Description
Prior research has highlighted the problem of dysfunctional reactions to control systems. Studies carried
out in auditing firms have revealed high levels of budget pressure leading to a variety of dysfunctional
behaviours which can be difficult to control. This study e xamines how subordinates’ reaction to control
systems is influenced by the leadership behaviour of supervisors.
Pergamon Accounting, Organizations and Society, Vol. 20, No. 5, PP. 405-420, 1995
Copyright 0 1995 Elsmier Science Ltd
Printed in Great Britain. AU rights reserved
0361-36132/95 $9.50+0.00
THE CONTROL PROBLEM IN PUBLIC ACCOUNTING FIRMS:
AN EMPIRICAL STUDY OF THE IMPACT OF LEADERSHIP STYLE*
DAVID T. OTLEY
University of Lancaster
and
BERNARD J. PIERCE
LWbl i n Ci ty Uni versi ty
Prior research has highlighted the problem of dysfunctional reactions to control systems. Studies carried
out in auditing firms have revealed high levels of budget pressure leading to a variety of dysfunctional
behaviours which can be difficult to control. This study e xamines how subordinates’ reaction to control
systems is influenced by the leadership behaviour of supervisors. Data were collected by means of a
questionnaire sent to all audit seniors in three Big Six audit firms. Two forms of dysfunctional behaviour
were measured: under-reporting of time and audit quality reduction behaviour. A leadership style
characterized by high structure and low consideration was found to be associated with the highest
level of dysfunctional behaviour for all behaviours examined, while the lowest level of dysfunctional
behaviour was associated with a style depicting low structure and high consideration. Perceived
environmental uncertainty was found to moderate these relationships, and to exercise a stronger
moderating effect for audit quality reduction behaviour than for under-reporting of time.
During the 198Os, a number of studies exam-
ined the workings of the formal control system
in large audit firms (e.g. Alderman h Deitrick,
1982; Lightner et al ., 1983; Cook & Kelley,
1988). The conclusion was that there was a
disturbingly high level of dysfunctional be-
haviour associated with the control system,
and in particular, auditors’ perceptions of
what constitutes acceptable performance.
Other evidence suggests that auditor behaviour
can be significantly influenced by the leadership
style of in-charge seniors (Pratt & Jiambalvo,
1981, 1982; Jiambalvo & Pratt, 1982; Kelley %
Margheim, 1990). This study extends this line
of research by examming the effects of audit
managers’ leadership style on the behaviour of
incharge seniors. Given that managers are not
in a position to observe seniors’ fieldwork
directly, and that, as a consequence, seniors
have considerable freedom to conceal undesir-
able behaviour, the manager-senior effect
might be expected to be more pronounced
than the senior-staff effect examined in earlier
studies. Testing of the moderator variable per-
ceived environmental uncertainty (PEU) repre-
sents a further extension of previous work.
PEU is of particular importance in an audit
context, in view of the changing and uncertain
nature of the auditors’ environment. Finally, the
study is the tirst to examine auditor leadership
issues in a non-U.S. audit setting.
The management control literature has fre-
l We gratefully acknowledge the cooperation and support given by the three participating lirms, and helpful suggestions
horn two anonymous reviewers.
405
406 D. T. OTLEY and B. J . PIERCE
quently reported instances of dysfunctional authors. Likert (l%l), for example, referred to
behaviours which have been linked to the con- “job centred” vs “employee centred” manage-
trol system. It is argued that perceptions ment. Blake & Mouton (1964) distinguished
regarding how managers use control informa- between “concern for production” and “con-
tion are important in determining subordinates’ cern for people”, while Fiedler (1967) con-
responses to a control system (Hopwood, trasted “directive” and “permissive” styles.
1974a; Emmanuel et al ., 1990). It has also Although some researchers (e.g. Apostolou et
been suggested (Argyris, 1952; Hopwood, al ., 1993) have advocated the use of more spe-
1974b) that the manner in which a manager ciflc categories of leadership behaviour, YukI
uses control information is related directly to (1989) concluded, following an extensive
the general leadership style of that manager. review of the literature:
Although leadership style has received con-
siderable attention in the industrial psychology
literature (Bass, 1990) it has been largely
ignored in the management control and
accounting literature, with the exception of a
small number of studies in the U.S.A. From the
limited evidence available, it appears that the
leadership style of managers may be an impor-
tant variable in developing a better understand-
ing of the inthtence of formal control systems
over individual behaviour. Given the serious
nature of auditor dysfunctional behaviour
identified in earlier studies, and the fact that
audit manager leadership style is a variable
over which management may be able to exer-
cise considerable control through appropriate
selection and training policies, a study of the
influence of managers’ leadership styles on
dysfunctional behaviours by audit seniors,
together with the moderating effects of situa-
tional variables, is likely to be beneficial in
helping to reduce the incidence of those
behaviours.
Except for taskariented and relationship-oriented
behaviour, there has been little agreement about what
categories of behaviour are meaningM.
Much leadership research has focused on
these categories of leader behaviour through
the two dimensions of consideration and ini-
tiating structure. These dimensions have been
defined as follows (Fleishman & Peters, 1962):
Structure reflects the extent to which an individual is
likely to define and sttucture his own role and those of
his subordinates towards goal attainment. A high score
on this dimension characterizes individuals, who play a
more active role in direct@ group activities through
pluming, communicating information, scheduling,
trying out new ideas, etc.
Consideration reflects the extent to which an individual
is likely to have job relationships chanctcriz.4 by
mutual trust, respect for subordi.aaWs ideas, and con-
sideration of their feelings. A high score is indicative of
a climate of good rapport and txv*way communication.
A low score indicates the supervisor is likely to be more
impersonal in his relations with group members.
LEADERSHIP
Leadership acts occur when one individual.
whether or not that person is in a formally
designated “leadership position”, attempts to
influence the behaviour of others towards
some goal (Fleishman & Peters, 1962). The
range of possible leadership styles can be
viewed as a continuum, extending from a
highly structured, autocratic approach to a
people-oriented, considerate style. These
extremes have been identitled by a number of
Early research has shown that leadership
behaviour can influence subordinate perfor-
mance. For example, Lewin et al (1939) con-
ducted a series of experiments involving
different leadership styles and concluded that
there was a direct relationship between leader-
ship style, group atmosphere and individual
behaviour. More recent work provides further
evidence that leadership behaviour influences
subordinate satisfiction, motivation and per-
formance (Preston & Helntx, 1949; Katz &
Kahn, 1952; Stogdill, 1974; Barrow, 1977).
Whereas the consideration dimension has
THE CONTROL PROBLEM IN PUBLIC ACCOUNTING FIRMS 407
consistently been found to have a beneficial
effect on subordinate satisfaction levels, other
Gmiings have been less consistent, and suggest
that the impact of leadership style may be heav-
lly dependent on situational variables (Yukl,
1989; Apostolou et al ., 1993).
Contingency-based approaches to leadership
research have therefore increased ln popularity
(Lee & Lawrence, 1985), based on the belief
that different leadership styles may be required
as circumstances change, and that no single
style is best in all situations. Adalr (1983) iden-
tified three separate, and sometimes confllct-
ing, sets of requirements ln relation to
effective leadership: (i) task needs; (ii) group
needs; and (lil) indlvldual needs. Effective lea-
dership requires a balancing of these requlre-
ments, and recognition of the inter-
dependencies which exist between them.
Furthermore, the relative attention required by
each set of needs to achieve this balance will
vary depending on the circumstances.
In the context of an audit setting, leadership
is exercised at many different levels, reflecting
the hlerarchlcal structure of audit firms. Senior
management is comprised of partners who pro-
vide overall direction and leadership, and who
maintaln particularly close working relation-
ships with the next lower level of management
(usually entitled managers). Managers provide
guidance and leadership on an overall basis to a
group of audit staff, while maintaining a close
working relationship with audit seniors, who
supervise fieldwork and provide leadership
for the audit team. As members of middle man-
agement, therefore, audit managers are in a
position to be influenced by the leadership
style of top management, and to influence a
number of subordinates through their own
leadership style. Unlike many middle manage-
ment positions, however, they are not in a
position to inspect quallty or monitor sub
ordinates’ behavlour, and as a result may be
in a weak position to prevent or even detect
dysfunctional behavlour which may be related
to leadership style.
PRIOR LEADERSHIP RESEARCH
A number of early studies examined leader-
ship issues ln the context of accounting control
systems (e.g. DeCoster & Fertakis, 1968;
Hopwood, 1974b; Ansari, 1976). Of these, the
Endings of Hopwood (1974b) are particularly
relevant to the current study. Following on
from his earlier work (Hopwood, 1972) which
showed that the manner in which managers
use accounting information for control pur-
poses (i.e. evaluation style) can influence the
response of individuals to the control system,
Hopwood sought to establish whether evalua-
tion style could be linked to the manager’s
general style of leadership. The study pro-
vided evidence that perceptions of a budget-
constrained style of evaluation (a rigid form
of evaluation which he found to be associated
with undesirable consequences such as high
levels of job-related stress and tension, poor rela-
tionships and manipulation) were more likely to
arise when a manager’s style was characterized
by high structure and low consideration.
Brownell (1983) examined the impact of lea-
dership style ln the context of a budgetary con-
trol system ln a large manufacturing company.
He found that the interaction between struc-
ture and consideration had a significant effect
on performance, and also reported a significant
positive association between consideration and
job satisfaction. Evidence of a significant lnter-
action between consideration and budgetary
participation affecting performance, and of an
interaction between structure and budgetary
participation affecting job satisfaction was
also observed. Finally, evidence of a three-
way interaction between consideration, struc-
ture and budgetary participation affecting job
satisfaction was also present.
In the context of audit firms, there is now
considerable evidence of high levels of dysfunc-
tional behaviour by auditors (CAR Report, 1978;
Alderman & Deitrlck, 1982; Kelley & Seiler,
1982; Iightner et al ., 1983; Kelley & Margheim,
1987; Cook & Kelley, 1988; Raghunathan,
1991). However, relatively little attention has
408 D. T. OTLBY and B. J. PIERCE
been given to the study of leadership issues in
audit firms.
Recognizing the importance of formal per-
formance evaluations in an auditing envlron-
ment, and findings reported ln the psychology
literature &lores, 1966) that these evaluations
can be significantly influenced by leadership
style (defined ln terms of structure and consid-
eration), Kida (1984) investigated the perform-
ance evaluation process in five international
CPA firms. The study concluded that leader-
ship style affected the relative emphasis placed
on specific dimensions of performance, and
furthermore, that seniors tended to emphasize
the same dimensions as managers. Consistent
with findings in the psychology literature, it
was also found that highly considerate man-
agers tended to give higher overall ratings,
although the difference was not significant.
Pratt & Jiambalvo (1981) concluded that
audit team performance was related to the
leadership behaviour of lncharge auditors.
They reported evidence that audit team per-
formance was positively related to the level of
consideration shown by the supervisor, and
that elements of structuring behavlour (in par-
ticular the degree of emphasis on meeting
budgets) also varied positively with audit team
performance. Considerate leader behaviour
was also found to be positively related to inter-
personal relationships, satisfaction and motiva-
tion. However, emphasis on budgets was found
either to have a negative relationship or no
significant relationship with these variables.
The researchers concludes that the impact of
this variable was somewhat complex and
unclear.
The importance of leadership behaviour in
an audit environment was again recognized
by Pratt & Jlambalvo (1982) who investigated
the determinants of leadership style. This study
provides evidence supporting the hypothe-
sized relationship between the leadership style
of in-charge audit seniors and two situational
variables: (i) the match between task complex-
ity and audit assistant experience and (ii) audit
assistant’s intolerance of ambiguity; and one
leader-person variable: attraction to the leader-
ship role. The researchers concluded, how-
ever, that a more complex model may be
required to explain leadership behaviour
adequately, involving additional independent
variables and recognizing interactions between
some of the independent variables.
Jiambalvo & Pratt (1982) used path-goal the-
ory to analyse the effect of in-charge leader
behaviour on assistant satisfaction and motiva-
tion. They found that considerate behaviour by
lncharge seniors had beneficial effects, and
that task complexity moderates the impact of
leadership behaviour on staff motivation and
satisfaction. It was also found that the effect
of structuring behaviour as an independent
variable was weak. However, a significant inter-
action effect between consideration and struc-
ture was found, leading to the conclusion that
unless structuring behaviour is accompanied
by considerate behaviour, the result will be
low levels of satisfaction. A general conclusion
from these studies is that whereas considerate
leader behaviour tends to be associated with
desirable outcomes, the consequences of struc-
turing behaviour are less predictable. It appears
that high levels of structuring behaviour may
be beneficial in certain circumstances (such
as clarifying the subordinate’s role when the
task is complex or not well defined), and that
a high score on both dimensions may be
indicative of effective leadership behaviour
(Hopwood, 1974b).
Kelley & Margheim (1990) presented evi-
dence that the average number of audit quality
reduction acts declined as a supervisor’s style
was perceived to reflect high levels of both
consideration and structure. This appears to
support the conclusion that high levels of both
dimensions may constitute effective leadership.
However, the researchers noted that as struc-
ture increased, under-reporting of time also
increased, suggesting that higher levels of
structuring behavlour may have resulted in
tighter control and less opportunity for audit
quality reduction behaviour, this forcing audi-
tors to engage ln a less detectable form of dys-
functional behaviour. One implication is that
high levels of structuring behaviour may help
THE CONTROL PROBLEM IN PUBLIC ACCOUNTING FIRMS 409
perpetuate excessively tight budgets which
may, in the long term, give rise to highly un-
desirable behaviour, such as premature sign-off.
Research into the causes of audit senior turn-
over has not focused specifically on the dimen-
sions of structure and consideration. However,
a number of variables which have been identi-
fied as perceived causes of turnover (Rhode et
al., 1977) can be related directly to those
dimensions, for example “not knowing what
your supervisors expected of you”, “having
some understanding people around you . . .
upon whom you can rely for assistance and
advice”, “feeling that most of the people in
the tirm really did not care about you or your
career with the firm”. These findings provide
further conformation of the importance of a
manager’s leadership style in influencing audit
senior behaviour.
HYPOTHESES
This study focused on the relationship
between leadership style of audit managers
and dysfunctional behaviour by audit seniors.
The measurement of dysfunctional behaviour
was based on previous studies, and included
(a) under-reporting of time (i.e. the deliberate
under-statement of chargeable time, usually by
working on personal time, thereby improving
the impression created by budget versus actual
performance reports and perpetuating art&
cially low budgets which may eventually lead
to quality reduction) and (b) a number of audit
quality reduction acts, including premature
signoff.
It was expected that in an audit environ-
ment, a structured style of leadership by man-
agers would be associated with perceptions of
tight control and inflexible use of time budgets.
Accordingly, budget difficulties are less likely
to be reported, but will tend to be resolved at
individual auditor level. When this occurs, the
likelihood of dysfunctional behaviour
increases. A more considerate style of leader-
ship, on the other hand, tends to be associated
with mutual trust and respect, helpfulness and
friendliness. In this environment, budget diffi-
culties are more likely to be reported and
resolved through open discussion without fear
of adverse consequences and, accordingly, are
less likely to be associated with dysfunctional
behaviour.
Prior accounting research (e.g. Hopwood,
1974b; Jiambalvo & Pratt, 1982; Brownell,
1983) has suggested a need to focus on the
combined effects of structure and considera-
tion. It is therefore hypothesized that the high-
est levels of dysfunctional behaviour will be
reported by seniors who perceive the leader-
ship style of their immediate supervisor to be
characterized by a combination of high struc-
ture and low consideration. The lowest levels
are expected to be found when the style
reflects a combination of low structure-high
consideration, and moderate levels are expected
when the style is either high structure-high
consideration or low structure-low considera-
tion (see Fig. 1). Although these general rela-
tionships have been supported by research
evidence from other work settings, they have
not been previously tested in the context of
the manager-senior relationship in an audit
environment.
The hypothesized relationships between
leadership style and dysfunctional behaviour
are as follows:
Hl When audit manager leadership style is perceived
to be characterized by high structure and low con-
sideration, audit seniors will report high levels of audit
quality reduction bebaviour and high levels of under-
reporting of time.
H2. When audit manager leadership style is perceived
to be characterized by either high structure and high
consideration or low smtctme and low consideration,
audit seniors will report moderate levels of audit quality
reduction behaviour and moderate levels of under-
reporting of time.
H3. When audit manager leadership style is perceived
to be characterized by low structure and high con-
sideration, audit seniors will report low levels of audit
quality reduction behaviour and low levels of under-
reporting of time.
410
D. T. OTL.EY and B. J. PIERCE
LOW
CONSIDERATION
High
INITIATION OF STRUCTURE
High LOW
Hypothais I
High levels of
Dysfunctional Behaviour
(Croup A)
Hypothesis 2
Moderate levels of
Dysfunctional Behaviour
(Croup W
Hypothesis 2 Hypothesis 3
Moderate levels of Low levels of
Dysfunctional Behaviour Dysfunctional Behaviour
(Group Cl (Group D)
Fig. 1. Hypothesized relationships between leadership style and dysfunctional behaviour.
Contingency-based approaches to the study
of leadership have sought to establish which
leadership style ls most effective in circum-
stances. Michaelson (1973) stated:
The ptimary question of research in leadership has
shifted from “what is the best kind of kadership” to
“what kind of leadership works best in what kind of
situation”.
This approach is reflected in the fact that
many variables have been examined in pre-
vious studies of leadership style, relating to
the organization, the work group and the task
(Bass, 1990). The specific variable of interest in
the present study was perceived environmental
uncertainty (PEU), which appears to be of par-
ticular relevance in an auditing environment.
PEU refers to a state in which an individual
makes decisions and engages in behaviours
without complete knowledge of his relation-
ship with his environment. Its particular rele-
vance to an audit senior arlses from the
presence of many unpredictable, volatile and
sometimes conflicting variables in his/her
work environment, such as clients, outside
agencies, supervisors, other departments in
the firm, the audit team, professiorud standards
and ethics, and the firm’s own rules and pro-
cedures. PEU has been linked to role conflict
and role ambiguity in other settings (Van Sell
et al., 1981) and has been found to be signifi-
cantly related to role conflict, role ambiguity,
job satisfaction and job performance in an audit
setting (Rebele & Michaels, 1990).
Findings regarding the influence of task
uncertainty on the structure-performance rela-
tionship have been mixed (Bass, 1990).
Whereas, in situations of high uncertainty,
high levels of structure would be expected to
have a role-clarifying effect and lead to higher
satisfaction and productivity, this has not
always been supported by research tindings
(e.g. Stinson & Johnson, 1975). A possible
reason for this is that, although a structured
style of leadership may clarify certain aspects
of the task, it may, in some circumstances, fall
to resolve major ambiguities, and may even
exacerbate existing problems. A highly struc-
tured style seems particularly likely to lead to
increased ambiguity and uncertainty in an audit
context, where previous research has shown
that although auditors are usually clear about
how to complete audit programmes, no clear
guidelines are given to staff to help deal with
the conthcts which can arise from being
expected to complete a highly structured audit
programme within the limits of an equaIIy well-
structured time budget (McNair, 1991). In this
situation, a high level of structuring behaviour
by audit managers may combine with a high
THE CONTROL PROBLEM IN PUBLIC ACCOUNTING FIRMS
level of PEU, which in itself helps to maintain
or even increase, to create conditions likely
to result in dysfunctional behaviour by indivi-
dual auditors. Accordingly, it is hypothesized
that PEU will moderate the relationship
between initiating structure and dysfunctional
behaviour, as follows:
H4. As the level of PBU increases, the positive relation-
ship between initiating structure and (i) audit quality
reduction behaviour and (ii) under-reporting of time
will become stronger.
In the situation of an audit assignment, a
highly considerate audit manager, displaying
qualities of friendliness and approachability, is
more likely to encourage audit staff to resolve
major dilemmas (McNair, 1991) in a functional
manner, by making the undiscussabIe discuss-
able, than a manager displaying a less consider-
ate style. The effect of this is likely to be more
pronounced in highly uncertain situations,
when the most acute dilemmas are faced by
auditors. It is therefore hypothesized that
PEU will moderate the relationship between
consideration and dysfunctional behaviour as
follows:
H5. As the level of PEU increases, the negative rela-
tionship between consideration and (i) audit quality
reduction behaviour and (ii) under-reporting of time
will become stronger.
THE STUDY
A questionnaire was sent to all audit seniors
ln three of the Big Six firms in Ireland, by prior
411
agreement with a senior partner. From the 356
questionnaires distributed, for return by post
directly to the researchers, 257 usable
responses were obtained, giving an overall
response rate of 72%. Questionnaires were
anonymous, and respondents were given an
undertaking that the contents of completed
questionnaires would remain confidential to
the researchers, except for the reporting of
aggregate results. Details of variable measure-
ment are presented below, together with
descriptive statistics.
Leadership style
The two dimensions of leader&p style, initi-
ating structure and consideration, were mea-
sured using an instrument adapted for an
audit setting by Pratt & Jiambalvo (1981), based
on Stogdill’s (1963) Leader Behaviour Descrip
tion Questionnaire. This measure was chosen
because it has been found elsewhere to show
acceptable levels of reliability (Bass, 1990), and
it has also been found suitable in an audit setting
(Jiambalvo & Pratt, 1982; Kida, 1984). The
instrument requests respondents to indicate
on a five-point scale the frequency of each of
20 listed behaviours by their manager(s). Ten of
the items reflect structuring behaviour, while
the remainder are indicative of a considerate
style. Summary statistics are shown in Table 1.
Audit quality reduction bebaviour
(AQRN
The most serious form of quality reduction
behaviour identilied in previous studies was
premature sign-off, which occurs when an
auditor signs off on a required audit step with-
TABLE 1. -hip style, AQRB and PEU - summary statistics
Theoretical range Actual range Mean S.D. Cronbach alpha
Stiucturr 10-M 16-50 35.15 7.08 0.08
Consideration 10-50 14-50 32.82 8.94 0.93
Audit *tY
reduction behaviour
(AQ-1 6-30 6-27 12.57 4.92 0.89
Perceived
envimnmental
uncertainty (PEU) 10-50 15-44 27.27 5.72 0.85
412 D. T. OTLEY and B. J. PIERCE
out completing the work or noting the omis
sion (CAR Report, 1978). The incidence of
this type of behaviour, together with other
specific forms of quality reduction behaviour
was measured by adding responses to the six
items in Questions 1 and 2 (shown in the
Appendix), which are based on the work of
Alderman & De&rick (1982) and Kelley & Mar-
gheim (1987). Although the possibility of social
desirability bias can never be completely ruled
out when collecting sensitive data of this nat-
ure, measures were taken to minimize the like-
lihood of such bias occurring (e.g. anonymity,
confidentiality, mailing of responses direct to
researchers). To the extent that such bias
does exist, reported levels of dysfunctional
behaviour are likely to be understated, and
thus represent conservative estimates of the
incidence of such behaviours.
Results revealed a high incidence of AQRB,
with 89% of respondents admitting to engaging
in at least one form of audit quality reduction
behaviour. Summary statistics for AQRB are
shown in Table 1.
Under-reporting of time CURTI
The frequency of URT was measured using
responses to the question “How often do you
under-report time? ” (scale 1-5; mean 2.87;
S.D. 1.42). The number of hours which seniors
estimated they under-reported was, on average,
12.2% of hours accurately recorded. In general,
these linclmgs suggest a greater incidence of
URT than that reported in the U.S.A. (Kelley
& Margheim, 1990).
Perceived environmental uncertafnty
(PEW
Measurement of this variable was based on
an instrument developed by Rebele & Michaels
(1990) form earlier work by Ferris (1977) and
Duncan (1972) (see Appendix, Question 3).
The measure seemed particularly suitable for
an audit setting, and achieved favourable psy-
chometric evaluations in previous studies. Sum-
mary statistics are shown in Table 1.
RESULTS
Respondents were classified into four groups
according to the perceived leadership style of
their manager(s), based on the median scores
for (i) structure and (ii) consideration. High
structure was defined as greater than or equal
to the median score of 35, and high considera-
tion was a score greater than or equal to the
median score of 32. This procedure was
selected, partly to cope with problems of multi-
collinearity, but also to aid interpretation by
allowing mean values for cells to be com-
puted. Table 2 shows descriptive statistics for
each group and in total, in relation to (i) audit
quality reduction behaviour and (ii) under-
reporting of time.
The most noteworthy feature of this analysis
is that a consistent pattern emerged in relation
to both forms of dysfunctional behaviour. In
both cases, the highest frequency was
reported by the high structure/low considera-
tion group (A). The second highest frequency
TABLE 2. Mean and S.D. of dy&mctionaI bebaviours for four leadership groups and in totaI
Variable
AQRB (scale 6-30) Under-reporting of time
(scale l-5)
Group structure Consideration n Mean S.D. Mean SD.
A me LOW 51 17.84 5.35 4.27 1.13
B LOW 63 13.05 3.72 2.83 1.18
C Ii@ 79 11.04 3.90 2.54 1.24
D LOW 64 9.94 3.21 2.20 1.31
Total 257 12.57 4.92 2.87 1.42
THE CONTROL PROBLEM IN PUBLIC ACCOUNTING FIRMS
413
TABLE 3. Ranking of leadership groups by dysfunctional
behaviour (rank 1 = highest dysfunctionaI behaviour)
Rank Group Structure Consideration
1 A
H&h
LOW
2 B LOW LOW
3 C
Hkh
High
4 D LOW High
was reported by the low structure/low con-
sideration group (B). This was followed by
the high structure/high consideration group
(C), and the lowest frequency for the specified
dysfunctional behaviours was reported by the
low structure/high consideration group (D).
The ranking of the four leadership groups in
terms of frequency of dysfunctional behaviours
is shown in Table 3. Although as explained in
the Discussion section, these findings were
expected (except that no prior hypothesis
compared group B with group C), such a
high level of consistency was surprising.
One-way ANOVA tests were then carried out
to establish which of the differences between
groups were statistically significant. The
results, which are summarized in Table 4,
showed that there were sign&ant differences
in the incidence of both types of behaviour
across the four leadership groups. Tukey ana-
lysis was used to identify which groups
differed, and the result of this is shown on
the right-hand side of Table 4.
A notable feature of the result is that the
frequency of both forms of dysfunctional beha-
viour was signilicantIy higher in the high/low
group (A) than in the two groups where there
was highly considerate manager behaviour (C
and Dj. Group A also reported signilkantly
TABLE 4. One-way ANOVA results for four leadership
groups
ANOVA results Signikant
differences
Varlabie F rati o Prob. between groups
AQm
39.04 0.000 A > D, C, B
B>D,C
Under-repotting
of time
30.40 0.000 A > D, C, B
B>D
higher figures than group B, where low con-
sideration was accompanied by low levels of
structuring behaviour. Furthermore, group B
(low/low) reported signiticantly higher levels
of AQRB than both the high consideration
groups (D and C), and significantly higher
levels of URT than the low/high group (D).
These results underline the importance of con-
siderate behaviour on the part of the audit
manager, and suggest that high levels of con-
siderate behaviour are associated with rela-
tively low levels of dysfunctional behaviour,
even when it is accompanied by high levels
of structuring behaviour.
In order to test hypotheses 4 and 5, the cate-
gories of high and low structure and high and
low consideration described earlier were
retained. In addition, respondents were cate-
gorized into three PEU groups of roughly simi-
lar size: (i) low uncertainty = PEU < 25 (n =
78; mean PELJ = 21.11; SD. = 2.55); (ii) med-
ium uncertainty = PEU 25-28 (n = 87; mean
PEU = 26.32; S.D. = 1.06); and (iii) high uncer-
tainty = PEU > 28 (n = 92; mean PEU = 33.27;
S.D. = 3.99). The data were analysed using a 2
X 2 X 3 factorial ANOVA analysis for each of
the two dependent variables, the results of
which are shown in Tables 5 (AQRB) and 7
(URT).
With regard to AQRB, Table 5 provides evi-
dence of a significant main effect for three
variables. Cell means indicate that this effect
is positive for structure and PEU and negative
for consideration. The findings support an
interactive effect between structure and PEU
and between consideration and PEU. Inspec-
tion of cell means (Table 6) indicates that in
general, as structure (consideration) increases,
AQRB increases (decreases), and that as the
level of PEU increases, this effect grows
stronger. This supports hypotheses 4 and 5,
as applied to AQRB. An exception arises in
relation to the high structure/low PEU group.
This is the only situation where high structure
is associated with a fall in AQRB, and, in accor-
dance with House’s (1971) path-goal theory of
leadership, suggests that if structuring beha-
viour is effective in reducing uncertainty,
414 D. T. OTLEY and B. J. PIERCE
TABLES. Resultsofa X 2 X 3ANOVAanaWslsforAQRB
Source of variation sum of squares d.f. Mean square P sig. of P
Main Effects
smcture
Consideration
PEU
Tweway ilwractions
Structure X consideration
structufe x PEU
Consideration X PEU
Thee-way interactions
Structure X consideration X PEU
Explained 3360.58 11
Residual 2820.89 242
464.49 1 464.49 39.85 0.000
563.09 1 563.09 48.31 0.000
972.57 2 486.28 41.72 0.000
12.59 1 12.59 1.08 0.300
188.4s 2 94.22 8.08 0.000
159.73 2 79.87 6.85 0.001
193.99 2 %.99 8.32 0.000
3OSSl
11.66
26.21 0.000
Total 6181.47 253 24.43
then it may help generate a favourable
response in terms of individual performance.
The implication is that where a high level of
structuring behaviour fails to resolve uncertain-
ties in the work environment (medium and
high PEU groups) the outcome in terms of
undesirable behaviour is considerably worse
than for a low structure leadership style.
The results summarized in Table 5 also indi-
cate that, in addition to two 2-way interactions,
there is also a significant 3-way interaction
effect between the three variables structure,
consideration and PEU. This suggests that the
PEU in the auditor’s environment interacts with
the manager’s leadership style, defined in terms
of structure and consideration, to produce a
significant effect on the level of AQRB
behaviour in which seniors engage. Implica-
TABLE 6. Interaction of PBU with (i) strocture and
(ii) consideration: mean AQRB by cell
PBU
LOW Medium
Low r3tNcN-e 9.29 (38) 11.77 (44) 13.04 (45)
High stNcture 8.97 (38) 12.9’3 (42) 18.38 (47)
Low consideration 10.86 (22) 12.71 (34) 17.27 (67)
Higb consideration 8.43 (54) 12.12 (52) 11.76 (25)
Note: numbers in parentheses indicate cell &es.
tions of PEU in the operating environment of
audit seniors are considered in the Discussion
section.
The analysis in relation to URT (Table 7)
provided evidence of a significant main effect
for structure and PEU (both positive) and con-
sideration (negative), similar to that found for
AQRB. However, of the two hypothesized
interaction effects, only structure X PEU was
statistically significant at the 0.05 level, and an
inspection of cell means showed this effect to
be in the direction set out in hypothesis 4.
Neither the consideration X PEW interaction
(hypothesis 5) nor the 3-way interaction
between structure, consideration and PEU
(not hypothesized but present in the AQRB
analysis) were significant for URT. A noticeable
feature of the results, however, was the signifi-
cance of the interactive term structure X con-
sideration. Tbe pattern of URT behaviour
associated with different combinations of
structure and consideration was set out earlier
(see Table 2), indicating the importance of
consideration if high levels of structure are
used.
The combined effect of these results is that
hypothesis 4 was supported, but whereas hypo-
thesis 5 was supported for AQRB, the evidence
did not support that hypothesis in relation to
Source of varhthm
THE CONTROL PROBLEM IN PUBLIC ACCOUNTING FIRMS
415
TABLE7.Rcsultsofa2X2X3~OVAvllhrsisforURT
sum of SaIlares d.f. Mean salare P sig. of F
Maine&cts
s- 38.15 1 38.15 31.29 0.000
CAXlSidC~tiOll 38.05 1 38.05 31.21 0.000
PEU 64.16 2 32.08 26.31 0.000
Twmvay intexactions
Structure X consideration 5.28 1 5.28 4.33 0.038
strum x PEU 7.40 2 3.70 3.03 0.050
consideration x PBU 5.70 2 2.85 2.34 0.099
Three-way lnuractions
Wwture X consideration X
PEU
4.71 2 2.35 1.93 0.147
Bxplaiaed 215.96 11 19.63 16.10 0.000
Residual 295.05 242 1.22
TotAl 511.01 253
2.02
URT. These results, taken together with the
finding of a 3-way interaction for AQRB but
not for URT and a slgnilicant structure X
consideration interaction effect for URT but
not AQRE suggest an overall conclusion that,
whereas PEU exercises a very signifkant
moderating effect on the leadership style-
AQRB relationship, the moderating effect of
PEU on the leadership style-URT relationship
is noticeably weaker.
DISCUSSION
The most important finding is the role of
considerate lea&r behaviour. High considera-
tion is associated with relatively low levels of
dysfunctional behaviour, regardless of struc-
turing behaviour (see Table 2). This result is
consistent across all forms of dysfuntitional
behaviour included in the study. Between the
extremes of groups A and D, groups B and C
ranked second and third, respectively. This
highhghts the importance of considerate be-
haviour in the fact that group C @gh/h@h)
reported signlfIcantly lower levels of quality
reduction behaviour than group B (low/law).
These findings suggest that audit fnms might
successfully tackle the problem of dysfunc-
tional behaviour by directing recruitment, pro
motion and training policies towards the
development of a high consideration/low struc-
ture style of leadership among managers and
partners. However, as previous research has
shown that no single style of leadership may
be ideal in all circumstances, the analysis was
extended to examine possible moderating
effects of the variable perceived environmental
uncertainty.
Analysis of auditor behaviour under different
conditions of uncertainty showed that in a rela-
tively certain environment, the level of AQRB
decreased as leadership style became more
structured Gable 6). This may be because
highly structured manager behaviour has suc-
ceeded in chuifying for audit seniors their
paths to achieving rewards, and communicat-
ing the value of those rewards, and as a result,
individual performance has become more
effective. It is also possible, however, that in
a relatively certain environment with a highly
structured style of management, the scope for
undetected AQRB is much reduced and audi-
tors turn to another form of dysfunctional be-
haviour, such as URT, to resolve conflicts and
dilemmas. The finding that URT increased as
416 D. T. OTLBY and B. J. PIERCE
leadership style became more structured in a that environment will moderate the effects of
relatively certain environment suggests this is the leadership dimensions structure (positive)
a more likely explanation. and consideration (negative) on AQRR.
The moderating effect of perceived environ-
mental uncertainty was noticeably stronger for
the dependent variable AQRR than for URT.
The URT analysis showed no evidence of a
moderating effect on the consideration-URT
relationship, and although there was some
evidence of a moderating effect on the struc-
ture-URT relationship, it barely achieved signi-
ficance at the 0.05 level. Furthermore, there
was evidence of a significant interaction effect
between structure and consideration for the
dependent variable URT, but not for AQRR.
These lindings reflect the fact that, although
the two dependent variables are similar in
some respects (Pearson correlation coefficient
= 0.56), there are also important differences.
Although most auditors are aware that URT is
forbidden by their firms, it is not a form of
behaviour that results in a direct and immedi-
ate threat to audit quality (although it may do
so at a future date by causing auditors to com-
promise quality in order to meet artificially low
budgets). Many respondents indicated that
audit managers sometimes encourage them to
engage in URT in order to avoid budget over-
runs. An auditor’s perception of the manager’s
attitude towards budget over-runs may there-
fore be the dominant consideration in making
the URT decision. AQRB, on the other hand,
is a potentlally very serious form of mls-
behaviour, involving a deliberate reduction in
quality through falsification of audit records.
Although it does not seem to be openly dis-
cussed or formally communicated, the penalty
for engaging in such behaviour would be
expected to be severe, and to carry major
implications for future career prospects. It is
likely that a decision to engage in such be-
havlour will not be taken lightly and will
involve an appraisal, not only of the manager’s
attitudes to various possible budget outcomes,
but also of the broader environment and the
possible ramilications of such behaviour.
Having formed an assessment of his environ-
ment, the perceived level of uncertainty in
An alternative possible explanation lies in
the essential difference between the two cate-
gories of behaviour, AQRR and URT. The find-
ings suggest that when a leadership style which
is associated with dysfunctional responses (i.e.
high structure-low consideration) is combined
with a highly uncertain environment, the indi-
vidual auditor response tends to reflect an
increase in AQRR rather than URT. This trend
underlines the high level of interdependence
between the two categories of behaviour, by
suggesting a shift towards quality reduction
behaviour in conditions of high uncertainty.
In these circumstances, it may be very difhcult
for the individual auditor to predict in advance
how many hours would have to be under-
reported in order to complete the job within
budget. There is therefore uncertainty regard-
ing the likely success of URT as a strategy for
meeting tight budgets and strategies involving
quality reduction become more attractive.
With regard to leadership style, the evidence
supports the view that a considerate style tends
to result in lower levels of dysfunctional beha-
viour. One interpretation of this is that a man-
ager displaying high consideration tends to be
more open and approachable when uncertain-
ties arise. This style of leadership is therefore
likely to result in lower levels of perceived
environmental uncertainty. The evidence
supports this, as the high consideration group
reported significantly lower levels of PEU than
the low consideration group (t = 7.53; prob.
< 0.001). These arguments are generally in line
with path-goal theory (House, 1971; House &
Mitchell, 1974), which holds that performance
and motivation will improve if paths leading to
valued rewards are made clear.
This suggests that the uncertainties in an
audit senior’s work environment are so diverse
and dynamic that a highly considerate style is
more helpful than a structured style ln inter-
preting goals and objectives. The auditor is
exposed to boundary-spanning activities to a
far greater degree than many other profes
THE CONTROL PROBLEM IN PUBLIC ACCOUNTING FIRMS 417
sionals (Rebele & Michaels, 1990) giving rise
to additional layers of uncertainty which may
be unaffected or even increased by a highly
structured leadership style. Such a style may,
however, offer little or no positive contribu-
tion in terms of role clarification, since seniors
are generally highly trained professional audi-
tors who are engaged ln relatively routine work
and are guided by a detailed audit programme.
A considerate style, on the other hand, helps
promote a supportive work environment in
which the varied sources of audit senior
uncertainty can more effectively be addressed.
CONCLUSION
The study has some important limitations
which may restrict the inferences to be drawn
from the findings. Only three of the Big Six
firms were included, and the results are not
therefore necessarily applicable to other
lirms. However, the study achieved broad
coverage, since a high response was obtained
from the sample of all audit seniors in the par-
ticipating lirms, whose structures and pro-
cedures are known to be broadly similar to
other large audit Iirms. A second limitation is
that a cross-sectional study cannot support con-
clusions regarding causality. Reverse causal
linkages are a distinct possibility, since seniors
who are suspected of engaging in dysfunctional
behaviour may attract a highly structured
response from the manager, whereas seniors
who are not suspected of such behaviour may
attract a more considerate response. It is also
possible that perceptions of leadership style
may vary depending on the senior’s inclination
towards dysfunctional behaviour, where a
senior who is seeking opportunities to conceal
such behaviour may be more conscious of
structuring behaviour by managers.
Despite these limitations, the llndings high-
light the importance of leadership style in an
audit setting, and have important implications
from both theoretical and practical perspec-
tives. Two serious forms of dysfunctional be-
haviour were found to be significantly related
to the two leadership dimensions, considera-
tion (negatively) and structure (positively),
and these relationships were found to be stron-
ger when perceived environmental uncertainty
increased. Although generally consistent flnd-
ings emerged for both types of dysfunctional
behaviour, the moderating effect of perceived
environmental uncertainty was found to be
stronger for AQRB than URT.
The evidence suggests that structured leader-
ship by audit managers often creates ambiguity
and conflict in the minds of individual auditors.
This occurs when an auditor discovers that the
time budget is too tight to allow proper
completion of the audit programme. The audit
programme is a highly structured plan, which
clarifies the tasks to be carried out. It is often
accompanied by high levels of uncertainty,
however, regarding how the tasks can be
completed within the time allowance.
In terms of practical implications, the find-
ings suggest that audit firms have considerable
scope to influence dysfunctional behaviour
through the recruitment, promotion and train-
ing of managers and partners. The study has
highlighted the important role of considerate
and supportive leader behaviour, in providing
an effective framework for resolving uncer-
tainty. Without this, the manager may become
an additional source of uncertainty in the
already complex and uncertain operating
environment of the audit senior.
BIBLIOGRAPHY
Adair, J., h”eelfve Lmadesbip (Aldershot: Gower, 193).
Alderman, C. W. & De&rick, J. W., Auditors’ Perceptions of Time Budge Pressures and Premature Sign-
offs: a Replication and Extension, Audfting: a J oumul of Practfce 6 Tbeo?y (1982) pp. 54-68.
418 D. T. OTLEY and B. J. PIERCE
Ansari, S. L., Behavioral Factors in Variance Controls l&port on a Laboratory Expcrimcn& Journ& of
Accounting Resea*cb (1976) pp. 189-211.
Apostolou, B., Pascwah, W. R & Strawscr, J. R, The Effects of Senior Internal Auditor Bchaviour on Staff
Perfotxnancc and SaMaction, Accounting and Business Research (1993) pp. 110-122.
Argyris, C., l&z Impact of Budgets on Peopk (New York: The Controllership Foundation, 1952).
Barrow, J. C., The Vatiables of hdership: a Beview and Conceptual Framework, Awdemy of
Management Revkw (1977) pp. 231-251.
Bass, B. M., Bass C Stogdilh Handbook of Leadership. fbeory, Research and Manag& Applications
(New York: The Free Press, 1990).
Blake, R R. & Mouton, J. S., IIx? Manage&l Grfd (Houston: Gulf Publishing, 1964).
Brownell, P., Leadership Style, Budgetary Participation and Managerial Behaviour, Accounting,
OrganFzatfons and Society (1983) pp. 307-321.
Commission on Auditors’ Responsibilities (CAR), Report, Conclusions and Recommendutions (New
York AI CPA, 1978).
Cook, E. % K&q, T., Auditor Stress and Time Budgets, 7h CPAJ ournui (1988) pp. 83-86.
JkCostcr, D. & Fertakis, J., Budget-induced Pressure and its Rehtionship to Supcrvhry Bchaviour,
J ournal of Accounting Research (1968) pp. 237-246.
Duncan, R B., chvlctciistics of Organhtional Environments and Perceived Bnvironmental Uncertainty,
Administrative Scknce Quarterly (1972) pp. 313-327.
Emmanuel, C., Otlcy, D. & Merchant, K., Accounting for Management Confrol (London: Chapman and
Hall, 1990).
Ferris, K. R, Perceived Uncertlinty and Job Satisfaction in the Accounting Environment, Accounting,
Organizations and Sock@ (1977) pp. 23-28.
Ficdler, F. E., A lbeory of ih&zsbip Effectfveness (New York: McGraw Hill, 1%7).
Fleishman, E. A. & Petm, D. R, Intcqcrxmal Values, Leadership Attitudes and Managerial “Success”,
Personnel Psychology (1%2) pp. 127-143.
Hopwood, A. G., An Empirical Study of the Role of Accounting Data in Perfomtancc Evaluation, J ournal
of Accounting Research (1972) pp. 156-182 (Suppl.).
Hopwood, A. G., Accounthg and Human Bebaviour (London: Prentice-Hall, 1974a).
Hopwood, A. G., Leadership Climate and the Use of Accounting Data in Performance Evaluation, ZYbe
Accounting Revkw (1974b) pp. 485-495.
House., R J., A Path-goal Theory of Leader Effectiveness, Admfnfstr&ue Science Quatier& (1971) pp.
321-338.
House, R J. % Mitchell, T. R., Path-goal Theory of Lea&tip, J ournal of Contemporary Business (1974)
pp. 81-97.
Jiaatbalvo, J. % Pratt, J., Task Complexity and Leadership Effectiveness in CPA Firms, ZIw Accounting
Reukw (1982) pp. 734-750.
Katz, D. % Kahn, R. L., Some Recent Findings in Human Belations Research in Industry, in S wanson, G. V.,
Newcomb, T. M. % Hardy, E. L. (eds), Readings k Sodal Ps@okgy @J ew Yoik: Holt, 1952).
Kellcy, T. % Marghcim, L., The Effect of Audit Billing Amngcment on Under-reporting of Time and Audit
Quality Bcduction Acts, Aduawes fn Accounting (1987) pp. 221-233.
Kelky, T. % Mar&ehn, L., The Impact of Time Budget F’reswe, Personality and Leadership Variables on
Dysfunctional Behaviour, AudiZing: A J ou& of PTactke 6 lbeory (1990) pp. 21-41.
Kelley, T. % Seller, S. E., Auditor Stress and Time BudBets, lbe CPA J oumal(1982) pp. 24-34.
Kida, T. E., F’erformance Won and Beview Meeting Characteristics in PubBc Accounting Firms,
Accounting, Organtzalions and Society (1984) pp. 137-147.
Klores, M., Bater Bias in ForceddkMb ution Fwformance Ratings, Personnel l?ycbology (1566) pp.
411-421.
Lee, R t Lawrcncc, P., Organfxutional Behtiour: Politfcs at Work (London: Hutchinson, 1985).
Lewin, K., Lippitt, R % White, R. K., Patterns of Aggressive Bchaviour io Fixpcrimentally Created Social
Envlrwmen*, I nternational J ournal of Social I ?sycbology (1939) pp. 271-299.
l&htner, S. M., Leisenriq, J. J. & Winters, A. J., Underreporting Chargeable Time, J ourrud of Accoun-
tancy (1983) pp. 52-57.
Likert, R, New Patterns of Manugement (New York: McGraw Hill, l%l>.
McNair, C. J., Proper Compromixs: the ManaBemat Control Dilemma ill Public Accounting and ita
Impact on Auditor Behavior, Accounting, O*gan&atJonF and So&@ (1991) pp. 635-653.
Michaelson, L. K., Leader Orientation, Leader Behavior, Group Effectiveness and Situational FavorabIBty:
THE CONTROL PROBLEM IN PUBLIC ACCOUNTING FIRMS
419
an EmpIrical Extension of the Contingency Model, Otgunkatkmul Eebavfour and Human Perfii-
muncle (1973) pp. 226-245.
Pratt, J. & Jiambaivo, J., Relationships between Leader Rcbaviors and Audit Team Performance, Aczoun&
ing, Otgantaations and Society (1981) pp. 133-142.
Pratt, J. & Jiambalvo, J., Detcnnimn ts of Leader Behavior in an Audit Rnvironment, Accountfng,
Organtzations and Society (1982) pp. 369-379.
Preston, M. G. & Hcintz, R. K., Rffects of Partkipatory Versus Supervisory LeadershIp on Group
Judgment, J ournal of Abnormd and Social Pycbology (1949) pp. 345-355.
Ra&nathan, B., Premature Signing-off of Audit Procedures: an Analysis, Accounting H0rlzon.r (1991)
pp. 71-79.
Rebele, J. E. & Michaels, R. E., Independent Auditors’ Role Stress: Antecedent, Outcome, and Moderating
Variables, Bebavloral Research fn Accounting (1990) pp. 124-153.
Rhode, J. G., Sorensen, J. E. % Lawler, E. E. III., Sources of ProfessionaI Staff Turnover in Public Account-
ing Fhms ReveaIed by the Exh Interview, Accountfng, Orguntiations andSocfe@ (1977) pp. 165-175.
.%inwn, J. E. % Johnson, T. W., The Path-goal Theory of Leadership: a PartiaI Test and Suggested
Refinement, Academy of Management J oumal(1975) pp. 242-252.
Str@Ul, R. M., Manual for tbe Leader Bebatiour Descrtptfon Questionnaire Form XI I (Columbus:
Bureau of Business Research, The Ohio State University, 1%3).
Stogdill, R. M., Handbook of Leader&p: a Survey of Tbeoy and Researcb (New York: The Free Press,
1974).
Van Sell, M., Brief, A. P. & Schuler, R. S., Role Conikt and Role Ambiguity: Integration of the Literature
and Directions for Future Research, Human Relations (1981) pp. 43-71.
Yukl, G., ManageriaI Leadership: a Review of Theory and Research,]ourn& of Management (1989) pp.
251-289.
APPENDIX
Relevant questions reproduced from tbe Survey Questionnaire
QlBu3tion 1
Please tick one answer on each line.
NA = nearly always 0 = often S = sometimes R = rarely N = never
(Scored)
In your work as auditor, do you ever sign off
a reqired audit step not covered by another
audit step, without completing the work or
noring the omission of procedures?
Question 2
During the past year, how often have you acted
in the following manner when carryirq out an
audit?
(a) accepted weak client explanations
(b) made superficial reviews of client documents
(c) failed to research an accounting principle
(d) reduced the amount of work performed on an
audit step below what you considered
reasonable
NA
5
0
4
s
3
R
2
N
1
0 0 0 0 0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
420 D. T. OTL.EY and B. J. PIERCE
(e) signed off an audit programme step without
completing the work or noting the omission
0 0 0 0 0
Question 3
The following questions relate to the h-equency with which you encounter certain situations in
your work as audit senior.
NA = nearly always 0 = often S = sometimes R = rarely N = never
a. How often are you sure about what approach
is best for dealing with job-related problems?
b. How o&en are you sure about how to deal with
a change in social, economic or technical
conditions outside the firm?
c. When dealing with people in other areas of the
firm, how often do you feel you have enough
detail to know how they want the job done?
d. How often are you sure about the best way to
approach job-related problems that arise in
other departments?
e. How often is it difficult for you to determine if a
job-related decision is a correct one?
f. How often are you in doubt about how to obtain
the necessary information for making job-related
decisions?
g. How often are you able to predict accurately the
outcomes of your job-related decisions?
h. How often are you working on clients that you
have worked on before and with whom you are
familiar?
i. How often are you working on first-time audits,
i.e. clients that are new to your firm?
j. How often are you working on clients that belong
to a particular industry in which you have
specialized and developed particular expertise as
an auditor?
NA
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
S
0
0
0
0
0
0
0
0
0
0
R
0
0
0
0
0
0
0
0
0
0
N
0
0
0
0
0
0
0
0
0
0
doc_172023483.pdf
Prior research has highlighted the problem of dysfunctional reactions to control systems. Studies carried
out in auditing firms have revealed high levels of budget pressure leading to a variety of dysfunctional
behaviours which can be difficult to control. This study e xamines how subordinates’ reaction to control
systems is influenced by the leadership behaviour of supervisors.
Pergamon Accounting, Organizations and Society, Vol. 20, No. 5, PP. 405-420, 1995
Copyright 0 1995 Elsmier Science Ltd
Printed in Great Britain. AU rights reserved
0361-36132/95 $9.50+0.00
THE CONTROL PROBLEM IN PUBLIC ACCOUNTING FIRMS:
AN EMPIRICAL STUDY OF THE IMPACT OF LEADERSHIP STYLE*
DAVID T. OTLEY
University of Lancaster
and
BERNARD J. PIERCE
LWbl i n Ci ty Uni versi ty
Prior research has highlighted the problem of dysfunctional reactions to control systems. Studies carried
out in auditing firms have revealed high levels of budget pressure leading to a variety of dysfunctional
behaviours which can be difficult to control. This study e xamines how subordinates’ reaction to control
systems is influenced by the leadership behaviour of supervisors. Data were collected by means of a
questionnaire sent to all audit seniors in three Big Six audit firms. Two forms of dysfunctional behaviour
were measured: under-reporting of time and audit quality reduction behaviour. A leadership style
characterized by high structure and low consideration was found to be associated with the highest
level of dysfunctional behaviour for all behaviours examined, while the lowest level of dysfunctional
behaviour was associated with a style depicting low structure and high consideration. Perceived
environmental uncertainty was found to moderate these relationships, and to exercise a stronger
moderating effect for audit quality reduction behaviour than for under-reporting of time.
During the 198Os, a number of studies exam-
ined the workings of the formal control system
in large audit firms (e.g. Alderman h Deitrick,
1982; Lightner et al ., 1983; Cook & Kelley,
1988). The conclusion was that there was a
disturbingly high level of dysfunctional be-
haviour associated with the control system,
and in particular, auditors’ perceptions of
what constitutes acceptable performance.
Other evidence suggests that auditor behaviour
can be significantly influenced by the leadership
style of in-charge seniors (Pratt & Jiambalvo,
1981, 1982; Jiambalvo & Pratt, 1982; Kelley %
Margheim, 1990). This study extends this line
of research by examming the effects of audit
managers’ leadership style on the behaviour of
incharge seniors. Given that managers are not
in a position to observe seniors’ fieldwork
directly, and that, as a consequence, seniors
have considerable freedom to conceal undesir-
able behaviour, the manager-senior effect
might be expected to be more pronounced
than the senior-staff effect examined in earlier
studies. Testing of the moderator variable per-
ceived environmental uncertainty (PEU) repre-
sents a further extension of previous work.
PEU is of particular importance in an audit
context, in view of the changing and uncertain
nature of the auditors’ environment. Finally, the
study is the tirst to examine auditor leadership
issues in a non-U.S. audit setting.
The management control literature has fre-
l We gratefully acknowledge the cooperation and support given by the three participating lirms, and helpful suggestions
horn two anonymous reviewers.
405
406 D. T. OTLEY and B. J . PIERCE
quently reported instances of dysfunctional authors. Likert (l%l), for example, referred to
behaviours which have been linked to the con- “job centred” vs “employee centred” manage-
trol system. It is argued that perceptions ment. Blake & Mouton (1964) distinguished
regarding how managers use control informa- between “concern for production” and “con-
tion are important in determining subordinates’ cern for people”, while Fiedler (1967) con-
responses to a control system (Hopwood, trasted “directive” and “permissive” styles.
1974a; Emmanuel et al ., 1990). It has also Although some researchers (e.g. Apostolou et
been suggested (Argyris, 1952; Hopwood, al ., 1993) have advocated the use of more spe-
1974b) that the manner in which a manager ciflc categories of leadership behaviour, YukI
uses control information is related directly to (1989) concluded, following an extensive
the general leadership style of that manager. review of the literature:
Although leadership style has received con-
siderable attention in the industrial psychology
literature (Bass, 1990) it has been largely
ignored in the management control and
accounting literature, with the exception of a
small number of studies in the U.S.A. From the
limited evidence available, it appears that the
leadership style of managers may be an impor-
tant variable in developing a better understand-
ing of the inthtence of formal control systems
over individual behaviour. Given the serious
nature of auditor dysfunctional behaviour
identified in earlier studies, and the fact that
audit manager leadership style is a variable
over which management may be able to exer-
cise considerable control through appropriate
selection and training policies, a study of the
influence of managers’ leadership styles on
dysfunctional behaviours by audit seniors,
together with the moderating effects of situa-
tional variables, is likely to be beneficial in
helping to reduce the incidence of those
behaviours.
Except for taskariented and relationship-oriented
behaviour, there has been little agreement about what
categories of behaviour are meaningM.
Much leadership research has focused on
these categories of leader behaviour through
the two dimensions of consideration and ini-
tiating structure. These dimensions have been
defined as follows (Fleishman & Peters, 1962):
Structure reflects the extent to which an individual is
likely to define and sttucture his own role and those of
his subordinates towards goal attainment. A high score
on this dimension characterizes individuals, who play a
more active role in direct@ group activities through
pluming, communicating information, scheduling,
trying out new ideas, etc.
Consideration reflects the extent to which an individual
is likely to have job relationships chanctcriz.4 by
mutual trust, respect for subordi.aaWs ideas, and con-
sideration of their feelings. A high score is indicative of
a climate of good rapport and txv*way communication.
A low score indicates the supervisor is likely to be more
impersonal in his relations with group members.
LEADERSHIP
Leadership acts occur when one individual.
whether or not that person is in a formally
designated “leadership position”, attempts to
influence the behaviour of others towards
some goal (Fleishman & Peters, 1962). The
range of possible leadership styles can be
viewed as a continuum, extending from a
highly structured, autocratic approach to a
people-oriented, considerate style. These
extremes have been identitled by a number of
Early research has shown that leadership
behaviour can influence subordinate perfor-
mance. For example, Lewin et al (1939) con-
ducted a series of experiments involving
different leadership styles and concluded that
there was a direct relationship between leader-
ship style, group atmosphere and individual
behaviour. More recent work provides further
evidence that leadership behaviour influences
subordinate satisfiction, motivation and per-
formance (Preston & Helntx, 1949; Katz &
Kahn, 1952; Stogdill, 1974; Barrow, 1977).
Whereas the consideration dimension has
THE CONTROL PROBLEM IN PUBLIC ACCOUNTING FIRMS 407
consistently been found to have a beneficial
effect on subordinate satisfaction levels, other
Gmiings have been less consistent, and suggest
that the impact of leadership style may be heav-
lly dependent on situational variables (Yukl,
1989; Apostolou et al ., 1993).
Contingency-based approaches to leadership
research have therefore increased ln popularity
(Lee & Lawrence, 1985), based on the belief
that different leadership styles may be required
as circumstances change, and that no single
style is best in all situations. Adalr (1983) iden-
tified three separate, and sometimes confllct-
ing, sets of requirements ln relation to
effective leadership: (i) task needs; (ii) group
needs; and (lil) indlvldual needs. Effective lea-
dership requires a balancing of these requlre-
ments, and recognition of the inter-
dependencies which exist between them.
Furthermore, the relative attention required by
each set of needs to achieve this balance will
vary depending on the circumstances.
In the context of an audit setting, leadership
is exercised at many different levels, reflecting
the hlerarchlcal structure of audit firms. Senior
management is comprised of partners who pro-
vide overall direction and leadership, and who
maintaln particularly close working relation-
ships with the next lower level of management
(usually entitled managers). Managers provide
guidance and leadership on an overall basis to a
group of audit staff, while maintaining a close
working relationship with audit seniors, who
supervise fieldwork and provide leadership
for the audit team. As members of middle man-
agement, therefore, audit managers are in a
position to be influenced by the leadership
style of top management, and to influence a
number of subordinates through their own
leadership style. Unlike many middle manage-
ment positions, however, they are not in a
position to inspect quallty or monitor sub
ordinates’ behavlour, and as a result may be
in a weak position to prevent or even detect
dysfunctional behavlour which may be related
to leadership style.
PRIOR LEADERSHIP RESEARCH
A number of early studies examined leader-
ship issues ln the context of accounting control
systems (e.g. DeCoster & Fertakis, 1968;
Hopwood, 1974b; Ansari, 1976). Of these, the
Endings of Hopwood (1974b) are particularly
relevant to the current study. Following on
from his earlier work (Hopwood, 1972) which
showed that the manner in which managers
use accounting information for control pur-
poses (i.e. evaluation style) can influence the
response of individuals to the control system,
Hopwood sought to establish whether evalua-
tion style could be linked to the manager’s
general style of leadership. The study pro-
vided evidence that perceptions of a budget-
constrained style of evaluation (a rigid form
of evaluation which he found to be associated
with undesirable consequences such as high
levels of job-related stress and tension, poor rela-
tionships and manipulation) were more likely to
arise when a manager’s style was characterized
by high structure and low consideration.
Brownell (1983) examined the impact of lea-
dership style ln the context of a budgetary con-
trol system ln a large manufacturing company.
He found that the interaction between struc-
ture and consideration had a significant effect
on performance, and also reported a significant
positive association between consideration and
job satisfaction. Evidence of a significant lnter-
action between consideration and budgetary
participation affecting performance, and of an
interaction between structure and budgetary
participation affecting job satisfaction was
also observed. Finally, evidence of a three-
way interaction between consideration, struc-
ture and budgetary participation affecting job
satisfaction was also present.
In the context of audit firms, there is now
considerable evidence of high levels of dysfunc-
tional behaviour by auditors (CAR Report, 1978;
Alderman & Deitrlck, 1982; Kelley & Seiler,
1982; Iightner et al ., 1983; Kelley & Margheim,
1987; Cook & Kelley, 1988; Raghunathan,
1991). However, relatively little attention has
408 D. T. OTLBY and B. J. PIERCE
been given to the study of leadership issues in
audit firms.
Recognizing the importance of formal per-
formance evaluations in an auditing envlron-
ment, and findings reported ln the psychology
literature &lores, 1966) that these evaluations
can be significantly influenced by leadership
style (defined ln terms of structure and consid-
eration), Kida (1984) investigated the perform-
ance evaluation process in five international
CPA firms. The study concluded that leader-
ship style affected the relative emphasis placed
on specific dimensions of performance, and
furthermore, that seniors tended to emphasize
the same dimensions as managers. Consistent
with findings in the psychology literature, it
was also found that highly considerate man-
agers tended to give higher overall ratings,
although the difference was not significant.
Pratt & Jiambalvo (1981) concluded that
audit team performance was related to the
leadership behaviour of lncharge auditors.
They reported evidence that audit team per-
formance was positively related to the level of
consideration shown by the supervisor, and
that elements of structuring behavlour (in par-
ticular the degree of emphasis on meeting
budgets) also varied positively with audit team
performance. Considerate leader behaviour
was also found to be positively related to inter-
personal relationships, satisfaction and motiva-
tion. However, emphasis on budgets was found
either to have a negative relationship or no
significant relationship with these variables.
The researchers concludes that the impact of
this variable was somewhat complex and
unclear.
The importance of leadership behaviour in
an audit environment was again recognized
by Pratt & Jlambalvo (1982) who investigated
the determinants of leadership style. This study
provides evidence supporting the hypothe-
sized relationship between the leadership style
of in-charge audit seniors and two situational
variables: (i) the match between task complex-
ity and audit assistant experience and (ii) audit
assistant’s intolerance of ambiguity; and one
leader-person variable: attraction to the leader-
ship role. The researchers concluded, how-
ever, that a more complex model may be
required to explain leadership behaviour
adequately, involving additional independent
variables and recognizing interactions between
some of the independent variables.
Jiambalvo & Pratt (1982) used path-goal the-
ory to analyse the effect of in-charge leader
behaviour on assistant satisfaction and motiva-
tion. They found that considerate behaviour by
lncharge seniors had beneficial effects, and
that task complexity moderates the impact of
leadership behaviour on staff motivation and
satisfaction. It was also found that the effect
of structuring behaviour as an independent
variable was weak. However, a significant inter-
action effect between consideration and struc-
ture was found, leading to the conclusion that
unless structuring behaviour is accompanied
by considerate behaviour, the result will be
low levels of satisfaction. A general conclusion
from these studies is that whereas considerate
leader behaviour tends to be associated with
desirable outcomes, the consequences of struc-
turing behaviour are less predictable. It appears
that high levels of structuring behaviour may
be beneficial in certain circumstances (such
as clarifying the subordinate’s role when the
task is complex or not well defined), and that
a high score on both dimensions may be
indicative of effective leadership behaviour
(Hopwood, 1974b).
Kelley & Margheim (1990) presented evi-
dence that the average number of audit quality
reduction acts declined as a supervisor’s style
was perceived to reflect high levels of both
consideration and structure. This appears to
support the conclusion that high levels of both
dimensions may constitute effective leadership.
However, the researchers noted that as struc-
ture increased, under-reporting of time also
increased, suggesting that higher levels of
structuring behavlour may have resulted in
tighter control and less opportunity for audit
quality reduction behaviour, this forcing audi-
tors to engage ln a less detectable form of dys-
functional behaviour. One implication is that
high levels of structuring behaviour may help
THE CONTROL PROBLEM IN PUBLIC ACCOUNTING FIRMS 409
perpetuate excessively tight budgets which
may, in the long term, give rise to highly un-
desirable behaviour, such as premature sign-off.
Research into the causes of audit senior turn-
over has not focused specifically on the dimen-
sions of structure and consideration. However,
a number of variables which have been identi-
fied as perceived causes of turnover (Rhode et
al., 1977) can be related directly to those
dimensions, for example “not knowing what
your supervisors expected of you”, “having
some understanding people around you . . .
upon whom you can rely for assistance and
advice”, “feeling that most of the people in
the tirm really did not care about you or your
career with the firm”. These findings provide
further conformation of the importance of a
manager’s leadership style in influencing audit
senior behaviour.
HYPOTHESES
This study focused on the relationship
between leadership style of audit managers
and dysfunctional behaviour by audit seniors.
The measurement of dysfunctional behaviour
was based on previous studies, and included
(a) under-reporting of time (i.e. the deliberate
under-statement of chargeable time, usually by
working on personal time, thereby improving
the impression created by budget versus actual
performance reports and perpetuating art&
cially low budgets which may eventually lead
to quality reduction) and (b) a number of audit
quality reduction acts, including premature
signoff.
It was expected that in an audit environ-
ment, a structured style of leadership by man-
agers would be associated with perceptions of
tight control and inflexible use of time budgets.
Accordingly, budget difficulties are less likely
to be reported, but will tend to be resolved at
individual auditor level. When this occurs, the
likelihood of dysfunctional behaviour
increases. A more considerate style of leader-
ship, on the other hand, tends to be associated
with mutual trust and respect, helpfulness and
friendliness. In this environment, budget diffi-
culties are more likely to be reported and
resolved through open discussion without fear
of adverse consequences and, accordingly, are
less likely to be associated with dysfunctional
behaviour.
Prior accounting research (e.g. Hopwood,
1974b; Jiambalvo & Pratt, 1982; Brownell,
1983) has suggested a need to focus on the
combined effects of structure and considera-
tion. It is therefore hypothesized that the high-
est levels of dysfunctional behaviour will be
reported by seniors who perceive the leader-
ship style of their immediate supervisor to be
characterized by a combination of high struc-
ture and low consideration. The lowest levels
are expected to be found when the style
reflects a combination of low structure-high
consideration, and moderate levels are expected
when the style is either high structure-high
consideration or low structure-low considera-
tion (see Fig. 1). Although these general rela-
tionships have been supported by research
evidence from other work settings, they have
not been previously tested in the context of
the manager-senior relationship in an audit
environment.
The hypothesized relationships between
leadership style and dysfunctional behaviour
are as follows:
Hl When audit manager leadership style is perceived
to be characterized by high structure and low con-
sideration, audit seniors will report high levels of audit
quality reduction bebaviour and high levels of under-
reporting of time.
H2. When audit manager leadership style is perceived
to be characterized by either high structure and high
consideration or low smtctme and low consideration,
audit seniors will report moderate levels of audit quality
reduction behaviour and moderate levels of under-
reporting of time.
H3. When audit manager leadership style is perceived
to be characterized by low structure and high con-
sideration, audit seniors will report low levels of audit
quality reduction behaviour and low levels of under-
reporting of time.
410
D. T. OTL.EY and B. J. PIERCE
LOW
CONSIDERATION
High
INITIATION OF STRUCTURE
High LOW
Hypothais I
High levels of
Dysfunctional Behaviour
(Croup A)
Hypothesis 2
Moderate levels of
Dysfunctional Behaviour
(Croup W
Hypothesis 2 Hypothesis 3
Moderate levels of Low levels of
Dysfunctional Behaviour Dysfunctional Behaviour
(Group Cl (Group D)
Fig. 1. Hypothesized relationships between leadership style and dysfunctional behaviour.
Contingency-based approaches to the study
of leadership have sought to establish which
leadership style ls most effective in circum-
stances. Michaelson (1973) stated:
The ptimary question of research in leadership has
shifted from “what is the best kind of kadership” to
“what kind of leadership works best in what kind of
situation”.
This approach is reflected in the fact that
many variables have been examined in pre-
vious studies of leadership style, relating to
the organization, the work group and the task
(Bass, 1990). The specific variable of interest in
the present study was perceived environmental
uncertainty (PEU), which appears to be of par-
ticular relevance in an auditing environment.
PEU refers to a state in which an individual
makes decisions and engages in behaviours
without complete knowledge of his relation-
ship with his environment. Its particular rele-
vance to an audit senior arlses from the
presence of many unpredictable, volatile and
sometimes conflicting variables in his/her
work environment, such as clients, outside
agencies, supervisors, other departments in
the firm, the audit team, professiorud standards
and ethics, and the firm’s own rules and pro-
cedures. PEU has been linked to role conflict
and role ambiguity in other settings (Van Sell
et al., 1981) and has been found to be signifi-
cantly related to role conflict, role ambiguity,
job satisfaction and job performance in an audit
setting (Rebele & Michaels, 1990).
Findings regarding the influence of task
uncertainty on the structure-performance rela-
tionship have been mixed (Bass, 1990).
Whereas, in situations of high uncertainty,
high levels of structure would be expected to
have a role-clarifying effect and lead to higher
satisfaction and productivity, this has not
always been supported by research tindings
(e.g. Stinson & Johnson, 1975). A possible
reason for this is that, although a structured
style of leadership may clarify certain aspects
of the task, it may, in some circumstances, fall
to resolve major ambiguities, and may even
exacerbate existing problems. A highly struc-
tured style seems particularly likely to lead to
increased ambiguity and uncertainty in an audit
context, where previous research has shown
that although auditors are usually clear about
how to complete audit programmes, no clear
guidelines are given to staff to help deal with
the conthcts which can arise from being
expected to complete a highly structured audit
programme within the limits of an equaIIy well-
structured time budget (McNair, 1991). In this
situation, a high level of structuring behaviour
by audit managers may combine with a high
THE CONTROL PROBLEM IN PUBLIC ACCOUNTING FIRMS
level of PEU, which in itself helps to maintain
or even increase, to create conditions likely
to result in dysfunctional behaviour by indivi-
dual auditors. Accordingly, it is hypothesized
that PEU will moderate the relationship
between initiating structure and dysfunctional
behaviour, as follows:
H4. As the level of PBU increases, the positive relation-
ship between initiating structure and (i) audit quality
reduction behaviour and (ii) under-reporting of time
will become stronger.
In the situation of an audit assignment, a
highly considerate audit manager, displaying
qualities of friendliness and approachability, is
more likely to encourage audit staff to resolve
major dilemmas (McNair, 1991) in a functional
manner, by making the undiscussabIe discuss-
able, than a manager displaying a less consider-
ate style. The effect of this is likely to be more
pronounced in highly uncertain situations,
when the most acute dilemmas are faced by
auditors. It is therefore hypothesized that
PEU will moderate the relationship between
consideration and dysfunctional behaviour as
follows:
H5. As the level of PEU increases, the negative rela-
tionship between consideration and (i) audit quality
reduction behaviour and (ii) under-reporting of time
will become stronger.
THE STUDY
A questionnaire was sent to all audit seniors
ln three of the Big Six firms in Ireland, by prior
411
agreement with a senior partner. From the 356
questionnaires distributed, for return by post
directly to the researchers, 257 usable
responses were obtained, giving an overall
response rate of 72%. Questionnaires were
anonymous, and respondents were given an
undertaking that the contents of completed
questionnaires would remain confidential to
the researchers, except for the reporting of
aggregate results. Details of variable measure-
ment are presented below, together with
descriptive statistics.
Leadership style
The two dimensions of leader&p style, initi-
ating structure and consideration, were mea-
sured using an instrument adapted for an
audit setting by Pratt & Jiambalvo (1981), based
on Stogdill’s (1963) Leader Behaviour Descrip
tion Questionnaire. This measure was chosen
because it has been found elsewhere to show
acceptable levels of reliability (Bass, 1990), and
it has also been found suitable in an audit setting
(Jiambalvo & Pratt, 1982; Kida, 1984). The
instrument requests respondents to indicate
on a five-point scale the frequency of each of
20 listed behaviours by their manager(s). Ten of
the items reflect structuring behaviour, while
the remainder are indicative of a considerate
style. Summary statistics are shown in Table 1.
Audit quality reduction bebaviour
(AQRN
The most serious form of quality reduction
behaviour identilied in previous studies was
premature sign-off, which occurs when an
auditor signs off on a required audit step with-
TABLE 1. -hip style, AQRB and PEU - summary statistics
Theoretical range Actual range Mean S.D. Cronbach alpha
Stiucturr 10-M 16-50 35.15 7.08 0.08
Consideration 10-50 14-50 32.82 8.94 0.93
Audit *tY
reduction behaviour
(AQ-1 6-30 6-27 12.57 4.92 0.89
Perceived
envimnmental
uncertainty (PEU) 10-50 15-44 27.27 5.72 0.85
412 D. T. OTLEY and B. J. PIERCE
out completing the work or noting the omis
sion (CAR Report, 1978). The incidence of
this type of behaviour, together with other
specific forms of quality reduction behaviour
was measured by adding responses to the six
items in Questions 1 and 2 (shown in the
Appendix), which are based on the work of
Alderman & De&rick (1982) and Kelley & Mar-
gheim (1987). Although the possibility of social
desirability bias can never be completely ruled
out when collecting sensitive data of this nat-
ure, measures were taken to minimize the like-
lihood of such bias occurring (e.g. anonymity,
confidentiality, mailing of responses direct to
researchers). To the extent that such bias
does exist, reported levels of dysfunctional
behaviour are likely to be understated, and
thus represent conservative estimates of the
incidence of such behaviours.
Results revealed a high incidence of AQRB,
with 89% of respondents admitting to engaging
in at least one form of audit quality reduction
behaviour. Summary statistics for AQRB are
shown in Table 1.
Under-reporting of time CURTI
The frequency of URT was measured using
responses to the question “How often do you
under-report time? ” (scale 1-5; mean 2.87;
S.D. 1.42). The number of hours which seniors
estimated they under-reported was, on average,
12.2% of hours accurately recorded. In general,
these linclmgs suggest a greater incidence of
URT than that reported in the U.S.A. (Kelley
& Margheim, 1990).
Perceived environmental uncertafnty
(PEW
Measurement of this variable was based on
an instrument developed by Rebele & Michaels
(1990) form earlier work by Ferris (1977) and
Duncan (1972) (see Appendix, Question 3).
The measure seemed particularly suitable for
an audit setting, and achieved favourable psy-
chometric evaluations in previous studies. Sum-
mary statistics are shown in Table 1.
RESULTS
Respondents were classified into four groups
according to the perceived leadership style of
their manager(s), based on the median scores
for (i) structure and (ii) consideration. High
structure was defined as greater than or equal
to the median score of 35, and high considera-
tion was a score greater than or equal to the
median score of 32. This procedure was
selected, partly to cope with problems of multi-
collinearity, but also to aid interpretation by
allowing mean values for cells to be com-
puted. Table 2 shows descriptive statistics for
each group and in total, in relation to (i) audit
quality reduction behaviour and (ii) under-
reporting of time.
The most noteworthy feature of this analysis
is that a consistent pattern emerged in relation
to both forms of dysfunctional behaviour. In
both cases, the highest frequency was
reported by the high structure/low considera-
tion group (A). The second highest frequency
TABLE 2. Mean and S.D. of dy&mctionaI bebaviours for four leadership groups and in totaI
Variable
AQRB (scale 6-30) Under-reporting of time
(scale l-5)
Group structure Consideration n Mean S.D. Mean SD.
A me LOW 51 17.84 5.35 4.27 1.13
B LOW 63 13.05 3.72 2.83 1.18
C Ii@ 79 11.04 3.90 2.54 1.24
D LOW 64 9.94 3.21 2.20 1.31
Total 257 12.57 4.92 2.87 1.42
THE CONTROL PROBLEM IN PUBLIC ACCOUNTING FIRMS
413
TABLE 3. Ranking of leadership groups by dysfunctional
behaviour (rank 1 = highest dysfunctionaI behaviour)
Rank Group Structure Consideration
1 A
H&h
LOW
2 B LOW LOW
3 C
Hkh
High
4 D LOW High
was reported by the low structure/low con-
sideration group (B). This was followed by
the high structure/high consideration group
(C), and the lowest frequency for the specified
dysfunctional behaviours was reported by the
low structure/high consideration group (D).
The ranking of the four leadership groups in
terms of frequency of dysfunctional behaviours
is shown in Table 3. Although as explained in
the Discussion section, these findings were
expected (except that no prior hypothesis
compared group B with group C), such a
high level of consistency was surprising.
One-way ANOVA tests were then carried out
to establish which of the differences between
groups were statistically significant. The
results, which are summarized in Table 4,
showed that there were sign&ant differences
in the incidence of both types of behaviour
across the four leadership groups. Tukey ana-
lysis was used to identify which groups
differed, and the result of this is shown on
the right-hand side of Table 4.
A notable feature of the result is that the
frequency of both forms of dysfunctional beha-
viour was signilicantIy higher in the high/low
group (A) than in the two groups where there
was highly considerate manager behaviour (C
and Dj. Group A also reported signilkantly
TABLE 4. One-way ANOVA results for four leadership
groups
ANOVA results Signikant
differences
Varlabie F rati o Prob. between groups
AQm
39.04 0.000 A > D, C, B
B>D,C
Under-repotting
of time
30.40 0.000 A > D, C, B
B>D
higher figures than group B, where low con-
sideration was accompanied by low levels of
structuring behaviour. Furthermore, group B
(low/low) reported signiticantly higher levels
of AQRB than both the high consideration
groups (D and C), and significantly higher
levels of URT than the low/high group (D).
These results underline the importance of con-
siderate behaviour on the part of the audit
manager, and suggest that high levels of con-
siderate behaviour are associated with rela-
tively low levels of dysfunctional behaviour,
even when it is accompanied by high levels
of structuring behaviour.
In order to test hypotheses 4 and 5, the cate-
gories of high and low structure and high and
low consideration described earlier were
retained. In addition, respondents were cate-
gorized into three PEU groups of roughly simi-
lar size: (i) low uncertainty = PEU < 25 (n =
78; mean PELJ = 21.11; SD. = 2.55); (ii) med-
ium uncertainty = PEU 25-28 (n = 87; mean
PEU = 26.32; S.D. = 1.06); and (iii) high uncer-
tainty = PEU > 28 (n = 92; mean PEU = 33.27;
S.D. = 3.99). The data were analysed using a 2
X 2 X 3 factorial ANOVA analysis for each of
the two dependent variables, the results of
which are shown in Tables 5 (AQRB) and 7
(URT).
With regard to AQRB, Table 5 provides evi-
dence of a significant main effect for three
variables. Cell means indicate that this effect
is positive for structure and PEU and negative
for consideration. The findings support an
interactive effect between structure and PEU
and between consideration and PEU. Inspec-
tion of cell means (Table 6) indicates that in
general, as structure (consideration) increases,
AQRB increases (decreases), and that as the
level of PEU increases, this effect grows
stronger. This supports hypotheses 4 and 5,
as applied to AQRB. An exception arises in
relation to the high structure/low PEU group.
This is the only situation where high structure
is associated with a fall in AQRB, and, in accor-
dance with House’s (1971) path-goal theory of
leadership, suggests that if structuring beha-
viour is effective in reducing uncertainty,
414 D. T. OTLEY and B. J. PIERCE
TABLES. Resultsofa X 2 X 3ANOVAanaWslsforAQRB
Source of variation sum of squares d.f. Mean square P sig. of P
Main Effects
smcture
Consideration
PEU
Tweway ilwractions
Structure X consideration
structufe x PEU
Consideration X PEU
Thee-way interactions
Structure X consideration X PEU
Explained 3360.58 11
Residual 2820.89 242
464.49 1 464.49 39.85 0.000
563.09 1 563.09 48.31 0.000
972.57 2 486.28 41.72 0.000
12.59 1 12.59 1.08 0.300
188.4s 2 94.22 8.08 0.000
159.73 2 79.87 6.85 0.001
193.99 2 %.99 8.32 0.000
3OSSl
11.66
26.21 0.000
Total 6181.47 253 24.43
then it may help generate a favourable
response in terms of individual performance.
The implication is that where a high level of
structuring behaviour fails to resolve uncertain-
ties in the work environment (medium and
high PEU groups) the outcome in terms of
undesirable behaviour is considerably worse
than for a low structure leadership style.
The results summarized in Table 5 also indi-
cate that, in addition to two 2-way interactions,
there is also a significant 3-way interaction
effect between the three variables structure,
consideration and PEU. This suggests that the
PEU in the auditor’s environment interacts with
the manager’s leadership style, defined in terms
of structure and consideration, to produce a
significant effect on the level of AQRB
behaviour in which seniors engage. Implica-
TABLE 6. Interaction of PBU with (i) strocture and
(ii) consideration: mean AQRB by cell
PBU
LOW Medium
Low r3tNcN-e 9.29 (38) 11.77 (44) 13.04 (45)
High stNcture 8.97 (38) 12.9’3 (42) 18.38 (47)
Low consideration 10.86 (22) 12.71 (34) 17.27 (67)
Higb consideration 8.43 (54) 12.12 (52) 11.76 (25)
Note: numbers in parentheses indicate cell &es.
tions of PEU in the operating environment of
audit seniors are considered in the Discussion
section.
The analysis in relation to URT (Table 7)
provided evidence of a significant main effect
for structure and PEU (both positive) and con-
sideration (negative), similar to that found for
AQRB. However, of the two hypothesized
interaction effects, only structure X PEU was
statistically significant at the 0.05 level, and an
inspection of cell means showed this effect to
be in the direction set out in hypothesis 4.
Neither the consideration X PEW interaction
(hypothesis 5) nor the 3-way interaction
between structure, consideration and PEU
(not hypothesized but present in the AQRB
analysis) were significant for URT. A noticeable
feature of the results, however, was the signifi-
cance of the interactive term structure X con-
sideration. Tbe pattern of URT behaviour
associated with different combinations of
structure and consideration was set out earlier
(see Table 2), indicating the importance of
consideration if high levels of structure are
used.
The combined effect of these results is that
hypothesis 4 was supported, but whereas hypo-
thesis 5 was supported for AQRB, the evidence
did not support that hypothesis in relation to
Source of varhthm
THE CONTROL PROBLEM IN PUBLIC ACCOUNTING FIRMS
415
TABLE7.Rcsultsofa2X2X3~OVAvllhrsisforURT
sum of SaIlares d.f. Mean salare P sig. of F
Maine&cts
s- 38.15 1 38.15 31.29 0.000
CAXlSidC~tiOll 38.05 1 38.05 31.21 0.000
PEU 64.16 2 32.08 26.31 0.000
Twmvay intexactions
Structure X consideration 5.28 1 5.28 4.33 0.038
strum x PEU 7.40 2 3.70 3.03 0.050
consideration x PBU 5.70 2 2.85 2.34 0.099
Three-way lnuractions
Wwture X consideration X
PEU
4.71 2 2.35 1.93 0.147
Bxplaiaed 215.96 11 19.63 16.10 0.000
Residual 295.05 242 1.22
TotAl 511.01 253
2.02
URT. These results, taken together with the
finding of a 3-way interaction for AQRB but
not for URT and a slgnilicant structure X
consideration interaction effect for URT but
not AQRE suggest an overall conclusion that,
whereas PEU exercises a very signifkant
moderating effect on the leadership style-
AQRB relationship, the moderating effect of
PEU on the leadership style-URT relationship
is noticeably weaker.
DISCUSSION
The most important finding is the role of
considerate lea&r behaviour. High considera-
tion is associated with relatively low levels of
dysfunctional behaviour, regardless of struc-
turing behaviour (see Table 2). This result is
consistent across all forms of dysfuntitional
behaviour included in the study. Between the
extremes of groups A and D, groups B and C
ranked second and third, respectively. This
highhghts the importance of considerate be-
haviour in the fact that group C @gh/h@h)
reported signlfIcantly lower levels of quality
reduction behaviour than group B (low/law).
These findings suggest that audit fnms might
successfully tackle the problem of dysfunc-
tional behaviour by directing recruitment, pro
motion and training policies towards the
development of a high consideration/low struc-
ture style of leadership among managers and
partners. However, as previous research has
shown that no single style of leadership may
be ideal in all circumstances, the analysis was
extended to examine possible moderating
effects of the variable perceived environmental
uncertainty.
Analysis of auditor behaviour under different
conditions of uncertainty showed that in a rela-
tively certain environment, the level of AQRB
decreased as leadership style became more
structured Gable 6). This may be because
highly structured manager behaviour has suc-
ceeded in chuifying for audit seniors their
paths to achieving rewards, and communicat-
ing the value of those rewards, and as a result,
individual performance has become more
effective. It is also possible, however, that in
a relatively certain environment with a highly
structured style of management, the scope for
undetected AQRB is much reduced and audi-
tors turn to another form of dysfunctional be-
haviour, such as URT, to resolve conflicts and
dilemmas. The finding that URT increased as
416 D. T. OTLBY and B. J. PIERCE
leadership style became more structured in a that environment will moderate the effects of
relatively certain environment suggests this is the leadership dimensions structure (positive)
a more likely explanation. and consideration (negative) on AQRR.
The moderating effect of perceived environ-
mental uncertainty was noticeably stronger for
the dependent variable AQRR than for URT.
The URT analysis showed no evidence of a
moderating effect on the consideration-URT
relationship, and although there was some
evidence of a moderating effect on the struc-
ture-URT relationship, it barely achieved signi-
ficance at the 0.05 level. Furthermore, there
was evidence of a significant interaction effect
between structure and consideration for the
dependent variable URT, but not for AQRR.
These lindings reflect the fact that, although
the two dependent variables are similar in
some respects (Pearson correlation coefficient
= 0.56), there are also important differences.
Although most auditors are aware that URT is
forbidden by their firms, it is not a form of
behaviour that results in a direct and immedi-
ate threat to audit quality (although it may do
so at a future date by causing auditors to com-
promise quality in order to meet artificially low
budgets). Many respondents indicated that
audit managers sometimes encourage them to
engage in URT in order to avoid budget over-
runs. An auditor’s perception of the manager’s
attitude towards budget over-runs may there-
fore be the dominant consideration in making
the URT decision. AQRB, on the other hand,
is a potentlally very serious form of mls-
behaviour, involving a deliberate reduction in
quality through falsification of audit records.
Although it does not seem to be openly dis-
cussed or formally communicated, the penalty
for engaging in such behaviour would be
expected to be severe, and to carry major
implications for future career prospects. It is
likely that a decision to engage in such be-
havlour will not be taken lightly and will
involve an appraisal, not only of the manager’s
attitudes to various possible budget outcomes,
but also of the broader environment and the
possible ramilications of such behaviour.
Having formed an assessment of his environ-
ment, the perceived level of uncertainty in
An alternative possible explanation lies in
the essential difference between the two cate-
gories of behaviour, AQRR and URT. The find-
ings suggest that when a leadership style which
is associated with dysfunctional responses (i.e.
high structure-low consideration) is combined
with a highly uncertain environment, the indi-
vidual auditor response tends to reflect an
increase in AQRR rather than URT. This trend
underlines the high level of interdependence
between the two categories of behaviour, by
suggesting a shift towards quality reduction
behaviour in conditions of high uncertainty.
In these circumstances, it may be very difhcult
for the individual auditor to predict in advance
how many hours would have to be under-
reported in order to complete the job within
budget. There is therefore uncertainty regard-
ing the likely success of URT as a strategy for
meeting tight budgets and strategies involving
quality reduction become more attractive.
With regard to leadership style, the evidence
supports the view that a considerate style tends
to result in lower levels of dysfunctional beha-
viour. One interpretation of this is that a man-
ager displaying high consideration tends to be
more open and approachable when uncertain-
ties arise. This style of leadership is therefore
likely to result in lower levels of perceived
environmental uncertainty. The evidence
supports this, as the high consideration group
reported significantly lower levels of PEU than
the low consideration group (t = 7.53; prob.
< 0.001). These arguments are generally in line
with path-goal theory (House, 1971; House &
Mitchell, 1974), which holds that performance
and motivation will improve if paths leading to
valued rewards are made clear.
This suggests that the uncertainties in an
audit senior’s work environment are so diverse
and dynamic that a highly considerate style is
more helpful than a structured style ln inter-
preting goals and objectives. The auditor is
exposed to boundary-spanning activities to a
far greater degree than many other profes
THE CONTROL PROBLEM IN PUBLIC ACCOUNTING FIRMS 417
sionals (Rebele & Michaels, 1990) giving rise
to additional layers of uncertainty which may
be unaffected or even increased by a highly
structured leadership style. Such a style may,
however, offer little or no positive contribu-
tion in terms of role clarification, since seniors
are generally highly trained professional audi-
tors who are engaged ln relatively routine work
and are guided by a detailed audit programme.
A considerate style, on the other hand, helps
promote a supportive work environment in
which the varied sources of audit senior
uncertainty can more effectively be addressed.
CONCLUSION
The study has some important limitations
which may restrict the inferences to be drawn
from the findings. Only three of the Big Six
firms were included, and the results are not
therefore necessarily applicable to other
lirms. However, the study achieved broad
coverage, since a high response was obtained
from the sample of all audit seniors in the par-
ticipating lirms, whose structures and pro-
cedures are known to be broadly similar to
other large audit Iirms. A second limitation is
that a cross-sectional study cannot support con-
clusions regarding causality. Reverse causal
linkages are a distinct possibility, since seniors
who are suspected of engaging in dysfunctional
behaviour may attract a highly structured
response from the manager, whereas seniors
who are not suspected of such behaviour may
attract a more considerate response. It is also
possible that perceptions of leadership style
may vary depending on the senior’s inclination
towards dysfunctional behaviour, where a
senior who is seeking opportunities to conceal
such behaviour may be more conscious of
structuring behaviour by managers.
Despite these limitations, the llndings high-
light the importance of leadership style in an
audit setting, and have important implications
from both theoretical and practical perspec-
tives. Two serious forms of dysfunctional be-
haviour were found to be significantly related
to the two leadership dimensions, considera-
tion (negatively) and structure (positively),
and these relationships were found to be stron-
ger when perceived environmental uncertainty
increased. Although generally consistent flnd-
ings emerged for both types of dysfunctional
behaviour, the moderating effect of perceived
environmental uncertainty was found to be
stronger for AQRB than URT.
The evidence suggests that structured leader-
ship by audit managers often creates ambiguity
and conflict in the minds of individual auditors.
This occurs when an auditor discovers that the
time budget is too tight to allow proper
completion of the audit programme. The audit
programme is a highly structured plan, which
clarifies the tasks to be carried out. It is often
accompanied by high levels of uncertainty,
however, regarding how the tasks can be
completed within the time allowance.
In terms of practical implications, the find-
ings suggest that audit firms have considerable
scope to influence dysfunctional behaviour
through the recruitment, promotion and train-
ing of managers and partners. The study has
highlighted the important role of considerate
and supportive leader behaviour, in providing
an effective framework for resolving uncer-
tainty. Without this, the manager may become
an additional source of uncertainty in the
already complex and uncertain operating
environment of the audit senior.
BIBLIOGRAPHY
Adair, J., h”eelfve Lmadesbip (Aldershot: Gower, 193).
Alderman, C. W. & De&rick, J. W., Auditors’ Perceptions of Time Budge Pressures and Premature Sign-
offs: a Replication and Extension, Audfting: a J oumul of Practfce 6 Tbeo?y (1982) pp. 54-68.
418 D. T. OTLEY and B. J. PIERCE
Ansari, S. L., Behavioral Factors in Variance Controls l&port on a Laboratory Expcrimcn& Journ& of
Accounting Resea*cb (1976) pp. 189-211.
Apostolou, B., Pascwah, W. R & Strawscr, J. R, The Effects of Senior Internal Auditor Bchaviour on Staff
Perfotxnancc and SaMaction, Accounting and Business Research (1993) pp. 110-122.
Argyris, C., l&z Impact of Budgets on Peopk (New York: The Controllership Foundation, 1952).
Barrow, J. C., The Vatiables of hdership: a Beview and Conceptual Framework, Awdemy of
Management Revkw (1977) pp. 231-251.
Bass, B. M., Bass C Stogdilh Handbook of Leadership. fbeory, Research and Manag& Applications
(New York: The Free Press, 1990).
Blake, R R. & Mouton, J. S., IIx? Manage&l Grfd (Houston: Gulf Publishing, 1964).
Brownell, P., Leadership Style, Budgetary Participation and Managerial Behaviour, Accounting,
OrganFzatfons and Society (1983) pp. 307-321.
Commission on Auditors’ Responsibilities (CAR), Report, Conclusions and Recommendutions (New
York AI CPA, 1978).
Cook, E. % K&q, T., Auditor Stress and Time Budgets, 7h CPAJ ournui (1988) pp. 83-86.
JkCostcr, D. & Fertakis, J., Budget-induced Pressure and its Rehtionship to Supcrvhry Bchaviour,
J ournal of Accounting Research (1968) pp. 237-246.
Duncan, R B., chvlctciistics of Organhtional Environments and Perceived Bnvironmental Uncertainty,
Administrative Scknce Quarterly (1972) pp. 313-327.
Emmanuel, C., Otlcy, D. & Merchant, K., Accounting for Management Confrol (London: Chapman and
Hall, 1990).
Ferris, K. R, Perceived Uncertlinty and Job Satisfaction in the Accounting Environment, Accounting,
Organizations and Sock@ (1977) pp. 23-28.
Ficdler, F. E., A lbeory of ih&zsbip Effectfveness (New York: McGraw Hill, 1%7).
Fleishman, E. A. & Petm, D. R, Intcqcrxmal Values, Leadership Attitudes and Managerial “Success”,
Personnel Psychology (1%2) pp. 127-143.
Hopwood, A. G., An Empirical Study of the Role of Accounting Data in Perfomtancc Evaluation, J ournal
of Accounting Research (1972) pp. 156-182 (Suppl.).
Hopwood, A. G., Accounthg and Human Bebaviour (London: Prentice-Hall, 1974a).
Hopwood, A. G., Leadership Climate and the Use of Accounting Data in Performance Evaluation, ZYbe
Accounting Revkw (1974b) pp. 485-495.
House., R J., A Path-goal Theory of Leader Effectiveness, Admfnfstr&ue Science Quatier& (1971) pp.
321-338.
House, R J. % Mitchell, T. R., Path-goal Theory of Lea&tip, J ournal of Contemporary Business (1974)
pp. 81-97.
Jiaatbalvo, J. % Pratt, J., Task Complexity and Leadership Effectiveness in CPA Firms, ZIw Accounting
Reukw (1982) pp. 734-750.
Katz, D. % Kahn, R. L., Some Recent Findings in Human Belations Research in Industry, in S wanson, G. V.,
Newcomb, T. M. % Hardy, E. L. (eds), Readings k Sodal Ps@okgy @J ew Yoik: Holt, 1952).
Kellcy, T. % Marghcim, L., The Effect of Audit Billing Amngcment on Under-reporting of Time and Audit
Quality Bcduction Acts, Aduawes fn Accounting (1987) pp. 221-233.
Kelky, T. % Mar&ehn, L., The Impact of Time Budget F’reswe, Personality and Leadership Variables on
Dysfunctional Behaviour, AudiZing: A J ou& of PTactke 6 lbeory (1990) pp. 21-41.
Kelley, T. % Seller, S. E., Auditor Stress and Time BudBets, lbe CPA J oumal(1982) pp. 24-34.
Kida, T. E., F’erformance Won and Beview Meeting Characteristics in PubBc Accounting Firms,
Accounting, Organtzalions and Society (1984) pp. 137-147.
Klores, M., Bater Bias in ForceddkMb ution Fwformance Ratings, Personnel l?ycbology (1566) pp.
411-421.
Lee, R t Lawrcncc, P., Organfxutional Behtiour: Politfcs at Work (London: Hutchinson, 1985).
Lewin, K., Lippitt, R % White, R. K., Patterns of Aggressive Bchaviour io Fixpcrimentally Created Social
Envlrwmen*, I nternational J ournal of Social I ?sycbology (1939) pp. 271-299.
l&htner, S. M., Leisenriq, J. J. & Winters, A. J., Underreporting Chargeable Time, J ourrud of Accoun-
tancy (1983) pp. 52-57.
Likert, R, New Patterns of Manugement (New York: McGraw Hill, l%l>.
McNair, C. J., Proper Compromixs: the ManaBemat Control Dilemma ill Public Accounting and ita
Impact on Auditor Behavior, Accounting, O*gan&atJonF and So&@ (1991) pp. 635-653.
Michaelson, L. K., Leader Orientation, Leader Behavior, Group Effectiveness and Situational FavorabIBty:
THE CONTROL PROBLEM IN PUBLIC ACCOUNTING FIRMS
419
an EmpIrical Extension of the Contingency Model, Otgunkatkmul Eebavfour and Human Perfii-
muncle (1973) pp. 226-245.
Pratt, J. & Jiambaivo, J., Relationships between Leader Rcbaviors and Audit Team Performance, Aczoun&
ing, Otgantaations and Society (1981) pp. 133-142.
Pratt, J. & Jiambalvo, J., Detcnnimn ts of Leader Behavior in an Audit Rnvironment, Accountfng,
Organtzations and Society (1982) pp. 369-379.
Preston, M. G. & Hcintz, R. K., Rffects of Partkipatory Versus Supervisory LeadershIp on Group
Judgment, J ournal of Abnormd and Social Pycbology (1949) pp. 345-355.
Ra&nathan, B., Premature Signing-off of Audit Procedures: an Analysis, Accounting H0rlzon.r (1991)
pp. 71-79.
Rebele, J. E. & Michaels, R. E., Independent Auditors’ Role Stress: Antecedent, Outcome, and Moderating
Variables, Bebavloral Research fn Accounting (1990) pp. 124-153.
Rhode, J. G., Sorensen, J. E. % Lawler, E. E. III., Sources of ProfessionaI Staff Turnover in Public Account-
ing Fhms ReveaIed by the Exh Interview, Accountfng, Orguntiations andSocfe@ (1977) pp. 165-175.
.%inwn, J. E. % Johnson, T. W., The Path-goal Theory of Leadership: a PartiaI Test and Suggested
Refinement, Academy of Management J oumal(1975) pp. 242-252.
Str@Ul, R. M., Manual for tbe Leader Bebatiour Descrtptfon Questionnaire Form XI I (Columbus:
Bureau of Business Research, The Ohio State University, 1%3).
Stogdill, R. M., Handbook of Leader&p: a Survey of Tbeoy and Researcb (New York: The Free Press,
1974).
Van Sell, M., Brief, A. P. & Schuler, R. S., Role Conikt and Role Ambiguity: Integration of the Literature
and Directions for Future Research, Human Relations (1981) pp. 43-71.
Yukl, G., ManageriaI Leadership: a Review of Theory and Research,]ourn& of Management (1989) pp.
251-289.
APPENDIX
Relevant questions reproduced from tbe Survey Questionnaire
QlBu3tion 1
Please tick one answer on each line.
NA = nearly always 0 = often S = sometimes R = rarely N = never
(Scored)
In your work as auditor, do you ever sign off
a reqired audit step not covered by another
audit step, without completing the work or
noring the omission of procedures?
Question 2
During the past year, how often have you acted
in the following manner when carryirq out an
audit?
(a) accepted weak client explanations
(b) made superficial reviews of client documents
(c) failed to research an accounting principle
(d) reduced the amount of work performed on an
audit step below what you considered
reasonable
NA
5
0
4
s
3
R
2
N
1
0 0 0 0 0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
420 D. T. OTL.EY and B. J. PIERCE
(e) signed off an audit programme step without
completing the work or noting the omission
0 0 0 0 0
Question 3
The following questions relate to the h-equency with which you encounter certain situations in
your work as audit senior.
NA = nearly always 0 = often S = sometimes R = rarely N = never
a. How often are you sure about what approach
is best for dealing with job-related problems?
b. How o&en are you sure about how to deal with
a change in social, economic or technical
conditions outside the firm?
c. When dealing with people in other areas of the
firm, how often do you feel you have enough
detail to know how they want the job done?
d. How often are you sure about the best way to
approach job-related problems that arise in
other departments?
e. How often is it difficult for you to determine if a
job-related decision is a correct one?
f. How often are you in doubt about how to obtain
the necessary information for making job-related
decisions?
g. How often are you able to predict accurately the
outcomes of your job-related decisions?
h. How often are you working on clients that you
have worked on before and with whom you are
familiar?
i. How often are you working on first-time audits,
i.e. clients that are new to your firm?
j. How often are you working on clients that belong
to a particular industry in which you have
specialized and developed particular expertise as
an auditor?
NA
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
S
0
0
0
0
0
0
0
0
0
0
R
0
0
0
0
0
0
0
0
0
0
N
0
0
0
0
0
0
0
0
0
0
doc_172023483.pdf