusting a client’s management
Auditors’ Satisfaction with a Client’s Management during a Prior Audit Engagement
may impair auditors’ professional responsibility of
maintaining professional skepticism, the decision to
trust a client’s management should be, at least in
part, an ethical judgment. Therefore, we expect
higher levels of moral reasoning to be related to
lower levels of trust in a client’s management.
Stated formally:
H2: Auditors’ moral reasoning will be negatively
related to their trust in the client’s manage-
ment.
To examine the two hypotheses and to explore the
research question, we will test the research models
presented in Panels A and B of Figure 1 using path
analysis.
Methods
Variables
Satisfaction with a client’s management during
a prior audit engagement
Auditors’ satisfaction with a client’s management
during a prior audit engagement was measured with
a single item stating: ‘‘I am satisfied with my expe-
rience with XYZ Corporation’s [the audit client]
management.’’ Participants responded to this item
on a seven-point Likert-type scale with anchors at
each point (1 – ‘‘Strongly Disagree,’’ 2 –
‘‘Disagree,’’ 3 – ‘‘Slightly Disagree,’’ 4 –
‘‘Neither Agree nor Disagree,’’ 5 – ‘‘Slightly
Agree,’’ 6 – ‘‘Agree,’’ 7 – ‘‘Strongly Agree’’).
Panel A: Auditors’ Trust in the Client’s Management
H1: RQ1:
+ ?
with a Client’s Management during a Prior Audit Engagement
Auditors’ Perceived Risk of Management
Fraud
Auditors’ Trust in the Client’s Management
Panel B:
H2: Auditors’ Perceived Risk of Management
– Fraud
Auditors’
Moral
Reasoning
Figure 1. Research models for main analyses.
Trust
Because we were unable to identify an existing scale
designed to measure an auditor’s trust in a client’s
management, we developed our own scale. This
section describes the process we undertook to
develop such a scale. We first identified existing
measures of trust through a review of the psychol-
ogy, management, and marketing literature. Nine
scales were selected because their items could be
modified and adapted to an auditing context (see
Appendix A for the source of these scales). Items
from these nine preexisting scales were used to
generate an initial sample of 32 items intended to
measure auditors’ trust in a client’s management.
Two pilot studies were then conducted with the
main goal of developing a reliable and valid measure
of trust. Each pilot study required participants to
complete the entire experimental instrument with all
the sections described later.5
The first pilot study utilized 24 MBA students
and 39 senior-level auditing students. An
examination of the participants’ responses to the
trust items (same seven-point Likert-type scale used
in the final instrument) led to the elimination of 24
of the initial 32 items for one of two reasons:
participant com- ments (e.g. participants indicated
on the experimental
 Auditors’
instrument that they did not understand what an
item was asking) or insufficient factor loadings. 6, 7
The remaining eight items loaded on a single factor,
which explained over 60 percent of the variance and
had a Cronbach’s Alpha of 0.90.
The second pilot study was conducted to further
refine the trust scale. This pilot study also required
participants to complete the entire experimental
instrument which included the remaining eight trust
items identified in the first pilot study. Participants
for this pilot study consisted of 58 practicing
auditors from each of the Big Four firms and nine
Non-Big Four firms. Of the 58 subjects, 30 were
from Non-Big Four firms, 25 were from Big Four
firms, and three did not indicate. They had an
average of
7.39 years of auditing experience and included ten
(17.2 percent) staff, 18 (31.0 percent) in-charge, 22
(37.9 percent) managers, and eight (13.8 percent)
partners. There were 36 male auditors and 22 female
auditors. Lastly, 50 participants indicated they had
experience assessing fraud risk while eight did not
have experience. Participant responses to the trust
items were analyzed, and based on a PCA with
varimax rotation, three items were eliminated. The
remaining five items loaded on a single factor, which
explained over 69 percent of the variance and had a
Cronbach’s Alpha of 0.89.8 These psychometric
properties of the five-item trust scale were deemed
satisfactory; hence, the scale was included in our final
instrument (see Table II for final five items). Par-
ticipants’ responded to the five items on a seven-
point Likert-type scale with anchors at each point
(1 – ‘‘Strongly Disagree,’’ 2 – ‘‘Disagree,’’ 3 –
‘‘Slightly Disagree,’’ 4 – ‘‘Neither Agree nor
Disagree,’’ 5 – ‘‘Slightly Agree,’’ 6 – ‘‘Agree,’’ 7
– and
1
‘‘Strongly Agree’’). To test the two hypotheses
and investigate the research question, we utilize
auditors’ mean response to the five items as our
measure of trust in the client’s management.
Perceived risk of management fraud
As discussed previously, management fraud in this
study refers to the intentional misstatement of the
financial statements. Participants assessed the risk of
management fraud for the current year’s audit. This
was based on a five-point Likert-type scale with
anchors on the endpoints and midpoint (1 – ‘‘Very
Low Risk,’’ 3 – ‘‘Medium Risk,’’ 5 – ‘‘Very High
Risk’’).
Moral reasoning
The Defining Issues Test (DIT), developed by Rest
(1979), measures an individual’s moral reasoning
‘‘based on the premise that people at different
levels of moral development interpret moral dilemmas
differently, define the critical issues of the dilemmas
differently, and have different intuitions about what
is right in a situation’’ (Kaplan et al., 1997, pp. 45–
46). The DIT used consists of three standard
hypothetical moral dilemmas with each dilemma
followed by twelve statements of issues reflecting the
different stages of moral reasoning.9 Participants read
the di- lemma, make an action choice, and then rank
which of the given issues are important in their
decision making. The analysis of the DIT yields the
indi- vidual’s level of ‘‘principled’’ thinking
(stages 5 and 6), hereafter called the p-score.10 The p-
score, which ‘‘summarizes the ranking data and is
defined as the weighted sum of the ranked principled
issues’’ (Thoma et al., 1991, p. 664), is calculated

TABLE II
Trust scale items

Item abbreviation Item

T1 I believe that XYZ Corporation’s management are thoroughly dependable

people T2 XYZ Corporation’s management will do everything possible to help me
T3 XYZ Corporation’s management are like my friends
T4 I do not think that XYZ Corporation’s management is completely open in dealing with me. (R)
T5 XYZ Corporation’s management plans to be helpful during future audits
(R) – denotes negatively worded item.
Subject responses on a seven-point, Likert-type scale with anchors at each point (1: strongly disagree, 2: disagree,
3: slightly disagree, 4: neither agree nor disagree, 5: slightly agree, 6: agree, and 7: strongly agree).
116 William A. Kerler III and Larry N.
Killough
transformed into a percentage ranging from 0 to
95.11
Control variables
As discussed previously, we controlled the compe-
tence of the client’s management. To verify the
control was successful, auditors responded to the
following statement: ‘‘XYZ’s [the audit client]
man- agement is competent.’’ Responses were on a
seven- point Likert-type scale with anchors at each
point (1 – ‘‘Strongly Disagree,’’ 2 –
‘‘Disagree,’’ 3 – ‘‘Slightly Disagree,’’ 4 –
‘‘Neither Agree nor Dis- agree,’’ 5 – ‘‘Slightly
Agree,’’ 6 – ‘‘Agree,’’ 7 – ‘‘Strongly Agree’’).
Participants also provided the following demographic
variables: gender, firm (Big Four or Non-Big Four),
years of auditing experience (in years), rank in their
firm (staff, in-charge, manager, partner), and whether
they have previously partici- pated in an audit that
detected financial statement fraud committed by
client management (yes or no).
Sample
Based on the experimental task, auditors with
experience assessing the risk of fraud were deemed
to be the appropriate participant group. Data were
collected in two ways. First, experimental cases
were provided to all attendees of a southeastern
state Accounting and Auditing Conference. The
confer- ence provided updates and training to
Certified Public Accountants seeking to stay
current in their field and seeking to fulfill hours for
their required continuing education requirements.
Participation was encouraged by providing raffle
tickets to those who completed the instrument.
During the last session of the conference, four
tickets were ran- domly selected and prizes were
12
distributed to win- ners. Sixty-eight cases were
completed and returned. Second, 60 experimental
instruments were mailed to contact partners at
offices (located in three eastern states) of four Big
Four audit firms (15 per firm). The contact partner
distributed the instru- ments to auditor subjects of
his or her choosing. Once completed, the subjects
returned the cases to the contact partner, who then
mailed them to one of the researchers. Of the 60
instruments sent, 45 (75.0 percent) were returned,
with responses from firms ranging from seven to
13
15. Combining the two
groups, a total of 113 cases were completed.14 Of the
113 instruments completed, 23 were excluded
because subjects indicated they had no experience
assessing the risk of fraud.15 One additional subject
was excluded because he/she failed to provide a
fraud risk assessment. Thus, our analysis is based
on responses from 89 auditors.
Table III provides a number of descriptive statis-
tics, which show that 51 (57.3 percent) and 38 (42.7
percent) of the participants were from Non-Big Four
firms and Big Four firms, respectively. Further, the 51
Non-Big Four auditors were employed at firms of
varying sizes, including national, regional, and local.
The subjects included a variety of position levels
within the firms: six (6.7 percent) staff, 11 (12.4 per-
cent) in-charge, 32 (36.0 percent) managers, and 40
(44.9 percent) partners. Subjects had an average of
14.1 years of auditing experience, ranging from one
year to 38 years of experience. Thirty-six (40.4 per-
cent) auditors reported having directly participated
in at least one audit that eventually detected
financial statement fraud committed by client
management. The sample consisted of 67 (75.3
percent) males and 22 (24.7 percent) females.
Seventy-nine participants completed the entire DIT
and had an average p-score of 30.9 (median of
33.3). Lastly, 43 (48.3 percent) subjects received
the satisfying previous experience case and the
other 46 (51.7 percent) received the unsatisfying
previous experience case.
Instrument
The experimental case (see Appendix B) developed
for this study consists of four sections and can be
completed in approximately 20 minutes. The first
section instructed the participants to assume the role
of a supervising auditor and contained information
pertaining to the planning of last year’s audit. All
participants received the same information describ-
ing the planning of last year’s audit. Auditors were
instructed that they worked for the first time on the
audit of XYZ Corporation, but it was the fourth
year their firm had audited XYZ. 16 The auditors
were then presented information describing the size
and firm type of XYZ (medium-sized furniture
manufacturer). To control for competence trust,
participants also received a fairly extensive descrip-
tion of XYZ’s management. In this description,
 Auditors’ 1
TABLE III
Sample descriptive statistics

Count Percent (%)

Firm type
Big 4 38 42.7
Non-Big 4 51 57.3
Position
Staff 6 6.7
In-Charge 11 12.4
Manager 32 36.0
Partner 40 44.9
Involved in audit(s) which detected financial statement fraud
Yes 36 40.4
No 53 59.6
Gender
Male 67 75.3
Female 22 24.7
Instrument version
Satisfying experience 43 48.3
Unsatisfying experience 46 51.7

Mean Median S.D. Minimum Maximum

Years of auditing
auditors receivedexperience
numerous(n =cues
89) indicating
14.11that
P-Score (n = was
management 79) qualified and competent to make
30.89
management decisions and to prepare financial
statements. The information also included the results
of a red flag checklist, which identified whether
certain fraud indicators are present. 17 The checklist
contained 20 items,18 three of which were marked as
present. The main purpose of the red flag checklist
in this instrument was to control for the potential
confound of changes in XYZ’s operations (i.e. the
same three items were indicated as present in
the current year). Finally, auditors were told that at
the end of last year’s planning phase they assessed
the risk of management fraud at 3 – ‘‘Medium
Risk.’’ This was based on a five-point scale
anchored 1 – ‘‘Very Low Risk’’ and 5 – ‘‘Very
High Risk,’’ with the above-mentioned midpoint.
The second section of the instrument described the
results of last year’s audit. Here we manipulated the
auditor’s satisfaction with XYZ’s management
during the previous audit engagement. The
manipulation
12.00 9.36 1 38
was intended to create either an overall satisfying or
33.33
unsatisfying 14.54
experience working0 with XYZ’s 70 man-
agement. While we realize an auditor’s satisfaction
with a client’s management is influenced by a
variety of interrelated factors (e.g. personality of
manage- ment, behaviors of management, the
environment, etc.) we focused our manipulation on
management’s behaviors. We chose this factor
because we believed it would create the most salient
manipulation, and therefore, it would lead to
significantly different levels of satisfaction. The
overall goal of the manipulation was to create one
version where auditors ‘‘enjoy’’ their interactions
with management and another version where auditors
are ‘‘frustrated’’ with their interactions with
management. This was done by providing details
regarding last year’s audit work. These details were
developed by the researchers, based in part, on
personal experiences, but mainly by uti- lizing
previously developed job and customer satis- faction
scales from the psychology and marketing literature.
We adapted scale items from numerous
118 William A. Kerler III and Larry N.
Killough
studies (Ironson et al., 1989; King, 1960; Nicholls
et al., 1998; Twery et al., 1958) and incorporated
them into the results of the audit work. For example,
one of Nicholls et al.’s (1998) components of cus-
tomer satisfaction is timely service. It seems likely
that a timely (untimely) response by client
management to auditors’ request would increase
(decrease) the audi- tors’ satisfaction with working
with the client’s management during the audit
engagement. There- fore, in the satisfying
(unsatisfying) audit engagement manipulation we
included a phrase describing how management made a
recommended adjustment without hesitation (put off
responding to the adjust- ment). The complete
satisfaction manipulations are provided below.
The satisfying prior audit engagement manipula-
tion states:
During the audit work, XYZ’s management always
had a respectful attitude and never engaged in disputes
with you. They always appeared to be well prepared
for your meetings. When you were on location, you
often joined them for lunches. During your review of
their accounting records the only issue you had was
that you believed they immaterially understated their
estimate of Allowance for Bad Debt, and hence,
immaterially understated their operating expenses.
Although the amount was immaterial, you proposed
an adjustment and XYZ’s management made it
without hesitation. The end result of the audit was an
unqualified opinion for XYZ’s financial statements.
The unsatisfying prior audit engagement manipula-
tion states:
During the audit work, XYZ’s management had a
hostile attitude and frequently engaged in disputes
with you. They never appeared to be well prepared for
your meetings. When you were on location, you were
never invited to join them for lunches. During your
review of their accounting records the only issue you
had was that you believed they immaterially under-
stated their estimate of Allowance for Bad Debt, and
hence, immaterially understated their operating
expenses. Although the amount was immaterial, you
proposed an adjustment. XYZ’s management put off
responding to your proposed adjustment until you
mentioned it again a few days later. At this point they
refused to make the adjustment. Because your pro-
posed adjustment did not materially affect the financial
statements, the end result of the audit was an
unqualified opinion for XYZ’s financial statements.
The third section of the instrument provided all
participants with the same information pertaining to
the planning phase of the current year’s audit. Here
the auditors were told that XYZ’s management
consisted of the same individuals and that the results
of the red flags checklist were the same as last
year’s. The final section of the instrument elicited
responses to the previously discussed items measuring
per- ceived risk of fraud, competence of client’s
man- agement, satisfaction with the client’s
management, trust in the client’s management, as
well as the demographic control variables. Lastly, the
partici- pants completed the DIT.
Results
Descriptive statistics and correlations
Table IV shows descriptive statistics (means and
standard deviations) of the measured variables for
each experimental version and in total, while Table V
shows the correlations between all measured and
manipulated variables. Pearson’s product-moment
correlation of 0.693 between ‘‘Sat’’ and
‘‘Trust’’ provides preliminary evidence that
auditors’ trust may be related to their satisfaction
working with the cli- ent’s management during a
prior audit engagement. Further, the Pearson’s
product-moment correlation of )0.554 between
‘‘Trust’’ and ‘‘Risk’’ provides initial evidence
suggesting auditors’ trust may be re- lated to their
fraud risk assessments.19 However, the insignificant
Pearson’s product-moment correlation of 0.051
between ‘‘P-score’’ and ‘‘Trust’’ provides
preliminary evidence that auditors’ trust may not be
related to their moral reasoning.
Manipulation checks
The experimental manipulation of this study is the
auditors’ satisfaction with the client’s management
during a previous audit engagement. We manipu-
lated this as either an overall satisfying or an overall
unsatisfying experience working with XYZ’s man-
agement. To test whether the manipulation was
successful, we included a single test item stating that
the participant is satisfied with their experience with
management and asked the participants to indicate
 Auditors’ 1
TABLE IV
Descriptive statistics of measured variables

Version Comp Sat T1 T2 T3

Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev.

Satisfying (N = 43) 5.95 0.68 5.88 0.762 5.26 0.928 5.05 0.950 3.60 1.348
Unsatisfying (N = 46) 5.91 1.071 4.15 1.577 4.15 1.414 2.78 1.114 2.09 0.890
Combined (N = 89) 5.93 0.902 4.99 1.519 4.69 1.319 3.88 1.536 2.82 1.361

T4 T5 Trust Risk Moral

Reasoninga

Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev.

Satisfying (N = 43) 3.19 1.200 5.49 0.798 4.842 0.734 2.84 0.485 31.08 14.331
Unsatisfying (N = 46) 5.17 1.305 3.41 1.087 3.05 0.914 3.37 0.679 30.68 14.931
Combined (N = 89) 4.21 1.599 4.42 1.413 3.917 1.222 3.11 0.647 30.89 14.537
a
For moral reasoning, N = 40 and 39 for the satisfying and unsatisfying versions, respectively.
Variables:
Version: Experimental case completed: 1 if overall satisfying experience working with the client’s management; 0 if
overall unsatisfying experience.
Comp: Subjects’ assessment of the competence of the client’s management (higher assessment, more competent). Mea-
sured on a seven-point, Likert-type scale with anchors at each point (1: strongly disagree, 2: disagree, 3: slightly disagree,
4: neither agree nor disagree, 5: slightly agree, 6: agree, and 7: strongly agree).
Sat: Subjects’ assessment of how satisfied he or she was with their experience working with the client’s management
(higher assessment, more satisfied). Measured on a seven-point, Likert-type scale with anchors at each point (1: strongly
disagree, 2: disagree, 3: slightly disagree, 4: neither agree nor disagree, 5: slightly agree, 6: agree, and 7: strongly agree).
T1–T5: See Table II.
Trust: Subjects’ trust in the client’s management (higher score, greater trust). Average score of five items: T1, T2, T3, T4
and T5.
Risk: Subjects’ assessment of the risk of management fraud for the current year’s audit. Measured on a five-point,
Likert-type scale with anchors at 1: very low risk, 3: medium risk, and 5: very high risk.
Moral reasoning: Subjects’ moral reasoning measured as participant’s p-score from the DIT.

their level of agreement on a seven-point scale. The
mean level of agreement to this item was 4.15 and
5.88 for participants receiving the unsatisfying and
satisfying manipulation, respectively. This
difference is statistically significant (p < 0.001,
one-tailed) and suggests that our experimental
manipulation suc- cessfully created varying levels
of satisfaction work- ing with XYZ’s
management during the prior audit engagement.20
As explained previously, we controlled the com-
petence level of management. We tested the ade-
quacy of this control by including a single test item
stating that the management is competent and
participants indicated their level of agreement on a
seven-point scale. The mean level of agreement to
this item was 5.93, which is significantly greater than
the midpoint (p < 0.001, one-tailed). We also tested
whether there is a significant difference in partici-
pants’ perception of management’s competence
depending on the satisfaction manipulation. The
mean agreement to the competence item was 5.91
and 5.95 for participants receiving the unsatisfying
and satisfying condition, respectively. These are not
statistically different (p = 0.834, two-tailed) and
suggest our control for management competence
was successful.21
 1
TABLE V
Correlation matrix of measured and manipulated variables (n = 89 for all except for correlations with p-score where n = 79)

Version Comp Sat T1 T2 T3 T4 T5 Trust Risk P-score Gender CPA Firm Rank Fraud Years

Version 1 )0.072 0.572** 0.409** 0.752** 0.554** )0.632** 0.746** 0.751** )0.418** 0.010 0.189 0.032 0.029 )0.096 0.074 0.011
Comp 0.023 1 0.377** 0.381** 0.046 0.013 )0.257* 0.058 0.174 )0.217* )0.171 )0.111 0.033 )0.406** )0.070 )0.150 0.053
Sat 0.573** 0.389** 1 0.618** 0.649** 0.455** )0.636** 0.578** 0.702** )0.392** 0.020 0.035 0.107 )0.205 0.039 0.062 0.110
T1 0.420** 0.431** 0.628** 1 0.586** 0.453** )0.640** 0.570** 0.746** )0.435** )0.052 0.002 0.214* )0.142 )0.038 )0.095 0.056
T2 0.741** 0.101 0.633** 0.603** 1 0.670** )0.745** 0.808** 0.918** )0.447** 0.040 0.086 0.132 )0.026 0.021 0.087 0.104
T3 0.560** 0.083 0.450** 0.455** 0.663** 1 )0.595** 0.646** 0.790** )0.436** 0.003 0.157 0.122 )0.072 0.065 0.141 0.154
T4 )0.625** )0.313** )0.626** )0.652** )0.725** )0.588** 1 )0.636** )0.871** 0.483** 0.060 )0.062 0.040 0.007 0.130 )0.017 )0.014

William A. Kerler III and Larry N.

T5 0.738** 0.111 0.580** 0.601** 0.804** 0.648** )0.633** 1 0.866** )0.490** 0.067 0.124 0.120 )0.075 0.021 )0.011 0.053
Trust 0.736** 0.244* 0.693** 0.779** 0.905** 0.792** )0.862** 0.874** 1 )0.530** 0.008 0.120 0.127 )0.088 )0.010 0.042 0.108
Risk )0.413** )0.220* )0.380** )0.490** )0.432** )0.415** 0.493** )0.511** )0.554** 1 )0.017 0.049 0.053 )0.025 0.195 0.108 0.154
P-score 0.014 )0.074 0.040 0.019 0.067 0.051 0.013 0.095 0.051 )0.037 1 0.030 )0.015 )0.003 0.022 )0.052 )0.001
Gender 0.189 )0.101 )0.004 )0.018 0.090 0.174 )0.054 0.114 0.098 0.060 0.045 1 0.123 )0.190 0.263* 0.101 0.374**
CPA 0.032 )0.022 0.108 0.184 0.140 0.115 0.039 0.116 0.117 0.051 0.045 0.123 1 )0.338** 0.479** 0.071 0.439**
Firm 0.029 )0.340** )0.189 )0.139 )0.019 )0.104 0.013 )0.061 )0.076 )0.045 )0.023 )0.190 )0.338** 1 )0.408** )0.017 )0.502**
Rank )0.080 )0.110 0.043 )0.006 0.042 0.102 0.121 0.017 0.004 0.177 0.135 0.238* 0.620** )0.386** 1 0.203 0.813**
Fraud 0.074 )0.142 0.052 )0.099 0.082 0.160 )0.024 )0.016 0.038 0.105 )0.040 0.101 0.071 )0.017 0.207 1 0.231*
Years 0.034 0.061 0.092 0.046 0.112 0.209* )0.056 0.076 0.117 0.157 0.055 0.365** 0.369** )0.514** 0.738** 0.217* 1

Pearson’s product-moment correlations below diagonal.

Spearman’s rho correlations above diagonal.
**Correlation is significant at the 0.01 level (2-tailed).
*Correlation is significant at the 0.05 level (2-tailed).
Variables:
Gender: 1 if male, 0 if female.
CPA: 1 if subject is a CPA, 0 otherwise.
Firm: 1 if subject is employed at a Big 4 firm, 0 otherwise.
Rank: Subjects’ level in firm: 3 if partner, 2 if manager, 1 if in-charge, 0 if staff.
Fraud: 1 if subject has directly participated in an audit that eventually detected financial statement fraud committed by client management, 0 otherwise.
Years: Subjects’ years of auditing experience.
T1–T5: See Table II.
See Table IV for all other variable definitions.
 Auditors’
Test of model
We first confirmed, utilizing the final sample of
auditors, the factor structure of the five-item trust
scale. A confirmatory factor analysis verified the
five items load on a single factor (all factor
loadings greater than 0.77) that explains over 71
percent of the variance and has a Cronbach’s
Alpha of 0.90.
To examine the first hypothesis and the research
question, we performed path analysis for the model
shown in Figure 2. The model was tested using path
analysis in AMOS 4.01 with maximum likelihood
estimation.22 Model fit is evaluated with the fol-
lowing measures: the chi-square statistic, chi-square
per degree of freedom, the comparative fit index
(CFI), and the root mean square error of approxi-
mation (RMSEA). Acceptable fits are indicated by
the following: the chi-square statistic should not be
significant, chi-square per degree of freedom should
be less than three (Bollen and Stine, 1993), CFI
should be should be greater than 0.90 (Hair et al.,
1998), and RMSEA should be less than 0.08
Trust (b)
Model Fit Indices
0.69 Chi-square 0.004
p < 0.001 df
p
1
0.953 -0.55
Chi-square / df 0.004 p < 0.001
CFI 1.000
RMSEA 0.000
Fraud Risk (c)
Figure 2.(a) Results of research model – path analysis (H1
Satisfaction
and research question). Paths are standardized coeffi-
cients; p-values are one-tailed when direction is hypoth-
esized (satisfaction to trust) and two-tailed when no
direction is hypothesized (trust to risk). (a)Subjects’
response to item indicating they are satisfied with their
previous experience with the client’s management.
Measured on a seven-point, Likert-type scale with
anchors at each point (1: strongly disagree, 2: disagree,
3: slightly disagree, 4: neither agree nor disagree, 5:
slightly agree, 6: agree, and 7: strongly agree). (b)Sub-
jects’ trust in client’s management. Average score of
five items: T1, T2, T3, T4, and T5 (defined in Table II).
Measured on seven-point, Likert-type scale described
above in ‘‘a.’’ (c)Subjects’ assessment of the risk of man-
agement fraud for the current year’s audit. Measured on
a five-point, Likert-type scale with anchors at 1: very
low risk, 3: medium risk, and 5: very high risk.
1
(Browne and Cudeck, 1993). As shown in Figure 2,
the model’s fit indices are acceptable and the results
of the path analysis support the first hypothesis.
Auditors’ satisfaction with the client’s
management during a prior audit engagement was
positively related to their trust in that management
(standard- ized path coefficient = 0.69, p < 0.001,
one-tailed). This suggests auditors develop trust in a
client’s management after a previous satisfying
experience working with the client and they develop
low trust after an unsatisfying experience. Figure 2
also pro- vides evidence regarding the research
question of whether auditors’ trust in the client’s
management is related to their fraud risk assessments.
Specifically, the results show a negative relationship
between auditors’ trust and their fraud risk
assessment (standardized path coefficient = )0.55, p <
0.001, two-tailed). This suggests lower levels of trust
are associated with higher levels of perceived fraud
risk; one would logically expect an auditor to be very
skeptical of a client’s management if the auditor
does not trust that management. However, the
negative relationship also suggests higher levels of
trust are associated with lower levels of perceived
risk. This finding provides evidence, consistent with
recent criticisms, that auditors’ decisions may be
affected by their trust in the client’s management,
and thus, auditors may not be able to maintain their
profes- sional skepticism, as required by SAS No. 99,
regardless of their beliefs about the honesty and
trustworthiness of the client’s management.
We further explored the first hypothesis and the
research question by investigating whether the results
of the path analysis were consistent across both
manipulations. Specifically, we examined whether
auditors’ satisfaction with a client’s management
during a prior audit engagement was positively related
to auditors’ trust in that management after both a
satisfying and unsatisfying experience (i.e. higher
satisfaction associated with higher trust and lower
satisfaction associated with lower trust, respectively).
We also examined whether auditors’ trust was neg-
atively related to their fraud risk assessments after
both a satisfying and unsatisfying experience (i.e.
higher trust associated with lower risk and lower
trust associated with higher risk). This analysis was
performed by re-estimating the model separately for
each group of subjects receiving the satisfying
manipulation and the unsatisfying manipulation.
 1 William A. Kerler III and Larry N.
Panel A of Figure 3 displays the results of the Panel A: Unsatisfying experience with the client’s management during
path analysis for auditors receiving the unsatisfying a prior audit engagement
previous experience manipulation. The results of Trust (b)
the path analysis for auditors receiving the satisfy- 0.56
-0.55
p < 0.001
ing previous experience manipulation are presented p < 0.001
Model Fit Indices
in Panel B.23 The models’ fit indices are acceptable Chi-square 0.205
df 1
and the results of the path analysis support the first p 0.651
hypothesis. As shown in Panel A, after an unsat- Chi-square / df 0.205
CFI 1.000 (c)
isfying experience with the client’s management Satisfaction (a) Fraud Risk
RMSEA 0.000

during a prior audit engagement, auditors’ satis-
faction with their experience was positively related
to their trust in that management (standardized
path coefficient = 0.56, p < 0.001, one-tailed). As
shown in Panel B, after a satisfying experience,
auditors’ satisfaction was also positively related to
their trust (standardized path coefficient = 0.33,
p = 0.011, one-tailed). Taken together, these
results support the first research hypothesis that
auditors’ satisfaction is positively related to their
trust. This suggests that auditors do indeed develop
trust in a client’s management after working with
the client during a previous audit engagement.
Further, the extent to which auditors trust the
client’s management depends, in part, on how
satisfied auditors are with their previous experience
with the client – more satisfaction leads to higher
trust and vice versa.
Looking at Figure 3, we can again see the
models’ fit indices are acceptable but the results of
the path analyses provide mixed results for the
research question. As shown in Panel A, after an
unsatisfying experience, auditors’ trust in a client’s
management was negatively related to their perceived
risk of management fraud (standardized path
coeffi- cient = )0.55, p < 0.001, two-tailed). This
suggests that lower levels of trust are associated with
higher levels of perceived risk. This is consistent with
what one would expect from auditors performing
their job with due care because auditors who do not
trust a client’s management should be more
skeptical and should perceive a higher risk of fraud.
However, as shown in Panel B, after a satisfying prior
experience with the client’s management, there was
no rela- tionship between auditors’ trust and their
fraud risk assessment (standardized path
coefficient = )0.14, p = 0.356, two-tailed). This
suggests auditors’ per- ceived risk of fraud is not
affected by auditors’ trust in a client’s
management developed after a satisfying experience
working with the client during a prior
Panel B: Satisfying experience with the client’s management during a prior audit
engagement
Trust (b)
0.33
-0.14
p = 0.011
p = 0.356
Model Fit Indices
Chi-square 0.168
df
p
1
0.682
Chi-square / df 0.168 Fraud Risk (c)
Satisfaction (a)
CFI
RMSEA
1.000
0.000
Figure 3. Results of research model for each manipula-
tion group – path analysis (H1 and research question).
Paths are standardized coefficients; p-values are one-
tailed when direction is hypothesized (satisfaction to
trust) and two-tailed when no direction is hypothesized
(trust to risk). (a)Subjects’ response to item indicating
they are satisfied with their previous experience with the
client’s management. Measured on a seven-point,
Likert-type scale with anchors at each point (1: strongly
disagree, 2: disagree, 3: slightly disagree, 4: neither
agree nor disagree, 5: slightly agree, 6: agree, and 7:
strongly agree). (b)Subjects’ trust in client’s
management. Average score of five items: T1, T2, T3, T4,
and T5 (defined in Table II). Measured on seven-point,
Likert- type scale described above in ‘‘a.’’ (c)Subjects’
assessment of the risk of management fraud for the current
year’s audit. Measured on a five-point, Likert-type scale
with anchors at 1: very low risk, 3: medium risk, and 5:
very high risk.
audit engagement. Contrary to recent criticisms of
the profession, this finding supports that auditors are
able to maintain their professional skepticism, as
required by SAS No. 99, regardless of their beliefs
about the honesty and trustworthiness of the client’s
management.
To examine the second research hypothesis we
performed path analysis for the model presented in
Figure 4.24 The model was tested using path
analysis in AMOS 4.01 with maximum likelihood

Trust (b)
0.71 -0.55
p < 0.001 p < 0.001
0.02 Fraud Risk (c)
Satisfaction (a)
p = 0.781
Model Fit Indices
Chi-square
df
Moral Reasoning
(d)
p 0.969
Chi-square / df 0.032
CFI
RMSEA
Figure 4. Results of research model – path analysis
(H2). Paths are standardized coefficients; p-values are one-
tailed when direction is hypothesized (satisfaction to
trust) and two-tailed when no direction is hypothe- sized
(trust to risk) or when the direction is in the opposite
direction of that hypothesized (moral reasoning to trust).
(a)
Subjects’ response to item indicating they are satisfied
with their previous experience with the client’s
management. Measured on a seven-point, Likert-type
scale with anchors at each point (1: strongly disagree, 2:
disagree, 3: slightly disagree, 4: neither agree nor dis-
agree, 5: slightly agree, 6: agree, and 7: strongly agree).
(b)
Subjects’ trust in client’s management. Average
score of five items: T1, T2, T3, T4, and T5 (defined in
Table II). Measured on seven-point, Likert-type scale
described above in ‘‘a.’’ (c)Subjects’ assessment of the
risk of management fraud for the current year’s audit.
Measured on a five-point, Likert-type scale with
anchors at 1: very low risk, 3: medium risk, and 5: very
high risk. (d)Subjects’ moral reasoning measured as sub-
jects’ p-score from the DIT.
estimation. As shown in Figure 4, the model’s fit
indices are acceptable but the results of the path
analysis do not support the second hypothesis.
Auditors’ moral reasoning was not related to their
trust in that management (standardized path coef-
ficient = 0.02, p = 0.781, two-tailed). We also
tested the model separately for each prior audit
engagement experience manipulation group (not
tabulated). For both the satisfying experience
group and the unsatisfying experience group,
auditors’ moral reasoning was not significantly
related to auditors’ trust.25 A potential explanation
for this finding will be discussed in the concluding
section.
Auditors’ 1
Supplemental analysis
The most interesting finding from our main analyses
is that auditors’ trust in a client’s management
developed after a satisfying experience working with
the client’s management during a prior audit
engagement does not affect their fraud risk assess-
ments. This suggests auditors are able to maintain
their professional skepticism despite their beliefs
regarding management’s trustworthiness. We attri-
bute this finding to auditors’ expertise, training, and
0.063
2 the renewed emphasis on maintaining professional
skepticism, in part brought on by the issuance of
1.000 SAS No. 99. One potential counter argument is that
0.000
the experimental scenario was too clear cut to allow
auditors’ perceived risk of fraud to be decreased
because of their trust in the client’s management. To
help alleviate this potential issue, we distributed
the same experimental case to Certified Public
Accountants (CPAs) and received 83 completed
cases from subjects with an average of 14.2 years of
accounting experience (standard deviation of 9.2).
The current profession of this group of CPAs does
not include auditing; rather, their duties involve
other accounting areas (e.g. internal audit, business
consulting, tax, financial planning, financial services,
regulating, and other duties performed by controllers
and private industry accountants). Being practicing
accountants and holding a CPA license, these sub-
jects have a sufficient understanding of the auditing
profession, auditing standards, and the complexities
of the auditor–client relationship. As shown in Panel
A of Figure 5, the results for the CPAs that com-
pleted the overall unsatisfying case were very similar
to that of the auditor subjects. However, as shown in
Panel B of Figure 5, for CPAs that completed the
overall satisfying experience case, CPAs’ trust was
negatively and significantly related to their fraud risk
assessments (standardized path coefficient = )0.42,
p = 0.002, two tailed). This finding provides sup-
port that the experimental instrument was not so
clear cut so as to prohibit subjects from allowing
their trust to decrease their perceived risk of fraud.
The results of our first hypothesis and research
question suggest auditors’ trust is affected by their
satisfaction with the client’s management during a
prior audit engagement, and this trust affects their
fraud risk assessments after an unsatisfying
experience with the client. To eliminate the
possibility that
 1 William A. Kerler III and Larry N.
Panel A: Unsatisfying experience with the client’s management during
a prior audit engagement (n = 35)
Trust (b)
0.69 -0.40
p < 0.001 p = 0.010
Satisfaction (a) Fraud Risk (c)
Panel B: Satisfying experience with the client’s management during a prior
audit engagement (n = 48)
Trust (b)
0.70 -0.42
p < 0.001 p = 0.002
Satisfaction (a) Fraud Risk (c)
Figure 5. Model for supplemental analysis – non-audi-
tor certified public accountants. Paths are standardized
coefficients; p-values are one-tailed when direction is
hypothesized (satisfaction to trust) and two-tailed when
no direction is hypothesized (trust to risk). (a)Subjects’
response to item indicating they are satisfied with their
previous experience with the client’s management.
Measured on a seven-point, Likert-type scale with an-
chors at each point (1: strongly disagree, 2: disagree, 3:
slightly disagree, 4: neither agree nor disagree, 5: slight-
ly agree, 6: agree, and 7: strongly agree). (b)Subjects’
trust in client’s management. Average score of five
items: T1, T2, T3, T4, and T5 (defined in Table II).
Measured on seven-point, Likert-type scale described
above in ‘‘a.’’ (c)Subjects’ assessment of the risk of man-
agement fraud for the current year’s audit. Measured on
a five-point, Likert-type scale with anchors at 1: very
low risk, 3: medium risk, and 5: very high risk.
these results are incorrect and our experimental
manipulations simply affected auditors’ perceived
risk of fraud we estimated an alternative model,
presented in Figure 6, where we included the
manipulated experimental version as the first
variable in the model as opposed to auditors’
satisfaction. Further, we included a direct path from
the version variable to auditors’ fraud risk
assessments. As shown in Figure 6, the path from
the version to fraud risk was insignificant
(standardized path coeffi- cient = )0.01, p =
0.930, two-tailed). This provides
Trust (b)
-0.55
p < 0.001
0.74
p < 0.001
-0.01 Fraud Risk (c)
p = 0.930
Version (a)
Figure 6. Model for supplemental analysis – Did the
experimental manipulation directly affect auditors’ fraud
risk assessments? (a)Experimental version, where 1 =
overall satisfying experience with the client’s manage-
ment during a prior audit engagement, and, 2 = overall
unsatisfying experience. (b)Subjects’ trust in client’s man-
agement. Average score of five items: T1, T2, T3, T4,
and T5 (defined in Table II). Measured on seven-point,
Likert-type scale described above in ‘‘a.’’ (c)Subjects’
assessment of the risk of management fraud for the cur-
rent year’s audit. Measured on a five-point, Likert-type
scale with anchors at 1: very low risk, 3: medium risk,
and 5: very high risk.
evidence that our manipulations of satisfaction did
not inadvertently manipulate auditors’ perceived risk
of fraud.
Our first research hypothesis and research ques-
tion examine whether trust completely mediates the
effect of auditors’ satisfaction on auditors’
perceived risk of management fraud. We also tested
an alter- native model, presented in Figure 7, where
trust only partially mediates the effect of satisfaction
on risk (i.e. satisfaction with a previous experience
has both a direct and indirect effect, through trust, on
auditors’ perceived risk of fraud). This analysis was
performed by estimating the new model separately
for the two groups of subjects created by the
manipulation – one path analysis for those receiving
the satisfying manipulation and one path analysis for
those receiving the unsatisfying manipulation. The
path from satisfaction to fraud risk was not significant
when included in the models (both p > 0.650, two-
tailed). Further, the inclusion of this path did not
qualitatively affect the path coefficients reported in
the main analyses and their related p-values.
Therefore, the predicted complete mediation model
is more appropriate.
The results of our second research hypothesis
indicate auditors’ moral reasoning was not related to
 Auditors’
Trust (b)
Fraud Risk (c)
Satisfaction (a)
Figure 7. Model for supplemental analysis of alternative
partially mediating model. (a)Subjects’ response to item
indicating they are satisfied with their previous experi-
ence with the client’s management. Measured on a seven-
point, Likert-type scale with anchors at each point (1:
strongly disagree, 2: disagree, 3: slightly dis- agree, 4:
neither agree nor disagree, 5: slightly agree, 6: agree, and
7: strongly agree). (b)Subjects’ trust in client’s
management. Average score of five items: T1, T2, T3,
T4, and T5 (defined in Table II). Measured on seven-
point, Likert-type scale described above in ‘‘a.’’ (c)Sub-
jects’ assessment of the risk of management fraud for the
current year’s audit. Measured on a five-point, Likert-
type scale with anchors at 1: very low risk, 3: medium
risk, and 5: very high risk.
their trust in the client’s management. However, it is
possible auditors’ moral reasoning may affect their
satisfaction with the client’s management and/or
their fraud risk assessment. To test this, we included
to the model presented in Figure 4 direct effects of
moral reasoning on both satisfaction and fraud risk
(not tabulated). This model was calculated for the
entire sample and separately for each prior audit
engagement experience manipulation group. The
effects of moral reasoning on satisfaction and fraud
risk were not significant in any of the three models
(all p > 0.05, two-tailed).
The correlations discussed previously indicated
the auditors’ assessment of the competence of the
client’s management (hereafter competence) may be
associated with auditors’ trust in the management
and their fraud risk assessments. To investigate the
potential relationships, we calculated the three path
analyses utilized to examine the first hypothesis and
the research question (Figures 2 and 3) with com-
petence included as an additional variable having a
direct effect on both trust and fraud risk. The effect
of competence on trust and fraud risk was not sig-
nificant in any of the models (all p > 0.05).
1
We also tested whether any of the demographic
control variables affect the results of the main path
analyses examining the first hypothesis and the
research question (Figures 2 and 3) for the two groups
created by the satisfaction manipulation. The four
control variables of interest include the participants’
firm (Big Four or Non-Big Four), the participants’
years of auditing experience (in years), the partici-
pants’ rank in their firm (staff, in-charge, manager,
partner), and whether the participants have previously
participated in an audit that detected financial state-
ment fraud committed by client management (yes or
no). To do this, we performed eight new path analyses
(two manipulation groups for each of the four control
variables) for the model presented in Figure 8.
With respect to the relationships between satis-
faction, trust, and fraud risk, the results of the eight
models were qualitatively the same (standardized
path coefficients and statistical significance) as those
presented in our main analyses. This suggests the
results of the main analyses and the conclusions
drawn from the results are not affected by the
inclusion of the demographic variables. With respect
to the demographic variables, all four were signifi-
cantly related to one variable (satisfaction, trust, or
fraud risk) in the unsatisfying manipulation group,
while none was significantly related to the variables
in the satisfying manipulation group. Participants’
firm type was negatively related (standardized path
coefficient = )0.28, p = 0.050, two-tailed) to par-
ticipants’ satisfaction with the client’s
management in the unsatisfying manipulation group.
This indi-
cates participants from Big Four firms were more
unsatisfied with the client’s management after an
unsatisfying experience than Non-Big Four partici-
pants. Participants’ years of auditing experience
(standardized path coefficient = 0.23, p = 0.050, two-
tailed) and rank in their firm (standardized path
coefficient = 0.25, p = 0.034, two-tailed) were
both positively related to their fraud risk assessments
in the unsatisfying manipulation group. These two
suggest more experienced auditors assessed a higher
risk of fraud after an unsatisfying experience than did
less experienced auditors. Lastly, whether partici-
pants have previously worked on an audit that
detected fraud was positively related (standardized
path coefficient = 0.25, p = 0.040, two-tailed) to
their fraud risk assessments in the unsatisfying
manipulation group. This finding suggests that
 1 William A. Kerler III and Larry N.
Trust (b)
Satisfaction (a)
Fraud Risk
Control Variable (d)
Figure 8. Model for supplemental analysis of control
variables. (a)Subjects’ response to item indicating they
are satisfied with their previous experience with the
client’s management. Measured on a seven-point, Lik-
ert-type scale with anchors at each point (1: strongly
disagree, 2: disagree, 3: slightly disagree, 4: neither
agree nor disagree, 5: slightly agree, 6: agree, and 7:
strongly agree). (b)Subjects’ trust in client’s management.
Average score of five items: T1, T2, T3, T4, and T5
(defined in Table II). Measured on seven-point, Likert-
type scale described above in ‘‘a.’’ (c)Subjects’ assessment
of the risk of management fraud for the current year’s
audit. Measured on a five-point, Likert-type scale with
anchors at 1: very low risk, 3: medium risk, and 5: very
high risk. (d)Four control variables include: subjects’
firm (0: Non-Big Four, 1: Big Four), subjects’ years of
auditing experience (in years), subjects’ rank in their
firm (0: staff, 1: in-charge, 2: manager, 3: partner), and
whether subjects have participated on an audit that
detected financial statement fraud committed by client
management (0: no, 1: yes).
auditors who have participated in one or more
audits in which fraud was detected became more
skeptical of another client that provided them with
an unsatisfying experience.
Finally, we tested whether any of the same four
demographic control variables affect the results of
the main path analysis examining the second
hypothesis (Figure 4). Specifically, we performed
twelve new path analyses (for each control variable
we performed three path analyses – one utilizing
the entire sample and one for each of the
manipulation groups) for the model presented in
Figure 9. Consistent with our main analysis, the
path between moral reasoning and trust was not
significant in any of the twelve models that
included the control variable. Furthermore, in
Trust (b)
Fraud Risk (c)
(c) Satisfaction (a)
Moral Reasoning (d) Control Variable (e)
Figure 9. Model for supplemental analysis of control
variables. (a)Subjects’ response to item indicating they
are satisfied with their previous experience with the
client’s management. Measured on a seven-point, Lik- ert-
type scale with anchors at each point (1: strongly disagree,
2: disagree, 3: slightly disagree, 4: neither agree nor
disagree, 5: slightly agree, 6: agree, and 7: strongly agree).
(b)
Subjects’ trust in client’s management. Average score
of five items: T1, T2, T3, T4, and T5 (defined in Table II).
Measured on seven-point, Likert- type scale described
above in ‘‘a.’’ (c)Subjects’ assessment of the risk of
management fraud for the current year’s audit. Measured
on a five-point, Likert-type scale with anchors at 1: very
low risk, 3: medium risk, and 5: very high risk.
(d)
Subjects’ moral reasoning measured as sub- jects’ p-
score from the DIT. (e)Four control variables include:
subjects’ firm (0: Non-Big Four, 1: Big Four),
subjects’ years of auditing experience (in years), subjects’
rank in their firm (0: staff, 1: in-charge, 2: manager, 3:
partner), and whether subjects have participated on an
audit that detected financial statement fraud committed
by client management (0: no, 1: yes).
each of the models, the path from the control variable
to moral reasoning was not significant. This suggests
that the results of our main path analysis testing the
second hypothesis are robust to the inclusion of a
variety of control variables.
Discussion
Both the press (e.g. Herrick and Barrionuevo, 2002)
and academics (e.g. Johnstone et al., 2001) have
suggested that close relationships between auditors
and their clients (and presumably the trust that devel-
ops) may impair auditors’ professional
skepticism.
 Auditors’
The purpose of this study is to examine whether
auditors develop trust in a client’s management
and whether that trust affects their audit decisions.
Further, this study examines whether auditors’ moral
reasoning affects their trust. This study finds that after
both an overall satisfying and unsatisfying experience
working with a client’s management during a
prior audit engagement, auditors’ trust in that
management is positively related to their satisfaction
(i.e. higher satis- faction is associated with higher trust
and lower satis- faction is associated with lower trust).
This suggests that auditors do indeed develop trust in
a client’s manage- ment and the extent of the trust
depends, at least in part, on how satisfied the auditors
are with their previous experience working with the
client. Further, the results show that after
an unsatisfying experience with a client’s
management, auditors’ perceived risk of
management fraud is negatively related to their trust.
In other words, the low level of trust developed from
the unsatisfying experience is related to a higher risk of
management fraud. This suggests, as we would hope,
that auditors become more skeptical of a client’s
management after that management displays
untrustworthy behaviors. However, this study also
finds that after an overall sat- isfying experience
working with a client’s management during a prior
audit engagement, auditors’ trust is not related to
their perceived risk of fraud. This finding is consistent
with the requirement of SAS No. 99 that auditors
‘‘should conduct the engagement with a mindset
that recognizes the possibility that a material
misstatement due to fraud could be present, regardless
of any past experience with the entity and regardless
of the
auditor’s belief about management’s honesty and integrity
[emphasis added]’’ (AICPA, 2002, paragraph 13, p.
10). The extent to which auditors’ trust a client’s
management should be an ethical decision because
excessive trust could impair auditors’ necessary
professional skepticism, which is required by their
professional standards (e.g. SAS No. 99). Based on
this expectation and findings in prior literature, we
predicted auditors’ moral reasoning would be
negatively related to auditors’ trust (i.e. higher
moral reasoning would be related to lower trust).
However, results of this study do not support our
supposition; moral reasoning was not related to
auditors’ trust in a client’s management. We be-
lieve the likely explanations for this finding are
auditors’ expertise and training, as well as the re-
newed emphasis on professional skepticism in
both
1
the press and auditing standards. It is likely auditors
are very aware of the potential negative conse-
quences of trusting a client’s management (e.g.
impaired skepticism), and they, as they should,
guard against developing excessive levels of trust.
Future research may wish to investigate whether
auditors’ moral reasoning affects auditors’ trust
when performing other audit tasks where the
negative consequences of trusting the client may
not be so salient.
The implications of these results are important
for all financial statement users. This study provides
evidence that, despite recent criticisms, auditors are
able to maintain their professional skepticism.
Further, auditors appear to be cognizant of poten-
tial fraud perpetrated by the client’s management,
even after prior satisfying experiences with the
client and regardless of any trust the auditor may
have in that management. Standard setters and
professionals should also be interested and encour-
aged by the results of this study. The results indi-
cate auditors are following one standard designed to
maintain values with which the profession was built
upon – integrity, objectivity, and professional
skepticism.
This study differs from previous auditing re-
search and contributes to the literature in three
primary ways. First, this study is the first to
investigate whether auditors develop trust in a
client’s management and to examine whether this
trust affects their audit decisions. Recent criticisms
of the auditing profession often revolve around
whether close auditor–auditee relationships devel-
oped over multiple years may be impairing audi-
tors’ professional skepticism (e.g. Herrick and
Barrionuevo, 2002). This study is an important
first step toward examining whether auditors do
develop trust in a client’s management and also
understanding how this trust may affect their audit
decisions. The second contribution of this study to
the literature is the development of a measure of
auditors’ trust in a client’s management that may
be useful for future auditing research. For exam-
ple, this measure can be utilized to develop an
understanding of what factors affect auditors’ trust
in a client. It can also be used to identify what
types of audit decisions are or are not affected by
auditors’ trust in the client’s management. The
final contribution of this study is the examination
 1 William A. Kerler III and Larry N.
of whether moral reasoning affects auditors’ trust.
To our knowledge, this is the first study to
investigate this potential relationship.
This study also provides future research with other
avenues for consideration. This study investigates the
relationships based on a single past experience. It
would be interesting to identify whether auditors’
trust in the client’s management changes after
differ- entially satisfying experiences with the same
client. For example, can one dissatisfying experience
offset the effects of four prior satisfying experiences?
Future research could also investigate what other
factors affect auditors’ trust. For example, do
personal characteris- tics of auditors make them more
prone to trust a client? This study also finds
auditors’ trust affects their fraud risk assessment
after an unsatisfying past experience, but not after a
satisfying past experience. Future re- search could
investigate whether this pattern is stable across other
audit decisions. For example, would trust have similar
effects when auditors evaluate the fairness of
individual account balances, as opposed to an overall
fraud risk assessment? Would auditors’ trust affect
their evaluation of a client’s control environ- ment?
It is also possible that goodwill trust and com- petence
trust interact in an auditing environment. Future
research could investigate the potential inter- active
effect of these two components of trust on auditors’
decisions. Lastly, future research could investigate
whether the relationships found in this study are
constant across clients of varying importance to the
audit firm.
We acknowledge that this study, like others
performed in an artificial setting, has limitations.
We find auditors’ trust is affected by their satis-
faction with a client during a previous experience
described on paper; however, a bond developed in
the ‘‘real world’’ may potentially be even
stronger. Future research could investigate this with
stronger manipulations (e.g. having a member of
manage- ment be a former trusted member of the
audit firm and a friend) or by utilizing a game theory
experiment requiring professional subjects (i.e.
auditors and company managers) to play the role
of auditors and client management. Also, while we
attempted to collect data from a diverse group of
auditors, it is possible that our results are not
representative of all external auditors. Future re-
search may look to replicate these findings under
different conditions.
Notes
1
Specifically assessing the risk of fraud has been
found to improve auditors’ ability to detect fraud (e.g.
Knapp and Knapp, 2001; Zimbelman, 1997).
2
As another example, although not testing the
effects of trust, recent research by Carey and Simnett
(2006) suggests long audit partner tenure (defined as
when the auditor has been the partner in charge for
greater than seven years) may be related to a lower level
of audit quality. Specifically, they find evidence that
long partner tenure is related to a lower likelihood of
issuing a going-concern opinion and the extent to
which earnings targets are just beaten (missed).
3
A distinction between a competence component
and one or more goodwill components has also been
made by other researchers (e.g. Ganesan, 1994; Kee and
Knox, 1970; Lieberman, 1981; Mayer et al., 1995).
4
Hereafter, ethical reasoning is used interchangeably
with moral reasoning.
5
The pilot studies were also conducted with the
goal of ensuring the clarity of the experimental case.
The responses led to some wording changes (e.g. origi-
nally developed manipulations were modified).
6
Multiple iterations of Principal Components Analy-
sis, or PCA, with varimax rotation were performed (i.e.
a new PCA was performed after each item, or group of
items, was eliminated).
7
During the development of the scale, great care
was taken to ensure items were not discarded without
just reason. For example, some items loaded highly on
factors that were obviously interpretable as tapping into
a construct other than goodwill trust (e.g. three items
consistently loaded on a factor tapping into the auditors’
perception of management’s loyalty to the corporation).
8
When the PCA was performed utilizing only the 50
participants with experience assessing fraud risk, the same
three items were eliminated. Further, the remaining five
items loaded on a single factor, which explained over 70
percent of the variance and had a Cronbach’s Alpha of 0.90.
9
We utilized the three-story DIT instead of the
six-story version because of participant time considerations.
10
The p-score is considered a reliable and valid mea-
sure of cognitive moral development based on Kohl-
berg’s model (see Rest, 1993).
11
Thorne (2000) created an Accounting Ethical
Dilemma Instrument which measures accountants’
con- text-specific moral reasoning (prescriptive and
delibera- tive reasoning). Prior research indicates that
accountants’ general moral reasoning (as measured by
the DIT) may differ from their context-specific moral
reasoning (e.g. Earley and Kelly, 2004); however, the
two are signifi- cantly correlated (Thorne, 2000).
 Auditors’
12
The four prizes included three desktop stereo/CD
players and a one-hundred dollar cash prize.
13
The response rate of 75 percent is in line with previ-
ous auditing research utilizing contact persons. For exam-
ple, Asare et al. (2000) had a response rate of 78.6 percent
and Taylor (2000) had a response rate of 74 percent.
14
Because data were collected utilizing contact per-
sons and conference volunteers, we are unable to test
for non-response bias.
15
This was based on subjects’ response (yes or no)
to the following question: ‘‘Have you had previous
experience assessing fraud risk.’’ The fairly large
num- ber of participants needing to be excluded was ex-
pected because we had no control over which specific
auditors the contact partners selected, nor did we have
control over which specific attendees of the confer-
ence completed the instrument. However, excluding
these subjects was deemed necessary because the main
experimental task centered on assessing the risk of
fraud. Including these participants in the analysis
would only serve to bias the results. Seven of the 23
excluded were from Big Four firms (this represents
about 16 percent of Big Four participants initially col-
lected) and 16 were from Non-Big Four firms (this
represents about 24 percent of Non-Big Four partici-
pants initially collected).
16
We used the fourth year audit for the firm because
we wanted to avoid the possible fluctuations in risk
assessments that may occur during the first couple of
years of auditing a new client.
17
While previous research generally shows that red flag
checklists are not as useful as other decision aids in
fraud risk assessment (e.g. Asare and Wright, 2004; Bell
and Carcello, 2000; Eining et al., 1997; Loebbecke et al.,
1989; Pincus, 1989), Shelton et al. (2001) report finding
three Big Five firms that utilize a red flag checklist to
identify the presence or absence of fraud risk factors.
18
Adapted from one or more of the following:
AICPA (1997); AICPA (2002); Apostolou et al. (2001);
Bell and Carcello (2000); Eining et al. (1997); Hacken-
brack (1993); Davidson (1994); Zimbelman (1997).
19
The significant Pearson’s product-moment correla-
tions between ‘‘Comp’’ and both ‘‘Trust’’ (0.244)
and ‘‘Risk’’ ()0.220) should be noted at this point
because they were not expected. These two correlations
indicate that the participants’ assessment of the
competence of the client’s management is associated with
the participants’ trust in the management and their fraud
risk assessment. This finding was unexpected because we
successfully controlled the competence level of
management, as dis- cussed in the following section. We
will further investi- gate the potential effects of competence
on trust and fraud risk assessment in our supplemental
analyses.
1
20
Not all participants responded in the expected direc-
tion. For example, three of the 46 participants who
received the unsatisfying manipulation rated their satisfac-
tion on the first point at the opposite end of the seven-
point scale (i.e. they ‘‘strongly agreed’’ that they
were satisfied), while none of the 43 participants who
received the satisfying manipulation rated their satisfaction
on the first point at the opposite end of the scale (i.e. none
‘‘strongly disagreed’’ that they were satisfied). Results
(path coefficients and p-values) of our main path analyses
utilized to examine the two hypotheses and the research
question (Figures 2–4) are qualitatively the same (not tab-
ulated) when these three participants are excluded from
the sample. Similarly, results (path coefficients and p-val-
ues) of our main path analyses utilized to examine the two
hypotheses and the research question (Figures 2–4) are
qualitatively the same (not tabulated) when partici- pants
are excluded using two additional cutoffs: (1) they rated
their satisfaction on either of the two points at the opposite
end of the seven-point scale (e.g. responded with a one or
a two when we expected a high response, or, responded
with a six or a seven when we expected a low response),
and (2) they rated their satisfaction on either of the three
points at the opposite end of the scale.
21
Not all participants responded in the expected direc- tion.
Two of the 89 participants ‘‘slightly disagreed’’ (the third
point at the opposite end of the seven-point scale) that
management was competent (no participants ‘‘dis-
agreed’’ or ‘‘strongly disagreed’’). Results (path
coeffi- cients and p-values) of our main path analyses
utilized to examine the two hypotheses and research
question (Figures 2–4) are qualitatively the same (not
tabulated) when these two participants are excluded from
the sample.
22
An assumption of path analysis is that the data have a
multivariate distribution. Lending some support to this
assumption, we find that all univariate skewness and kur-
tosis values for the measured variables, with the exception
of satisfaction’s skewness, are insignificant following
Tabachnick and Fidell’s (2001) recommended alpha level
of 0.01 with small-to-moderate samples. Furthermore,
multivariate kurtosis is also insignificant at an alpha level
of 0.05. Finally, using Hair et al.’s (1998) suggested alpha
level of 0.001, no multivariate outliers are identified with
the Mahalanobis D2 measure. Because of the potential
issue identified in the univariate skewness of the satisfac-
tion variable, we also perform the path analysis using the
asymptotically distribution-free method of estimation.
According to Hair et al. (1998), this ‘‘technique has
received particular attention recently due to its insensitiv-
ity to nonnormality of the data’’ (606). The results (path
coefficients and p-values) of our main path analyses uti-
lized to examine the two hypotheses and the research
question (Figures 2–4) are qualitatively the same as those
 1 William A. Kerler III and Larry N.
reported using the maximum likelihood method of esti-
mation and therefore not reported.
23
Results (path coefficients and p-values) of our main
path analyses utilized to examine the two hypotheses and
the research question (Figures 2–4) are qualitatively the
same (not tabulated) when the sample is tested excluding
the six staff-level subjects.
24
To test the second hypothesis, ten subjects needed
to be excluded from the analysis because they failed to
complete the DIT. The remaining 79 auditors were in-
cluded in the analysis.
25
A ‘‘meaningless’’ score, or m-score, is also
calculated
from the DIT and serves as an internal reliability check.
These items ‘‘are written in a pretentious and lofty
sounding [manner], but are really meaningless’’ (Rest,
1993, pp. 12–13). Rest (1993) recommends discarding a
participant’s data if his/her m-score is greater than three
on the three-story DIT. Results (path coefficients and p-
values) of the main path analyses utilized to test the sec-
ond hypothesis (Figure 4) are qualitatively the same when
we exclude auditors with m-scores greater than three.
Further, a less stringent m-score cutoff value of greater
than five has recently been recommended (Rest et al.,
1999b). Results (path coefficients and p-values) of the
main path analyses utilized to test the second hypothesis
(Figure 4) are qualitatively the same when we exclude
auditors with m-scores greater than five.
Appendix A
Existing scales utilized to develop trust scale
Author(s) (Year): Name of scale or name of construct
being measured
• Armitage and Conner (1999): Intentions
• Cook and Wall (1980): Interpersonal Trust at Work
• Craig and Gustafson (1998): Integrity
• Doney and Cannon (1997): Trust of Supplier Firm and
Trust of Salesperson
• Ganesan (1994): Vendor’s Benevolence
• Plank et al. (1999): Measures of Trust
• Rempel et al. (1985): Trust Scale (in close relationships)
• Rioux and Penner (2001): Motives
• Rotter (1967): Interpersonal Trust Scale
Appendix B: Experimental case
Year four’s audit
You are an audit manager for a large public
accounting
firm. Last year you worked for the first time on the
audit of XYZ Corporation. This was the fourth year
your firm has audited XYZ. The previous three
years’ audits all went smoothly. XYZ is a medium-
sized furniture manufacturing firm. Their management
consists of six individuals, all with over eight years of
management experience at XYZ. Two are Certified
Public Accountants (both with over four years of
auditing experience) who earned both a bachelor’s
and master’s degree in accounting. One, also a Cer-
tified Public Accountant (two years of auditing
experience), earned a bachelor’s degree in
accounting and a Master’s of Business
Administration. Two earned both a bachelor’s
degree in management and a Master’s of Business
Administration. One earned a bachelor’s degree in
industrial and systems engineer- ing and a doctorate
in manufacturing systems engi- neering.
During the planning phase for last year’s audit
you interviewed XYZ’s management and
determined that they were very competent; they had
the re- quired knowledge, skills, and training to make
management decisions and to prepare financial
statements for XYZ. You also became very familiar
with XYZ’s operations by observing and inter-
viewing numerous employees. Your firm’s policy
during the planning phase of each year’s audit is to
identify whether certain fraud indicators, or ‘‘red
flags,’’ are present. The presence of a red flag is an
indicator of potential financial statement fraud. After
becoming familiar with both the management and
operations, you used this Red Flag Checklist to
identify which of the 20 ‘‘red flags’’ were present
or absent. You found three items to be present. The
following page contains the Red Flag Checklist with
all 20 items; the three items you found are listed first
and indicated with an ‘‘X.’’
Red flag checklist
Check if Red flag indicators
present
X (1.) The industry is seeing increased competition
in increasingly tougher economic times
X (2.) The company has a high dependence on debt
financing
 Auditors’
Check if Red flag indicators
present
X (3.) Management has a significant accounting-
based incentive plan
(4.) There is weak monitoring of the company’s
internal controls
(5.) Management is inexperienced or unqualified
(6.) Management has a large stock ownership
interest
(7.) Management operation and financial deci-
sions are dominated by a single individual
(8.) There has been high management turnover
(9.) Management places undue emphasis on
meeting earnings projections
(10.) Management is under strong pressure to in-
crease the price of company’s stock
(11.) There is inadequate staffing of the
accounting department
(12.) There is inadequate use of budgeting process
and interim financial statements
(13.) The internal audit department is ineffective
(14.) There is a known history of securities law
violations or claims against the entity or its
management alleging fraud
(15.) The company lacks proper review and
approval procedures
(16.) The company has experienced a net loss at
least once in the past three years
(17.) The company has narrow margins of com-
pliance with debt covenants
(18.) The company is in a period of rapid growth
(19.) The company’s profitability is inconsistent
with most of the industry
(20.) The company has entered into transactions
with related parties
Your firm’s policy at the end of each year’s audit
planning phase is to make an overall assessment of
the risk of management fraud. Therefore, at the end
of year four’s audit planning phase, you used the
results of the Red Flags Checklist and your under-
standing of the client to assess the risk of management
fraud for year four’s audit at ‘‘3 – Medium
Risk,’’ based on the following scale.
Risk of management fraud
3 XYZ’s management consists of the same six
Medi
um
1 2
Risk
4 5
Very Low Very High
Risk Risk
1
Satisfying experience version
Results of year four’s audit
During the audit work, XYZ’s management always
had a respectful attitude and never engaged in
disputes with you. They always appeared to be well
prepared for your meetings. When you were on
location, you often joined them for lunches. During
your review of their accounting records the only issue
you had was that you believed they immaterially
understated their estimate of Allowance for Bad Debt;
and hence, immaterially understated their operating
expenses. Although the amount was immaterial, you
proposed an adjustment and XYZ’s management
made it without hesitation. The end result of the audit
was an unqualified opinion for XYZ’s financial
statements.
Unsatisfying experience version
Results of year four’s audit
During the audit work, XYZ’s management had a
hostile attitude and frequently engaged in disputes
with you. They never appeared to be well prepared for
your meetings. When you were on location, you were
never invited to join them for lunches. During your
review of their accounting records the only issue you
had was that you believed they immaterially under-
stated their estimate of Allowance for Bad Debt; and
hence, immaterially understated their operating ex-
penses. Although the amount was immaterial, you
proposed an adjustment. XYZ’s management put off
responding to your proposed adjustment until you
mentioned it again a few days later. At this point they
refused to make the adjustment. Because your pro-
posed adjustment did not materially affect the
financial statements, the end result of the audit was an
unqualified opinion for XYZ’s financial statements.
Year five’s audit
The next year you have again been assigned to work on
the audit of XYZ Corporation. This will be the second
year you have audited XYZ (your firm’s fifth year).
During the planning phase, you once again interviewed
management and other employees. You found that
individuals
as last year. Furthermore, XYZ operates and
functions the same as it did last year when you first
gained an understanding of the company. After
familiarizing
 1 William A. Kerler III and Larry N.
yourself again with both the management and the
operations, you used the same checklist as last year to
identify whether 20 ‘‘red flags’’ were present or

Strongly

Strongly
1. Using the results of the red flags checklist, and your understanding of the client, how would you assess the risk of management fraud for year five’s audit?
absent. You found the same three items to be present as
you did last year. Below is the Red Flag Checklist with

7
all 20 items; the three items you found are listed first
and indicated with an ‘‘X.’’

Agr

Agr
Red flag checklist

6
For each of the following statements, please indicate – by circling the corresponding number – how much you agree or disagree.
Check if Red flag indicators

Very High
present

Slightly

Slightly
X (1.) The industry is seeing increased competition
in increasingly tougher economic times

5
5
X (2.) The company has a high dependence on debt
financing
X (3.) Management has a significant accounting-
based incentive plan
(4.) There is weak monitoring of the company’s ***Please DO NOT refer back to the preceding pages when answering these questions***
internal controls

Neither Agree nor

Neither Agree nor

(5.) Management is inexperienced or unqualified
(6.) Management has a large stock ownership interest
(7.) Management operation and financial deci-
sions are dominated by a single individual
(8.) There has been high management turnover
(9.) Management places undue emphasis on
4

meeting earnings projections

4

4
(10.) Management is under strong pressure to in-

3. I am satisfied with my experience with XYZ Corporation’s

crease the price of company’s stock
(11.) There is inadequate staffing of the
accounting department
(12.) There is inadequate use of budgeting process
Medium

and interim financial statements

Slightly

Slightly
(13.) The internal audit department is ineffective
3

(14.) There is a known history of securities law

3

violations or claims against the entity or its

management alleging fraud
***Be sure to answer all the questions***

(15.) The company lacks proper review and ap-

Disagr

Disagr

proval procedures
(16.) The company has experienced a net loss at
2

least once in the past three years

(17.) The company has narrow margins of com-
2. XYZ’s management is competent.

pliance with debt covenants

(Please circle your risk assessment)

(18.) The company is in a period of rapid growth

(19.) The company’s profitability is inconsistent
Very Low

with most of the industry

Strongly

Strongly

(20.) The company has entered into transactions

with related parties
1

1
4. I believe that XYZ Corporation’s management are thoroughly dependable people.
1 2 3 4 5 6 7
Strongly Disagree Disagree Slightly Disagree Neither Agree nor Disagree Slightly Agree Agree Strongly Agree
5. XYZ Corporation’s management will do everything possible to help me.
1 2 3 4 5 6 7
Strongly Disagree Disagree Slightly Disagree Neither Agree nor Disagree Slightly Agree Agree Strongly Agree
6. XYZ Corporation’s management are like my friends.
1 2 3 4 5 6 7
Strongly Disagree Disagree Slightly Disagree Neither Agree nor Disagree Slightly Agree Agree Strongly Agree
7. I do not think that XYZ Corporation’s management is completely open in dealing with me.
1 2 3 4 5 6 7
Strongly Disagree Disagree Slightly Disagree Neither Agree nor Disagree Slightly Agree Agree Strongly Agree
8. XYZ Corporation’s management plans to be helpful during future audits.
1 2 3 4 5 6 7
Strongly Disagree Disagree Slightly Disagree Neither Agree nor Disagree Slightly Agree Agree Strongly Agree

Auditors’
General questions

**These questions should be answered based on your actual experiences and/or opinions.**
a. What is your gender? Male Female
b. Are you a Certified Public Accountant (CPA)? Yes No
c. How many years of auditing experience do you have?
d. Which of the following best describes your current job position? (CIRCLE ONE)
Staff In-Charge Manager Partner
e. How many audits, on which you have directly participated, eventually detected financial statement fraud committed by client management?
f. Have you had previous experience assessing fraud risk? Yes No

1
 1 William A. Kerler III and Larry N.
Defining Issues Test (DIT)
Not included: © 1979, James Rest
All Rights Reserved
References
American Institute of Certified Public Accountants
(AICPA): 1997, Consideration of Fraud in a
Financial Statement Audit. Statement on Auditing
Standards No. 82 (AICPA, New York).
American Institute of Certified Public Accountants
(AICPA): 2002, Consideration of Fraud in a
Financial Statement Audit. Statement on Auditing
Standards No. 99 (AICPA, New York).
Apostolou, B., J. Hassell, S. Webber and G. Sumners:
2001, ‘The Relative Importance of Management
Fraud Risk Factors’, Behavioral Research in
Accounting 13, 1–24.
Armitage, C. and M. Conner: 1999, ‘The Theory of
Planned Behaviour: Assessment of Predictive Validity
and ‘‘Perceived Control’’’, British Journal of
Social Psychology 38, 35–54.
doi:10.1348/014466699164022.
Asare, S., G. Trompeter and A. Wright: 2000, ‘The Effect
of Accountability and Time Budgets on Auditors’
Testing Strategies’, Contemporary Accounting Research
17(4), 539–560. doi:10.1506/F1EG-9EJG-DJ0B-
JD32.
Asare, S. and A. Wright: 2004, ‘The Effectiveness of
Alternative Risk Assessment and Program Planning
Tools in a Fraud Setting’, Contemporary Accounting
Research 21(2), 325–352. doi:10.1506/L20L-7FUM-
FPCB-7BE2.
Beaulieu, P.: 2001, ‘The Effects of Judgments of New
Clients’ Integrity upon Risk Judgments, Audit Evi-
dence, and Fees’, Auditing: A Journal of Practice &
Theory 20(September), 85–99.
Bell, T. and J. Carcello: 2000, ‘Research Notes: A
Decision Aid for Assessing the Likelihood of Fraudu-
lent Financial Reporting’, Auditing: A Journal of
Practice & Theory 19(Spring), 169–184.
Bernardi, R. A.: 1994, ‘Fraud Detection: The Effect of
Client Integrity and Competence and Auditor Cog-
nitive Style’, Auditing: A Journal of Practice and
Theory 13(Supplement), 68–84.
Bollen, K. and R. Stine: 1993, ‘Bootstrapping Good-
ness-of-Fit Measures in Structural Equation Models’,
in K. Bollen and J. Long (eds.), Testing Structural
Equation Model (Sage Publications, Newbury Park,
CA), pp. 111–135.
Browne, M. and R. Cudeck: 1993, ‘Alternative Ways of
Assessing Model Fit’, in K. Bollen and J. Long (eds.),
Testing Structural Equation Model (Sage Publications,
Newbury Park, CA), pp. 136–162.
Carey, P. and R. Simnett: 2006, ‘Audit Partner Tenure
and Audit Quality’, The Accounting Review 81(3),
653– 676.
Cook, J. and T. Wall: 1980, ‘New Work Attitude
Measures of Trust, Organizational Commitment and
Personal Need Non-Fulfillment’, Journal of Occupa-
tional Psychology 53, 39–52.
Craig, S. and S. Gustafson: 1998, ‘Perceived Leader
Integrity Scale: An Instrument for Assessing Employee
Perceptions of Leader Integrity’, Leadership
Quarterly 9(2), 127–145. doi:10.1016/S1048-
9843(98)90001-7.
Davidson, A. S.: 1994, ‘Discussion of Fraud Detection:
The Effect of Client Integrity and Competence and
Auditor Cognitive Style’, Auditing: A Journal of
Practice & Theory 13(Auditing Symposium), 85–89.
Doney, P. and J. Cannon: 1997, ‘An Examination of the
Nature of Trust in Buyer-Seller Relationships’, Journal
of Marketing 61(2), 35–51. doi:10.2307/1251829.
Earley, C. and P. Kelly: 2004, ‘A Note on Ethics Edu-
cational Interventions in an Undergraduate Auditing
Course: Is There an ‘‘Enron Effect’’?’, Issues in
Accounting Education 19(1), 53–71.
Eining, M., D. Jones and J. Loebbecke: 1997, ‘Reliance
on Decision Aids: An Examination of Auditors’
Assessment of Management Fraud’, Auditing: A Journal
of Practice & Theory 16(2), 1–19.
Flores, F. and R. Solomon: 1998, ‘Creating Trust’,
Business Ethics Quarterly 8(2), 205–232. doi:10.2307/
3857326.
Ganesan, S.: 1994, ‘Determinants of Long-Term Orien-
tation in Buyer-Seller Relationships’, Journal of Mar-
keting 58(2), 1–19. doi:10.2307/1252265.
Goodwin, J.: 1999, ‘The Effects of Source Integrity and
Consistency of Evidence on Auditors’ Judgments’,
Auditing: A Journal of Practice & Theory 18(2), 1–16.
Hackenbrack, K.: 1993, ‘The Effect of Experience with
Different Sized Clients on Auditor Evaluations of
Fraudulent Financial Reporting Indicators’, Auditing:
A Journal of Practice & Theory 12(Spring), 99–110.
Hair , J. F. Jr., R. E. Anderson, R. L. Tatham and
W. C. Black: 1998, Multivariate Data Analysis,
5th Edition (Prentice Hall, Upper Saddle River,
New Jersey).
Herrick, T. and A. Barrionuevo: 2002, ‘Were Auditor
and Client too Close-Knit?’, Wall Street Journal,
January 21, p. C1.
Ironson, G., P. Smith, M. Brannick, W. Gibson and
K. Paul: 1989, ‘Construction of a Job in General
Scale: A Comparison of Global, Composite, and
 Auditors’ Trust
Specific Measures’, Journal of Applied Psychology
74(2), 193–200. doi:10.1037/0021-9010.74.2.193.
Johnstone, K., M. Sutton and T. Warfield: 2001,
‘Antecedents and Consequences of Independence
Risk: Framework for Analysis’, Accounting Horizons
15(1), 1–18.
Kaplan, S. E., K. J. Newberry and P. M. J. Reckers: 1997,
‘The Effect of Moral Reasoning and Educational
Communications on Tax Evasion Intentions’, The
Journal of the American Taxation Association 19(2),
38– 54.
Kee, H. and R. Knox: 1970, ‘Conceptual and Method-
ological Considerations in the Study of Trust and
Suspicion’, The Journal of Conflict Resolution 14(3),
357– 366. doi:10.1177/002200277001400307.
King, D.: 1960, ‘A Multiplant Factor Analysis of
Employees’ Attitudes Toward Their Company’,
Jour- nal of Applied Psychology 40, 241–243.
doi:10.1037/ h0042484.
King, R.: 2002, ‘An Experimental Investigation of Self-
Serving Biases in an Auditing Trust Game: The Effect
of Group Affiliation’, The Accounting Review 77(2),
265–284.
Knapp, C. and M. Knapp: 2001, ‘The Effects of Expe-
rience and Explicit Fraud Risk Assessment in Detect-
ing Fraud with Analytical Procedures’, Accounting,
Organizations and Society 26, 25–37. doi:10.1016/
S0361-3682(00)00005-2.
Kohlberg, L.: 1969, ‘Stages and Sequences: The Cog-
nitive Developmental Approach to Socialization’, in
D. Goslin (ed.), Handbook of Socialization Theory
and Research (Rand McNally, Chicago), pp. 347–
480.
Latham, C., F. Jacobs and P. Roush: 1998, ‘Does Auditor
Tenure Matter?’, Research in Accounting Regulation
12, 165–177.
Lieberman, J.: 1981, The Litigious Society (Basic Books,
New York).
Loebbecke, J., M. Eining and J. Willingham: 1989,
‘Auditors’ Experience with Material Irregularities:
Frequency, Nature, and Detectability’, Auditing: A
Journal of Practice & Theory 9(Fall), 1–28.
Mayer, R., J. Davis and F. Schoorman: 1995, ‘An Inte-
grative Model of Organizational Trust’, Academy of
Management Review 20, 709–734. doi:10.2307/258792.
Messier, W. Jr., S. Glover and D. Prawitt: 2006, Auditing
& Assurance Services: A Systematic Approach, 4th Edition
(McGraw-Hill Irwin, New York).
Nicholls, J., G. Gilbert and S. Roslow: 1998, ‘Parsimo-
nious Measurement of Customer Satisfaction with
Personal Service and the Service Setting’, Journal of
Consumer Marketing 15(3), 239–253. doi:10.1108/
07363769810219116.
135
Nooteboom, B.: 1996, ‘Trust, Opportunism and Gov-
ernance: A Process and Control Model’, Organization
Studies 17(6), 985–1010. doi:10.1177/01708406
9601700605.
Peecher, M.: 1996, ‘The Influence of Auditors’
Justifica- tion Processes on Their Decisions: A
Cognitive Model and Experimental Evidence’,
Journal of Accounting Research 34(1), 125–140.
doi:10.2307/2491335.
Pincus, K.: 1989, ‘The Efficacy of Red Flags Question-
naire for Assessing the Possibility of Fraud’,
Accounting, Organizations, and Society 14(1–2),
153–163. doi:10. 1016/0361-3682(89)90039-1.
Plank, R., D. Reid and E. Bolman Pullins: 1999, ‘Per-
ceived Trust in Business-to-Business Sales: A New
Measure’, Journal of Personal Selling & Sales
Management 19(3), 61–71.
Ponemon, L. A.: 1992, ‘Auditor Underreporting of Time
and Moral Reasoning: An Experimental Lab Study’,
Contemporary Accounting Research 9(1), 171–189.
Ponemon, L. A. and D. R. L. Gabhart: 1993, Ethical
Reasoning in Accounting and Auditing (Canadian
Certi- fied General Accountants’ Research Foundation,
Vancouver).
Rempel, J., J. Holmes and M. Zanna: 1985, ‘Trust in
Close Relationships’, Journal of Personality and Social
Psychology 49, 95–112. doi:10.1037/0022-3514.49.1.
95.
Rest, J. R.: 1979, Development in Judging Moral Issues
(University of Minnesota, Minneapolis, MN).
Rest, J. R.: 1993, Guide for the Defining Issues Test –
How to Use the Optical Scan Forms and the Center’s
Scoring Service (University of Minnesota, Minneapolis,
MN).
Rest, J. R., D. Narvaez, M. J. Bebeau and S. J. Thoma:
1999a, Postconventional Moral Thinking: A Neo-Kohl-
bergian Approach, 4th Edition (Lawrence Erlbaum
Associates, Mahwah, New Jersey).
Rest, J. R., D. Narvaez, S. J. Thoma and M. J. Bebeau:
1999b, ‘DIT2: Devising and Testing a Revised
Instrument of Moral Judgment’, Journal of Educational
Psychology 91(4), 644–659. doi:10.1037/0022-0663.
91.4.644.
Ring, P. and A. Van De Ven: 1992, ‘Structuring
Cooperative Relationships Between Organizations’,
Strategic Management Journal 13(October), 483–498.
doi:10.1002/smj.4250130702.
Rioux, S. and L. Penner: 2001, ‘The Causes of Organi-
zational Citizenship Behavior: A Motivational Analy-
sis’, Journal of Applied Psychology 86(6), 1306–1314.
doi:10.1037/0021-9010.86.6.1306.
Rotter, J.: 1967, ‘A New Scale for the Measurement of
Interpersonal Trust’, Journal of Personality 35, 651–665.
doi:10.1111/j.1467-6494.1967.tb01454.x.
 1 William A. Kerler III and Larry N.
Shaub, M.: 1996, ‘Trust and Suspicion: The Effects of Twery, R., J. Schmid and C. Wrigley: 1958, ‘Some
Situational and Dispositional Factors on Auditors’ Factors in Job Satisfaction’, Educational and
Trust of Clients’, Behavioral Research in Accounting 8, Psychological Measurement 18, 189–202.
154–174. doi:10.1177/00131644580
Shelton, S., O. Whittington and D. Landsittel: 2001, 1800120.
‘Auditing Firms’ Fraud Risk Assessment Practices’,
Accounting Horizons 15(1), 19–33. Windsor, C. A. and N. M. Ashkanasy: 1995, ‘The Effect
Tabachnick, B. G. and L. S. Fidell: 2001, Using Multi- of Client Management Bargaining Power, Moral
variate Statistics, 4th Edition (Allyn & Bacon, Massa- Reasoning Development, and Belief in a Just World
on Auditor Independence’, Accounting,
chusetts).
Taylor, M.: 2000, ‘The Effects of Industry Specialization Organizations and Society 20(7/8), 701–720.
on Auditors’ Inherent Risk Assessments and Confi- doi:10.1016/0361- 3682(95)00018-5.
dence Judgments’, Contemporary Accounting Research Zimbelman, M.: 1997, ‘The Effects of SAS No. 82 on
17(4), 693–712. doi:10.1506/3LDH-AV52-0F4W- Auditors’ Attention to Fraud Risk Factors and Audit
H4BB. Planning Decisions’, Journal of Accounting Research
Thoma, S. J., J. R. Rest and M. L. Davison: 1991, 35(3), 75–97. doi:10.2307/2491454.
‘Describing and Testing a Moderator of the Moral
Judgment and Action Relationship’, Journal of William A. Kerler
Person- ality and Social Psychology 61(4), 659–669. III McGladrey Faculty
doi:10.1037/ 0022-3514.61.4.659. Fellow,
Thorne, L.: 2000, ‘The Development of Two Measures Department of Accountancy and Business Law,
to Assess Accountants’ Prescriptive and Deliberative
Moral Reasoning’, Behavioral Research in Accounting 12, University of North Carolina Wilmington,
139–169. Cameron Hall – Suite 230, 601 South College
Tsui, J. S. L. and F. A. Gul: 1996, ‘Auditors’ Behaviour Road, Wilmington, NC 28403-5901,
in an Audit Conflict Situation: A Research Note on the U.S.A.
Role of Locus of Control and Ethical Reasoning’,
Accounting, Organizations and Society 21(1), 41–51. E-mail: kerlerw@uncw.edu
doi:10.1016/0361-3682(95)00009-X.
Larry N.
Killough KPMG
Professor,
Department of Accounting and Information Systems,
Virginia Tech, 3007 Pamplin Hall,
Blacksburg, VA 24061, U.S.A.