Perceptions of Integrity, Competence and Trust in
Senior Management as Determinants of Cynicism Toward Change
Simon
L Albrecht
School
of Psychology
Curtin University of Technology
Abstract
As Public Sector
organizations continue to undergo transformational change, it is important to
identify organizational factors that impact on employee attitudes to change.
There is limited empirical evidence about the determinants of cynicism toward
change. In this paper, a model is proposed which identifies three key
trust-related antecedents of cynicism toward change: perceptions of integrity,
competence, and trust in senior management. Data were collected from two Public
Sector organizations to identify levels and correlates of cynicism toward
change and to test the proposed model. The results of confirmatory factor
analysis and structural equation modeling suggest that perceptions of integrity
and trust in senior management influence cynicism toward change. Integrity, as
an element of trustworthiness, was found to directly influence trust in senior
management as well as cynicism toward change. Contrary to expectations,
employee perceptions of the competence of senior management did not to have a
direct influence on trust nor on cynicism toward change. Collectively the
antecedent variables accounted for approximately forty percent of the variance
in cynicism toward change. In general terms, the findings will prove helpful to
human resource practitioners interested in diagnosing and managing
organizational trust and attitudes toward organizational change.
Introduction
Senior management clearly plays a
key role in organizational change. In both the private and the public sector,
it is senior management who take the primary responsibility and accountability
for designing, and managing organizational change. It is widely acknowledged
(Halligan & Power, 1992; Hughes, 1994; O'Neill, 2000; Pollitt, 1993) that
public sector executives have overseen fundamental changes in the size,
structure and administration of public sector functions over the past decade or
so.
Recent
literature suggests a strong relation between trust in senior management and
employee attitudes to change (Rousseau & Tijoriwala, 1999; Kramer, 1996;
Kanter & Mirvis, 1989). The research suggests that trust is a central
factor in the way that employees experience aspects of organizational change.
In this paper an attempt is made to empirically identify trust-related factors
that predispose employees to be less cynical toward change. Consideration will
be given to both trust and trustworthiness, with a particular focus on
integrity and competence.
Trust and Trustworthiness in Organizational
Settings
The influence of trust in organizational
settings has attracted increasing interest in recent years (Clark & Payne,
1997; Dirks & Ferrin, 2001; Kramer & Tyler, 1996; Mayer & Davis,
1999; Mayer, Davis & Schoorman, 1995).
An entire issue of the Academy of Management Review (1998) was devoted
to the topic, as was a chapter from a recent Annual Review of Psychology
(Kramer, 1999). Researchers and
practitioners continue to recognize that trust is an important factor in
determining organizational success, organizational stability and the well-being
of employees (Cook & Wall, 1980; Tyler & Kramer, 1996; Shaw, 1997).
Trust has been defined as a
“willingness to engage in risk-taking with a focal party” (Mayer & Davis,
1999, p. 124). Rousseau, et al. (1998) defined trust as “a psychological state
comprising the intention to accept vulnerability based on the positive
expectations of the intentions or behavior of another” (p. 395). McAllister (19950 defined trust “the extent
to which a person is confident in, and willing to act on the basis of, the
words, action, and decisions of another” (p. 25). A number of elements are
common to these definitions. First
trust is a psychological state represented by a ‘willingness’ to behave or
act. Second, vulnerability, uncertainty
and risk are important components of trust. Third, trust emerges from
‘conditions’ such as ‘positive expectations’ or performance of particular
actions.
The ‘conditions’ that enable trust, as
distinct from trust itself, define trustworthiness. While, as previously argued, trust is a psychological state
defined in terms of a willingness to act in the face of uncertainty,
trustworthiness refers to a set of attributions or beliefs about the motives,
intentions, qualities or ‘trustworthiness’ of some influential other. Mayer Davis and Schoorman (1995, p. 717)
argued that the “characteristics of the trustee that determine trustworthiness
... although not trust per se, ... help build the foundation for the
development of trust.” Evaluations of
trustworthiness are thus central to the conceptualisation of trust (Kramer,
1996).
A review of the literature reveals
that trustworthiness has been conceptualized in a variety of ways. It has been variously defined as having one
dimension (Barber, 1983), two dimensions (Cook & Wall, 1980 ), three
dimensions (Cummings & Bromiley, 1996),
four dimensions (Mishra, 1996), five dimensions (Clark & Payne, 1997), and
up to ten dimensions (Butler, 1991). While perceived competence, integrity, and
benevolence have consistently been cited as markers of trustworthiness (Mayer,
Davis & Schoorman, 1995; Tan & Tan, 2000), Butler (1981) argued that
integrity is the condition most similar to and important to trust. Davis,
Schoorman, Mayer and Tan (2000) found that trust was significantly related to
employees’ perceptions of their manager’s integrity and competence.
Integrity
refers to employees perceiving that trust referents espouse, demonstrate and
adhere to a set of principles and values that the trustor finds acceptable
(Mayer & Davis, 1999).
Trustworthiness can be indexed by care and concern for employee
well-being, truth telling, consistency, fairness and a willingness to share
information openly.
Competence refers to qualities such as
influence, impact, ability, expertness, knowledge and the ability to do what is
needed. Trustworthiness is, in part,
indexed by the extent to which trust referents are perceived as being
competent. Mayer and Davis (1999) described competence as “that group of
skills, competencies and characteristics that allow a party to have influence
within some domain” (p. 124). Different competence would be appropriate
depending on the trust referent.
Senior Management as a Focus for Trust
A wealth of anecdotal evidence abounds
as to the importance of establishing high levels of trust between senior
management and employees (Horton & Reid, 1991; Shaw, 1997). However, as previously noted, there has been
relatively little research focused on the antecedents and consequences of trust
in senior management, particularly with respect to the Public Sector. Cook and
Wall (1980), Mayer and Davis (1999), McCauley and Kuhnert (1992), Carnevale and
Wechsler (1992), Costigan, Ilter and Berman (1998), and Tan and Tan (2000), are
among the few researchers who have studied trust in ‘management’, ‘top management’,
‘senior management’, ‘the CEO and top management’ or the ‘organization’. Given
the impact that managerial decisions can have on the physical and psychological
conditions for employees (Nelson, Cooper & Jackson, 1995; Schweiger &
De Nisi, 1991), this seems an important area for investigation.
The trust referent for the present
research was clearly specified as ‘senior management’. Less specific referents,
such as ‘management’ or the ‘organization’, may lead to ambiguity in the minds
of survey respondents (Costigan, Ilter & Berman), and therefore introduce
measurement error when operationalising the construct and when investigating
relationships with other factors. Senior management, for the purposes of the
present study, was defined as members of the organizational strategic and
operational executive. The group would
normally be comprised of the CEO, Directors, Divisional Heads, and Functional
Heads, all of whom directly contribute to the organization’s key strategic and
business decisions.
Consistent with the conceptualization
of trust previously described, trust in senior management was hereby defined as
“the willingness of employees to act on the word, actions or decisions of
senior management under conditions of uncertainty or risk”. Mayer and Davis
(1999) also operationalised their measure of trust in senior management in
terms of ‘willingness’ or behavioural intentions. Their four item measure
included “I would be comfortable giving top management a task or problem that
was critical to me, even if I could not monitor their actions” and “I would be
willing to let top management have complete control over my future in this
company”. Although their measure had
acceptable levels of test-retest reliability, Mayer and Davis reported
Cronbach’s alphas of 0.59 and 0.60 over their two-wave study. These values are lower than those generally
regarded as acceptable (Nunnally, 1978). Mayer and Davis reported that
Schoorman et al. (1996), using the same measure, found a Cronbach’s alpha of
.82 in a different sample and setting. The variation in reliability across
studies raises questions as to the generalisability of the measure. Mayer and
Davis (1996), themselves, suggested that “further developmental work with this
measure of trust is warranted” (p. 133). Consequently, one of the primary aims
for the present research centered on developing a robust measure of trust in
senior management.
Beyond validation of a measure of
trust, the present research also focused on developing robust measures of
trustworthiness. Specifically, given
the calls for increased standards of governance and ethics in both private and
public sector domains, it is important that researchers and practitioners have
measures of constructs such as integrity and competence which have demonstrated
validity and reliability. Existing
conceptualizations of trustworthiness and trust for the most part, have been
validated using post hoc, exploratory factor analytic methods. For the present study, more rigorous
confirmatory factor analytic methods were used to establish the dimensionality
of trustworthiness and trust in senior management.
Cynicism Toward Change
Within the general field
of attitudes toward change, a number of researchers have focused particularly
on cynicism toward change (Abraham, 2000; Andersson & Bateman, 1997; Dean,
Brandes, & Dhwardkar, 1998; Reichers, Wanous, & Austin, 1997). Andersson and Bateman, for example, found
that high levels of executive compensation, poor organizational performance,
and insensitive downsizing strategies generated cynicism in white-collar
workers. Golombiewski (1995) argued that the bureaucratic structure of Public
Sector organizations generates conditions conducive to cynicism and mitigates
against the development of trust.
Wanous, Reichers, and Austin (2000)
defined cynicism about organizational change as “a pessimistic viewpoint about
change efforts being successful” (p. 133). Wanous et al. conceptualized
cynicism as consisting of two dimensions: a pessimistic outlook on the likely
success of change, and dispositional attributions about those responsible for
effecting successful change. Of the two dimensions, the “pessimism” dimension
was deemed particularly salient for the present research purposes because it
taps employees’ generalized attitudes to change. The dispositional dimension,
reflected in the item “the people responsible for making things better around
here do not care enough about their jobs”, could equally be based on
consideration of union representatives, supervisors, middle managers, or senior
management. The dimension therefore lacks the focal specificity (see Reichers,
1985) to be of practical use with respect to the understanding or management of
organizational change.
Model Development – Trustworthiness, Trust and Cynicism about
Organizational Change
As
previously noted, the literature suggests a strong relation between trust in
senior management and employee attitudes to change. For example, Rousseau and
Tijoriwala (1999) argued from a social accounts perspective (Sitkin & Bies,
1993), that trust in management leads to acceptance of organizational change.
Their findings suggested that “high trust creates a broad zone of acceptance to
the exigencies of complex organizational change” (p. 525). Similarly, Kramer
(1996) argued that management credibility, based on a history of good faith
relations, facilitates positive employee responses to change. Kanter and Mirvis
(1989) also suggested a link between trust in management and attitudes to
change. They argued that cynicism follows when employees lack trust in the
motives of senior management. On the basis of these findings trust is
positioned as a central factor in the way that employees experience aspects of
organizational change. Figure 1 shows the proposed relationships between the
trust-related constructs and cynicism toward change. As previously discussed, because integrity and competence are
conditions of trust, the model shows them having direct effects on trust as
well as direct effects on cynicism toward change.
Figure
1.
Proposed
model of antecedents and cynicism toward change.
Change Cynicism e
Method
Samples
In
order to conduct a preliminary test of the model, data was collected from two
large Public Sector organizations. The first was responsible for the
administration of land title (Organization 1). The second was responsible for
the administration of publicly funded library services, arts and theatres
(Organization 2). A total of 425 useable responses was received from
Organization 1, at a response rate of 42%. A total of 325 useable responses was
received from Organization 2, at a response rate of 33%. Both samples were
found to be representative of their overall organizational demographic in terms
of age and gender. Examination of the
overall distribution of the trait positive and trait negative affectivity
within the samples suggested that the respondents were neither atypically
positive nor negative in their dispositions (details available form the
author).
Procedure
Data was collected in both organizations
through an employee opinion survey. In both samples questionnaires were
attached to employees’ pay slips and accompanied by a cover letter, a
participant information sheet and a return envelope addressed to the
researcher. Employees were also provided with the option of posting their
completed responses in designated sealed collection boxes that were cleared by
the researcher on a regular basis. Participants were assured that the
researcher would retain all the data and would provide no information to the
host organizations that could identify any individual employee. The researcher
attended a number of meetings with employees to address any questions or
concerns they may have had. In addition to completing rating scales,
participants also had the opportunity to provide written comments on what the
organization needed to improve and how it could do so.
Measures
The criterion variable of interest was
cynicism toward change. Cynicism toward change was measured with a four item
sub-scale developed by Wanous, Reichers, and Austen (2000). Employees were
asked to indicate their agreement with change-related statements on a scale
ranging from 1 = Strongly Disagree to 7 = Strongly Agree. Lower scores suggest less cynicism or
'pessimism' toward change. Example items included: “suggestions on how to solve
problems will not produce much real change”, “plans for future improvements
will not amount to much” and “attempts to make things better around here will
not produce good results”. Wanous et al. reported an acceptable full scale
alpha reliability coefficient for the cynicism toward organizational change
(CAOC) scale (a = .86). They did not report subscale reliabilities.
Integrity
was measured with a 10 item scale developed by the author. The items focus on
the extent to which employees perceive senior management as exercising
integrity, being fair, being consistent, and telling the truth. Example items
included: ‘senior management demonstrate integrity in their dealings with
employees’, ‘senior management deliver on the promises they make to employees’
and ‘senior management act ethically toward employees’.
Competence
was measured with a four item scale focusing on the extent to which employees
perceive senior management as having clear strategy, requisite knowledge and
skills, and the drive and capacity to achieve objectives. Example items included: ‘senior management
have a clear idea about where this organization is headed’, ‘senior management
have the knowledge and skills to effectively lead this organization’, and
‘senior management can be trusted to make sensible decisions for the
organization’s future’. Responses alternatives for integrity and competence
were anchored on a seven-point scale ranging from 1 = Strongly Disagree to 7 =
Strongly Agree.
Trust in senior management was assessed with a 3
item scale developed by Albrecht and Sevastos (1999). The validity and reliability of the measure has been demonstrated
in private and public sector organizations using confirmatory factor analysis
and structural equations modeling. Example items included: "I will act on
the basis that senior management would honor any agreements made with
employees" and "I will act on the basis that senior management will
keep the promises they make to employees". Response alternatives were
anchored on a 7 point scale where 1 = Extremely Unlikely and 7 = Extremely
Likely. Acceptable alpha reliabilities have been reported using data from the
same two public sector organizations
(Albrecht & Travaglione, in press).
Analyses
As a first step, descriptives,
correlations and reliabilities were calculated using SPSS.
Next, the dimensionality of the
proposed measurement model was tested using confirmatory factor analysis (CFA).
That is an assessment was conducted of how well a four factor solution (with
the factors consisting of integrity, competence, trust and cynicism) would fit
the data. This analysis, using EQS 5.4 (Bentler, 1995) and the raw data as the
input, was conducted on the sample drawn from Organization 1 (n = 330). A number of recommended ‘fit indices’ (Browne & Cudeck, 1993;
Marsh, Balla, & Hau, 1996; MacCallum, Browne & Sugarawa, 1996) were
used to assess the overall model fit.
As a third step, the proposed
structural relations between the factors were tested using structural equations
modeling (SEM) (see Figure 1). In contrast to other methods such as path
analysis and regression analysis, SEM takes account of measurement error when
simultaneously estimating the direct and indirect causal relations amongst a
set of factors. The overall fit of the structural model, again, was evaluated
against a number of recommended ‘fit indices’.
The
fourth and final step in the analysis involved cross-validating the structural
model in an independent sample in order to assess the generalizability of the
model. It is important to
cross-validate because evidence of construct validity in one sample does not
guarantee construct validity in another (Burke et al., 1989; Cole &
Maxwell, 1985). Thus, the multi-sample procedure in EQS was used to assess the
statistical equivalence or ‘invariance’ of the proposed measurement and
structural models across the two data sets.
Results
Descriptive statistics and
reliabilities
The
means, standard deviations, alpha reliability coefficients and the correlation
matrix for integrity, competence, trust in senior management, and change
cynicism are shown in Table 1. The competence and cynicism scores corresponded
to a scale rating of ‘neither agree nor disagree’. The mean score for integrity
corresponded to scale rating falling between ‘slightly disagree’ and ‘neither
agree nor disagree’. The mean score for trust in senior management corresponded
to scale rating falling between ‘neither likely nor unlikely' and ‘slightly
likely’. All four alpha reliability coefficients exceeded Nunnally’s (1978)
criterion of a=0.80 and thus provide clear evidence in support of psychometric
internal consistency for each of the scales.
The factor correlation matrix shows
moderate to strong correlations among the factors. The correlations were all in
the expected direction and generally were not so strong as to suggest that
respondents could not discriminate between the constructs (Christiansen,
Lovejoy, Szynmanski & Lango, 1996).
The high correlation between integrity and competence (r = .84)
does however suggest the need to further assess the independence of these
constructs. (More rigorous
evidence in support of the independence of the four constructs, derived from a
procedure recommended by Anderson and Gerbing (1988), is available on request).
Table
1
Correlations between change cynicism
and trust-related variables, Organisation 1 (n=330)
|
|
Mean |
Possible Range |
Alpha |
1 |
2 |
3 |
4 |
|
1 Integrity |
3.6 |
1-7 |
0.98 |
- |
|
|
|
|
2 Competence |
4.1 |
1-7 |
0.94 |
.84** |
- |
|
|
|
3 Trust in Senior Management |
4.6 |
1-7 |
0.80 |
.66** |
.56** |
- |
|
|
4 Change Cynicism |
3.8 |
1-7 |
0.88 |
-.55** |
-.44** |
-.48** |
- |
|
|
|
|
|
|
|
|
|
Note: ** = significant at p < .01
(2-tailed);
Confirmatory Factor Analysis
The results of the CFA are shown in
Table 2. Fit indices for the null model (no structure within the items) and for
a one dimensional model (all items specified to load on a single factor) are
provided for comparison purposes.
Table 2 shows that the fit indices for
the proposed model were above criterion levels. The Bentler-Bonnett Normed Fit Index (.94), Bentler-Bonnett NonNormed
Fit Index (.95), and the Robust Comparative Fit Index (.96) were all above
.90. The Comparative Fit Index (.96)
exceeded the criterion value of .93. The RMSEA point estimate (.075) suggests
'reasonable fit', well below Browne and Cudeck’s (1993) criterion for a
“mediocre fit” (between .08 and 1.00). The RMSEA confidence intervals (.067 -
.082) do not however suggest a close fitting model (MacCallum, Browne &
Sugarawa, 1996). Neither the null model nor the one-dimensional model provided
acceptable fit to the data.
Table
2
Fit indices of the 4 factor measurement model,
Organisation 1
|
Model |
c2 |
DF |
NFI |
NNFI |
CFI |
Robust CFI |
RMSEA |
(90% CI) |
|
|
|
|
|
|
|
|
|
|
|
Null model |
7853.476 |
210 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 Factor |
1390.103 |
189 |
.823 |
.825 |
.843 |
.864 |
.140 |
(.137 - .147) |
|
|
|
|
|
|
|
|
|
|
|
4 Factor |
512.580 |
183 |
.935 |
.951 |
.957 |
.965 |
.075 |
(.067 - .082) |
|
|
|
|
|
|
|
|
|
|
Note: NFI= Normed Fit Index, NNFI= Non-Normed
Fit Index, CFI = Comparative Fit Index, RMSEA = Root Mean Square Error of
Approximation, 90%CI = RMSEA
90%Confidence Interval.
The
convergent validity of the items in the model was determined by examining the
factor loadings of the items on their specified dimensions. The standardized factor loadings (available
from the author on request) ranged from 0.55 to 0.94. All were statistically
significant, and apart from one exception, all loadings were greater than 0.75.
Given that standardized values greater than 0.5 demonstrate reasonably high
factor loadings (Kline, 1998), these results support the convergent validity of
the proposed four dimensional measurement model.
Structural Model
Having established a theoretically
viable measurement model with CFA, the next step in the analysis was to examine
the proposed structural relations between the constructs (see Figure 1). The
fit statistics suggested a good fitting model: c2 (183, n =
330) = 512.580, NFI = .94, NNFI = .95, CFI = .96, RCFI = .96, RMSEA = .075. The
path estimates, representing the strength and direction of influence between
the constructs, and their significance values are shown in Figure 2.
Consistent with the proposed model,
the paths from integrity to trust and from trust to cynicism were statistically
significant. The path from competence to trust was not significant. The path
from competence to cynicism just failed to reach statistical significance.
Overall, just over 38% of the variance in change cynicism was explained by the
model. Just under 50% of the variance in trust in senior management was
explained by integrity and competence.
Figure
2.
Standardised
parameter estimates and t-values (in parentheses) of proposed structural model
Change Information