Professional Documents
Culture Documents
Chong and Mahama 2014 - Interactive and Diagnostic Use of Budget
Chong and Mahama 2014 - Interactive and Diagnostic Use of Budget
a r t i c l e i n f o a b s t r a c t
This study investigates the impact of style of budget use on team-level motivation and
Keywords:
Biotechnology
team effectiveness. Specifically, we draw on Simons’ work (1994, 1995, 2000) to examine
Budgets how, in biotechnology firms, the extent of interactive and diagnostic uses of budgets affect
Perceived collective efficacy team effectiveness directly and indirectly through perceived collective efficacy. We used an
Extent of interactive use online survey to obtain data for this study and analysed the data using partial least squares
Extent of diagnostic use approach. The main results suggest that the extent of interactive use of budgets has direct
Team effectiveness positive effects on team effectiveness and that the positive effect of the interactive use of
budgets on team effectiveness is partially mediated by perceived collective efficacy. We did
not find support for the predicted positive effect of diagnostic use of budgets on perceived
collective efficacy and team effectiveness. We explain the implications of these results.
© 2013 Elsevier Ltd. All rights reserved.
1044-5005/$ – see front matter © 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.mar.2013.10.008
K.M. Chong, H. Mahama / Management Accounting Research 25 (2014) 206–222 207
Otley, 2004; Simons, 2000; Otley, 1994), we find it nec- use of budgets on collective efficacy. In addition, consis-
essary to examine how accounting is implicated in team tent with prior research (Chen and Bliese, 2002; Lent et al.,
motivation and team performance. 2006; Wang and Lin, 2007), we find a significant direct
Second, there is limited evidence on how account- positive relationship between perceived collective efficacy
ing controls are implicated in team motivation. Although and team effectiveness. Whilst the test of the link between
prior accounting studies have examined motivation at the perceived collective efficacy and team effectiveness is a
individual level, extending the results of these studies replication of prior studies, this test is necessary to exam-
to team level may not be appropriate. Research suggests ine how the relationship between the extent of interactive
that team-level motivation may differ from individual- use of budgets and team effectiveness is partially mediated
level motivation. This is because team-level motivation is by perceived collective efficacy. Our analysis provides sup-
not simply the sum of individual team member’s motiva- port for partial mediation with a significant indirect effect
tions (Bandura, 2001, 1997; Levine et al., 1993). Bandura between the interactive use of budgets and team effective-
(1997) argues that beyond individual-level motivation, ness through perceived collective efficacy.
people who work in teams also have additional consider- This study makes three broad contributions to the liter-
ations that contribute to team-level motivation such as: ature. First, it contributes to our understanding of how the
(a) beliefs about whether team members have the req- style of budget use (particularly the interactive use of budg-
uisite skill to execute assigned task, and (b) confidence ets) affects team-level motivation and team effectiveness.
that team members will collaborate in the execution of Teams are becoming central to the effective functioning
task. Zajonc (cited in Levine et al., 1993) also argues that of contemporary organisations (Jehn and Mannix, 2001),
team-level motivation differs from individual-level moti- and given the amount of organisational resources invested
vation in that, the presence of others in teams is likely to ensure their effectiveness, improving our knowledge
to elicit accessible cognitions that may facilitate or impair of the style of budget use in teams will enable managers
performance. In addition, Baron (1986) argues that the to improve team performance. In particular, this study is
presence of others in team settings is a distraction, which among the first to examine the relationship between the
causes either performance increments or decrements. The extent of interactive and diagnostic uses of budgets and
paucity of accounting research that focuses on team moti- team-level motivation and how that link translates into
vation and the recognition that team-level motivation team effectiveness. Second, we contribute to the account-
may differ from individual-level motivation, prompt us to ing literature that focuses on the use of accounting controls
examine the effect of accounting controls on team motiva- in innovation settings (Bisbe and Otley, 2004; Davila,
tion. 2000; Hertenstein and Platt, 2000) by examining (biotech)
Third, this study focuses on the style of budget use to teams that operate in an innovation environment. Third,
examine the role of accounting controls in a team con- Abernethy and Brownell (1999) argue that much of prior
text. We focus on budgets as they are a major feature research on budgets focuses on the diagnostic use of budg-
of organisations’ management control systems (Abernethy ets to the virtual exclusion of the interactive use of budgets.
and Brownell, 1999) and are among the most widely used This study responds to their call for more studies on the
accounting controls in practice (Bisbe and Otley, 2004). interactive use of budgets by examining both the interac-
Simons (1990) argues that there are distinctive differences tive and diagnostic uses of budgets in a team environment.
in the way accounting controls (including budgets) are used
by top management. Traditionally, the focus has been on 2. Literature review and hypotheses development
the diagnostic style of budget use where top management
evaluates a team’s performance by comparing it to the tar- 2.1. Work teams in organisational settings
gets. Prior studies have helped to enrich our understanding
of the effects of the diagnostic uses of accounting in team Kirkman and Rosen (1999, p. 58) define a work team
settings, for example, the influences of different incentive as “a group of individuals who work interdependently to
structures (Libby and Thorne, 2009; Towry, 2003; Young solve problems or carry out work.” Such groups of individ-
et al., 1993) and the use of comprehensive performance uals are interdependent because of the tasks they perform
measures on team performance (Scott and Tiessen, 1999). as members of a group within a larger organisational body.
We seek to complement the existing accounting litera- Social psychologists argue that when individuals are con-
ture by examining the impact of both the diagnostic and stituted into teams, their joint outcome is not merely the
interactive uses of budgets on team motivation and team product of their individual knowledge and skills but also
effectiveness. the interactive and synergistic dynamics of their working
We conduct a cross-sectional survey of members of relationships (Bandura, 2000; Guzzo and Dickson, 1996).
teams in biotech firms in the United States. The results Consequently, a work team in an organisational setting
support our prediction that the extent of interactive use is conceptualised as a social entity within which beliefs
of budgets in these teams will have a direct positive rela- and perceptions about the actions and behaviours of other
tionship with team effectiveness. However, the extent of members of the team can affect not only the level of moti-
diagnostic use of budgets has no significant relationship vation within the team, but also effective team outcomes
with team effectiveness. We also find support for our pre- (Bandura, 2000; Mulvey and Klein, 1998).
diction that the extent of interactive use of budgets will A major team motivational factor that is argued to
have a direct positive relationship with perceived collec- be conditioned by team members’ beliefs and percep-
tive efficacy but no support for the effect of diagnostic tions (and which have consequences for effective team
208 K.M. Chong, H. Mahama / Management Accounting Research 25 (2014) 206–222
outcomes) is perceived collective efficacy. Prior research et al., 1995; Little and Madigan, 1997; Wang and Lin,
on teams indicates that perceived collective efficacy is 2007; Wood and Bandura, 1989). Bandura (1999), in
significantly linked to team goal attainment (Goddard arguing for the motivational significance of collective
et al., 2004; Hoigaard et al., 2006; Little and Madigan, 1997; efficacy, states that team member’s collective efficacy
Sampson et al., 2000). It is also argued in prior research beliefs influence the type of futures they seek to achieve
that the level of this team motivational factor is a function through collective action, how they use their resources,
of organisational processes and controls. As accounting and their staying power when team efforts fail to produce
controls are an essential part of organisational processes results. Team members are less likely to initiate or persist
and control, we argue that the style of budget use will in a course of action if they feel a low sense of collective
affect team-level motivation. We therefore investigate efficacy (Bandura, 1986). In summarising the importance
the impact of the extent of interactive and diagnostic of collective efficacy perceptions to team motivation, Little
uses of budgets on perceived collective efficacy and team and Madigan (1997, p. 519) state that “collective efficacy
effectiveness in the context of biotech work teams. appears to offer the most clearly articulated framework
for examining group-level motivational effects”. It has also
2.2. Perceived collective efficacy been argued to strongly predict team performance (Jung
and Sosik, 2002) and to mediate the relationship between
Perceived collective efficacy was defined earlier as team contextual factors and team performance (Bandura, 1986;
members’ beliefs in their team’s collective ability to organ- Little and Madigan, 1997; Pethe, 2002).
ise and execute a given task. It is a perception of a team’s In summary, the existing team literature suggests that
performance capabilities in a specific situation (Kazub and perceived collective efficacy beliefs are shared by team
McDonnell, 2000; Prussia and Kinicki, 1996). This percep- members and these beliefs influence how teams function.
tion is a causal attribution made to the team rather than We draw on the concept of perceived collective efficacy to
the situation (Lindsley et al., 1995; Pethe, 2002). Lent et al. explain how the extent of interactive and diagnostic uses
(2006) argue that collective efficacy is a social cognitive of budgets will influence team performance.
element that may help to explain how individuals function
together as a team. 2.3. Interactive and diagnostic uses of budgets
Perceived collective efficacy reflects a team’s collective
belief that is, a belief shared by members of a team (Lindsley A significant amount of prior research on budgets has
et al., 1995; Watson et al., 2001). According to Lindsley et al. focused on their traditional diagnostic role, where budgets
(1995, p. 648), there are certain cognitions or beliefs that are used on an exception basis to monitor organisational
“group and organisational members have which are quite outcomes and to evaluate and reward performance based
different and distinguishable from beliefs they experience on preset budget standards (Abernethy and Brownell,
as individuals in isolation”. Such cognitions are argued to 1999; Henri, 2006). Whilst research that examined the
be collectively shared, group-based beliefs that emanate diagnostic use of budgets has made important contrib-
from the common exposure of team members to objec- utions to the literature, Abernethy and Brownell (1999)
tive external and/or internal stimuli and through processes argue for a need to examine how budgets are used interac-
of socialisation within the team (Lindsley et al., 1995; tively and their effects in practice.
Watson et al., 2001). Perceived collective efficacy is con- Simons (1990) argues that almost all organisations
sidered to be one such group or team cognition. Collective have similar management control systems (for example,
efficacy beliefs are shared because a team member’s inter- budgets) but there are distinctive differences in the way
pretations of stimuli are based on how other members these control systems are used by top management. He
perceive and interpret the stimuli (Watson et al., 2001). As argues that any control system can be used in one of two
a consequence, team members’ perception of their team’s ways: interactively (interactive controls) or diagnostically
collective efficacy should be similar (Lindsley et al., 1995; (diagnostic controls). Whilst almost all management con-
Watson et al., 2001). trol systems are used diagnostically, top management can
Empirical studies have found support for collective effi- decide which management control systems to use inter-
cacy as an attribute of a team. In assessing the extent to actively and the extent to which these controls are used
which collective efficacy beliefs are shared, prior research interactively (Simons, 1990). Our main focus in this paper
has examined within-group agreement by computing is to study teams that use budgets and to focus on the vari-
inter-rater reliability coefficients (Rwg ) for each of the ations in the extent of interactive and diagnostic uses of
teams they studied. For example, Watson et al. (2001) find these budgets to explain the levels of perceived collective
high Rwg of between 0.88 and 0.98 among the teams he efficacy and team effectiveness.
studied. Similarly, Jung and Sosik (2002) find average Rwg of According to Simons (2000), the interactive use of con-
0.9 across the teams in their sample. These high inter-rater trols occur when top management uses control systems to
reliability coefficients lend empirical support to collective personally and regularly engage themselves in the decision
efficacy as a group attribute; collectively shared by team activities of subordinates. In a team context, the interac-
members. tive use of control may involve superiors using the control
Collective efficacy beliefs are argued to be motivational system to personally and regularly engage with the activi-
as they have been found to affect goal setting, choice ties of work teams and/or team members using the control
of activities, level of effort exerted, and persistence of systems interactively among themselves. Simons (1994)
coping behaviour (Kazub and McDonnell, 2000; Lindsley argues that any diagnostic control (e.g., budgets) can be
K.M. Chong, H. Mahama / Management Accounting Research 25 (2014) 206–222 209
to discuss and evaluate the information they bring to the We expect that the extent of interactive use of budg-
team. The interactive dialogue and debates associated ets in a team context will be positively associated with
with the higher extent of interactive use of budgets will perceived collective efficacy. Greater extent of interactive
more likely create an environment that encourages team use of budgets will facilitate the building of perceived col-
members to discuss and evaluate each other’s judgement lective efficacy in at least three ways. First, Abernethy and
by challenging underlying data, assumptions and action Brownell (1999) argue that the interactive use of budgets
plans. We argue that this will lead to higher quality implies that superiors and subordinates participate in the
decisions, with greater focus on critical success factors and budget process and the on-going dialogue clarifies how
more efficient and effective use of resources. Prior research behaviours can be adapted to meet defined objectives.
also suggests that the diagnostic use of budgets can be used Henri (2006) and Bisbe and Otley (2004) also argue that
to achieve effective performance outcomes by providing the interactive use of controls focuses attention on what
motivation, resources and information to ensure that top management considers to be of strategic importance,
organisational strategies and goals are achieved (Widener, thereby guiding subordinates actions towards desired out-
2007; Bisbe and Otley, 2004; Davila, 2000; Simons, 2000; comes. Thus, when managers use budgets interactively,
Scott and Tiessen, 1999). Following from prior research we they continuously and regularly signal issues to which they
argue that the interactive and diagnostic uses of manage- want to direct team members’ attention. Managers use
ment control systems in a team context will be positively these budget signals to challenge team members and their
and directly related to team effectiveness. More formally: action plans with a view to refocussing the team on their
expected role. Consequently, greater extent of interactive
Hypothesis 1a. The extent of interactive use of budgets
use of budgets will have the effect of enhancing perceived
will have a direct positive relationship with team effective-
collective efficacy.
ness.
Second, the interactive use of budgets provides steer-
Hypothesis 1b. The extent of diagnostic use of budgets ing mechanisms such as meetings. These mechanisms
will have a direct positive relationship with team effective- facilitate interaction and encourage dialogue among team
ness. members. Prior research suggests that the dialogue pro-
cesses associated with the interactive use of controls
2.5. Interactive and diagnostic uses of budgets and motivate information gathering and knowledge shar-
perceived collective efficacy ing (Bisbe and Otley, 2004; Henri, 2006; Simons, 2000;
Widener, 2007). The gathering and sharing of information
Perceived collective efficacy was defined earlier as team enables individuals to share knowledge about their indi-
members’ collective shared belief about their team’s ability vidual capabilities and how such capabilities can be fused
to organise and execute a given task. Given the significance together synergistically. Understanding the capabilities of
of perceived collective efficacy to team behaviour and team members can have the effect of influencing an indi-
outcomes, researchers have investigated how it is devel- vidual’s belief in their team’s collective efficacy. Third, the
oped and enhanced. Prior research has identified at least participatory and dialogic nature of using budgets inter-
three issues that affect team members’ perceived collective actively enables debate about the underlying assumptions
efficacy beliefs that are, clear roles, facilitory interaction, of work processes and the associated interdependencies,
and coordination (Bandura, 1997; Chen and Bliese, 2002; and the development of plans to coordinate team work.
Gibson, 1999; Lee et al., 2002; Taggar and Seijts, 2003). This coordinative role of using budgets interactively is
‘Clear roles’ refers to how clearly team members under- expected to enhance team members’ collective efficacy
stand what is expected of the team as a whole and the beliefs.
expected role of each member. Clear roles are said to influ- Following from the above, we expect that greater extent
ence the cognitive representation processes through which of interactive use of budgets in work teams will enhance
members imagine their team’s ability to succeed and by individual members’ confidence in the conjoint capabilities
which these members develop beliefs in their team’s col- of their team. Thus, we predict that:
lective efficacy (Bandura, 1997; Chen and Bliese, 2002;
Hypothesis 2a. The extent of interactive use of budgets
Chen et al., 2000). In addition, facilitory interaction allows
will have a direct positive relationship with perceived col-
team members to actively communicate among them-
lective efficacy.
selves and in the process they are able to: (a) discover
each member’s creative abilities, (b) share task-specific Furthermore, prior research shows that collective
knowledge, and (c) engage in discussions about alternative efficacy may emerge from common exposure of team
problem solving strategies (Bandura, 1997; Gibson, 1999; members to objective stimuli such as outcomes of team
Taggar and Seijts, 2003). These outcomes from interac- performance (Pethe, 2002; Lepper et al., 1986). Outcomes
tion will then influence team members’ collective efficacy of team performance have directional effect on individual
beliefs. Furthermore, as team members depend upon each members’ beliefs about the conjoint capabilities of their
other in the processes of organising and executing assigned team and also enables members to identify areas where
tasks, coordination of such interdependencies is crucial they need to cooperate to improve future outcomes. One
to the way these members use collective resources and way in which such stimuli can be provided is through
is therefore essential in developing members’ collective the diagnostic use of controls. As indicated earlier, diag-
efficacy beliefs (Bandura, 1997; Gibson, 1999; Lindsley nostic controls are feedback systems used to monitor
et al., 1995; Taggar and Seijts, 2003). outcomes and to provide direction to subordinates. The
K.M. Chong, H. Mahama / Management Accounting Research 25 (2014) 206–222 211
Table 1
Survey participants.
Analyst 5 1 3 8 0 17
Director 4 0 12 2 0 18
Other 0 1 13 6 3 23
Research associate 6 2 15 5 0 28
Research fellow 2 3 5 4 0 14
Research officer 1 1 2 1 0 5
Research/clinical coordinator 3 0 7 2 0 12
Senior analyst 6 0 39 3 0 48
Senior research associate 6 0 8 1 0 15
Senior research fellow 2 0 4 0 0 6
Total 35 8 108 32 3 186
n = 186.
a
Bioinformatics is a field of research that focuses on developing and applying algorithms, computational and statistical methods for the analysis of
biological data and for solving biological problems.
b
Green biotechnology focuses on developing and using environmentally-friendly solutions for agriculture, horticulture, and animal breeding processes.
c
Red biotechnology involves the manipulation and use of organisms for the improvement of medical processes.
d
White biotechnology is the application of biotechnology for industrial purposes, including manufacturing, biomaterials, and bioenergy.
questions. The remaining 186 participants completed the An independent sample t-test reveals no statistically
survey. Table 1 (Panel A) provides a summary of the data significant differences in means across all the primary
collection process. constructs.
Biotech firms are noted to be innovation intensive, Each participant was asked to nominate one specific
wherein they use integrated processes of biochemistry, biotech team he/she was currently working in and to
microbiology, and chemical engineering to research and respond to the survey based on the nominated team. The
develop new and innovative products (Oliver, 2001; 186 respondents had a mean of 3.67 (std. dev. 2.95) years
Powell et al., 1996). The innovation processes in biotech working as part of the teams they nominated for the survey
firms occur mainly through research teams as “it is and a mean of 6.16 (std. dev. 4.87) years working for their
unusual to work alone” in these firms (American Chemical current employer. 52 (27.96%) respondents were female.
Society, 2010). Although biotech teams may be struc- Table 1 (Panel B) reports the descriptive statistics of the
tured differently, they mostly comprise a hierarchy respondents’ demographics. Table 1 (Panel C) summarises
of senior scientists/researchers, research associates and the respondents’ job descriptions and the biotech sub-
research assistants (American Chemical Society, 2010). industry classifications for their companies. We conducted
Our survey sample captures this broad hierarchy of a difference in means test to determine whether the biotech
biotech team membership, with the 186 respondents sub-industry was associated with our primary constructs
representing team membership from different levels of of interest. The largest sub-industry group in our sample
the hierarchy (see Table 1, Panel C). We conducted was “Red biotechnology (medical)” and we classified that
a difference in means test to determine whether the as one group and all other sub-industry groups were com-
respondents’ rank within the biotech team was asso- bined into a second group. We coded “Red biotechnology
ciated with our primary constructs of interest. Senior (medical)” as 1 and all other sub-industry groups as 2 and
researchers (i.e., senior analyst, senior research asso- compared the means of these two groups across all the
ciate, and senior research fellow) were treated as one primary constructs. We found no statistically significant
group and the others were classified into a second group. differences in means across all the primary constructs.
K.M. Chong, H. Mahama / Management Accounting Research 25 (2014) 206–222 213
We also collected information about the size of respon- formative indicators of the latent construct. Wilcox et al.
dents’ teams (SIZE) and the duration the respondents had (2008) argue that latent constructs themselves are nei-
being working as part of that team (TENURE). We antic- ther inherently formative nor reflective, which suggests
ipated that SIZE may have effects on all of our primary that a given construct can either be measured using for-
constructs of interest (i.e., the extent of interactive and mative or reflective indicators, depending on an informed
diagnostic uses of budgets, perceived collective efficacy, choice by the researcher. Such a choice should be con-
and team effectiveness) and that TENURE could also have ceptually driven. That is, it should depend on how the
effects on perceived collective efficacy and team effective- researcher conceptualises the relationship between the
ness. Even though we did not develop any hypothesis for latent construct and their respective indicators and the
the effects of SIZE and TENURE, we included them in our relationship among the indicators (Diamantopoulos et al.,
statistical model as covariates. 2008; Wilcox et al., 2008). Where the researcher meas-
Prior studies (e.g., Widener, 2007) also suggest opera- ures indicators that are manifestations or effects of the
tional risk as a possible covariate. Given that our sample is latent construct, then causality is assumed to be from
drawn from the same industry, we assume that the compa- the latent construct to the indicators and hence a reflec-
nies will experience relatively similar levels of operational tive indicator measurement model is appropriate (Bisbe
risk and therefore, operational risk will not be a confound- et al., 2007; Diamantopoulos et al., 2008; Mackenzie et al.,
ing factor. This assumption is consistent with Ittner et al. 2005). Alternatively, where the researcher measures indi-
(2003), who argue that drawing a survey sample from a cators that “represent the defining characteristics that
single industry helps to implicitly control for a myriad collectively explain the meaning of the construct, a for-
of confounding factors that could affect the results if the mative indicator measurement model should be specified”
sample is a cross section of industries. To confirm this (Mackenzie et al., 2005, p. 713). Also, whilst the indi-
assumption, we sought the views of two academics whose cators are expected to covary and to share the same
research focuses on the biotech industry and one practi- antecedents/consequences in a reflective indicator mea-
tioner in this industry about the differences in operational surement model, the indicators in a formative indicator
risk across the different biotech sub-industry types listed in measurement model are not necessarily expected to covary
Table 1. There is a consensus among them that there are no or have the same antecedents/consequences (Bisbe et al.,
significant differences in operational risk across the differ- 2007; Diamantopoulos et al., 2008; Mackenzie et al., 2005).
ent biotech sub-industries. Another possible confounding Informed by the above developments in the literature,
factor in studies of interactive use of control is culture, but we develop reflective indicators for all our theoretical con-
consistent with Ittner et al. (2003), we expect our focus on structs. We measured each construct using multi-item
a single industry in a single country (the U.S.) to implicitly indicators; adapting measures from established scales (see
control for the confounding effect of culture. Appendix B for the measures used for each construct).
Following Dillman (2007), the questions for the online Each indicator (question) was anchored on a 7-point Likert-
survey were presented in a format similar to those of type scale. The indicators for each construct are discussed
paper-based self-administered questionnaires. The ques- below:
tions were presented in a screen-by-screen format, with
each screen representing a separate section of the ques- 3.2.1. Interactive use of budgets
tionnaire. The design of each screen was such that all As indicated in the literature review and hypotheses
questions had to be completed before the respondent could development section, a budget that is used diagnostically
proceed to the next screen. Where the answers on a given can simultaneously be used more or less interactively. Con-
screen were incomplete, a pop-up message appears to sequently, we measure the interactive use of budgets and
highlight unanswered questions and to encourage the par- diagnostic use of budgets as separate constructs rather than
ticipant to answer those questions before proceeding to the two ends of a continuum. To measure the extent of inter-
next screen. As a result, there was no missing data in the active use of budgets, we adapted the seven-item scale
responses. developed by Henri (2006). Two of these seven items (INT
3 and INT4) are excluded from our final statistical model
3.2. Measurement model specification as we could not link them to the concept of interactive
control systems outlined in Section 2.3. Jarvis et al. (2003)
The constructs of theoretical interest in this paper are and Rossiter (2002) argue that scale items in a reflective
latent in nature that is, they cannot be observed or mea- construct are interchangeable and that adding or dropping
sured directly. We therefore measure them indirectly by an item does not change the conceptual domain of the
means of observable multi-item indicators. Recent devel- construct. Consistent with this, eliminating the two meas-
opments in social research shows that one of two strategies ures did not significantly affect the measurement model
may be used when constructing multi-item indicators for interactive use of budget and did not change the signif-
for latent constructs, depending on the researcher’s con- icance and direction of the results in the structural model.
ceptualisation of the focal construct (Bisbe et al., 2007; The remaining five items were used as they sought to cap-
Diamantopoulos and Siguaw, 2006; Diamantopoulos and ture interactive use of budgets such as senior managers
Winklhofer, 2001; Henseler et al., 2009; Mackenzie et al., use of budgets to: (i) involve themselves in the activi-
2005). The first strategy involves measuring the indica- ties of the teams through meetings; (ii) enable debate
tors as reflective indicators of the latent construct and and to encourage continual challenge of data, underlying
the second strategy involves measuring the indicators as assumptions and action plans of the team; and (iii) facilitate
214 K.M. Chong, H. Mahama / Management Accounting Research 25 (2014) 206–222
team-consensus building by focusing attention on common The PLS approach is suitable for small sample sizes. The
issues and critical success factors. The instrument also cap- minimum sample size for PLS modelling is ten times the
tures the involvement of team members. However, similar largest regression in the model (Chin and Newsted, 1999).
to Bisbe and Malagueno (2009), our instrument did not In this study, the construct requiring the most complex
capture the focus on strategic uncertainties. regression is team effectiveness (with four paths leading to
this construct), suggesting a minimum sample size of 40.
3.2.2. Diagnostic use of budgets Our sample size of 186 is therefore considered adequate for
We adapted the four-item scale developed by Henri PLS modelling. We used SmartPLS Release 2 (Ringle et al.,
(2006) to measure the extent to which budgets were used 2005) to simultaneously estimate the measurement and
diagnostically to manage teams. The measures focus on the structural models.
extent to which budgets are used to track team’s progress
towards goals, to monitor the team’s results, to compare 4.1. Results of the measurement model
the team’s outcome with expectations, and to review the
team’s key performance measures. Considering that the constructs in this study are latent
constructs, we evaluated the reliability and/or validity of
3.2.3. Perceived collective efficacy their indicators. We assessed the measurement models for
Whiteoak et al. (2004) argue that an important con- reliability (individual item reliability and composite reli-
sideration when using any of the measures of collective ability) and validity (convergent and discriminant validity)
efficacy is that the measure must have a group or team using traditional classical test theory (Hulland, 1999). We
referent, even if the group or team referent is implicit. As assessed individual item reliability by examining the factor
collective efficacy manifests itself through confidence in loading of each scale-item. Adequate reliability is demon-
a team’s capability and competence, we ensure that the strated when factor loadings for a construct are 0.70 or
measures used in this study had team referents and also higher (Nunnally, 1978). All the indicators for the measure-
captured respondents’ estimate of their belief in the teams’ ment models loaded higher than 0.78 on their respective
capability and competence. In this regard, the respondents’ constructs with p-values less than 0.0001 (see Table 2),
perceptions of their teams’ collective efficacy was mea- demonstrating adequate reliability. In addition, compos-
sured using a four-item scale adapted from Salanova et al. ite reliability of 0.70 or higher is considered acceptable.
(2003). Each of the measurement models had a composite reli-
ability higher than 0.87 (see Table 3), indicating acceptable
composite reliability.
3.2.4. Team effectiveness
We assessed the convergent validity of the mea-
Conceptually, team effectiveness is assessed based on
surement models by examining each construct’s average
the quality, accuracy and quantity of work done; the
variance extracted (AVE). An AVE of 0.5 indicates accept-
timeliness of, and satisfaction with, work performed; and
able convergent validity (Chin, 1998; Hulland, 1999). As
operating efficiency. We measure team effectiveness by
shown in Table 3, the AVEs for all constructs in our model
adapting the six-item scale from Kathuria and Davis (2001).
were higher than 0.64, demonstrating adequate conver-
Given that the respondents in this survey were anony-
gent validity. We also assessed discriminant validity by
mous, it was not possible to obtain actual or supervisor
comparing the square roots of each construct’s AVE to the
ratings of team effectiveness measures. We therefore relied
correlations among the construct. Table 3 reports the corre-
on self-ratings of team effectiveness. Though self-rating of
lation coefficients in the off-diagonal and the square roots
performance is considered subjective, existing accounting
of the AVE in the diagonal. All the square roots of AVE were
research suggests that subjective measures are not worse
greater than the correlation coefficients, suggesting accept-
than objective measures in survey studies (Abernethy and
able discriminant validity. The cross-loading of scale items
Stoelwinder, 1991).
also show that the items load higher on the construct they
are intended to measure than on any other construct (see
4. Statistical analysis and results Table 2), thus providing further evidence for discriminant
validity.1
We analysed the data obtained from the survey
using the partial least squares (PLS) approach. PLS is a 4.2. Common method bias
component-based latent variable modelling technique that
aims to maximise variance explained while minimising Common method bias is a concern in self-report surveys
measurement errors. It enables path models involving in which the same rater responds to all questionnaire items.
latent variables to be estimated, where the latent vari- This concern is more serious in situations where both the
ables are indirectly measured by multiple indicators (Chin independent and dependent latent constructs are percep-
et al., 2001). PLS is able to simultaneously examine theory tual measures derived from the same rater (Podsakoff and
(structural model) and measures (measurement model). Organ, 1986). Given that the independent and dependent
Unlike covariance-based structural equation modelling,
PLS makes no distributional assumption about data. There-
fore, the R2 is considered the most appropriate statistic for 1
Chin (1998) argues that discriminat validity is demonstrated in cross-
assessing the overall predictiveness of a model and boot- loadings when scale-items load higher with their respective constructs
strap re-sampling is used to test the significance of results. than any other scale-items from other constructs.
K.M. Chong, H. Mahama / Management Accounting Research 25 (2014) 206–222 215
Table 2
Measurement model: factors, descriptive statistics, cross-loadings, and t-statistics.
Latent constructs Scale items Actual range Mean Standard deviation Factor 1 Factor 2 Factor 3 Factor 4 t-statistics
All item loadings on their respective constructs are statistically significant (p < 0.0001, one-tailed).
n = 186.
Table 3
Composite reliability, Cronbach’s alpha, and average variance extracted (AVE), discriminant validity and correlations.
Latent construct Composite reliability Cronbach’s alpha AVE Correlations and SQRT of AVEa
n = 186.
a
Diagonal elements are square roots (SQRT) of AVE; off-diagonal elements are correlations between constructs.
variables in our survey instrument were perceptual meas- this approach using PLS, we followed the procedure out-
ures and that responses to these measure were derived lined and implemented in Liang et al. (2007) and Mahama
from the same rater, there is a possibility that common and Cheng (2013). We created the common method fac-
method bias exist in our data. We adopt two approaches tor by using all the indicators for the substantive (main)
suggested by Podsakoff et al. (2003) to address concerns constructs in our model. To include the common method
with common method bias. First, we conducted Harman’s factor in our PLS model, Liang et al. (2007) suggest that
single-factor test in which all the questions measuring our we model the indicators of our substantive constructs
constructs were entered into an unrotated exploratory fac- as single-indicator latent variables (first-order constructs)
tor analysis. A common measure bias is said to be present and the substantive constructs as second-order constructs
when (a) a single factor emerge from the factor analysis of their respective single-indicator variables. The common
or (b) one factor accounts for the majority of the covari- method factor is then included in the PLS model, with
ance among the variables (Podsakoff and Organ, 1986). The links to all single-indicator variables in the model. The
test did not find evidence of common method bias since we path coefficients between the substantive constructs and
found a multi-factor rather than a single factor solution and the single-indicator variables as well as those between
the first factor explained only 34.6% of the total variance.2 the common method factor and the single-indicator vari-
Second, we included a latent common method fac- ables are then interpreted as the factor loadings of the
tor in our PLS model to assess the presence of common observed indicators (Liang et al., 2007; Marcoulides and
method bias and to partial out its effect. To implement Moustaki, 2002). To assess whether common method bias
is a concern, we examine the statistical significance of fac-
tor loadings of the method factor and also compared the
2
Though the Harman’s test is widely used, there are no general guide- percentage variance (measured as the squared values of
lines about the number of factors a researcher should expect from the the factor loadings) of each observed indicator explained
factor analysis. Therefore, the absence of a single factor is generally by its substantive construct and by the common method
deemed to be a sufficient measure of the absence of common measure
factor. As shown in Table 4, the method factor loadings
bias (Podsakoff and Organ, 1986).
216 K.M. Chong, H. Mahama / Management Accounting Research 25 (2014) 206–222
Table 4
Common method bias analysis with latent common method variance factor (LCMVF).
Scale Items Substantive factor Variance explained Method factor Variance explained
loading (R1) by substantive loading (R2) by method factor
factor (R12 ) loading (R22 )
n = 186.
*
p < 0.05 (one-tailed).
*****
p < 0.0001 (one-tailed).
are insignificant (except for two items) and the percent- This confirms a direct positive relationship and thus pro-
age variances of the observed indicators explained by the vides support for hypothesis 1a. Hypothesis 1b predicted a
substantive constructs (average variance = 0.725) are sub- positive direct relationship between the diagnostic use of
stantially greater than the percentage variances explained budgets and team effectiveness. This prediction is not sup-
by the common method factor (average variance = 0.012). ported as the path coefficient was negative (−0.102) and
The ratio of substantive variance to common method insignificant.
variance is about 60:1. This test provides evidence that Hypothesis 2a predicted a direct positive relationship
common method bias is unlikely to be a serious concern between the extent of interactive use of budgets and
for this study (Liang et al., 2007). perceived collective efficacy. We find support for this
In addition, the results of the PLS structural model prediction. The path coefficient between the extent of
(when the common method factor is included) are very interactive use of budgets and perceived collective efficacy
similar to those of our substantive model. In particular, is positive (ˇ = 0.322) and significant at the 0.05 level. This
the R2 s for the two models are the same and this indicates significant result implies that the dialogue and knowledge
that the inclusion of the common method factor does not sharing associated with a greater extent of interactive use
influence the predictiveness of our structural model, thus of budgets improves team members’ beliefs in their teams’
providing further evidence that our results are unlikely to conjoint capabilities to organise and execute a given task.
be affected by common method bias. Consequently, we dis- Hypothesis 2b predicted that the diagnostic use of budgets
cuss only the PLS results for the substantive model. The will be positively related to perceived collective efficacy.
results of the PLS structural model (when the common Unfortunately, this hypothesis is not supported as the path
method factor is included) are reported in Appendix C. coefficient is negative (−0.055) and insignificant.
We speculate that the negative and insignificant results
4.3. Results of the structural model for hypotheses 1b and 2b can be attributed to the nature of
the research setting (i.e., biotechnology firms). This specu-
We tested our hypotheses using the PLS structural lation is based on prior studies that find the diagnostic use
model. The results of the structural model are summarised of controls acts to constraint creativity in innovation sett-
in Table 5 (Panel A shows direct relationships, Panel B ings and thus tends to have negative impacts on motivation
shows indirect effects, and Panel C shows total effects). and performance (Henri, 2006; Simons, 2000). Also, in an
Hypothesis 1a predicted a direct positive relationship innovation environment, targets (or predicted outcomes of
between the extent of interactive use of budgets and team team performance) are unlikely to be as well fixed com-
effectiveness. The structural path coefficient leading from pared to a non-innovation environment (i.e., the targets are
the extent of interactive use of budgets to team effec- vague and ambiguous, and more likely to provide guidance
tiveness is positive (ˇ = 0.314) and significant (p < 0.005). rather than a definitive aim) thereby limiting the effects of
K.M. Chong, H. Mahama / Management Accounting Research 25 (2014) 206–222 217
Table 5
PLS structural model results.
Interactive use of budgets Diagnostic use of budgets Perceived collective efficacy Team effectiveness
Panel B: Indirect (mediated) effects – path coefficient and bootstrap confidence interval
Lower Upper
n = 186.
*
p < 0.05 (one-tailed).
**
p < 0.01 (one-tailed).
***
p < 0.005 (one-tailed).
****
p < 0.001 (one-tailed).
*****
p < 0.0001 (one-tailed).
the diagnostic use of controls. Given that the primary work step-wise approach. In the first step, we tested a model
in biotechnology settings is about creativity and innova- involving the direct relationship between the extent of
tion, we cannot rule out the effect of the research setting interactive use of budgets and team effectiveness. This
on the negative and insignificant path coefficients observed yielded a positive (ˇ = 0.532) and significant (t = 3.1727;
for the relationships involving the diagnostic use of budgets p < 0. 001) result, thus supporting a direct relationship.3 In
construct and our dependent variables (perceived col- the second step, we tested a model of a direct relation-
lective efficacy and team effectiveness). Future research ship between the extent of interactive use of budgets and
could examine the model tested in this paper in sett- perceived collective efficacy and found a direct positive
ings other than those primarily involved in innovation and (ˇ = 0.327) and significant (t = 2.1180; p < 0.05) relationship.
creativity. The third step examined a model involving a direct rela-
Hypothesis 3 predicted a direct positive relationship tionship between perceived collective efficacy and team
between perceived collective efficacy and team effective- effectiveness, while controlling for the effects of the extent
ness. This prediction is supported by the results of the of interactive use of budgets (and also diagnostic use
structural model where the path coefficient leading from of budgets) on team effectiveness. The results provided
perceived collective efficacy to team effectiveness is pos- support for a direct positive (ˇ = 0.703) and significant
itive (ˇ = 0.702) and significant (p < 0.0001). The highly (t = 13.0474; p < 0.0001) relationship. Given that the results
significant result and the relatively large path coefficient from these first three steps satisfied the conditions for the
(ˇ = 0.702) for this prediction suggest that collective effi- mediation test, we proceeded to test a model involving all
cacy has an important role in work teams. This result is the relationships among the three constructs. The results
consistent with the findings of team studies that have of this test are reported in Table 5 and in Fig. 2. Whilst all
examined the relationship between collective efficacy and three direct paths remain positive and significant, we find
team outcomes (Chen and Bliese, 2002; Lent et al., 2006; that the significance of the direct path leading from inter-
Wang and Lin, 2007). Given that in hypothesis 2 there is active use of budgets to team effectiveness is reduced when
a significant positive relationship between the extent of perceived collective efficacy is included in the model. The
interactive use of budgets and perceived collective efficacy, t-values fell from 3.1727 (ˇ = 0.532) to 2.6963 (ˇ = 0.314)
the result for hypothesis 3 makes the interactive use of and the associated p-values also changed from p < 0.001
budgets in a team context salient.
Following from the test of our main hypotheses, we
further examined whether perceived collective efficacy 3
The reliability and validity of the measurement models for the models
also mediates the positive relationship between the extent tested in the first three steps of the Baron and Kenny (1986) mediation
of interactive use of budgets and team effectiveness. To test procedure were consistent with the measurement model results of
the full PLS model. Therefore, we have reported only the reliability and
test this mediation, we follow Baron and Kenny’s (1986)
validity for the full model.
218 K.M. Chong, H. Mahama / Management Accounting Research 25 (2014) 206–222
B.3. Perceived collective efficacy (COLLECT) American Chemical Society, 2010. Biotechnology is the science
of the future. http://portal.acs.org/portal/acs/corg/content?
nfpb=true& pageLabel=PP ARTICLEMAIN&node id=1188&content id=
The response scale for the following items ranged from CTP 003377&use sec=true&sec url var=region1& uuid=b43431d5-
1 (strongly disagree) to 7 (strongly agree). 034a-44b3-a32b-fd200b219dd1. Downloaded on 21 November 2010.
Bandura, A., 1986. Social Foundations of Thought and Action. Prentice-
Hall, Englewood Cliffs, NJ.
1. I feel confident about the capability of my team to per- Bandura, A., 1997. Self-efficacy: The Exercise of Control. Freeman, New
form the tasks very well. York.
2. My team is able to solve difficult tasks if we invest the Bandura, A., 1999. Social cognitive theory: an agentic perspective. Asian
Journal of Social Psychology 2, 21–41.
necessary effort. Bandura, A., 2000. Exercise of human agency through collective efficacy.
3. I feel confident that my team will be able to manage Current Directions in Psychological Science 9, 75–78.
effectively unexpected troubles. Bandura, A., 2001. Social cognitive theory: an agentic perspective. Annual
Review of Psychology 52, 1–26.
4. My team is totally competent to solve the task.
Baron, R.M., Kenny, D.A., 1986. The moderator–mediator variable dis-
tinction in social psychological research: conceptual, strategic and
B.4. Team effectiveness (TEFFECT) statistical considerations. Journal of Personality and Social Psychology
51, 1173–1182.
Baron, R.S., 1986. Distraction/conflict theory: progress and problems. In:
The response scale for the following items ranged from Berkowitz, L. (Ed.), Advances in Experimental Social Psychology, vol.
1 (not satisfied) to 7 (very satisfied). 19. New York, Academic Press.
Bisbe, J., Batista-Forhuet, J.M., Chenhall, R., 2007. Defining management
How satisfied are you with your team’s performance on
accounting constructs: a methodological note on the risk of con-
the following: ceptual misspecification. Accounting, Organizations and Society 32,
789–820.
Bisbe, J., Malagueno, R., 2009. The choice of interactive control systems
1. Accuracy of work performed
under different innovation management modes. European Account-
2. Quantity of work performed ing Review 18, 371–405.
3. Quality of work performed Bisbe, J., Otley, D., 2004. The effects of the interactive use of management
4. Operating efficiency control systems on product innovation. Accounting, Organizations
and Society 29, 709–737.
5. Client satisfaction Bray, S.R., 2004. Collective efficacy, group goals, and group performance
6. Timeliness in meeting delivery schedules of a muscular endurance task. Small Group Research 35, 230–238.
Chen, G., Bliese, P.D., 2002. The role of different levels of leadership in pre-
dicting self- and collective efficacy: evidence for discontinuity. Journal
Appendix C. Results for PLS structural model (when
of Applied Psychology 87, 549–556.
latent common method factor is included) Chen, G., Gully, S.M., Whiteman, J.A., Kilcullen, R.N., 2000. Examina-
tion of relationships among trait-like individual differences, state-like
Latent Path to: individual differences, and learning performance. Journal of Applied
construct Psychology 85, 835–847.
Interactive Diagnostic Perceived Team Chenhall, R.H., 2008. Accounting for the horizontal organization: a review
use of use of collective effectiveness essay. Accounting, Organizations and Society 33, 517–550.
budgets budgets efficacy Chin, W.W., 1998. The partial least squares approach to structural
Interactive use 0.284 (1.8713)* 0.309 (2.8069)*** equation modeling. In: Marcoulides, G.A. (Ed.), Modern Methods
of budgets for Business Research. Lawrence Erlbaum Associates, NJ, pp. 295–
Diagnostic use −0.022 (0.1439) −0.104 (1.0489) 336.
of budgets Chin, W.W., Mathieson, K., Peacock, E., 2001. Extending the technology
Perceived 0.705 (13.2614)***** acceptance model: the influence of perceived user resources. The
collective
DATA BASE for Advances in Information Systems 32, 86–112.
efficacy
SIZE −0.026 −0.057 0.078 (0.9614) −0.019 (0.34518)
Chin, W.W., Newsted, P.R., 1999. Structural equation modelling analysis
(0.2904) (0.6542) with small samples using partial least squares. In: Hoyle, R.H. (Ed.),
TENURE 0.162 (2.3378)** −0.067 (0.5199) Statistical Strategies for Small Sample Research. Sage Publications,
R2 0.1 0.62 Thousand Oaks, pp. 307–341.
Davila, T., 2000. An empirical study on the drivers of management control
n = 186
*
systems’ design in new product development. Accounting, Organiza-
p < 0.05 (one-tailed). tions and Society 25, 383–409.
**
p < 0.01 (one-tailed). Diamantopoulos, A., Riefler, P., Roth, K.P., 2008. Advancing forma-
*** tive measurement models. Journal of Business Research 61, 1203–
p < 0.005 (one-tailed).
**** 1218.
p < 0.001 (one-tailed).
*****
p < 0.0001 (one-tailed). Diamantopoulos, A., Siguaw, J.A., 2006. Formative versus reflective indi-
cators in organizational measure development: a comparison and
empirical illustration. British Journal of Management 17, 263–282.
Appendix D. Supplementary data Diamantopoulos, A., Winklhofer, H.M., 2001. Index construction with for-
mative indicators: an alternative to scale development. Journal of
Supplementary data associated with this article can be Marketing Research XXXVIII, 269–277.
Dillman, D.A., 2007. Mail and Internet Surveys: The Tailored Design
found, in the online version, at http://dx.doi.org/10.1016/
Method, second ed. John Wiley & Sons, New York.
j.mar.2013.10.008. Earley, C., 1993. East meets west Mideast: further exploration of collec-
tivstic and individualistic work groups. The Academy of Management
Journal 36, 319–348.
References
Gibson, C.B., 1999. Do they do what they believe they can? Group efficacy
and group effectiveness across tasks and cultures. The Academy of
Abernethy, M.A., Brownell, P., 1999. The role of budgets in organizations Management Journal 42, 138–152.
facing strategic change: an exploratory study. Accounting, Organiza- Gigone, D., Hastie, R., 1993. The common knowledge effect: information
tions and Society 24, 189–204. sharing and group judgement. Journal of Personality and Social Psy-
Abernethy, M.A., Stoelwinder, J.U., 1991. Budget use, task uncertainty, chology 65, 959–974.
system goal orientation and subunit performance: a test of the ‘fit’ Gigone, D., Hastie, R., 1997. The impact of information on small group
hypothesis in non-for-profit hospitals. Accounting, Organizations and choice. Journal of Personality and Social Psychology 72, 132–140.
Society 16, 105–120.
K.M. Chong, H. Mahama / Management Accounting Research 25 (2014) 206–222 221
Goddard, R.D., Hoy, W.K., Hoy, A.W., 2004. Collective efficacy beliefs: Little, B.L., Madigan, R.M., 1997. The relationship between collective effi-
theoretical developments, empirical evidence, and future directions. cacy and performance in manufacturing work teams. Small Group
Educational Researcher 33, 3–13. Research 28, 517–534.
Gordon, J., 1992. Work teams – how far have they come? Training 29, Mackenzie, S.B., Podsakoff, P.M., Jarvis, C.B., 2005. The problem of model
59–65. misspecification in behavioural and organizational research and some
Gully, S.M., Incalcaterra, K.A., Joshi, A., Beaubien, J.M., 2002. A recommended solutions. Journal of Applied Psychology 90, 710–730.
meta-analysis of team-efficacy, potency, and performance: inter- Mahama, H., Cheng, M.M., 2013. The effect of managers’ enabling percep-
dependence and level of analysis as moderators of observed tions on costing system use, psychological empowerment, and task
relationships. Journal of Applied Psychology 87, 819–832. performance. Behavioral Research in Accounting 25, 89–114.
Guzzo, R.A., Dickson, M.W., 1996. Teams in organizations: recent research Marcoulides, G.A., Moustaki, I., 2002. Latent Variable and Latent Structure
on performance and effectiveness. Annual Review Psychology 47, Models. Lawrence Erlbaum Associates, Mahwah, NJ.
307–338. Mesmer-Magnus, J.R., DeChurch, L.A., 2009. Information sharing and team
Hayes, A.F., 2009. Beyond Baron and Kenny: statistical mediation anal- performance: a meta-analysis. Journal of Applied Psychology 94,
ysis in the new millennium. Communication Monographs 76, 408– 535–546.
420. Mohammed, S., Dumville, B.C., 2001. Team mental model in team
Henri, J., 2006. Management control systems and strategy: a resource- knowledge framework: expanding theory and measurement across
based perspective. Accounting, Organizations and Society 31, disciplinary boundaries. Journal of Organizational Behavior 22,
529–558. 89–106.
Henry, R.A., 1995. Improving group judgment accuracy: information shar- Mulvey, P.W., Klein, H.J., 1998. The impact of perceived loafing and col-
ing and determining the best member. Organizational Behavior and lective efficacy on group goal processes and group performance.
Human Decision Processes 62, 190–197. Organizational Behavior and Human Decision Processes 74, 62–87.
Henseler, J., Ringle, C.M., Sinkovics, R.R., 2009. The use of partial least Nunnally, J.C., 1978. Psychometric Theory. McGraw-Hill, New York.
squares path modelling in international marketing. Advances in Inter- Oliver, A.L., 2001. Strategic alliances and the learning life-cycle of biotech-
national Marketing 20, 277–319. nology firms. Organization Studies 22 (3), 467–489.
Hertenstein, J.H., Platt, M.B., 2000. Performance measures and manage- Otley, D., 1994. Management control in contemporary organizations:
ment control in new product development. Accounting Horizons 14, towards a wider framework. Management Accounting Research 5,
303–323. 289–299.
Hoigaard, R., Safvenbom, R., Tonnessen, F.E., 2006. The relationship Pescosolido, A.T., 2003. Group efficacy and group effectiveness: the effects
between group cohesion, group norms, and perceived social loafing of group efficacy over time on group performance and development.
in soccer teams. Small Group Research 37, 217–232. Small Group Research 34, 20–42.
Hulland, J., 1999. Use of partial least squares (PLS) in strategic manage- Pethe, S., 2002. Collective measure: development of a measure. Decision
ment research: a review of four recent studies. Strategic Management 29, 25–38.
Journal 20, 195–204. Podsakoff, P.M., MacKenzie, S.M., Lee, J., Podsakoff, N.P., 2003. Common
Ittner, C.D., Larcker, D.F., Randall, T., 2003. Performance implications method variance in behavioral research: a critical review of the liter-
of strategic performance measurement in financial service firms. ature and recommended remedies. Journal of Applied Psychology 88,
Accounting, Organizations, and Society 28, 715–741. 879–903.
Jarvis, C.B., Mackenzie,.S.B., Podsakoff, P.M., 2003. A critical review of Podsakoff, P.M., Organ, D.W., 1986. Self-reports in organizational research:
construct indicators and measurement model misspecification in problems and prospects. Journal of Management 12, 531–544.
marketing and consumer research. Journal of Consumer Research 30, Porter, C.O.L.H., Gogus, C.I., Yu, R.C.F., 2011. Does backing up behaviour
199–218. explain the efficacy-performance relationship in teams. Small Group
Jehn, K.A., Mannix, E.A., 2001. The dynamic nature of conflict: a longitudi- Research 42, 458–474.
nal study of intragroup conflict and group performance. Academy of Powell, W.W., Koput, K.W., Smith-Doerr, L., 1996. Interorganizational
Management Journal 44, 238–251. collaboration and the locus of innovation: networks of learning in
Jung, D.I., Sosik, J.J., 2002. Transformational leadership in work groups: biotechnology. Administrative Science Quarterly 41, 116–145.
the role of empowerment, cohesiveness and collective efficacy Prussia, G.E., Kinicki, A.J., 1996. A motivational investigation of group
on perceived group performance. Small Group Research 33, 313– effectiveness using socio-cognitive theory. Journal of Applied Psychol-
336. ogy 81, 187–198.
Kathuria, R., Davis, E.B., 2001. Quality and work force management Ringle, C.M., Wende, S., Will, A., 2005. SmartPLS Release 2. University of
practices: the managerial performance implication. Production and Hamburg, Hamburg.
Operations Management 10, 460–477. Rossiter, J.R., 2002. The C-OAR-SE procedure for scale development in
Kazub, S.A., McDonnell, J.F., 2000. Exploring the relationship between marketing. International Journal of Research in Marketing 19, 1–31.
cohesion and collective efficacy in rugby teams. Journal of Sports Salanova, M., Llorens, S., Cifre, E., Martínez, I.M., Schaufeli, W.B., 2003.
Behavior 23, 120–129. Perceived collective efficacy, subjective well-being and task perfor-
Kerr, N.L., 1983. Motivational losses in task-performing groups: a social mance among electronic work groups: an experimental study. Small
dilemma analysis. Journal of Personality and Social Psychology 45, Group Research 34, 43–73.
819–828. Sampson, R.J., Morenoff, J.D., Earls, F., 2000. Beyond social capital: spa-
Kirkman, B.L., Rosen, B., 1999. Beyond self-management: antecedents and tial dynamics of collective efficacy for children. American Sociological
consequences of team empowerment. Academy of Management Jour- Review 64, 633–660.
nal 42, 58–74. Schnake, M.E., 1991. Equity in effort: the sucker effect in co-acting groups.
Lee, C., Tinsley, C.H., Bobko, P., 2002. An investigation of the antecedents Journal of Management 17, 41–55.
and consequences of group-level confidence. Journal of Applied Social Scott, T.W., Tiessen, P., 1999. Performance measurement and managerial
Psychology 32, 1628–1652. teams. Accounting, Organizations and Society 24, 263–285.
Lent, R.W., Schmidt, J., Schmidt, L., 2006. Collective efficacy beliefs in stu- Simons, R., 1990. The role of management control systems in creating
dent work teams: relation to self-efficacy, cohesion, and performance. competitive advantage: new perspectives. Accounting, Organizations
Journal of Vocational Behavior 68, 73–84. and Society 15, 127–143.
Levine, J.M., Resnick, L.B., Higgins, E.T., 1993. Social foundations of cogni- Simons, R., 1991. Strategic orientation and top management attention to
tion. Annual Review Psychology 44, 585–612. control systems. Strategic Management Journal 12, 49–62.
Lepper, M.R., Ross, L., Lau, R.R., 1986. Persistence of inaccurate beliefs Simons, R., 1994. How new top managers use control systems as levers of
about the self: perseverance effects in the classroom. Journal of per- strategic renewal. Strategic Management Journal 15, 169–189.
sonality and social psychology 50, 482–491. Simons, R., 1995. Levers of Control. Harvard Business School Press, Boston,
Liang, H., Saraf, N., Hu, Q., Xue, Y., 2007. Assimilation of enterprise sys- MA.
tems: the effect of institutional pressures and the mediating role of Simons, R., 2000. Performance Measurement and Control Systems for
top management. MIS Quarterly 31, 59–87. Implementing Strategy. Prentice Hall, Upper Saddle River, NJ.
Libby, T., Thorne, L., 2009. The influence of incentive structure on group Taggar, S., Seijts, G.H., 2003. Leader and staff role-efficacy as antecedents
performance in assembly lines and teams. Behavioral Research in of collective efficacy and team performance. Human Performance 16,
Accounting 21, 57–72. 131–156.
Lindsley, D.H., Brass, D.J., Thomas, J.B., 1995. Efficacy – performance spi- Towry, K.L., 2003. Control in a teamwork environment: the impact of
rals: a multilevel perspective. Academy of Management Review 20, social ties on the effectiveness of mutual monitoring contracts. The
645–678. Accounting Review 78, 1069–1095.
222 K.M. Chong, H. Mahama / Management Accounting Research 25 (2014) 206–222
Wang, S., Lin, S.S.J., 2007. The effects of group composition of self-efficacy Widener, S.K., 2007. An empirical analysis of the levers of control frame-
and collective efficacy on computer-supported collaborative learning. work. Accounting, Organizations and Society 32, 757–788.
Computers in Human Behavior 23, 2256–2268. Wilcox, J.B., Howell, R.D., Breivik, E., 2008. Questions about formative
Watson, C.B., Chemers, M.M., Preiser, N., 2001. Collective efficacy: a measurement. Journal of Business Research 61, 1219–1228.
multilevel analysis. Personality and Social Psychology Bulletin 27, Wood, R., Bandura, A., 1989. Social cognitive theory of organizational man-
1057–1068. agement. Academy of Management Review 14, 361–384.
Whiteoak, J.W., Chalip, L., Hort, L.K., 2004. Assessing group efficacy: com- Young, S.M., Fisher, J., Lindquist, T.M., 1993. The effects of intergroup
paring three methods of measurement. Small Group Research 35, competition and intragroup cooperation on slack and output in a
158–173. manufacturing setting. The Accounting Review 68, 466–481.