z. c.

73...}:

73....‘4

.5 .
Q
7
1.1 if... .11:
51...... t... J .3...“
Hr- 1......

.g' A -
(5.1-5: .9. A .1

J 1.?
r.

I,

.
30:3.“ “a.
.f .

7.1... 4. i.

‘. Law-what.

 

 

3.: .v
WW: .3.

.‘
I.
r

. 0.9.1.. s.

 

. 9.... .03..\.w....u..
. . «fiflm

L m---

I:f\’{
‘VU

This is to certify that the
dissertation entitled

EFFECT OF ORGANIZATIONAL CONTROL SYSTEMS ON
ORGANIZATIONALLY DIRECTED DEVIANCE

presented by

JACLYN MARIE NOWAKOWSKI

has been accepted towards fulfillment
of the requirements for the

Ph.D. degree in PsychologL

 

 

 
 
 

‘7 . essor’s Signature

7//I

Date

MSU is an Alfinnafivo Action/Equal Opportunity Institution

_.__ _—-——'— ——.__..__. -m

LIBRARY
Michigan State
University

 

 

 

 

PLACE lN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 p:/ClRC/DateDue indd-p.1

 

EFFECT OF ORGANIZATIONAL CONTROL SYSTEMS ON
ORGANIZATIONALLY DIRECTED DEVIANCE

By

Jaclyn Marie Nowakowski

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Psychology

2006

ABSTRACT

EFFECT OF ORGANIZATIONAL CONTROL SYSTEMS ON
ORGANIZATIONALLY DIRECTED DEVIANCE

By
Jaclyn Marie Nowakowski

Previous research on workplace deviance has focused almost exclusively on the
antecedents and consequences of destructive or negative deviance, yet deviance can also
be beneficial and violate expectations in a positive, constructive direction. As both forms
of behavior are affected by policies and procedures which establish norms and
expectations, a field experiment was conducted to examine the effect of several forms of
organizational control (formal vs. informal and external vs. internal) on destructive and
constructive deviance. Specifically, the experiment tested the hypothesis that formal and
external control, as compared to informal and internal control, would lessen destructive
deviance but would also have negative effects on constructive deviance. The process
reactions of justice and trust were hypothesized to mediate this relationship and
conscientiousness and proactive personality were hypothesized to moderate the link
between process reactions and deviance. Two hundred twenty-five individuals were
employed in a one-time work opportunity where real time behavioral observations of
deviance were measured. Results revealed that while no forms of control were effective
in eliminating destructive deviance, external control was most likely to depress

constructive behavior. Implications and directions for future research are presented.

TABLE OF CONTENTS

LIST OF TABLES ........................................................................................................ v
LIST OF FIGURES ...................................................................................................... vii
INTRODUCTION ........................................................................................................ 1
LITERATURE REVIEW ............................................................................................. 4
Control in Organizations ......................................................................................... 4
Typology of Control ............................................................................................... 10
Workplace Deviance ............................................................................................... 13
Capturing Destructive and Constructive Deviance ................................................. 15
MODEL DEVELOPMENT AND HYPOTHESES ..................................................... 18
Control — Justice ..................................................................................................... l9
Affective Reactions — Deviance .............................................................................. 23
Moderators of Trust/Justice — Deviance Relationship ............................................ 25
Control, Reactions, and Deviance ........................................................................... 28
METHOD ..................................................................................................................... 30
Design ..................................................................................................................... 30
Sample ..................................................................................................................... 31
Procedure ................................................................................................................ 32
Manipulations ......................................................................................................... 38
Pilot Testing ............................................................................................................ 40
Measures ................................................................................................................. 42
Destructive Deviance ........................................................................................ 46
Constructive Deviance ...................................................................................... 49
RESULTS ..................................................................................................................... 56
Manipulation Checks .............................................................................................. 59
Hypotheses 1-4 ........................................................................................................ 63
Hypotheses 5-8 ........................................................................................................ 64
Hypotheses 9-10 ...................................................................................................... 71
Hypotheses 11-14 .................................................................................................... 78
Post-Hoe Analysis ............................................................................................. 79
DISCUSSION ............................................................................................................... 83
Deviance Behavior .................................................................................................. 83
Control — Process Reactions ................................................................................... 84
Process Reactions - Deviance ................................................................................. 86
Moderating Role of Conscientiousness and Proactive Personality ......................... 87
Control — Deviance ................................................................................................. 89

iii

Limitations and Future Directions ......................................................................... 92

Implementing Control ..................................................................................... 92
Defining and Assessing Norms ....................................................................... 93

Effect of Time ................................................................................................. 95
Rewards and Organizational Factors ............................................................... 95
Conclusions ........................................................................................................... 96
APPENDICES ............................................................................................................. 98
REFERENCES ............................................................................................................ 119

LIST OF TABLES

Table 1: Characteristics, Typology, and Examples of Control ..................................... 10
Table 2: Deviance Diagram .......................................................................................... 16
Table 3: Number of Employees Engaging in Destructive Deviance (N=225) ............. 48
Table 4: Number of Employees Engaging in Constructive Deviance (N=225) ........... 51

Table 5: Means, Standard Deviations, and Correlations of Experimental Variables 57
Table 6: Mean (SD) Values of Manipulation Check Items by Condition (N=225) ..... 60
Table 7: Mean (SD) Values of Process Variables by Condition (N=225) ................... 63
Table 8: Descriptive Statistics for Deviance (N=225) .................................................. 65

Table 9: Summary of Regression Analysis for Justice Variables Predicting
Constructive Deviance (N =225) ................................................................................... 67

Table 10: Summary of Regression Analysis for Justice Variables Predicting
Destructive Deviance (N =225) ................................................................................... 68

Table 11: Post-Hoe Regression Analysis for Justice Variables Predicting Destructive
Deviance (N=225) ........................................................................................................ 69

Table 12: Summary of Hierarchical Regression Analysis for Trust x
Conscientiousness Predicting Destructive Deviance (N=225) ..................................... 71

Table 13: Summary of Hierarchical Regression Analysis for Justice x
Conscientiousness Predicting Destructive Deviance (N=225) ..................................... 72

Table 14: Summary of Hierarchical Regression Analysis for Trust x
Conscientiousness Predicting Constructive Deviance (N =225) ................................... 73

Table 15: Summary of Hierarchical Regression Analysis for Justice x
Conscientiousness Predicting Constructive Deviance (N=225) ................................... 74

Table 16: Summary of Hierarchical Regression Analysis for Trust x Proactive
Personality Predicting Constructive Deviance (N=225) ............................................... 76

Table 17: Summary of Hierarchical Regression Analysis for Justice x Proactive
Personality Predicting Constructive Deviance (N=225) ............................................... 77

Table 18: Post-Hoc Tukey HSD Test for Interaction Term (N =225) ........................... 80

vi

LIST OF FIGURES

Figure 1: Experimental Model ...................................................................................... 18
Figure 2: Distribution of Destructive Deviance ............................................................ 66
Figure 3: Distribution of Constructive Deviance .......................................................... 67

Figure 4: Interaction Between Conscientiousness and Trust on Destructive Deviance 72

Figure 5: Interaction Between Conscientiousness and Procedural Justice on
Constructive Deviance .................................................................................................. 75

Figure 6: Interaction Between Extemal/Intemal Control and Conscientiousness on
Destructive Deviance .................................................................................................... 82

vii

INTRODUCTION

In the wake of several corporate scandals such as Enron, Worldcom, and
Adelphia, organizations are increasingly concerned with employees following proper
standards of conduct. It comes as no surprise to see that the academic literature has
recently focused on methods of stamping out improper behavior by such means as
electronic performance monitoring (Alder & Ambrose, 2005; Alge, 2001; Ambrose &
Alder, 2000; Kidwell & Bennett, 1994; Stanton & Weiss, 2000), control in team-based
environments (Barker, 1993; Ezzamel & Willmott, 1998; Sewell, 1998), and ethical
codes of conduct (Adam & Rachman-Moore, 2004; Stevens, Steensma, Harrison, &
Cochran, 2005; Trevino & Weaver, 2001; Weaver, Trevino, & Cochran, 1999). These
approaches stand in contrast to earlier work on control in organizations (Ouchi, 1977,
1979, 1980; Ouchi & Johnson, 1978; Ouchi & Maguire, 1975) that focused on the
broader purpose of control systems in directing employee behavior. While both recent
and past streams of research give greater insight into the characteristics of monitoring and
control systems that result in more positive employee reactions, there is an opportunity to
take a step back and understand how control systems designed to curb negative employee
deviance not only do as intended but also inadvertently result in fewer positive employee
behaviors as well.

A popular example of attempting to control undesirable behavior while also
restricting desirable responses is a university’s honor code designed to limit cheating and
dishonest behavior in college. Students are told that they are responsible for abiding by
the code and that violation of the rules will result in serious sanctions. In some cases,

codes will encourage students to report on the negative behavior of their peers to fimher

limit academic dishonesty. While cheating may decrease, an inadvertent consequence is
that students are hyper-vigilant, distrusting of one another, and leery of the process
designed to extract guilt or innocence (Hoover, 2002). Organizations may also fall prey
to similar problems when employees perceive control programs as overly harsh or unjust.
While these systems are designed to limit destructive behavior in the organization (e.g.,
employee theft), an unintended consequence is that employees may react negatively and
be no longer as willing to go above or beyond what is required of them. The willingness
to do more than expected is termed constructive deviance, defined as behavior that
deviates from the actions of an individual’s reference group but which conforms to a
higher order set of values and beliefs (Warren, 2003). An example of constructive
deviance is the employee who engages in innovative problem solving to benefit the
organization while his or her peers stick with tried and true methods.

As organizations implement control programs in organizations via computer
monitoring, electronic surveillance, work observation, and a variety of other methods
(Smith, 2005), it becomes more important to understand how elements of control systems
influence employee reactions and subsequent behavior, whether positive or negative acts
of deviance. Even the most effective system on paper may have unintended
consequences in practice. Because the foundation of these relationships is based largely
in organizational context, this study will use field experiment methodology which
combines the advantages of experimental methods with the external validity of an
organizational setting. Therefore, this study will rely on previous research on
organizational control systems that provides a rich starting point to understand how

different facets of control systems influence employee reactions. The purpose is

threefold: (1) to develop a framework highlighting two key facets of control programs —
formality of control and locus of responsibility, (2) employ field experiment methodology
to understand deviance behavior in various control environments, and (3) investigate
unintended as well as intended consequences of systems designed to reduce

organizationally directed deviant behavior at work.

Control in Organizations

In their review of dominant control ideologies, Barley and Kunda (1992) note that
the earliest evidence of control can be found in scientific management. This form of
control was characterized by managers attempting to control their employees by
specifying the precise physical movements required for increased productivity. F athered
by Frederick Taylor, this form of control was adopted by many businesses as the way to
modernize and create efficient workplaces. Scientific management lost some of its steam
following World War I when many efficiency experts noted that these systems were not
able to substantially reduce waste and cost, despite Taylor’s rhetoric (Barley & Kunda,
1992).

As scientific management fell out of favor, a wave of humanistic control systems
came into the spotlight, with an increased focus on normative control and improving
work environments and conditions for employees (Barley & Kunda, 1992). The work of
human relations scholars such as Oliver Sheldon, Mary Parker Follett, and Elton Mayo
gave rise to the Western Electric Hawthorne Studies which demonstrated that employees
were more productive when supervisors were physically monitoring their work than
when working unsupervised (Pennock, 1930). As a result of these findings, many
advances in employee-employer relations were developed, including innovations in
compensation systems, supervisory training, and job enrichment (Barley & Kunda, 1992).

As with prior popular forms of control, the human relations movement also fell
out of favor in the 19605 as scholars shified towards a more analytic orientation towards
control. In this approach to control, managers were thought to be more able to control

employees by manipulating decision structures and reward processes according to their

needs and the needs of the organization (Galbraith, 1977). This shifi in control was
captured by much of the work conducted by William Ouchi (Ouchi, 1977, 1979, 1980;
Ouchi & Johnson, 1978; Ouchi & Maguire, 1975). Ouchi’s work focused on
understanding the purpose of different types of organizational control, whether
determined by managerial needs (Ouchi & Maguire, 1975), stemming from variations in
organizational structure (Ouchi, 1977), or driven by the organizational environment
(Ouchi, 1979, 1980).

To understand control systems predicated on managerial needs, Ouchi and
Maguire (1975) focused on methods to control employees based on behavioral or output
measures. Managers employing behavioral control techniques closely and directly
monitor employees’ actions over time, while managers employing output control
techniques evaluate employee outcomes, whether number of defects, financial returns, or
sales volume (Ouchi & Maguire, 1975). Ouchi and Maguire (1975) found that the type
of control used by managers depended largely on task complexity and managerial
knowledge of job performance. Behavioral control was used most oflen with simple
tasks where managers were knowledgeable about individual on-the-job performance,
while output control was used most often with complex tasks that required
interdepartmental collaboration where managers were not as knowledgeable about the
performance and precise contributions of each employee towards the desired outcome.

While managerial needs can serve as one factor determining the appropriate type
of organizational control, Ouchi (1979, 1980) also examined the organization’s
environment as dictating appropriate measures of control. In his research, he

differentiated environments into three types: market, bureaucracy, or clan environment

(Ouchi, 1979, 1980). In a market environment, employee performance can be observed
and rewarded based on principles of exchange: employees who produce little are
rewarded little, while employees who produce a great amount are rewarded
proportionally (Ouchi, 1979). Organizations attempt to control employee behavior via
the reward process. In a bureaucracy, managers must closely watch, direct, and monitor
employee behavior to ensure compliance with company standards and rules. Managers
must assess whether employee performance satisfactorily complies with the rules and
employees must satisfy managerial expectations to be rewarded for their performance
(Ouchi, 1979). In a clan environment, individual performance is difficult to disentangle
from the performance of other individuals. The organization is able to control employee
behavior via socialization processes which commit employees to organizational
objectives and reaching performance rewards (Ouchi, 1979).

Ouchi and Johnson (1978) also discussed differences in organizational
environments and linked them to two main types of control systems that operate in
organizations, termed Type A or Type Z control systems. Type A systems establish
control through a tightly monitored system, similar in nature to what would be
implemented in a market environment. Employees are frequently evaluated based on
hard data collected through formal employee monitoring. Type Z systems operate
differently and establish control through acculturation and socialization of employees,
similar to control that would be implemented in a clan environment. These systems
emphasize coordination between units rather than high individual specialization and

evaluations are based on subtle, informal criteria.

The research of Sitkin and colleagues (Cardinal, Sitkin, & Long, 2004; Roth,
Sitkin, & House, 1994; Sitkin & George, 2005) also touches upon the formal and
informal nature of control systems. Most simply, they refer to formal control as anything
that is written while informal control is anything unwritten (Cardinal et al., 2004).
Formal control measures are visible, objective forms of control while informal control
relies on less objective, unofficial forms of control.
F ormality of Control

Ouchi’s research on Type A and Z control systems along with Sitkin and
colleagues’ perspective on written and unwritten control forms the basis for formality of
control, the first dimension of the control framework to be examined in this research.
Formal systems employ strict monitoring and frequent evaluation. They are
characterized by written procedures, rules, policies, rigid routines, and by their very
official nature. By contrast, informal systems are characterized by a casual or unofficial
nature, with employees’ evaluations being based on subtle, informal factors. In informal
control systems, employees develop a sense for doing what is expected of them via social
norms or examples set by more senior members of the organization (Adam & Rachman-
Moore, 2004).
Locus of Responsibility

Within informal and formal systems there is a need to identify who or what is
responsible for ensuring that employees comply with the system. It is also important to
link this distinction back to the purpose of the current research, which is to understand
how differences among control systems relate to both positive and negative deviance.

This second dimension will be referred to as locus of responsibility defined as the

location of the entity that will monitor employee behavior, whether it is located internal
or external to the person (e. g., I am responsible for monitoring myself vs. my
organization or supervisor will monitor my behavior). In the control research discussed
above, the locus of responsibility was external. The organization was responsible for
monitoring employee behavior, whether through behavioral, output, Type A, or Type Z
control. Although employees may have also kept track of their performance for personal
knowledge, final responsibility rested with the organization to monitor employee actions.
The notion of external responsibility has become particularly salient with the use
of electronic performance monitoring (EPM). EPM is defined as gathering information
about the performance of employees via technological means (Alge, 2001; Ambrose &
Alder, 2000). Ambrose and Alder (2000) reported that the use of EPM is extensive, with
the number of workers who have been monitored at work growing from more than 10
million in the early 19903 to over 20 million by 1996. EPM techniques can be used to
count keystrokes, read email, monitor intemet usage and phone calls, and track employee
locations via video surveillance (Alge, 2001 ). In situations where companies utilize EPM
or some other form of employee surveillance, the responsibility for monitoring employee
actions rests in the hands of the organization, external to the employee. As organizations
monitor employee behavior, Ambrose and Alder (2000) have identified ten dimensions
related to EPM that the organization can control, including: (a) employee participation in
system design, (b) disclosure of monitoring to employees, (c) tasks monitored, (d)
amount and length of monitoring, (e) use of monitoring data to develop performance
standards, (f) providing feedback to employees, (g) offering opportunities to employees

to challenge the system, (h) timeliness of feedback, (i) monitoring of group or individual

performance, and (j) use of monitoring data to inform performance reward decisions. It
is important to recognize that EPM systems can be implemented under a host of different
conditions, with each having implications for employee reactions and behavior. Recent
research has focused on employee justice reactions stemming from EPM techniques in
the workplace (Aiello & Kolb, 1995; Ambrose & Alder, 2000; Niehoff & Moorrnan,
1993; Stanton, 2000).

In contrast to control systems characterized by a high degree of external
responsibility are systems with a high degree of internal responsibility where employees
are held responsible for monitoring their own behaviors and subsequently their
compliance with the control system. These types of systems are often observed in team-
based work environments, where individuals self-monitor and observe each other to
ensure that each person is complying with the necessary work procedures (Barker, 1993;
Ezzamel & Willmott, 1998; Sewell, 1998). Programs designed to instill ethical values
are internal control systems when employees are tasked with monitoring their own
behavior related to organizational principles, values, and goals. When the organization
steps back and places responsibility in employee hands, either because it cannot
adequately monitor employee behavior or because it feels employees would be more
compliant when self-accountability is high, the control system takes on this internal
nature. These types of systems still fall under the broad category of organizational
control, since the ideals and values of the organization underlie the nature of the system
being implemented. However, the emphasis on individual employees to self-monitor is
expected to result in different employee behavior when compared to a system which is

based on complying with an external agent.

Table 1

Characteristics, Typology, and Examples of Control

 

Forrnality of Control

Formal
Written expectations of
employee behavior

Individual employees

Informal
Unwritten norms regarding
organizational expectations

Individual accountability for

 

 

Internal tracking own behavior with compliance
explicit documentation
Tacit commitment to
Locus of Work diary organizational values
Responsibility
Written expectations of Unwritten understanding
employee behavior between employee and
External . . organization
Monitoring by an external
organizational agent External individuals monitor
employee behavior
Electronic surveillance
Typology of Control

Considering the dimensions of formal/informal and extemal/intemal control, a

typology for understanding organizational control systems is offered (Table l).

The cell combining a formal control system where locus of responsibility is internal to

the employee is characterized by employees tracking their own behavior with explicit

documentation and self-reporting when they deviate from norms. An example would be

a work diary where employees are required to document their daily work activities

(Bonke, 2005; Hewitt, 1983; O'Reilly, 1973). To the extent that the employee is honest

and diligent about documenting their activities, any deviation from work procedures

(either positive or negative) could be identified from an examination of the diary entries.

10

The cell characterized by an informal control system where locus of responsibility
is internal to the employee requires the employee to self-monitor. In this environment, it
is assmned that employees recognize and accept that they are expected to follow the
organization’s norms and values. At the peer-group level, an example would be the
control that develops in team based work systems, either through norms that develop in
self-managed work teams or via total quality management initiatives (Barker, 1993;
Ezzamel & Willmott, 1998; Sewell, 1998). Another example are values-oriented ethics
codes that create an ethical culture and tacit commitment to organizational values,
thereby activating an employee’s inner sense of compliance (Stevens et al., 2005; Weaver
et al., 1999).

The cell combining formal control with an external locus of responsibility is
characterized by a precise system with written expectations of employee behavior.
Employees are watched in the workplace and their behaviors (whether destructive or
constructive) can be documented by the organization and/or their supervisor. An
example of this type of control is a budget system where fiscal goals are stated for the
forthcoming cycle and progress is assessed by upper management at key intervals
(Carlson, 1987; Rabinowitz, 1996). Another example is a code of conduct specified in
the employee handbook combined with a form of monitoring such as tracking employee
Internet usage (Swanson, 2002) or geographic location in a company vehicle via GPS
technology (Cadrain, 2005).

Finally, the cell containing an informal system with external locus of
responsibility is characterized by an unwritten understanding between the employee and

organization of what is expected from the employee. Supervisors and/or peers monitor

11

employee behavior. An example would be the politics that develop in organizations
which persuade individuals to action that are largely based on unwritten norms and tend
to emanate from higher authorities in the organization. Reports of employee behavior (or
misbehavior) tend to travel through the grapevines of organizations, which act as a
feedback loop to the supervisor (Voyer, 1994) who can then monitor employees’
behaviors.

The preceding discussion has focused on organizational level systems, and it is
now important to recognize the impact that these systems have on individual level
behavior. Organizational control systems are designed, in large part, to limit destructive
deviance in the workplace. Organizations want to coordinate and control the activities of
organizational members so that all behavior is oriented towards organizational goals
(Lawler & Rhode, 1976). It is hard to imagine any upstanding organization that would
want their employees to steal, come to work late, laze about while at work, be rude to
coworkers, or harass each other. Therefore, control systems try to influence employees
by specifying the type of behavior that is appropriate (Lawler & Rhode, 1976). An
unfortunate consequence of control systems is that the process of eliminating undesirable
behavior may limit more desirable responses (in the form of constructive deviance), as
employees perform only what is required of them and nothing more. As some
preliminary research has linked the frequency of monitoring (which is associated with
eliminating destructive deviance) to decreases in positive employee behavior (Niehoff &
Moorrnan, 1993), I will draw on these initial findings and further relate deviance

outcomes to the typology of organizational control proposed in this research.

12

Workplace Deviance

Early definitions of workplace deviance focus on the voluntary nature of behavior
of organizational members that violates significant organization norms, and, in doing so,
threatens the well being of the organization and/or its members (Robinson & Bennett,
1997). According to Robinson and Bennett (1997), there are several key components of
deviance behavior in the workplace. First, the behavior must be purposeful,
discretionary, and voluntary (e.g., not an accident). It does not necessarily have to be
intended to be deviant or harmful towards the organization for the behavior to be
considered counterproductive, as long as it results in negative organizational outcomes.
Second, the behavior violates significant organizational norms. These norms are
espoused by the dominant administrative members of the organization, as compared to
norms of a subgroup or subculture, and are pivotal or central to the well being of the
organization. Violation of these central norms would be known to organizational
members to result in negative sanctions. Examples of deviant behavior include theft,
shirking, aggression, and harassment.

More recently, Warren (2003) has offered a typology of employee deviance,
classifying behavior against two sets of norms: reference group norms and hypemorms.
Reference group norms are similar in nature to the norms discussed by Robinson and
Bennett (1995). They refer to standard behavior within a person’s group. Hypemorrns
are generally held beliefs about fundamental societal principles and expectations for
doing the right thing (e.g., nourishment, freedom, physical security). Behavior that
deviates from both reference group and hypemorrns is termed destructive deviance while

behavior that deviates from reference group norms but conforms to hypemorrns is termed

l3

constructive deviance (Warren, 2003). One example of constructive deviance would
include an employee exercising voice to initiate and encourage necessary change in a
work group that embraces the status quo. Here, the employee initiating change is
conforming to hypemorms by advocating for improvement while deviating from the
reference group norm. A second example is an employee who takes the organizationally-
authorized time for a break when everyone in the work group takes longer than permitted
breaks. Here, the employee deviates from the work group by taking a shorter break, but
also conforms to hypemorms of following organizational rules. Along similar lines are
whistleblowers who violate work group norms by speaking out but conform to
hypemorms by reporting wrongdoing. Because norm violations can occur in either
positive or negative directions, Warren’s (2003) definition more fully captures the scope
of deviance behavior. The behaviors classified by Robinson and Bennett (1995) only
typically fall into the category of destructive deviance. Therefore, Warren’s
conceptualization of deviance will serve as the basis for capturing deviance in the current
study.

It is important to recognize that constructive deviance is often associated with
creativity, prosocial behavior, and organizational citizenship behavior (OCB). Much
research has been conducted to examine the antecedents and consequences of this set of
behaviors (LePine, Erez, & Johnson, 2002; Organ & Ryan, 1995; Pearce & Herbik, 2004;
Podsakoff, MacKenzie, Paine, & Bachrach, 2000). However, constructive deviance can
be separated from this broader set of positive, helpful behavior by recognizing the role
that norms play in dictating what is deviant and what is not. A hallmark characteristic of

deviant behavior is that it violates norms. In the case of constructive deviance, the

14

behavior must violate reference group norms. Simply because a behavior is positive and
conforms to a higher sense of doing what is right does not warrant the constructive
deviance label: it must be contrary to the behavior of an individual’s reference group.
Therefore, if everyone in a reference group engages in OCB and a norm is established
where everyone is expected to do so, the target person who also engages in OCB is not
being constructively deviant but is simply doing something positive along with his or her
colleagues. Given the body of research that has been conducted on OCB, contrasted
against the small set of research on the behavior labeled constructive deviance, it is
reasonable to assert that constructive deviance may be a subset of OCB in that it shares
the overarching “helpful” distinction (Spreitzer & Sonenshein, 2004). What
differentiates these concepts is the context and norms operating in the environment that
graduate constructive deviance beyond the more general category of organizational
citizenship.
Capturing Destructive and Constructive Deviance

To account for the various forms of employee deviance that will be the focus of
this research, a diagram organizing the different dimensions can be found in Table 2.
This table depicts the primary dimension distinguishing destructive from constructive
deviance (deviation from reference group norms + conformity or deviation to
hypemorms). To capture deviance, it will be very important to understand the content of
both reference group norms, which is likely to differ across organizations, as well as
hypemorms which are expected to be more general in nature and stable across situations
as they capture global standards, values, and beliefs (Warren, 2003). Therefore, the

behaviors displayed in Table 2 should be viewed as reasonable examples, with the caveat

15

Table 2

Deviance Diagram

 

 

 

Hypemorms
Conform Deviate
Constructive Organizational Deviance Destructive Organizational Deviance
Volunteering or taking on extra work Theft
Innovative problem solving Vandalism

Participating in organizational committees Withholding effort or low productivity
that reference group norms may expand or shrink which behaviors fall into the categories.
For example, although theft is widely considered to be unacceptable, if everyone in your
work group steals and you do as well, the behavior is not deviant as compared to the
reference group but instead conforms to peers’ behavior.I

Robinson and Bennett (1995, 1997) also include two additional dimensions in
their typology of deviance behaviors: severity of deviance and target of deviance. For the
purpose of this research, understanding whether deviant behavior has a minor or major
impact on the target warrants a much deeper exploration into the environment and

characteristics of the actor and target. Obviously there are some destructive behaviors

which are undoubtedly severe such as murder, physical assault, or embezzlement.

 

' While this research will focus on behavior that deviates from reference group norms, it
is worth noting the other half of Warren’s (2003) typology which considers behavior that
conforms to reference group norms and either conformity or deviation to hypemorms.
Behavior which conforms to both reference group norms and hypemorms is called
constructive conformity while behavior which conforms to reference group norms but
deviates from hypemorms is termed destructive conformity. In this example, theft which
conforms to reference group norms but deviates from hypemorms is destructive
conformity.

16

Similarly, there are certain constructive behaviors which can have a major impact, such
as whistleblowing. However, despite the popular attention these behaviors attract, they
have very low base rates. For example, the Federal Bureau of Investigation (2004)
reported that homicide accounted for only 0.1% of all workplace victimization between
1993-1999, while whistle blowing may be an even less frequent activity, despite the
recent passage of legislation designed to protect whistleblowers from dismissal or other
discrimination (Glassman, 2005). Therefore, it is unlikely that these behaviors will be
captured as part of this research. There are also a set of behaviors that are generally mild
in nature, such as gossiping on the negative side and creativity on the positive side,
whose impact may be very strongly or weakly felt by the target depending on whether the
behavior is in retaliation, reciprocation, or motivated by some other means. Therefore,
because of the difficulty associated with making a clean distinction of deviance outcomes
along this major/minor dimension, it will not be a focus of the deviance dimensions
explored in this research.

I will also limit the scope of deviance behavior with regard to the target of
deviance and focus solely on organizationally-directed deviance. Robinson and Bennett
(1995, 1997) do make a strong distinction between organizational and interpersonal
deviance and note the different antecedents that motivate these behaviors, which is not to
be dismissed. However, the nature of the experimental task used in the current study will
restrict the interaction between participants, thereby reducing the opportunity for
interpersonal deviance while allowing for a more careful examination of organizational

deviance.

l7

MODEL DEVELOPMENT AND HYPOTHESES

As organizations attempt to control negative employee behavior, the current
research is focused on the intended outcomes of control to limit undesirable behavior as
well as the inadvertent consequences of control which restricts desirable behavior. The
typology of organizational control systems highlights two key facets of control -
formality and locus of responsibility - that are expected to influence destructive and
constructive employee deviance. The following section will explain the experimental
model and detail hypotheses linking the control dimensions to these behavioral outcomes.
Ideally, organizations would want to find a solution to organizational control that
minimizes destructive deviance while simultaneously maximizing constructive deviance.
This research will explore the simultaneous impact of control mechanisms on both forms
of deviance while also taking into consideration important mediating and moderating
influences impacting the primary relationship. The precise relationships that will be
tested in the current research are displayed in Figure 1.
Figure 1

Experimental Model

 

 

 

Conscientiousness

 

 

 

 

Destructive deviance

 

 

Formality of control —> Justice

 

 

 

 

 

 

 

<

 

 

 

Trust Constructive deviance

 

V

 

 

 

 

 

 

Locus of responsibility

 

 

 

 

 

 

 

Proactive personality

 

l8

Hypotheses

Control — Justice. Research on organizational justice has developed in the
psychological literature over the past 30 years, dating back to the work of on aspects of
equity, fair procedures, and process control. Early work on organizational justice
recognized the notion that individuals desire fairness in their work environments, whether
by equitable distribution of resources and rewards, termed distributive justice, or by use
of fair procedures that impact outcomes, termed procedural justice (Adams, 1965;
Leventhal, 1976; Thibaut & Walker, 1975). More recent research on organizational
justice has expanded its dimensionality to include a focus on the fairness of interpersonal
treatment received by individuals, termed interpersonal justice (Colquitt, 2001; J.
Greenberg, 1993). The current research will focus on these three dimensions of justice
reactions.

Alge (2001) found that participants perceived greater procedural justice when the
activities being monitored by the control system were relevant to key decisions affecting
employees and when employees had the opportunity to voice their opinion about the
monitoring system. Douthitt and Aiello’s (2001) research on monitoring also found that
employees who had the opportunity to provide their input during the monitoring process
had more positive procedural justice reactions. This notion of participation in monitoring
is related to the control dimension of locus of responsibility. When employees are
responsible for monitoring their own behavior, they play a large participatory role in the
process, whereas when the organization is responsible for monitoring behavior,

employees play less of a role. Therefore, as compared to control systems with an

19

external locus of responsibility, control systems employing an internal locus of
responsibility will result in more positive procedural justice reactions.

HI : Control systems with an internal locus of responsibility will result in more
positive procedural justice reactions compared to systems with an external locus of
responsibility.

Considering distributive justice, Niehoff and Moorman (1993) proposed that the
formality of monitoring (whether formal meetings or informal discussions between
supervisors and employees) should also affect distributive justice reactions. This relates
to the current research dimension of formality of control. They hypothesized a positive
relationship between both types of monitoring and distributive justice, stating that
employees whose managers meet with them to learn about their work behavior (and
subsequently use that information in performance appraisal) will perceive greater
distributive fairness than employees whose managers do not meet with them. In
actuality, they did not find any relationship between formal or informal meetings and
distributive justice. However, in their research it is likely that the frequency with which
managers met with employees also affected distributive justice perceptions and
confounded the relationship between formality of control and distributive justice.
Holding frequency constant, when more formal methods of control are used employees
may perceive that outcomes stemming from this type of monitoring are fairer. Because
the system gathers data in a routinized, structured manner, as compared to an informal
method where the system is less structured, it is less susceptible to a supervisor’s bias in

evaluation. Therefore,

20

H2: Formal control systems will result in more positive distributive justice
reactions compared to informal control systems.

While rooted in procedural justice, interpersonal justice is another arm of fairness
which deals with the quality of interpersonal treatment received during organizational
procedures (Bies & Moag, 1986). More specifically, interpersonal justice deals with the
respect and propriety associated with the type of treatment received (Colquitt, 2001).
Shapiro and Brett (1993) found that individuals perceive procedures and outcomes as
more fair when they are able to exercise voice. When authorities give greater
consideration to an employee’s views, procedural justice judgments also increase (Tyler,
1987). Extending these findings into the realm of interpersonal justice and control,
informal control systems may give employees the impression that their input matters as
the organization places greater responsibility for monitoring in individual hands. This
may further be seen as indicating greater respect for the employee as the employee is able
to express his or her “voice” by monitoring and controlling their own behavior in an
indirect manner (Shapiro & Brett, 2005). Therefore,

H3: Internal control systems will result in more positive interpersonal justice
reactions compared to external control systems.

Control - Trust. The literature on organizational control has also identified trust
as a necessary condition for improved organizational functioning. Trust has been defined
as “a psychological state comprising the intention to accept vulnerability based upon
positive expectations of the intentions or behavior of another” (Rousseau, Sitkin, Burt, &
Camerer, 1998, p. 345). In organizational environments characterized by a great amount

of uncertainty and where work tasks are ambiguous, dynamic, difficult to observe, or

21

result in outcomes that are hard to assess, employees must have faith in their coworkers
and organization that everything will proceed as expected. In the face of this kind of
uncertainty, employees must rely on one another to accomplish their work and trust that
responsibilities will be fulfilled in the give-and-take relationship (Monis & Moberg,
1994). Trust is not only something that can be developed between employees but also
between the employee and its organization, via work policies, procedures, and written
and unwritten contracts that explain responsibilities and outline expectations (Robinson
& Rousseau, 1994). This kind of trust is based on employee beliefs about the
organization’s integrity, motives, consistency, and openness (Gabarro & Athos, 1976).

Focusing on trust and locus of responsibility, the role that the organization plays
in the control system is expected to impact employees’ trust perceptions. An external
system, designed with organizational monitoring in mind, is likely to result in negative
trust reactions. Employees may believe that they are being held to extreme levels of
accountability, which have been shown to lead to negative outcomes (Frink & Klimoski,
1998). On the other hand, internal systems which place responsibility in the hands of
employees grant a certain level of freedom. Employees may see this as an indicator that
the organization believes they have integrity, which is a basis for the development of
interpersonal trust (Mayer, Davis, & Schoonnan, 1995). This freedom can result in
greater perceptions of security and stability between employees and the organization,
which will positively impact trust (Strutton, Pelton, & Lumpkin, 1993). Therefore,

H4: Control systems with an internal locus of responsibility will result in more

positive trust reactions compared to systems with an external locus of responsibility.

22

As part of employment with an organization, expectations are a natural part of the
work relationship whether presented via written or unwritten means. While previous
research by Malhan and Mumighan (2002) has shown that non-binding, or informal
contracts, are most likely to establish trust, their experiment was conducted so that the
overall trust evaluation was a function of several rounds of a simulation, where in each
round the conditions of trust had changed. Because the conditions in the current
experiment will remain the same throughout the work session, it is not expected that
employee trust reactions will differ whether control is implemented via formal or
informal methods. Therefore, no hypothesis is offered linking formality of control to
trust reactions.

Aflective Reactions and Deviance

Justice — Deviance. Research by Trevino and Weaver (2001) on corporate ethics
programs and organizational control suggests that positive employee justice perceptions
regarding ethics program follow-through leads to greater willingness to report problems.
Justice has also been linked to positive creativity in devising solutions to a workplace
problem (Clark & James, 1999) and helping behavior in response to procedural fairness
of budgetary procedures (Little, Magner, & Welker, 2002). While the norms in these
environments are not known, to the extent that willingness to report problems, help out,
or display creativity were considered constructively deviant, it is hypothesized that:

H5: The more positive the justice reactions, the greater the frequency of
constructive deviance.

The relationship between justice and destructive deviance has been suggested in

several previous studies on the topic (Ambrose, Seabright, & Schminke, 2002; Cortina &

23

Magley, 2003; Fox, Spector, & Miles, 2001). Increased perceptions of injustice have
been linked to a variety of organizationally-directed negative outcomes, including
sabotage, turnover, and retaliation (Conlon, Meyer, & Nowakowski, 2005). Of most
relevance are the results of Fox et al. (2001), which demonstrated a negative relationship
between both procedural and distributive justice and organizationally-directed deviance,
including behavior such as trying to look busy while doing nothing (withholding effort),
putting in to be paid for more hours than worked (theft), and coming to work late without
permission (tardiness). It is hypothesized that:

H6: The more negative the justice reactions, the greater the fiequency of
destructive deviance.

Trust — Deviance. One stream of research on trust in organizations has examined
the relationship between trust, cooperation, and contributions (de Cremer, Snyder, &
Dewitte, 2001). Individuals who trust more are expected to contribute more in a
relationship in expectation of reciprocation and cooperation. Individuals who have a high
sense of trust are also believed to have a strong sense of moral commitment (Kramer &
Goldman, 1995), which has been further linked to a lessened propensity for theft (J.
Greenberg, 2002). Conversely, individuals who are low in trust have negative
expectations about reciprocity and are less willing to engage in constructive, cooperative
interactions (de Cremer et al., 2001). Therefore, in the context of constructive deviance,
it is hypothesized that:

H7: The more positive the trust reactions, the greater the frequency of

constructive deviance.

24

While still a small body of literature, there is some evidence linking trust
perceptions to decreases in employee productivity, which is one example of destructive
deviance. Hultrnan (2004) suggests that systemic mistrust in an organization can have
negative effects on worker productivity. Along the same lines, Frenkel and Orlitzky
(2005) have suggested that increased trust lead to increases in productivity in a sample of
Australian laborers. As Frenkel and Orlitzky (2005) note, employees who feel trusted
and supported by their organization are going to be more likely to commit to working
hard and being more productive, and at the same time, less deviant. Therefore,

H8: The more positive the trust reactions, the lower the fi'equency of destructive
deviance.

Moderators of T rust/Justice-Deviance Relationship

Conscientiousness. Conscientiousness is an often examined concept in relation to
employee deviance. Conscientiousness refers to the extent to which someone is hard
working, responsible, dependable, reliable, and careful (Barrick & Mount, 1991). Results
consistently demonstrate that more conscientious individuals are less likely to be deviant
than their less conscientious counterparts (Ones & Viswesvaran, 2001), yet recent
research demonstrates that situational factors will interact with personality to produce
interesting effects. In Colbert, Mount, Harter, Witt, and Barrick (2004), the researchers
were focused on examining the interactive effects of personality and situational
characteristics on workplace deviance. They found that the relationship between
workplace perceptions and employee deviance (measured as withholding effort) was
strongest for low conscientious individuals. While the deviance behavior of highly

conscientious individuals did not vary as a function of their workplace perceptions,

25

individuals low in conscientiousness were found to withhold the greatest amount of effort
when they held strongly negative perceptions of the workplace. In a similar fashion, it
was expected that conscientiousness would interact with employee trust and justice
reactions to affect constructive and destructive deviance. Highly conscientious
individuals will refrain from destructive deviance despite feeling untrusted or unfairly
treated, whereas individuals low in conscientiousness may engage in deviance even more
so when motivated by feelings of distrust or injustice.

H9a: C onscientiousness will moderate the relationship between employee trust
reactions and destructive deviance, such that when conscientiousness is low, the
relationship will be stronger.

H9b: Conscientiousness will moderate the relationship between employee justice
reactions and destructive deviance, such that when conscientiousness is low, the
relationship will be stronger.

Research has also supported the link between conscientiousness and
organizational citizenship behavior (LePine et al., 2002). Organ’s (1988) model suggests
that conscientiousness is an integral part of OCB and Van Dyne, Graham, and Dienesch’s
(1994) model supports the inclusion of obedience (of which conscientiousness is a
subset) as a predictor of OCB. Because constructive deviance is being conceptualized as
a subset of organizational citizenship, past research supporting the link between
conscientiousness and OCB also provides support for the relationship between

conscientiousness and constructive deviance. Therefore,

26

H9c: Conscientiousness will moderate the relationship between employee trust
reactions and constructive deviance, such that when conscientiousness is low, the
relationship will be stronger.

H9d: Conscientiousness will moderate the relationship between employee justice
reactions and constructive deviance, such that when conscientiousness is low, the
relationship will be stronger.

Proactive personality. Proactive personality is defined as an individual’s
disposition to affect change in his or her environment (Bateman & Crant, 1993).
Proactive individuals seek out opportunities, demonstrate initiative, and persist in a
course of action until their desired change comes about. Proactive personality has been
linked to the tendency for individuals to shape their work environment and alter their
work methods, procedures, and task assignments (Bell & Staw, 1989). Seibert, Crant,
and Kraimer (1999) also suggest that individuals who are more proactive by nature are
better equipped to understand the contingencies in their environments. Despite feeling as
though they were treated unfairly or do not feel trusted, highly proactive individuals may
still engage in constructive behavior as they receive satisfaction from their demonstrated
initiative.

H10a: Proactive personality will moderate the relationship between employee
trust reactions and constructive deviance, such that when proactivily is high, the
relationship will be weaker.

H10b: Proactive personality will moderate the relationship between employee
justice reactions and constructive deviance, such that when proactivily is high, the

relationship will be weaker.

27

While Bateman and Crant (1993) have stated that not all proactive behaviors are
desirable, there has been very little conceptual or empirical work linking proactive
personality to negative behavior, including destructive deviance. Chan’s (2006) recent
research on proactive personality and situational judgment noted that in most all studies
of proactive personality, being proactive is considered a positive, adaptive trait which has
been linked to better job performance and positive work outcomes. Therefore, no
hypothesis is offered linking proactive personality to negative or destructive deviance as
the literature cannot support this relationship.

Control, Reactions, and Deviance

This final set of hypotheses relates the three key components of the focal
relationship: dimensions of control, trust, justice, and destructive and constructive
deviance. Organizations use formal control systems when they want to enhance their
perceived legitimacy (Roth et al., 1994). Law-like, written formal practices are objective
and provide a concrete tool for assessing appropriate employee behavior. These control
systems will be more effective in limiting destructive deviance overall than will informal
systems. Employees may perceive informal systems to be more transparent or open to
the interpretation of a particular supervisor or individual. With a concrete set of rules and
policies in place, and drawing on earlier hypotheses, it is hypothesized that:

H] 1 a: Formal control systems are more likely to limit destructive deviance than
informal control system, and

HI I b: The relationship between the formality of the control system and the

fi'equency of destructive deviance will be mediated by justice reactions.

28

Formal control systems also better specify the responsibilities of employees
compared to informal systems (Roth et al., 1994). In doing so, employees will be made
aware of what they should and should not do. These specifications may make the
environment seem to favor strict, explicit, and expected behavior. This will have
negative consequences for constructive deviance as employees are limited by formal
responsibilities. Therefore, it is hypothesized that:

H12a: Formal control systems will be more likely to restrict constructive
deviance as compared to informal control systems, and

H12b: The relationship between the formality of control and the fiequency of
constructive deviance will be mediated by justice reactions.

External control systems are regulated by the organization and its agents.
Perceptions of “big brother” may dissuade employees from wavering outside of the
bounds of acceptable behavior as they know they are being watched (Crossen, 1993). In
contrast, the responsibility for monitoring internal systems lies with the employee
himself, and he may be less vigilant in both detecting and reporting of norm deviations.
This has consequences for both destructive and constructive deviance. With internal
systems, the actual lack of accountability to an external source can make breaches of
norms less detectable and more likely (e.g., If I can get away with coming to work late,
and no one will notice it for a while, I am going to keep doing it).

H13a: External control systems will be more likely to restrict destructive deviance
as compared to internal control systems, and

The relationship between locus of responsibility and the fi'equency of destructive

deviance will be mediated by (H13b) justice and (HI 3c) trust reactions.

29

As mentioned, deviation from what is acceptable will be more readily noted by an
external control system. When deviations indicate that the employee is not doing what
they are supposed to (whether for better or for worse), employees may feel compelled to
comply with the rules to avoid potential punishment. In contrast, informal systems may
give employees more latitude to engage in constructive behavior as the individual
employee is responsible for ensuring that they comply with the system and may be more
committed to the system as a whole (Crossen, 1993). Whether the individual chooses
strict or lax compliance is in the eyes of the employee him or herself. Therefore, internal
control systems will promote constructive deviance more than external control systems.

H14a: External control systems will be more likely to restrict constructive
deviance as compared to internal control systems, and

The relationship between locus of responsibility and the frequency of constructive
deviance will be mediated by (H14b) justice and (H14c) trust reactions.

This last set of hypotheses highlights the dilemma faced by organizations wanting
to control positive and negative employee behavior. While formal and external systems
may be beneficial for limiting destructive deviance, it is hypothesized that these systems
are also most likely to restrict constructive deviance. The proposed study is designed to
explore the intended and unintended consequences of control and elaborate on the
complicated nature of discouraging negative deviance at work while at the same time
encouraging positive deviance.

METHOD
Design. This study employed a 2 x 2 fully crossed between subjects design,

varying the formality of control (formal vs. informal) and locus of responsibility (external

30

vs. internal). The field experiment took place in a single 3-hour session where
participants were asked to perform a research task for an organization. The use of field
experiment methodology was particularly important for the nature of the outcome being
studied. Deviance is a sensitive, affectively loaded topic, which can inhibit honest
responding. In addition, self-report methodology commonly used to study deviance
raised validity issues and concerns with biased retrospective reporting of behavior. A
field experiment presented an opportunity to more carefully consider the cause-and-effect
relationship between control and deviance behavior. Most previous research on deviance
(e.g., Robinson & Bennett, 1995, 1997) has used correlational data, which made it
difficult to assess causal relationships and draw strong conclusions about the antecedents
and connectedness between variables. Therefore, a field experiment was conducted.

This method combined the advantages of true experiment methodology, including control
over extraneous variables, along with the enhanced realism of an organizational
environment.

Sample. A total of 251 participants from a large Midwestern university
community participated in the present study. Four participants did not give permission
for their data to be used as part of the research, resulting in a sample of 247 participants.
Due to missing data, 22 participants were further eliminated from the data set for a final
total of 225 participants. This resulted in a sample with 52 individuals in the formal-
extemal group, 59 individuals in the forrnal-intemal group, 53 individuals in the
informal-extemal group, and 61 individuals in the inforrnal-intemal group. The sample
was 61% female and 68% Caucasian. All participants were between the ages of 18-53

(M = 21.56 years) and had an average of 5.62 years of work experience. Everyone was

31

paid for their time during the 3-hour session and sessions were conducted with groups of
people consisting from 2-12 participants. Because data was gathered in a group setting,
group membership was recorded to assess any group-based dependencies affecting the
research outcomes. Beginning with the number one, groups were assigned a number in
the order in which data was collected and 31 groups were run altogether.

To determine appropriate sample size, power analysis was conducted prior to data
collection. Results of the power analysis indicated each cell should contain 60
individuals, for a total of 240 participants. To arrive at this estimate, a literature search
was conducted for studies which examined similar variables to those included in the
current research. Recently, Colbert, Mount, Harter, Witt, and Barrick (2004) conducted a
study examining moderated hierarchical regression models predicting aggression, a form
of destructive behavior, from the interaction between perceptions of the work
environment and individual differences including conscientiousness. Their findings
generated an effect size of R2 = .12. Because the variables being studied by Colbert et al.
(2004) were similar to those in the current study, the effect size served as the benchmark
for the desired effect size included in this power analysis calculation. To calculate
sample size Tabachnick and F idell (2001) recommended using the following formula: N
>= {8/[R2/(1- R2)]} + (m — 1), where m = number of independent variables and N =
number of participants per cell. For the current research, m = 2 and R2 was input at .12
from Colbert et al.’s (2004) research.

Procedure. Participants were recruited from the Michigan State University
campus and surrounding area. A series of flyers advertising a work opportunity through

“Dynamic Research Group” were posted around campus (Appendix A). In addition, the

32

job was advertised in local newspapers as well as on electronic job boards accessible to
college students in the area.

Dynamic Research Group (DRG) was a fictional company created for the purpose
of this research. It is important that participants felt as though they were true employees
of an organization, with real opportunities to engage in workplace behavior. Every effort
was made to ensure that the fictional company appeared real and viable to all
participants. A website and email address were created for this company as a way to
recruit participants and communicate the details of the work opportunity. The
advertisements directed interested participants to DRG’s webpage
(www.dvngmicresearchgroup.com) which explained that DRG was a consulting firm
hired by a client organization whose founder was a graduate of Michigan State. The
premise of the job was that the client was considering using skilled external labor to
outsource the research component of his work process rather than paying full time
employees salary and benefits to do this same kind of work.

Participants were hired by DRG for a 3-hour block of time where they performed
some Internet research, wrote a short report about their findings, and answered some
surveys about their experience. The reports were to inform a decision regarding
outsourcing. Participants worked independently and were paid $10/hour. Applicants
filled out a short application and emailed it to the company (Appendix B). This
application included individual difference measures of the Big Five and proactive
personality as well as a financial knowledge and intemet skills test. All applicants who
successfully completed the application were emailed back with a selection of dates and

times for the work sessions.

33

Prior to the session, participants were emailed a copy of the DRG Terms of
Employment and Employee Expectations (Appendix C). Although participants were only
employees of DRG for a short amount of time, they were expected to abide by company
policies. This l-page document outlined company norms, expectations, and terms of
employment. Participants learned that on the upcoming day of their session, they would
be doing some Internet research and writing a four-page report about their findings. They
were told that they would not be expected to do anything beyond what was outlined in the
Terms of Employment.

Embedded in this document was the manipulation of the type of control system
that was to be used to direct employees' behavior. Work sessions were randomly
assigned to the four experimental conditions. Applicants were assigned to a work session
based on their personal availability. Participants were sent, via email, the Terms of
Employment document that corresponded to their condition and they were asked to read
this document prior to attending their session.

On the day of the work session, employees were welcomed into the work space
and seated at a computer terminal. The sessions were scheduled in MSU campus
computer labs to allow for easy access by MSU students and community members.
Participants were placed at least 1 seat away from the next participant to give room for
people to work as well as reduce the interaction between individuals. The administrator
of the session introduced herself and her assistant as employees of DRG and reminded
employees of the terms of their employment and the type of control that was used to

monitor their behavior.

34

Next, the task for the day was handed out to participants (Appendix D).
Participants were told that they were to visit a series of websites to conduct some
research on the topic of “Advice for the New Investor.” Five investing websites were
pre-selected for employees to visit to complete their research: The Securities and
Exchange Commission (www.sec.gov), The Motley Fool (www.fool.com), Merrill Lynch
(askmerrilmlcom/investmentsfl, Investment U (www.investmentu.coanex.html), and
The Investment FAQ (invest-faq.com). These websites were selected from a wide variety
of available investing websites to provide a mix of information based on investing facts
(e.g., SEC) and investment advice (e. g., Merrill Lynch). Beyond telling employees to
search for information on the general topic of “Advice for the New Investor,” employees
had full discretion to view as much or as little information on each of the web pages as
they felt necessary to complete their work. Employees were told that these five websites
were the only websites they needed to visit during the session and that their responsibility
was to write a report summarizing the information available from each of the websites on
the topic, comparing and contrasting the investment information, and critically evaluating
the information. After viewing and writing about each of the sites, employees were asked
to make a recommendation to the client on the quality of the web resources they had just
investigated. Employees were told they could include a table or figure in their report as a
way to help organize their ideas but that this was not required.

Employees were asked to structure their written report into four sections,
including an opening statement which summarized the report’s contents, 3 research
evaluation section where they were to summarize and compare/contrast each of the

websites, and closing paragraph where they were to make their recommendation based on

35

the quality of information they had gathered. The fourth and final section of the report
was an opportunity to evaluate the prospects associated with investing in a particular
company based on the information employees had just gathered about investing in
general. Each employee was provided with a company name and its stock ticker symbol.
Company names included Ford Motor Company, Apple Computers, Proctor and Gamble,
Verizon, Exxon Mobil, McDonald’s, The Gap, and Eastman Kodak. Company
information could be accessed from several of the project websites (Merrill Lynch, The
Motley Fool) or through other financial related websites employees sought out on their
own (e.g., finance.yahoo.com, moneycentral.msn.com, or the company’s website).
Employees were told that their reports were expected to be 4 pages in length, double
spaced, l2-point font, 1.25” margins, and written using Microsoft Word.

As part of the instructions, employees were reminded that this was an individual
task and any conversation between employees was not permitted. All participants were
given the opportunity to take two 5 minute breaks during the session which took place
approximately 1 and 2 hours after the start of the session. The breaks occurred at uneven
minute intervals (example 9:08 vs. 9:10, 4:07 vs. 4:05) to make it more difficult for
participants to easily tell when the 5 minute break was up. The work began
approximately 10 minutes into the start of the session and all participants were given until
15 minutes prior to the session’s end to complete their work. All participants were told to
use the entire work period to create the highest quality report possible. The last 15
minutes were reserved for participants to complete a series of questionnaires and to
distribute payment. This resulted in approximately 2 hours and 35 minutes of total work

time. As participants completed their written report and indicated they were finished,

36

they were asked to answer a series of questionnaires. If an individual finished the report
before the work time was up, they were asked if their report was complete and then they
were escorted to a separate computer to complete the questionnaires. Otherwise, when
the 2 hours and 35 minutes was up, all participants were asked to save their work to a
floppy disk provided by DRG and were seated at a different computer to respond to the
questionnaires. All questionnaires were administered online and consisted of a
manipulation check, trust, justice, effort, attitudes about the Internet, attitudes about
outsourcing, and demographics questions (Appendix E).

Following these required questions were two questionnaires about teamwork and
financial investing that participants were told were optional and not required. When
finished, participants were escorted away from their work station and debriefed. The
employees were told about the true nature of this work experience, informed that they
had been participants in a research exercise, and had not been performing a real job. They
were given some background information on the purpose of the research and had the
opportunity to ask questions. Participants were asked to give permission to use the data
that had been gathered on them and to sign a permission form if they were willing to do
so (Appendix F). Employees were told that they would be paid whether or not they
granted this permission. After reading and signing the form, all participants were paid for
their time and dismissed.

Throughout the work session, the session administrator and her assistant were
monitoring participant behavior. The session administrator was seated at the front of the
room to give the appearance of being the authority figure without imposing on the work

space of any one individual. She stayed seated throughout the entire work session except

37

in the forrnal-in'temal control condition where she circulated around the room several
times to check in on participants, consistent with that condition’s control manipulation.
The session administrator covertly timed and recorded the length of break that each
participant took, as well as the minutes of participants’ late arrivals or early departures (if
applicable). The frequency of any other observable off-task behavior or chat between
participants was also recorded. After participants began working, the assistant quietly
seated herself in the back of the computer lab and covertly observed participants’
behavior for things that could not be tracked via the computer’s Internet Explorer history,
such as playing Solitaire or using an Instant Messaging program. At the conclusion of
the session, any observations made by the assistant were noted in the participant’s
observation file. After all participants left the room, both the session administrator and
assistant were responsible for removing data from participants’ computers. If a
participant did not give permission for his or her data to be used as part of the research
project, the data was not downloaded from the computer to the participant’s data file.
Manipulations

The manipulations employed in this study were generated from the two
dimensions of control: formality and locus of responsibility. The formality manipulation
underscored whether the type of control being used was formal (written) or informal
(unwritten) in nature. The locus of responsibility manipulation indicated whether an
external agent was monitoring behavior or whether an internal agent (the individual) was
responsible for monitoring their own behavior. One of the following manipulations was

included in the Terms of Employment form sent to participants:

38

F ormal—External

To ensure that employees meet our work standards, we will be monitoring your
workplace behavior. Our code of conduct states that employees must be on task, work
efficiently, and conduct themselves in a professional manner at all times. We will be
monitoring your Internet activity, tracking the web pages you look at, and checking in
with you throughout your work session. You will be provided with these expectations
again in writing when you arrive at your work session.
F ormal-Internal

To ensure that employees meet our work standards, you will be responsible for
keeping a work diary. This diary should be a list of all of your activities during your
work session today. Our code of conduct states that employees must be on task, work
efficiently, and conduct themselves in a professional manner at all times. You will be
expected to commit to making a list of the web pages you visit and of all work-related
behaviors that you do. You will be provided with these expectations again in writing
when you arrive at your work session.
Informal-External

To ensure that employees meet our work standards, the administrator will be
walking through the room to ensure that you are on task, working efficiently, and
conducting yourself in a professional manner at all times. Ethical conduct is a defining
characteristic of all DRG employees. Everyone is expected to conduct themselves
appropriately and in the past, our employees have had a strong desire to comply with

these principles.

39

Informal-Internal

To ensure that employees meet our work standards, we trust that you will commit
to our values of keeping on task, working efficiency, conducting yourself in a
professional manner at all times. Ethical conduct is a defining characteristic of all DRG
employees. Everyone is expected to conduct themselves appropriately and in the past,
our employees have had a strong desire to comply with these principles.

A more elaborate version of this manipulation was emphasized to employees at
the start of their work session, reiterating their Terms of Employment and the type of
control system that was in place during that day. Out of the four conditions of control,
the formal-intemal condition required some extra work of participants in that they were
asked to keep a work diary. Participants were provided with a sheet to track their
workplace activity and were asked to record their actions in a fair amount of detail.
During pilot testing, participants said that the time for them to complete the work diary
was negligible (under 5 minutes) and did not affect their ability to complete their work
within the given time.

Pilot Testing

Prior to full data collection, a sample of 43 individuals (who did not participate in
the full experiment) was used to conduct extensive pilot testing for several reasons,
including: verifying psychological fidelity of control manipulations, gathering evidence
that participants would engage in destructive and constructive behaviors as
conceptualized and measured, verifying reliability of the deviance behavioral measures
(e. g., will Internet Explorer history provide the necessary information or are participants

clearing their history), assessing variance on destructive and constructive behavioral

40

measures, questioning their knowledge of norms (both work group and hypemorms), and
understanding participant reactions to the experience itself, as a whole and in part. The
quality of many of the experimental measures depended on being able to demonstrate the
validity of the assessed manipulations.

Pilot testing examined the effects of all four conditions on behavior and beliefs
about control. While the results did not reach significance, there was evidence that
participants in the different conditions were sensitive to the control manipulations
through the manipulation checks. For example, participants in the external condition (11 =
14) more strongly endorsed the items referencing external control (“Dynamic Research
Group was monitoring my behavior” and “An administrator walked through the room
during the session”) than did participants in the internal condition. On a scale of 1-5
where l=strongly disagree and 5=strongly agree, means were 4.00 (external) and 3.12
(internal) for the first item and 4.21 (external) and 3.29 (internal) for the second item.

During pilot testing, participants were interviewed about their reactions to the
experience and about each aspect of the research process at the conclusion of the work
session. Data from these interviews also indicated that participants were sensitive to the
control manipulations. A participant from the formal-extemal condition said that she
thought the control seemed a little harsh but that the written expectations did impact what
she did and did not do during the session. One participant from the inforrnal-intemal
condition indicated that the norms presented at the beginning of the session were strong
and if they had been told that the Dynamic Research Group was monitoring their
behavior then they would not have looked at other websites. A second participant from

this same session said that since the expectations were pretty relaxed and the

41

administrator was not walking around looking at what he was doing that he felt freer to
check other websites and do what he wanted during the work session. These and other
participants’ responses were used to alter the pilot procedures which resulted in the final
version described earlier. Their responses also gave evidence that the measures of
deviance were justified and effective in this environment and that the expectations and
amount of the work required of employees was not too much or too little for the time
allotted.

Measures

To capture destructive and constructive deviance, several behavioral observations
were made during data collection. These observations included: length of break 1
(minutes), length of break 2 (minutes), length of any extra breaks (minutes), chatting with
other employees (frequency), any off task behavior such as playing Solitaire or using
Instant Messaging software (frequency), arriving late to the session (minutes), length of
written report (number of pages), presence of “extras” such as a table or figure in the
written report (number), and the extent to which the participant responded to the optional
questionnaires at the end of the session (percent complete). The rationale for the
inclusion of these variables as deviant is described below.

In addition, participants’ web activity was tracked via the History function within
Microsoft Internet Explorer. The History function captured the name of each site visited
(e.g., www.msu.edu) as well as the number of pages of information viewed within that
site. An individual could visit www.msu.edu and simply view the home page (which

would record 1 page of information) or someone could view the home page, click on

42

Academics and Research (http://www.msu.edu/academics_research/index.html), and
further explore Academic Units by College (http://www.msu.edu/common/academic/
units.html). This would result in a total of 3 pages viewed under the www.msu.edu
umbrella. If a participant went to another website such as www.cnn.com and viewed a
one page story about the weather and another one page story about politics, this would
result in a total of 2 pages viewed under the www.cnn.com umbrella. In total, this
participant has visited 2 different sites but viewed 5 total pages of information.

For each of the five websites required as part of the written report, an observation
was made for whether or not the participant visited the site, followed by the number of
pages of information that were viewed. All other websites participants visited were
categorized into either extra financial-related websites or non-work related websites.
Examples of extra financial websites included websites that provided general financial
information such as money.cnn.com or the New York Stock Exchange (www.mysecom)
or the website of the organization that corresponded to the company each participant was
asked to make an investment recommendation about (e.g., Proctor & Gamble,
wwwfipgcom). Examples of non-work websites included any email sites (mail.msu.edu),
e-commerce sites (www.cbay.com), or online communication sites (www.myspace.com).
Observations were taken for whether or not each participant visited either extra financial-
related websites or non-work websites. If the participant visited either, the total number
of sites visited was recorded by category, along with the total number of pages viewed
within each site.

To determine which behaviors to include as deviant, the norms and expectations

established from the Terms of Employment served as the baseline for judging behavior as

43

deviant or not deviant. While the definitions of deviance provided by Warren (2003) cite
group norms as the proximal referent, in the current study, there was not enough time for
norms to develop at the group level separate from the norms and expectations set at the
organizational level within each control condition. Recognizing that norms can operate
at multiple levels, in the current study the norms set at the organizational level were a
better referent for judging deviant behavior and were used as the benchmark for assessing
deviance.

For destructive deviance, any behaviors which violated both a priori work group
norms and hypemorms were considered. In the Terms of Employment, the work group
norms were established as all participants needing to keep on task, work efficiently, and
conduct themselves in a professional manner at all times. The generally accepted
hypemorm was similar in nature, with employees being expected to meet the
organization’s work standards as part of their employment. For constructive deviance,
the Terms of Employment stated that employees were not expected to do anything
beyond what was originally asked of them. This established the a priori work group
norm. By providing opportunities to employees to go above and beyond expectations in
adding a table or figure to their written report or completing optional questionnaires,
opportunities to comply with the hypemorm of general good will were created.

Intercorrelations among deviance (destructive and constructive) and performance
appear in Appendix G. The behaviors were standardized to accommodate the different
measurement scales (e. g., frequency, minutes). To be considered either destructively or
constructively deviant, the pattern of correlations demonstrate some content validity in

that destructive behaviors were positively correlated and constructive behaviors were

44

positively correlated. For example the length of break 1 and break 2 correlated r = .21;
people who took a long (or short) break at time 1 also took a long (or short) break at time
2. Participants who were more likely to chat with others also took a longer break at time
2 (r = .13). The number of non-work related websites visited also positively correlated
with taking an extra break (r = .18) and off task behavior (r = .22). Employees who
wrote a longer report were less likely to take a long break 1 (r = -.15) but were more
likely to include extras in their report (r = .19) and complete the extra teamwork
questionnaire (r = .20) and financial questionnaire (r = .15). These correlations support
the expected connections between destructive and constructive behaviors and legitimate
their inclusion in an overall measure of deviance. Interestingly, individuals who were
more likely to arrive late were less likely to visit extra financial websites (r = -.l3), but
individuals who viewed more extra financial websites also visited more non-work related
websites (r = .22) suggesting that the relationship between destructive and constructive
behavior warrants further exploration.

With little opportunity for employees to talk, interact, or communicate, it is not
likely that separate norms developed in individual work sessions related to the time on
task, websites visited, et cetera. However, because individuals were clustered in sessions,
a group level (session effect) was possible. To account for this potential effect, group
session membership was considered as a level of analysis in a hierarchical linear model.
The modeling of group membership took into consideration any group-specific effects
that may have emerged in the experimental setting. However, because the focus of this

research is on the overall effect of the control manipulation on deviance and not of group

45

session membership on deviance, the definition of deviance will stem from the
overarching expectations that were set for all employees.

Destructive deviance. Several indicators of destructive deviance were gathered
from behavioral observation data. Spending time checking email, visiting ESPN.com, or
any other non-work related website violated these norms. One indicator of destructive
deviance was whether or not participants visited non-work related websites, the total
number of these websites visited, and the number of pages of information viewed.
Because the amount of time spent on each page could not be tracked, the amount of pages
viewed served as an indicator of the degree of destructive deviance.

The length of employees’ breaks was judged by the experimenter. By telling
participants that they had 5, and only 5 minutes, for their breaks, any participant who
took longer than a 5 minute break was violating the norm of keeping on task as well as
not meeting the organization’s standards. In this case, the measure of destructive
deviance was whether or not the participant took longer than a 5 minute break and the
total length of their break. The experimenter who sat in the room recorded the time on
individual observation sheets when each employee stopped working and stood up fi'om
their computer. When the employee sat back down at their computer, the time was
recorded again. Rather than leaving the room, some employees took their break while
sitting at their work station. This was evident when an employee stopped typing, pushed
his or her chair back from their computer, and sat and relaxed for the break time. For
these employees, the start of the break time was recorded when they stopped typing. As
soon as the employee began working at the computer again, the time was again recorded.

A clock on the experimenter’s computer was used to measure start and stop times and

46

when multiple employees took their break simultaneously, the experimenter recorded
each person’s time on his or her individual observation sheet so as not to confuse the
length of break between people.

Participants were observed for additional behaviors that also violated work
standards and kept them off task. These behaviors included chatting with other
employees, arriving to the session late, and any other off task behavior. Off task behavior
included instances of playing computer games such as Solitaire, using an Instant
Messaging program, sending or receiving text messages with a cell phone, or
daydreaming. The experimenter who recorded break times also observed employees for
these off task behaviors. These observations were supplemented by observations from
the experimenter’s assistant and where the observations overlapped, only one instance of
the off task behavior was recorded in the participant’s data file. Table 3 provides a
summary of all destructive deviance behaviors and the number of employees engaging in
each behavior.

Destructive deviance was based on seven behaviors. These were: length of break
1, length of break 2, length of extra break, number of non-work related websites visited,
number of times chatted with others, number of times off task, and length of lateness.
For chat, off task, and number of non-work websites visited, the raw scores were the sum
of the number of occurrences. For length of break 1 and 2, the raw score was time in
minutes when the break exceeded 5 minutes because the instructions indicated that a 5
minute break was allowed for each of these; destructive deviance occurred only when
employees took longer than 5 minutes. For the extra break and length of lateness, the raw

score was time in minutes as no extra breaks or lateness was permitted. Finally, for the

47

Table 3

Number of Employees Engaging in Destructive Deviance (N =225 )

 

Destructive Behavior # Yes # No

Taking longer than 5 minute breaks

Break 1

R = 6 — 9 minutes 92 133
Break 2

R = 6 — 9 minutes 42 183
Extra break

R=1-5minutes 8 217

Visiting non-work related websites 210 15
R = 1 — 28 total sites
R = 1 — 157 total pages

Chatting with other workers 24 201
R = l — 8 instances

Off task behavior 23 202
R = l - 4 instances

Arriving late to session 16 209
R = 5 - 25 minutes

overall destructive deviance score, each raw score was standardized and the seven
standard scores were summed as a measure of an individual’s level of destructive
deviance. Coefficient alpha for this scale was a = .41. While alpha is low, it bears
mentioning that for the domain of destructive deviance the link for any individual
behavior to the overall domain is not meant to be strong because there are many ways to
display deviance. Because actual behavior was measured, being deviant in one aspect of
behavior (e. g, chatting with others) may mean that employees were not doing another

(e. g., off task playing Solitaire). Therefore, while each behavior contributes to the overall

48

measure of deviance, the individual behaviors are not intended to be highly positively
correlated, which has resulted in a low coefficient alpha.

It is important to note that a large percentage of individuals did visit non-work
related websites (210/225 participants) yet this behavior was not readily obvious to other
participants in the room. Therefore, while the majority of participants engaged in this
behavior, the subtle nature of the deviance did not disrupt the a priori norm and this
behavior was still considered to be destructive in nature. Furthermore, since it was a
destructive behavior in which all employees were likely to have thought they could do
without being observed, it warrants inclusion in the destructive category as the action
kept people from doing their work.

Constructive deviance. Several indicators of constructive deviance were gathered
from behavioral observation data. Spending time visiting any additional financial-related
websites that could be used to enhance their research report, such as financeyahoocom.
violated the expectations that no other websites were to be visited, yet the outcome is
beneficial because sites like these contain information which can enhance the quality of
the written report. Therefore, one indicator of constructive deviance was whether or not
participants visited additional financial-related websites, the total number of these
websites visited, and the number of pages of information viewed. Because the amount of
time spent on each page could not be tracked, the amount of pages viewed served as an
indicator of the degree of constructive deviance.

The length of employees’ breaks was judged by the experimenter. By telling
participants that they had 5, and only 5 minutes, for their breaks, any participant who

took shorter than a 5 minute break was violating work expectations but to the benefit of

49

the organization as the employee got back to work sooner. In this case, the measure of
constructive deviance was whether or not the participant took shorter than a 5 minute
break and the total length of their break.

Participants were observed for additional behaviors that also violated work
standards but went above and beyond expectations. These behaviors included writing a
report longer than the 4-page requirement and adding any non-required tables or figures
to their written report. The experimenter checked each written report for page length as
well as the number of tables or figures included and recorded these variables in
participants’ observation files.

One final indicator of constructive deviance was employees’ willingness to
complete some optional questionnaires at the conclusion of the work session.
Participants who did complete additional questionnaires complied with the generally
accepted hypemorm of helping out or having a good will and at the same time violated
the work group norm of not being expected to do anything beyond the original terms. At
the very end of the work session after employees completed the questionnaire items about
justice and trust, two optional questionnaires were given to participants. The first
optional questionnaire was a 2-item essay questionnaire about teamwork experiences and
the second was a lO-item multiple choice/short answer questionnaire about financial
investing. The teamwork questionnaire was presented before the financial investing
questionnaire with the thought that the essay questionnaires would take some additional
effort and truly require individuals to go “above and beyond” if they employees to fully
complete both. Because the additional questionnaires contained multiple items, the

indicator of constructive deviance was the percentage of questions completed (ranging

50

from 0-100%). Table 4 provides a summary of all destructive deviance behaviors and the

number of employees engaging in each behavior.

Table 4

Number of Employees Engaging in Constructive Deviance (N =225)

 

 

Constructive Deviance # Yes # No
Taking shorter than 5 minute breaks
Break 1 102 123
R = 0 — 4 minutes
Break 2 168 57
R=0—4mmM%
Visiting extra financial-related websites 197 28

R= l - 19 total sites
R = 1 — 139 total pages

Responding to extra team questionnaire

Responding to extra financial questionnaire

Writing longer than 4-page report
R = 4.25 — 8 pages

Adding extra (graph/figure) to report
R = l — 3 extras

l4= 50% 107

104= 100%
6= 10% 64
2=20%
1=30%
5 =40%
2=50%
15 =70%
l3 =80%
l9=90%
98=100%
72 153
43 182

Constructive deviance was based on seven behaviors. These were: length of

break 1, length of break 2, number of additional financial-related websites visited, length

51

of written report, number of tables and figures added to report, and percent complete of
optional questionnaires. For number of tables/figures and number of additional financial-
related websites visited, the raw scores were the sum of the number of occurrences. For
length of break 1 and 2, the raw score was time in minutes when the break was less than 5
minutes because the instructions indicated that a 5 minute break was allowed for each of
these; constructive deviance occurred only when employees took less than 5 minutes.

For the length of the written report, the raw score was the number of pages when the
page length was longer than 4 as the instructions stated that a 4 page report was expected
as part of the work. Constructive deviance occurred only when employees wrote a longer
report. For the percent complete of optional questionnaires, the raw score was the
percentage of questions answered. Finally, for the overall constructive deviance score,
each raw score was standardized and the seven standard scores were summed as a
measure of an individual’s level of constructive deviance. Coefficient alpha for this scale
was or = .35. As with the measure of destructive deviance, while alpha is low, it bears
mentioning that for the domain of constructive deviance the link for any individual
behavior to the overall domain is not meant to be strong because there are many ways to
display deviance. Because actual behavior was measured, being deviant in one aspect of
behavior (e. g, visiting additional financial related websites) may mean that employees
were not doing another (e.g., responding to optional questionnaires). Therefore, while
each behavior contributes to the overall measure of constructive deviance, the individual
behaviors are not intended to be highly positively correlated, which has resulted in a low

coefficient alpha.

52

Similar to the behavior of individuals in the destructive category, it is important to
note that a large percentage of individuals did visit extra financial related websites
(197/225 participants) yet this behavior was subtle in nature and was not readily obvious
to other participants in the room. Therefore, while the majority of participants engaged in
this behavior, the subtle nature of the deviance did not disrupt the a priori norm and this
behavior was still considered to be constructive in nature.

Conscientiousness. As part of the employment application, conscientiousness was
measured using a 10-item measure from the IPIP (Goldberg, 1999). Coefficient alpha for
this scale was 0.81.

Proactive personality. As part of the employment application, a modified version
of Bateman and Crant’s (1993) 17 item scale of proactive personality was used. A
sample item is “Wherever I have been, I have been a powerful force for constructive
change.” Coefficient alpha was 0.87.

Financial knowledge. Included in the employment application was a 15 item
financial knowledge test used to assess participant’s knowledge and ability. The items
were taken from two financial investments websites (http://www.investoredca and
http://www.investorwords.com). Items were scored 0 for incorrect and l for correct.

Manipulation check Using the manipulations as a basis, several items were
created. Items are listed in Appendix E. Three items were created to check the
effectiveness of the external control manipulation (Items 1, 5, and 8). An example item
was: “I knew that my company Dynamic Research Group was charged with monitoring

my behavior.” These items were combined into a single external manipulation check

scale with a = .55.

53

Two items were created to check the effectiveness of the internal control
manipulation (Items 4 and 7). An example item was: “I knew that I, and not Dynamic
Research Group, was charged with monitoring my own behavior.” These items were
analyzed independently because the alpha for the scale was too low (or = .12).

To assess the efficacy of the formal/informal control manipulation, three
manipulation check items were created, one of which (Item 2) addressed participants’
sensitivity to the formal manipulation: “At the start of today’s session, I was given a
written, formal document explaining how I was expected to keep on task, work
efficiently, and conduct myself in a professional manner at all times.”

The remaining two items (Items 3 and 6) specifically related to the informal
control manipulation addressed the notion that participants were told in an informal,
unwritten manner about the work expectations of the day and that DRG expected
participants to commit to their values and comply with their principles. These items were
combined into a scale for the informal condition manipulation check (a=.49).

Justice. Organizational justice was measured using Colquitt’s (2001) 4 item
measure of distributive justice, 7 item measure of procedural justice, and 4 item measure
of interpersonal justice. Coefficients alpha for these scales were 0.88, 0.67, and 0.91 ,
respectively.

Trust. Robinson and Rousseau’s (1994) 7 item scale of trust was used to measure
employee perceptions of the organization’s trust in them. A sample item is “Dynamic
Research Group is open and upfront with me.” Coefficient alpha was 0.74.

Additional measures. Several additional measures were included in the set of

final questionnaires to which participants responded to further conceal the nature of the

54

experiment. Because it was important for employees to believe in the realism of the work
experience, these measures were included in the final set of questionnaires as a way to
enhance the face validity of this portion of the experiment. Without these additional
questionnaires, employees may have been suspicious about the psychological nature of
many of the manipulation check, justice, and trust items.

These additional questionnaires included attitudes about the Internet (Lavoie &
Pychyl, 2001), attitudes about outsourcing, and a 5-item measure of withholding effort
modified fiom Kidwell and Robie (2003). The measure of withholding effort (or = 0.79)
was used to assess the construct validity of the destructive and constructive deviance
behavioral measures. Consistent with expectations, withholding effort was positively
correlated with the standardized composite of destructive deviance (r = 0.20, p < .01)
suggesting that as participants reported withholding more effort, more destructive
deviance was observed. Likewise, withholding effort was negatively correlated with the
standardized composite of constructive deviance (r = -0.16, p < .05) suggesting that as
participants reported withholding less effort, more constructive deviance was observed.

Several measures of performance were also gathered as a way to understand
baseline expectations for intemet activity and general work conduct. Performance is not
a variable that will be tested in the experimental model but nevertheless it does help to
understand the types of things employees were doing while they were working.
Performance measures included whether or not participants visited all five required
websites and the number of pages of material viewed across all websites, including
required and non-required pages. Two hundred twenty one participants visited all

required sites with an average of 60.29 pages viewed across these required sites. When

55

participants visited other financial and content-related sites, they viewed on average
12.00 pages of information; when visiting non-work related websites, an average of 14.16
pages of information were viewed. In total, participants viewed an average of 86.45
pages (across required and non-required sites). There was a wide range of activity within
this performance dimension (R = 14-195 pages viewed), suggesting that the number of
pages viewed is just one aspect of performance but not a substitute for the thoroughness
of the job completed. To assess the thoroughness of performance would require
evaluation of each of the written reports, which is beyond the scope of this study.
However, with the expectation of the written report set at 4 pages, performance was also
assessed via the length of the written report. Participants did fulfill this performance
expectation as reports were, on average, 3.96 pages long. Judging from intemet activity,
it is reasonable to say that participants were largely on task (60 out of 80 pages viewed,
on average, came from required sites) and met expectations for the length of the written
report.
RESULTS

A summary of the results of all hypothesis tests is provided in Appendix H. The
means, standard deviations, intercorrelations, and where appropriate, internal consistency
estimates of reliability, are provided for all variables in Table 5. The sections below
describe the effects of the experimental manipulations on the hypothesized relationships
as well as additional analyses used to test the experimental model in more detail.

Hierarchical linear modeling (HLM, Raudenbush, Bryk, & Congdon, 2000) was
used for many analyses. HLM was appropriate for these data because individuals were

nested within groups when the data was collected. Because of this nested structure, as

56

.oREomnN
c3er_ geese 825-3 .32 8.8 3 383333 833:8 r. 323$ 35 8.8 3 33332.. 83338 _. .mmmuz .eez

 

Ed- 86 mod- cod wed- wed- smfid mod ma; ofim 353% 03585800.:
.42.? Ed- :6 8.0.. 3.0.. mod- 36 mod- Sam mod 8.33% o>zo=bmoa.w_

 

8.8 8.8 8.8 8.8 8.? 8.? :88 :88 8.» 8.: eaeeeaee 820.2
8.? 8.8 8.8 8.8 8.? 8.? 8.? 8.8 8: 83 28333 359:
8.? :.? 8.? .2? 8.8 z? 8.? 8.8 own New 85385 325345.:
8.8 8.? 8.8 8.? 8.8 .28 8.8 8.8 as 8; 33.5.2
8.? .2? 8.? 2.? 8.8 8.? 8.? 8.8 8.4 8.: 8.3:
8.8 8.8 .2? 8.? 8.8 z? 8.? 8.8 83 Se 38285333592
:2? tom? :8? :8? :8? :8? 8.8 8.8 8.8 8: some 833.538
:5 :38 :38 :88 .28 :88 8.8 8.8 8.8 8.4 25.3
:3 :32 :88 .28 :28 8.? 8.8 28 8.4 easea 3833935 .5
£3 :38 .28 :88 8.? 8.8 88 m3. 85.. 333535 .e
33 8.8 :88 8.? 8.8 $8 8.4. 833 35385 .3
E: :38 8.? 8.? one 8.3. 83:88 388$ .4
:3 8.? 8.8 8.8 :3 3838:8388 .m
-- 8.? -- -- 3.55-385 .N
-- -- -- 3§ea_-3§ee ._

w e e m e m N _ am 2 2.33;

 

$38.6; ~S§Etee~xmxo 3833.989 we: .EotSSQ 38328.0 .83»:

h e335

57

 

 

.2? 8.8 8.? 8.? 8.8 8.? .28 .2? 8.338 838380.:
8.? 8.? 8.? 8.? 8.? 2? :88 8.33% 33835.2
8.8 8.8 :8 8.? 8.8 8.8 3385...: 9.80.2
8.8 8.? 2.? .28 8.8 28333 ”seams.
8.? :38 2.8 8.8 888%.. 3e53§ .2
8.8 :2? 8? 3380 .2
2.8 8.? am... .:
8.? 832322 33.35 .2
85 some mazeefia .e
2 2 3 2 2 : 2 8 235.5

 

833.8; ~S§Etemxm~8 3888th 38 .azo.:8.§eQ 28828285 .88»:

«8.38.6» m. 8385

58

well as unequal sample size and unequal variance between groups, HLM analysis was
more appropriate for the data than AN OVA analysis, which is commonly used to test
similar between-group differences, minus levels of analysis. When interpreting HLM
results, there are several effects to consider. To determine the effect of the manipulations
on behavior, a significant effect should be present for the level 2 predictor of the
formal/informal control manipulation (701) or the extemal/ internal control manipulation
(702) as indicated by the hypotheses. Proponents of HLM suggest that one precondition to
testing for level 2 predictors is sufficient between-group variance (to) at level 1
(Hoffman, 1997). However, for the current study, between-group variance at level 1 is
less of an issue because the focal predictions are based on the effect of the manipulations
on behavior rather than the effects of group membership on behavior. While significant
between-group variance at level 1 will be reported, analyses still proceeded when
between-group variance was not significant and level 2 tests for the effects of the
manipulations were conducted.
Manipulation Checks

Mean values of the manipulation check items/scales by condition are presented in
Table 6. Item numbers correspond to the scale presented in Appendix E. In the event
that group size was to influence behavior, size was included as a grand mean centered
predictor in all level 2 analyses. Grand mean centering is recommended when both
categorical and continuous predictors are used in HLM analysis (Hoffman & Gavin,
1998) as was the case with categorical dichotomous manipulations and continuous group

size.

59

Table 6

Mean (SD) Values of Manipulation Check Items by Condition (N =225 )

 

 

Formal Informal External Internal
External (Items 1, 5, 8) 3.55 (0.72) 3.36 (0.84) 3.72 (0.70) 3.22 (0.78)
Internal
Item 4: Work diary 3.57 (1.56) 2.07 (1.01) 2.18 (1.10) 3.36 (1.60)
Item 7: Self-responsible 3.60 (1.50) 3.96 (0.88) 3.62 (1.09) 3.93 (0.86)
Formal (Item 2) 4.69 (0.69) 4.53 (0.85) 4.65 (0.69) 4.57 (0.84)
Informal (Items 3, 6) 3.96 (0.87) 4.18 (0.84) 4.07 (0.85) 4.08 (0.88)

 

External control. Results revealed significant between-group variance at level 1,
to = 0.08, df= 30, x2 = 65.05, p < .01, ICC = 0.15. Further analysis revealed that
participants in the external condition more strongly endorsed these items than did
participants in the internal control condition, (yo; = —0.51, se = 0.11, t = -4.52, p < .01),
providing support for this manipulation. Afier adding these level 2 predictors, the
between-group variance which was previously significant at level 1 was no longer
significant, To = 0.03, df= 27, x2 = 34.82, p = .14.

Internal control. For the item “I was asked to keep a work diary, including a list
of the web pages I visited and all the work-related behaviors I did,” results revealed
significant between-group variance at level 1, to = 1.36, df= 30, x2 = 340.79, p < .01 ,
ICC = 0.59. Further analysis revealed that participants in the internal condition more
strongly endorsed these items than did participants in the external control condition, (702

= 1.24, se = 0.27, t = 4.64, p < .01) (Table 6). Afier adding these level 2 predictors, the

60

between-group variance which was previously significant at level I remained significant,
to = 0.46, df= 27, x2 = 126.45, p < .01.

Similarly, participants in the internal condition also more strongly endorsed the
second internal condition manipulation check, “I knew that I, and not Dynamic Research
Group, was charged with monitoring my own behavior,” (702 = 0.30, se = 0.13, t = 2.41, p
< .05). Results revealed marginally significant between-group variance at level 1, to =
0.06, df= 30, x2 = 41.97, p = .07, ICC = 0.06, which was not significant after entering the
level 2 predictors, to = 0.01, df= 27, xz = 27.71 , p = .43. These analyses combined
provide support for the success of the internal control manipulation.

Interestingly, participants in the informal control condition also more strongly
endorsed the belief that they were individually more responsible for monitoring their
behavior [Item 7, (701 = 0.36, se = 0.13, t = 2.88, p < .01)] and participants in the formal
condition more strongly endorsed the belief that they were asked to keep a work diary
[Item 4, (701 = -l .48, se = 0.27, t = -5.58, p < .01)], despite these items not reflecting
content unique to the formal or informal control condition. This is likely due to the
nature of the between subjects design which fully crossed the extemal/intemal
manipulation with the formal/informal control manipulation; participants in the formal
and informal cells were also exposed to the external or internal manipulation per their
respective condition.

Formal control. Results revealed no significant between group variance at level 1,
to = 0.01, df= 30, x2 = 32.70, p = .33, ICC = 0.02, but a marginal effect for the
formal/informal condition on this item at level 2, 701 = -0.17, se = 0.10, t = -1 .72, p = .10,

suggesting that participants in the formal control condition were somewhat more likely to

61

perceive they had been given a written, formal document outlining the specifications of
control. In retrospect, while participants in the formal condition did receive special
written expectations outlining the control system that would be in place, all participants
were provided with both written instructions as to how to complete the intemet search
task as part of their job for the day, which could have been interpreted in the same way as
the manipulation for control.

Informal control. Results at level 1 revealed no significant between—group
variance, to = 0.01, df= 30, x2 = 29.33, p > .50, ICC = 0.01. Despite low alpha (ct=.49),
results at level 2 revealed a significant effect for the formal/informal condition on the
informal manipulation check scale (701 = 0.23, se = 0.10, t = 2.17, p < .05) such that
participants in the informal condition more strongly endorsed these items than
participants in the formal control condition. This suggested that the informal control
manipulation was successful.

While results from this overall analyses demonstrated that the formal
manipulation was weaker than the other control manipulations, significant differences
observed for all other items and/or scales warranted further exploration of the effects of
both the formal/informal and extemal/intemal control manipulations on process reactions
and deviance behavior.

Hypotheses 1-4

To examine the effects of the manipulations on process reactions, HLM was

performed. The covariates of demographics, financial knowledge, and years of work

experience did not correlate significantly with the manipulations and were therefore not

62

included in the HLM analysis. Group size was included as a level 2 predictor. Mean

values for all justice and trust variables are presented in Table 7.

Table 7

Mean (SD) Values of Process Variables by Condition 0V=225)

 

 

Formal Informal External Internal
Procedural Justice 3.87 (0.52) 3.88 (0.45) 3.92 (0.49) 3.84 (0.49)
Distributive Justice 4.09 (0.69) 4.15 (0.69) 4.17 (0.62) 4.08 (0.74)
Interpersonal Justice 4.63 (0.50) 4.69 (0.49) 4.67 (0.50) 4.65 (0.49)
Trust 4.03 (0.53) 4.05 (0.52) 4.00 (0.53) 4.08 (0.52)

 

Hypothesis 1. Hypothesis 1 suggested that internal control would result in more

positive procedural justice reactions than would external control. Despite significant

between-group variance at level 1, to = 0.02, df= 30, x2 = 51.41, p < .01, ICC = 0.08,

results revealed no significant effect of external or internal control on procedural justice,

702 = -0.06, se = 0.08, t = -0.70, p = .49. Therefore hypothesis 1 was not supported.

Hypothesis 2. Hypothesis 2 suggested that formal control would result in more

positive distributive justice reactions than would informal control. There was no

significant between-group variance at level 1, to = 0.01, df= 30, x2 = 36.16, p = .20, ICC

= 0.02. Further results revealed no significant effect of formal or informal control on

distributive justice at level 2, 701 = -0.08, se = 0.09, t = -0.89, p = .38. Therefore

hypothesis 2 was not supported.

Hypothesis 3. Hypothesis 3 suggested that external control would result in less

positive interpersonal justice reactions than would internal control. With no significant

between-group variance at level 1, To = 0.01, df= 30, x2 = 32.45, p = .35, ICC = 0.04,

63

further results revealed no significant effect of external or internal control on
interpersonal justice at level 2, 702 = -0.00, se = 0.07, t = -0.06, p = .95. Therefore
hypothesis 3 was not supported.

Hypothesis 4. Hypothesis 4 suggested that internal control would result in more
positive trust reactions than would external control. There was no significant between-
group variance at level 1, to = 0.01, df= 30, x2 = 36.21, p = .20, ICC = 0.04. Further
results revealed no significant effect of external or internal control on trust at level 2, yo;
= 0.09, se = 0.07, t = 1.16, p = .26. Therefore hypothesis 4 was not supported.

Summary. Contrary to expectations, the control manipulations did not result in
any differences in justice or trust reactions. Post-hoe analysis suggested that restricting
the sample to participants in the 10th and 90th percentile on justice (11 = 13 — 37
individuals, depending on form of justice) and trust (n = 35) reactions still did not result
in significant differences as predicted by the hypotheses. For this type of work, the
presence of control expectations may not have been so extraordinary that it incited
negative reactions. Employees’ ability to do their work was not greatly hampered by any
one form of control, so control may not have been as offensive or limiting. Without
knowing any differently, the type of control seemed reasonably fair and trustworthy and
the strongly negative reactions were absent. Because the manipulations were not
significantly associated with the process reactions, the mediation hypotheses (H 11b — H
14b) cannot be supported and will not be tested.

Hypotheses 5 -8
This set of hypotheses suggested that the process reactions would affect the two

dependent variables in a variety of ways. To examine these hypotheses, simple

64

regression analyses were conducted. Group membership was not included as a covariate
because its influence was already accounted for in the relationship between the
independent variables and process measures prior to this set of analyses. For all analyses,
the appropriate justice or trust variable was entered in Step 1 and the dependent variable
was the standardized composite of either destructive or constructive deviance.

Before describing the hypothesis tests for destructive and constructive deviance, it
is important to understand the distribution, range, and nature of the standardized deviance
composite scales. These values are displayed in Table 8.

Table 8

Descriptive Statistics for Deviance (N =225 )

 

Constructive Deviance Destructive Deviance

Mean 3.19 2.08
Standard Deviation 1.95 2.41
Range 0.0—ll.6l 0.0—13.13
Kurtosis 1 .5 l 4.95
Skewness 1 .00 2. 10

 

Both distributions were positively skewed but this was to be expected as deviance is

typically a low base rate behavior (Figures 2 and 3).

65

Figure 2

Distribution of Destructive Deviance

 

 

 

0.00 2.00 4.00 6.00 8.00 10.001200

Destructive Deviance (standardized
composite)

The constructive deviance composite also had positive kurtosis, indicating that a
large number of individuals engaged in a small amount of constructive behavior. While
some scholars advocate for the transformation of non-normal data, it was important to
retain the low base rate distribution which is somewhat unique to deviance behavior.
Therefore, the common square root data transformation for positively skewed data and
logarithm for positive kurtosis was not used to normalize the distribution as these
transformations can sometimes hamper the interpretation of variables, especially when

data are grouped (Tabachnick & Fidell, 2001).

66

Figure 3

Distribution of Constructive Deviance

 

 

 

 

 

 

 

 

 

0.00 2.00 4.00 6.00 8.00 10.001200

Constructive Deviance (standardized
composite)

Hypothesis 5. Hypothesis 5 suggested that more positive justice reactions would
result in greater fiequency of constructive deviance. Results revealed no significant
effects for any of the justice variables on constructive deviance, and therefore Hypothesis
5 was not supported (Table 9).

Table 9

Summary of Regression Analysis for Justice Variables Predicting Constructive
Devigpce (N =225)

 

variable B SE8 0 R2
Procedural 0.04 0.30 0.01

Distributive -0. 19 0.22 -0.07

Intemnal 0.21 0.29 0.05 0.01

 

67

Hypothesis 6. Hypothesis 6 suggested that more negative justice reactions would
result in greater frequency of destructive deviance. Results revealed a significant effect
for both distributive and interpersonal justice on destructive deviance. As expected,
lower interpersonal justice reactions resulted in greater destructive deviance, yet higher
distributive justice reactions also resulted in greater destructive deviance (Table 10).
Therefore Hypothesis 6 was only partially supported.

Table I 0

Summary of Regression Analysis for Justice Variables Predicting Destructive
Deviafl'e (N=225)

 

mm B SE3 0 R2
Procedural -O.23 0.37 -0.05

Distributive 0.66 0.26 0. l 9* *

Interpersonal -0.95 0.36 -0.1 7* 0.04*

 

Note. "‘ p < .05; ** p < .01.

Hypothesis 7. Hypothesis 7 suggested that more positive trust reactions would
result in greater frequency of constructive deviance. Results revealed no significant
effect for trust on constructive deviance (B = -.01, p = ns.) and therefore Hypothesis 7
was not supported.

Hypothesis 8. Hypothesis 8 suggested that more positive trust reactions would
result in lower frequency of destructive deviance. Results revealed a significant effect for
trust on destructive deviance, B = -.13, p < .05, R2 = 0.02, such that higher trust reactions
resulted in lower destructive deviance. Therefore Hypothesis 8 was supported.

Summary. Consistent with expectations, more positive interpersonal justice and

trust reactions lead to lower destructive deviance, yet more positive distributive justice

68

reactions lead to increased destructive deviance. Post-hoe analysis was conducted to
examine if any one item from the distributive justice scale was driving results in this
unexpected direction. Table 11 displays the result of this analysis.

Table I I

Post-Hoe Regression Analysis for Justice Variables Predicting Destructive
Deviance (N =225 )

 

mble B SE8 0 R2
Distributive_l 0. 10 0.27 0.04

Distributive_2 0.54 0.35 0. 16+

Distributive_3 -0. 15 0.32 -0.05

Distributive_4 0.20 0.34 0.06

Procedural —0.22 0.38 -0.05

Interpersonal -0.88 0.37 -0.18* 0.05+

 

Note. + p < .10; * p < .05. Distributive_l=Item l of distributive justice scale;
Distributive_2=1tem 2; Distributive_3=1tem 3; Distributive_4=ltem 4.

Results of this post-hoe analysis showed that Item 2, “My pay is appropriate for
the work I have completed,” was marginally significant. In the context of destructive
behavior, individuals who thought that their pay was appropriate were more likely to
engage in destructive behavior and slack off their responsibilities. Because pay was
distributed equally and not based on equity, individuals who worked less were not
punished with lesser pay. Conversely, those individuals who were working hard (not
engaging in destructive behavior) did not think their pay was fair. Because the pay
structure rewarded everyone regardless of effort, this group of individuals was likely
under-rewarded. Overall, paying everyone $30 for 3 hours work without consideration of

performance seemed to have inadvertently driven negative behavior.

69

Another explanation for these findings stems from the equality rule of distributive
justice. Pay in this environment was allocated based on the equality rule (versus the
equity rule) where everyone was paid the same amount for their time and effort. For the
type of work being done, employees may have seen the pay as fair but they may have
also taken advantage of the situation and engaged in greater destructive deviance because
they knew if they were deviant they would not encounter any negative consequences such
as receiving less pay. Had pay been distributed based on an equity rule where greater (or
fewer) contributions led to greater (fewer) pay, the positive relationship between
distributive justice and destructive deviance would not likely exist.

It is also worth noting that the high mean levels of justice reactions and the
pattern of correlations between the distributive justice, destructive deviance, and
withholding effort suggest that the relationship between distributive justice and
destructive deviance may be a spurious effect. For example, distributive justice and
withholding effort correlated r = -.32 (p = .01) and withholding effort and destructive
deviance correlated r = .20 (p = .01), yet distributive justice and destructive deviance
correlated r = .11 (p = .11). This pattern suggested that greater distributive justice was
negatively related to withholding effort, which was positively related to destructive
deviance, yet the regression results indicated a positive relationship between distributive
justice and destructive deviance, which would not be supported by the correlations alone.
Because of this pattern and the marginal correlation between distributive justice and
destructive deviance, there is also the possibility that sampling variation and

measurement error may explain these unexpected findings.

70

Hypotheses 9-1 0

This set of hypotheses predicted that two individual difference variables,
conscientiousness and proactive personality, would affect the relationships proposed in
Hypotheses 5-8. Hypotheses 9-10 were tested using moderated hierarchical regression.
The appropriate process reaction and individual difference were entered in Step 1
followed by the interaction between process and individual difference in Step 2.

Hypothesis 9a. This hypothesis predicted that conscientiousness would moderate
the relationship between trust and destructive deviance. Following on the significant
relationship between trust and destructive deviance supported in Hypothesis 8, the
interaction between trust and conscientiousness was also significant ([3 = 1.33, p < .01,
AR2 = 0.03) (Table 12). A graph of this interaction reveals when conscientiousness was
low, high trust depressed destructive behavior (Figure 4). Therefore, Hypothesis 9a was
supported.
Table 12

Summary of Hierarchical Regression Analysis for Trust x C onscientiousness
Predicting Destructive Deviance (N =225)

 

Mk B SE B 13 AR2
Step 1

Trust (T) -0.54 0.32 -0. 12

Conscien (C) -0.16 0.27 -0.04 0.02
Step 2

T x C 1.33 0.49 2.06" 0.03"
ml R’- 0.05

 

Note. ** p < .01. Conscien=Conscientiousness.

71

Figure 4

Interaction Between C onscientiousness and Trust on Destructive Deviance

 

 

 

 

 

Conscientiousness
5.00—
— Low
---Hmh

4.00-
o
0
l:
.2
g 3.00—
o
>
‘3 2.00—
E
o
o 1.00—

0.00—

I 1
Low High
Trust

Hypothesis 9b. This hypothesis predicted that conscientiousness would moderate
the relationship between justice reactions and destructive deviance. While some
relationships between justice and destructive deviance were supported in Hypothesis 6,
the interaction between conscientiousness and justice was not significant (Table 13).
Therefore, Hypothesis 9b was not supported.

Table 13

Summary of Hierarchical Regression Analysis for Justice x C onscientiousness
Predicting Destructive Deviance (N=225)

 

Variable 8 SE B 0 AR2
Step 1
Distributive (DJ) 0.68 0.26 0.10"

72

Table 13 (cont ’d.)

 

Mble B SE B B AR2

Procedural (PJ) -0.17 0.37 -0.03

Interpersonal (II) -0.81 0.34 -0.l7*

Conscien (C) -0.29 0.26 -0.08 0.05*
Step 2

DJ x C -0.02 0.35 -0.04

PJ x C 0.82 0.62 1.15

U x C -0.03 0.60 -0.05 0.01
mat R2 0.06

 

Note. * p < .05; ** p < .01. Conscien=Conscientiousness.

Hypothesis 9c. This hypothesis predicted that conscientiousness would moderate
the relationship between trust and constructive deviance. Following from the lack of a
relationship between trust and constructive deviance as hypothesized in Hypothesis 7, the
interaction between trust and conscientiousness was also not significant (Table 14).
Hypothesis 9c was not supported.
Table I 4

Summary of Hierarchical Regression Analysis for Trust x Conscientiousness
Predicting Constructive Deviance (N=225)

 

ya_riable B SE B [3 AR2
Step 1

Trust (T) 0.02 0.26 0.01

Conscien (C) -0.1 1 0.22 -004 0.00

73

Table 1 4 (cont ’d)

 

mac B SE B B AR2
Step 2

T x C 0.09 0.41 0.17 0.00
11.411 R2 0.01

 

Note. Conscien=Conscientiousness.

Hypothesis 9d. This hypothesis predicted that conscientiousness would moderate
the relationship between justice and constructive deviance such that when
conscientiousness was low, the relationship would be stronger. While there were no
main effects for justice on constructive deviance, the interaction between procedural
justice and conscientiousness on constructive deviance was significant ([3 = 1.98, p < .05,
AR2 = 0.03) (Table 15). A graph of this interaction revealed when procedural justice was
low, individuals high in conscientiousness were most likely to engage in constructive
behavior (Figure 5). Because the relationship was in the opposite direction than
hypothesized, Hypothesis 9d was not supported.

Table 15

Summary of Hierarchical Regression Analysis for Justice x C onscientiousness
Predicting Constructive Deviance flV=225)

 

\_Igi_able B SE8 13 AR2
Step 1

Distributive (DJ) -0.19 0.22 -0.07

Procedural (PJ) 0.06 0.31 0.02

Interpersonal (IJ) 0.22 0.30 0.06

Conscien (C) -0.12 0.22 -0.04 0.00

74

Table I5 (cont ’d. )

 

Mic B SE B 0 AR2
Step 2

DJ x C -043 0.29 -0.87

PJxC 1.14 0.50 198*

U x C -0.38 0.49 -072 0.03
Mal R2 0.03

 

Note. " p < .05. Conscien=Conscientiousness.

Figure 5

Interaction Between C onscientiousness and Procedural Justice on Constructive Deviance

Constructive Deviance

 

4.50-

4.00-1

3.504

3.00d

 

 

Conscientiousness
— Low
- I - High

 

 

l 1
Low High

Procedural Justice

75

Hypothesis I 0a. This hypothesis predicted that proactive personality would
moderate the relationship between trust and constructive deviance. Just as the
relationship between trust and constructive deviance was not supported in Hypothesis 7,
the interaction between proactive personality and trust was also not significant (Table
16). Therefore, Hypothesis 10a was not supported.

Table I 6

Summary of Hierarchical Regression Analysis for Trust x Proactive Personality
Predicting Constructive Deviance (N=225)

 

Lariable B SE B 13 AR2
Step 1

Trust (T) -0.01 0.25 0.00

Proactive (P) -0. l 6 0.26 -004 0.00
Step 2

T x P -0.66 0.49 -1.04 0.01
m R2 0.01

 

Hypothesis 10b. This hypothesis predicted that proactive personality would
moderate the relationship between justice reactions and constructive deviance. Following
on the null results from Hypothesis 5 linking justice constructive deviance, there was also
no significant interaction between constructive deviance and proactive personality (Table

17). Therefore, Hypothesis 10b was not supported.

76

Table 17

Summary of Hierarchical Regression Analysis for Justice x Proactive Personality
Predicting Constructive Deviance (N=225)

 

Mble B SE B 0 AR2
Step 1

Distributive (DJ) -0.18 0.22 -0.07

Procedural (PJ) 0.02 0.30 0.01

Interpersonal (II) 0.23 0.30 0.06

Proactive (P) -0.16 0.26 -0.04 0.00
Step 2

DJ x P -0.52 0.40 -0.92

P] x P 0.59 0.63 0.82

U x P -0.41 0.57 -0.31 0.02
Mal R2 0.02

 

Summary. Despite a lack of main effect relationships between justice and trust
and constructive and destructive deviance, the interactions between conscientiousness
and trust on destructive deviance and between procedural justice and conscientiousness
on constructive deviance were significant. While individuals low in conscientiousness
may otherwise be inclined to engage in destructive deviance, feelings of high trust
diffused this inclination. A sense of procedural fairness, however, did not enhance the
likelihood of constructive behavior for individuals low in conscientiousness. In this
situation, despite feeling procedural injustice, highly conscientious individuals still
engaged in constructive deviance. This finding is consistent with LePine et al.’s (2002)

results linking high conscientiousness to constructive behavior, yet the under the

77

influence of procedural justice, the relationship is not as clear. Because of the high mean
on procedural justice (M = 3.88, SD = 0.49) and conscientiousness (M = 3.81, SD = 0.63),
one standard deviation below the mean is still above the midpoint on either 5-point scale,
suggesting that “low” procedural justice or conscientiousness is still relatively high. In
fact, a post-hoe planned comparison did not reveal any significant differences in means
between groups, mainly because when creating groups containing individuals scoring
lower (higher) than 1 stande deviation below (above) the mean for both procedural
justice and conscientiousness, the sample size was severely restricted to 19 individuals.
Therefore, the apparent negative impact of high procedural justice on highly
conscientious individuals warrants carefill interpretation as the differences between
individuals on these variables is not very large and may be due to a measurement artifact
similar to a ceiling effect.
Hypotheses I I -I 4

Hypotheses 11-14 were designed to fully examine the mediating and moderating
effects operating in the experimental model. Because Hypotheses 1-4 were not
supported, the mediating relationships proposed in Hypotheses 11b, 12b, 13b, 13c, 14b,
and 14c will also not be supported and will not be tested. However, the direct
relationship between the control manipulations and the different forms of deviance was
still examined using HLM.

Hypothesis 1 Ia. Hypothesis 11a predicted that formal control systems were more
likely to limit destructive deviance than informal control systems. Despite significant
between-group variance at level 1, to = 1.21, df= 30, x2 = 83.56, p < .01, ICC = 0.21,

results revealed no significant effect for formal or informal control on destructive

78

deviance, Yon = -0.21, se = 0.48, t= -0.46, p = .65. Therefore Hypothesis 11a was not
supported.

Hypothesis 12a. Hypothesis 12a predicted that formal control systems were more
likely to limit constructive deviance than informal control systems. Significant between
group variance was observed at level 1, to = 0.68, df= 30, xz = 74.46, p < .01, ICC =
0.18. Further analysis, however, did not reveal a significant effect for formal/informal
control on constructive deviance, 701 = 0.3564, se = 0.34, t = 1.04, p = .32. Therefore
Hypothesis 12a was not supported.

Hypothesis 13a. Hypothesis 13a predicted that external control systems were
more likely to limit destructive deviance than internal control systems. Despite
significant between-group variance at level 1, to = 1.21, df= 30, x2 = 83.56, p < .01, ICC
= 0.21, results revealed no significant effect for external or internal control on destructive
deviance, 702 = 0.25, se = 0.48, t = 0.52, p = .61. Therefore Hypothesis 13a was not
supported.

Hypothesis 14a. Hypothesis 14a predicted that external control systems were
more likely to limit constructive deviance than internal control systems. Significant
between group variance was observed at level 1, to = 0.68, df= 30, x2 = 74.46, p < .01,
ICC = 0.18. Further analysis revealed a marginally significant effect for extemal/intemal
control on constructive deviance, 702 = 0.64, se = 0.33, t = 1.92, p = .06, such that the
frequency of constructive deviance was lower under external control than internal control
(Ms = 3.45 and 3.92, respectively). Therefore Hypothesis 14a was supported.

Post hoc analysis. While the overall interaction between formal/informal control

and extemal/intemal control was not significant for constructive deviance, 703 = 0.15, se

79

= 0.68, t = 0.23, p = .82, the means between conditions suggested that post hoc analysis
was in order to examine the observed difference between the formal-external condition
and the informal-intemal condition. Table 18 displays the results of Tukey’s HSD post
hoc analysis, which revealed a significant difference in the expected direction.
Individualsin the formal-external condition engaged in significantly less constructive
deviance (M = 3.25) than did individuals in the informal-intemal condition (M = 4.10).
Table 18

Post-Hoc Tykey HSD Test for Interaction Term (N=225)

Mean
Condition (A) Condition (B) Difference (A-B) SE

Formal/Extemal Formal/Intemal -0.49 0.27
(#3 .25) (M=3.74)
Informal/Extemal -0.40 0.28
(M=3.65)
Informal/Intemal -0.85** 0.27
(M=4. 10)

 

Note. " p < .01.

While results for the overall destructive deviance construct did not support the
hypotheses, filrther exploration of destructive deviance suggests that the negative
behavior may be further split into two categories, that being deviance that is subtle versus
obvious. Subtle forms of deviance included the number of times a participant was off
task (daydreaming, playing computer games) and the number of non-work related
websites visited. More obvious forms of deviance included the number of minutes when
breaks lasted longer than 5 minutes, the number of minutes a participant arrived late, and

the number of times a participant chatted with another participant. Individuals in the

80

external control condition may have felt that while they knew the organization was
monitoring their behavior, deviance of a subtle nature would be less noticeable than
would overtly long breaks, chatting with another participant, or arriving late. Results of
HLM analysis revealed a marginally significant effect for extemal/intemal control on
subtle destructive deviance, 702 = -0.3 8, se = 0.20, t = -1.92, p = .06, such that individuals
in the external condition engaged in more subtle destructive deviance (M = 1.07) than did
individuals in the internal condition (M = 0.68). There was no effect for formal/informal
control on subtle deviance, but it appears that the presence of an external monitoring
agent was not enough to suppress this subtle form of negative behavior. When
examining between group differences on obvious forms of destructive deviance, there
were no effects for any type of control on this form of destructive deviance.

Another set of post hoc analyses was conducted to explain the role of individual
differences in moderating the relationship between control and deviance. Because there
was no support for the role of process reactions mediating the control-deviance pathway,
it was possible that conscientiousness and proactive personality moderated the control-
deviance relationship supported in Hypothesis 14a such that when conscientiousness and
proactive personality were low, the relationship would be stronger. HLM analyses were
conducted to examine these relationships but no support was found. Interestingly, results
revealed that conscientiousness played a marginal moderating role in the relationship
between extemal/intemal control and destructive deviance, 710 = -l .33, se = 0.80, t = -
1.66, p = .10. A graph of the interaction is displayed in Figure 6. Results show that the
limitations of low conscientiousness are enhanced in internal control environments,

perhaps where a lack of careful attention on the part of individual employees leads to a

81

greater frequency of destructive behavior. In external environments where the presence

of an outside monitoring agent is made salient, the likelihood of destructive behavior was

lessened.
Figure 6

Interaction Between External/Internal Control and C onscientiousness on Destructive
Deviance

 

6.0
Conscientiousness
— Low

5-0 — — High

:A
o

Destructive Deviance
N 0)
'o o

.s
O

 

 

 

.9
0

External lntemal
Type of Control

Summary. While it was not possible to show support for the mediation
hypotheses, support was found for the relationship between internal control and
constructive deviance. There was also support for the post hoc comparison between the
two dimensions of control on constructive deviance such that individuals in the informal-
intemal control condition were more likely to engage in constructive deviance than
individuals in the formal-external condition. Although the hypotheses linking control to
destructive deviance were not supported, the results suggest that there are forms of

control which do enhance constructive behavior. Further post hoc analyses also revealed

82

subtle forms of destructive deviance are more commonly observed in participants in the
external control condition. The benefits of external control were also observed for low
conscientiousness individuals such that destructive deviance was lessened under external
control than under internal control.
DISCUSSION

As counterproductive behavior continues to be an issue for many organizations, a
normative approach to understanding deviance was used to explore the nature of both
positive and negative behavior as well as the interplay between constructive and
destructive deviance. The creative use of field experiment methodology in the current
study allowed for an interesting look at behavioral criteria of deviance, as opposed to
typical self-report methods used in deviance research. As such, this research was
conducted with careful consideration of the cause, effect, mediating processes, and
moderating influences that ultimately affect deviance behavior.
Deviance Behavior

A key contribution of this research comes from the field experiment methodology
used to gather behavioral observations of deviance behavior. Only a small handful of
field experiments conducted in an organizational context are published each year, perhaps
because of the many challenges they present. Field experiments also present many
opportunities associated with the advantages of laboratory experiments, including greater
control over the situation to better isolate the causal relationships between variables. The
“field” in field experiment also helps maintain external validity and generalizability.

The measure of deviance was created through several steps, all of which were

designed to legitimate and validate the individual behaviors combined into the deviance

83

composites. Intercorrelations between behaviors served as one judgment point while the
expectations set forth to establish norms in the session served as another. To combine
individual behaviors, each behavior was standardized and summed into its respective
deviance composite (destructive or constructive). Employees’ self-reported withholding
effort served as another tool to validate the interconnectedness between destructive,
constructive, and effort behaviors. With few exceptions, most research on deviance has
used self-report, retrospective accounts of counterproductive behavior. The use of
observational data takes a step away from the biases inherent in self-report questionnaires
of behavior and moves forward in capturing deviance in context.
Control-Process Reactions

Previous research on control and justice reactions has found that employees tend
to respond more positively when they are allowed to give their input and feel that the
control will not impede their ability to be successful in the organization (Alge, 2001;
Douthitt & Aiello, 2001). Furthermore, when control was more structured and unbiased,
employees have responded with a greater sense of justice as well (N iehoff & Moorman,
1993). Control systems also communicate a sense of trust (or distrust) to employees as
employees read-in to the control and gain a sense of the organization’s motives and
openness (Gabarro & Athos, 1976). To explore three dimensions of justice (distributive,
procedural, and interpersonal), a series of hypotheses was offered that predicted that
control systems with an internal locus of responsibility would result in more positive
procedural and interpersonal justice reactions as well as more positive trust reactions.
Formal control systems were expected to result in more positive distributive justice

reactions.

84

The results did not support any of the hypotheses linking types of control to either
justice or trust reactions. This was possible for several reasons. Alge’s (2001) research
on reactions to monitoring found that concerns about invasion of privacy was a key driver
of procedural justice reactions. In this context, a single exposure to more obtrusive forms
of control may not have stirred a strong sense that there was an invasion of privacy,
especially given that the task being performed was a computer-based task and monitoring
of computer activity was not beyond the boundaries of what would be reasonable in this
environment.

Research by Tabak and Smith (2005) suggests that an employee’s disposition to
trust, in combination with their experience at the current organization and prior
experience at other organizations, will affect an employees’ decision about managerial
trustworthiness and subsequent trust formation. With only a single experience with
Dynamic Research Group, the foundation for trust formation was in its formative stages
and could also have contributed to a lack of variability on trust reactions. In ongoing
employment experiences, the link between control and trust (or justice) is an iterative
cycle such that employee experience informs how they evaluate their organization and at
the same time supervisor or organization experience informs how employees are
evaluated and further trusted or distrusted. Therefore, while a lack of support for the
hypotheses signaled some doubt that control was appropriately linked to justice and trust
reactions, a more comprehensive look at the temporal dynamics associated with the
formation of trust and justice would provide a better setting to understand these

relationships.

85

Process Reactions-Deviance

Research linking justice and trust reactions to both constructive and destructive
deviance has typically found that more positive trust justice reactions lead to increased
constructive behavior (de Cremer et al., 2001; Little et al., 2002) and more negative
justice and trust reactions lead to increased destructive behavior (Fox et al., 2001;
Hultrnan, 2004). The second set of hypotheses reflected the likelihood of these
relationships in the current setting but there was limited support linking process reactions
to deviance behavior.

Results showed that greater interpersonal justice and trust led to lessened
destructive deviance, which is a hopeful message for employers looking for ways to
dampen destructive behavior. To take advantage of this relationship in an employment
setting, organizations may wish to implement strategies to enhance interpersonal justice
and/or trust. For example, for organizations wishing to enhance trust, research has shown
that trust can be further developed when there is equal power between parties in an
employment relationship and where employees feel as though they have full knowledge
of their supervisor’s intentions in an evaluation setting (Enzle & Anderson, 1993).

Contrary to expectations, no effects were observed for any of the justice or trust
reactions on constructive deviance. One likely explanation stems from the limited
variability of these reactions, which decreases the likelihood that reactions would further
predict differences in behavior. An examination of the mean levels of distributive,
procedural, and interpersonal justice and trust demonstrated that on a five point scale, all
means are above the mid-point of the scale and in all cases, a large percentage of

individuals scored at 4.0 (agree) or higher [procedural justice > 4.0 = 25%, distributive

86

justice > 4.0 = 40%, interpersonal justice > 4.0 = 75%, trust > 4.0 = 50%], which further
limited the variation in measures. The demand characteristics in the current study were
unlike characteristics in other psychological studies of deviance (e.g., J. Greenberg, 2002)
where individuals are provoked to react negatively, which may further explain the lack of
variance and effects for process reactions on deviance behavior observed in the current
research.

Moderating Role of Conscientiousness and Proactive Personality

The research linking conscientiousness and proactive personality to deviance has
found support for the relationship between high conscientiousness, decreased destructive
deviance (Ones & Viswesvaran, 2001), and organizational citizenship (Van Dyne et al.,
1994) and between high proactive personality and increased constructive behavior (Bell
& Staw, 1989). Therefore, to understand the effect of conscientiousness and proactive
personality on deviance behavior, several hypotheses were generated to suggest that
conscientiousness would moderate the relationship between process reactions and
destructive deviance and constructive deviance and that proactive personality would
moderate the relationship between process reactions and constructive deviance.

Support was found for Hypothesis 9a; conscientiousness did moderate the
relationship between trust and destructive deviance such that when conscientiousness was
low, a trusting environment helped to lessen destructive behavior. This is particularly
interesting for managers who are faced with pressure to do more with less and perhaps do
not have many resources or time to devote to constant monitoring of employees. If

establishing a trust-based relationship with employees (who might otherwise be inclined

87

to engage in off-task behavior) helps to decrease destructive deviance, there are many
low-cost, practical avenues for the manager to pursue in developing that trust.

While support was not found for Hypothesis 9d, procedural justice did interact
with conscientiousness in an unexpected direction. Procedural justice is defined by
actions that are consistent and ethical (Thibaut & Walker, 1975). When procedural
justice was lower, highly conscientious individuals may have still felt that the situation
was flexible and would allow them to go beyond the rules and expectations of the work
and engage in constructive behavior. Highly conscientious individuals often have a
internal locus of control (Judge, Erez, Bono, & Thoresen, 2003), meaning they believe
that rewards and outcomes are driven by personal ability and effort (O'Brien, 1984).
Because the benefits of helping or going above and beyond are often rewarding in itself,
the extra effort to go above and beyond (even in situations where the procedures were
seen as unfair) is not uncharacteristic of highly conscientious individuals.

For highly conscientious individuals who perceived high procedural justice, the
results are less intuitive but are consistent with some research on consistency, procedural
justice, and self-esteem. For example, Gilliland (1994) found that self-esteem was
lowered for individuals who were treated consistently in a selection system but were
rejected. Similarly, Schroth and Pradhan Shah (2000) suggest that knowing procedures
were fair does not balance out the disappointment or negative affect associated with
receiving an undesirable bad outcome. These two studies shed some light on the current
findings that procedural fairness essentially had a negative effect on constructive
deviance for highly conscientious individuals. For highly conscientious individuals who

operate largely according to the rules, knowing that the procedures in the work

88

environment were fair, applied consistently, and were free of bias may have meant that
they were less apt to engage in constructive deviance because they knew that the set
expectations were already fair — why go above and beyond in an environment that already
seemed reasonable? To build on this preliminary explanation for why procedural justice
had a negative effect on constructive behavior, future research should explore this
relationship more carefully, especially in a sample with greater variability on
conscientiousness.
Control-Deviance

The final set of hypotheses suggested that the different forms of control would
affect employee deviance and that this relationship would be mediated by process
reactions. Because the link between control and process reactions was not supported, the
mediating pathways could not be tested, but some very interesting results from the
control - deviance pathway were still explored. On the constructive deviance side, both
informal and internal control systems were expected to promote constructive behavior as
employees are not limited either by formal responsibilities (Roth et al., 1994) but they are
given more latitude regarding their behavior (Crossen, 1993). For similar reasons,
control of a formal and external nature was expected to have effects on destructive
deviance as employees were more aware of precise expectations and would have an
external agent monitoring their behavior.

While results did not support the relationship between control and destructive
deviance, support was found for the hypothesized relationship between internal control
and constructive deviance. This suggests that while control in this environment may have

been unable to suppress destructive behavior, there were forms of control which

89

simultaneously affected constructive behavior. Individuals in the external condition were
less likely to engage in constructive deviance than individuals in the internal control
condition. (This same pattern of means was observed in the formal and informal
conditions, Ms = 3.03 and 3.34, respectively.) Post hoc analysis of the interaction
between control dimensions revealed that individuals in the formal-external condition
were significantly less likely to engage in constructive behavior than individuals in the
informal-internal condition.

A premise of this research was that prior work in this area has failed to consider
the simultaneous occurrence of destructive and constructive behavior and the results of
this research support that there are some forms of control which are detrimental towards
positive behavior. Research has supported the many benefits of counter normative,
constructive behavior including a desire to reciprocate help for help (El-Alayli & Messe,
2004) and the development of a “good neighbor” mentality (Heckert & Heckert, 2004).
Depending on the perspective of who is evaluating the benefits of constructive behavior,
there are some examples of counter normative behavior such as rate busting which have
not been viewed as favorably, especially by labor unions and employees who are satisfied
with maintaining the status quo (Heckert & Heckert, 2004). In any organization wishing
to enhance constructive deviance, it remains important for employees to feel as though
they can engage in this behavior without fear of punishment for not following the rules or
ostracism fiom their work group for bucking the trend (e.g., Barker, 1993).

Additional analysis of destructive deviance showed that individuals in the external
control condition were somewhat more likely to engage in subtle destructive deviance

than were individuals in the internal control condition, despite the presence of an external

90

monitoring agent. Despite no overall findings suggesting that there are control systems
which are more likely to limit destructive deviance, a closer look at the destructive
deviance construct revealed that there are several ways employees can be deviant, not all
of which may be as salient or as destructive as the other. Further follow up analysis
revealed that while the mediating path between control and deviance could not be tested,
there were moderating effects for the individual difference of conscientiousness on the
relationship between extemal/intemal control and destructive deviance. In an internal
environment, low conscientiousness individuals were more likely to engage in negative
behavior. This detrimental outcome does not bode well when employees who are low in
conscientiousness and not careful to begin with are placed in an environment where they
are responsible to self-monitor, as more negative behavior is likely to slip through the
cracks.

The process of studying both destructive and constructive deviance comes at a
time when both entitlement perceptions and the push for innovation are seemingly at
odds with one another. When performance expectations become more ambiguous and
employees are encouraged to do less with more, the boundaries between ethics,
performance, and deviance are inevitably blurred (Litzky, Eddieston, & Kidder, 2006).
As such, employees may be more willing to go above and beyond, but at negative long
term costs to the organization and its customers. The one time nature of the work
experience studied here did not allow for a longitudinal look at deviance, but as
performance expectations and the psychological contract are re-negotiated and re-
interpreted overtime, the dilemma between constructive and destructive deviance is

bound to grow. Here, the first glance at the overall relationship linking control to

91

deviance suggests that control has the potential to impact both positive and negative
behavior and that further examination of this relationship is warranted in future research
studies.

Limitations and Future Directions

Despite the noteworthy findings linking control to constructive deviance and the
careful consideration given to the theory and methods used to examine deviance, a
majority of this study’s hypotheses were not supported. I took away from this experience
an appreciation for how the context truly matters in understanding deviance (and related
processes). A post hoc review of the context reveals several key factors that will help
further explain the results of the study and offer directions for future research, namely:
the ability to effectively implement control, defining and assessing norms of deviance,
the effect of time, and the impact of rewards and other organizational factors that
influence employee behavior.

Implementing control. In designing the framework of control, there were
challenges associated with a creating a clear and distinct boundary between cells that also
allowed for a realistic manipulation of the control system. For example, when
distinguishing formal from informal control, simply relying on one type to be defined by
written instructions and the other to be defined by unwritten instructions was limited in
scope. Given that all participants were provided with written instructions to outline the
job task, the strength of the formal/informal control manipulation was inadvertently
weakened. Manipulating the other dimension of control was more clear cut in that the
boundaries separating the organization (external monitor) from oneself (internal monitor)

are distinct. Rather than treating the cells as distinct from one another, one possible

92

direction is to treat these control systems on a continuum where control varies in its
strength on these key dimensions. Another avenue for future research would also not
abandon these dimensions but but rather work to better elucidate the distinctions between
formal and informal control and take advantage of the role of the administrator/supervisor
in implementing the control. For example, in the current research the administrator was
dressed professionally at all times to give credence to the work experience. In the
informal control condition, if the control is truly more casual in nature, the administrator
could more thoroughly “act the part” and dress down or use less formal speech when
delivering task instructions.

Defining and assessing norms. Gaining a more thorough understanding of the
normative nature of deviance was one of the main drivers of this research effort. In a
nutshell, destructive deviance was conceptualized as behavior which violated norms and
did harm while constructive deviance was conceptualized as behavior which violated
norms but was beneficial. In real organizations, history and context also provide insight
into normative behavior. Without any history or context, the norms operating in this
environment were established from the expectations laid out for employees at the
beginning of each work session. These norms were essentially established at the
organizational level via the control system rather than at the group level, which for
defining deviance, would be the appropriate level to consider. However, because norms
can operate at many levels, it was a choice to consider norms at the organizational level
as the benchmark for judging deviance given the current context paired with the lack of

opportunity for norms to truly develop within each individual work session.

93

With these expectations came norms for behavior, but anecdotal observations
suggest that there may have been slight group effects which further affected the norms for
that session. That is why it was important to properly account for group membership in
all data analysis by using hierarchical linear modeling. For example, if no one got up
fi'om their work station within a minute of the break being announced, it seemed to be
much less likely that the entire group would take a break than if several people got up,
made a little noise, and took a few minutes away from work. This has implications in
terms of understanding the stability and content of norms. The point of view of the
source also impacts norms. Norms in this environment were assessed from the
perspective of the organization, rather than from the employees. Noise in the channel
between the delivery of expectations, establishing a norm, and execution of behavior
definitely creates a more complicated picture.

Further complication arises from understanding whether “countemorrnative”
behavior is truly deserving of the destructive or constructive label or if the labels and
definitions of norm-based deviance are too strict. This research raises the question of
whether violating a norm to do extra work is actually constructive deviance or is it more
similar to OCB? Is the behavior so unexpected and beyond extra role expectations that it
fits the definition of deviant? Upon observing the behavior of these one-time employees,
the waters have become murky in distinguishing constructive deviance from
organizational citizenship. The helpful, beneficial behaviors observed of employees may
be more a reflection of natural behavior you would expect out of highly conscientious,

proactive individuals (which is characteristic in this sample). Future research should

94

work to more clearly distinguish citizenship behavior from constructive deviance while
taking into consideration the content and stability of norms operating in the environment.

Eflect of time. As evidenced by a lack of support for Hypotheses 1-4, the impact
of control on justice and trust reactions was lacking. Previous research on justice has
shown that the effects of company procedures on justice reactions tend to unfold over
time (Mesch & Dalton, 1992). In the current context, a one-time exposure to a particular
method of control did not result in immediate reactions of unfairness or distrust. In the
short term, more explicit forms of control might even be expected as new employees
learn the job and require more guidance. As the control seems to be more intrusive and
less desirable over time, it can serve as a stressor which may result not only in reactions
of injustice but also other psychological symptoms leading to greater problems in the
organization (Snow, Swan, Raghavan, Connell, & Klein, 2003). Therefore, future
research in this area should understand how the repeated exposure to control as part of an
employee’s daily routine may more slowly provoke frustration and erode feelings of trust
and justice, rather than attempting to capture injustice in a single snapshot of the work
experience.

Rewards and organizational factors. In real organizations, employees have a
past, present, and future, all of which contribute to their behavior. An employee’s history
in the organization is likely to have a large influence on the likelihood that they will or
will not engage in deviance. For example, Greenberg and Barling (1999) found that an
employee’s past history of aggression in the organization served as a strong predictor of
future aggression against other employees. An employee who is trying to impression

manage and who has recently been praised for going above and beyond is probably more

95

likely to do so again in the future (Bolino, 1999). Similarly, an employee who gets away
with coming to work late is likely to continue that behavior as well. While the current
study has made inroads into understanding the relationship between control and deviance
(with interesting results), future avenues in this area should consider the impact of an
employee’s history and tenure, along with his or her relationships with peers and
supervisors, all of which will influence the likelihood and frequency of deviance
behavior.

Rewards in the workplace also drive employee behavior. A reward in the form of
a paycheck can enhance employee performance when pay is tied to performance
outcomes. Pay which is not tied to performance not only has the potential to depress any
constructive behavior but it may also encourage, to a slight degree, destructive behavior
(Litzky et al., 2006). Because there were no between-group differences for control on
destructive deviance, one explanation is that knowing the 830 would be rewarded at the
end of the session, regardless of on-task or off-task performance, could have promoted
negative behavior in the form of general laziness or off task behavior. The type and
fi‘equency of organizational rewards, whether pay or praise, can serve as a subtle form of
control in and of itself. As a way to manage deviance, future research should investigate
these factors which may inadvertently encourage (or discourage) deviance behavior.
Conclusions

This field experiment exploring the relationship between organizational control
and organizationally directed deviance was a first glance into a behavior-based
examination of destructive and constructive behavior. With the advantages of field

experiment methodology, there were still many challenges in conceptualizing and

96

operationalizing destructive and constructive deviance as the nature of this type of
behavior depended on the strength of norms in the work environment as well as employee
reactions regarding fairness and trust. Building on the results of the current study,
organizations seeking to enhance constructive behavior may wish to experiment with
external and internal forms of control while future research in this area should seek to
examine control in real world settings and its short- and long-tenn effects on employee

behavior.

97

APPENDIX A

 

IIElI' WANTED!

Earn $30 for 3 hours work!

 

Dynamic Research Group

 

 

- We’re looking for any 18-or-older MSU student or community
member with some Internet and computer skills. All majors
are welcome!

- You’ll be doing some internet research,
writing a brief report about your findings,
and responding to some short surveys in your
3-hour session.

 

- Best of all, you’ll be paid $30 in cash at the end of your work
session!

- All work sessions will be held on the MSU Campus at various
times throughout the semester.

www.dynamicresearchgoup.com
Spaces are quickly filling up!

 

 

 

.Z‘fE fife be 31‘5 31‘s BfE Z‘fE
”oo "-00 *oo '00 "0° “0° "'°°
c r: r: l: c t: c
:32. :3; 33:.- 333 :32 :32. 33:.
Elf: :3: £2: E3: ‘33:: t3: g3:
0.32 3.38 r132 0.32 1:39 8.32 0.32
n.°en o.° n.° c1.°nn noun a.° a°en
0‘1: 0‘2" 0‘2” 0‘1: 0": 0‘5" 0":
xmu mu mu mu mu mu mu
I- x r. x r. x I. a: \— x r. x L
LB“ 55“ 5|.“ LL“ 1.5“ LI.“ hl-u
3°83: 0.23 9.99,» 023 0.23 0,93 0.29:
ct» 30o 3our 3cm 3our goal 3°11:
comb arml- crank omb on; can‘- on;
540E Em}! Ems! Ed»... 8‘”? 5‘03 Ew'fi
’3 a 55 '55 ES 2 a ‘55 '3 2
r:
o 3. 8 3. 5 3. 8 3. 8 9. 8 3 8 *3

\O
00

 

APPENDIX B

ERRQE GROUP

Employment Application

DYNAMIC

 

Please enter all of your responses directly on this electronic application. You will be asked to send this
application as an email attachment as part of your pre-employment test. After receipt of your application.
you will be contacted by Dynamic Research Group and notified of your status. Thank you for your interest

in this position!

Name

Street Address
State

Phone Number

Email

weekends

 

 

     

 

 

 

 

ucatron ~‘ 2:9 .......... .
__H_ig School Year Graduated: School:
College Major: Institution:
Year in School/Year of Graduation:

 

 

 

 

Dates of

Job Title

Street Address
State

Phone Number

Dates of

Job Title

Street Address
State

Phone Number

Dates of

Job Title

Street Address
State

Phone Number

 

99

Applicant Background Questionnaire and Pro-Employment Test
A. Employee Behavior
The following questions are a series of phrases describing people's behaviors. Use the 5 point rating scale
where:

         

i=Very ” ' "”"2='Sbm'ewhal "”’»'-’Neither ‘ "
f; Inaccurate inaccurate ' Accurate nor

 

 

 

 

 

., .._,lnaccurate, .

 

 

We would like for you to describe how accurately each statement describes you. Describe yourself as you
generally are now, not as you wish to be in the future. Describe yourself as you honestly see yourself, in
relationtootherpeople you knowofthesamesexasyou are, and roughlyyoursameage. Sothatyoucan
describe yourself in an honest manner. your responses will be kept in absolute confidence. Please read
each statement carefully. Enter your response to the left of each question.

 

100

 

B. Approach to Problem Solving

The following questions ask about your approach to problem solving. Your answers will be kept
confidential. Please answer each of the following items honestly using a 1-5 scale, where:

 

101

C. Knowledge of Investments

The following questions ask about your knowledge of investments and investment terminology. Pleas
answer each question without the assistance of the Internet or any other information source. Please enter
yourresponses inthecellstothe Ieftofeach question.

 

. : Question 1: To be a safe investor, where should you put your money?
‘ 1. A bank account and/or treasury bills
2. Conservative bonds
3. Bluechipstocksmdconservativemuhralfunds
4. Quality growth stocks
5. All of the above
‘ Question 2: Which of these investments has the lowest risk?
1. Money market funds
2. Quality growth stocks
3. High tech mutual funds
4. High yield bonds
. ,, 5. Balanced mutual funds
' ’ Question 3: Which of these investments will help your money grow with moderate risk?
', 1. Large-cap stocks
2. Blue chip stocks
3. Balanced mutual funds
4. Quality growth stocks
5. All of the above
' Question 4: When making investment decisions, the most important thing is:
1. Timing the market
2. Buying and selling regularly
3. Asset allocation
4. Sticking to one strategy
5. None of the above
Question 5: Securities laws help protect investors by requiring companies to:
1. Show profits before they can sell stocks
2. Give investors important information about the company's financial position
3. Pay dividends
4. R .. investors who have lost mone

 

' a

‘ Question 6: If you had extra money, what would be the best investment choice?
' 1. Invest in the stock market

2. Invest in mutual funds

3. Invest in bonds

,- 4. Place it in a savints account
Question 7: Stacey will retire in 40 years and wants to start saving for her retirement now.
Stacey has a number of investment choices. Which of the following would be a good choice
'. . for Stacey to meet her long-tenn goal?

' 1. Asavings accountatthebank

2. A Guaranteed Investment Certificate (GIG)

3. A mutual fund that invests in a range of stocks

4. Thestockofonecom n

 

 

 

102

 

. Question 8: Please indicate how strongly you agree or disagree with the folowing

. it statement:
‘1 'There are some things about investing that you just learn bytrial and enor.‘

1. Agree

2. Agree Slightly

3. No Opinion

4. Disagree Slightly
5. Dis_agree

~' Question 9: What is the calculation of the approximate sales volume mulled tojust cover

costs, below which produclion would be unprofitable and above which it would be profitable?
1. Break-even analysis
2. Stock actions
3. Debt service coverage
4. Market 80-1 ~'~

 

Question 10: Whatisthetaxrate paidonthelastdoliarofone's income?
1. Capped tax rate
2. Marginal tax rate
3. Investment tax rate
4. Wealth tax rate
- Question 11:Whatisthemleofthumbthat20%ofapopulafionean1380%ofltsirrcome
referred to as?
1. Domino’s Law
2. Law of Returns
3. Pareto's Law
, 4. Market Value
‘15 Question 12: What is the term for a financial obligation. debt. claim, or potential loss?
"' 1. Asset
2. Equity
3. Ratio
‘ 4. Lighifl
Question 13: Which of the following is not a commodity that can be traded?
‘ 1. Food products
2. Petroleum
3. Bonds
4. Intellectual . . . -

 

Question 14: What type of money manager are you?
, 1. Exceptional, a wiz
2. Better than most
3. Okay but not great
4. Not very good
5. I spend more and save less than I could
6. Neverseem to havee . -h cash

" . Question 15: Have you ever invested in any of the following?

 

1. Stocks

2. Bonds

3. Mutual Funds

4. Certificates of Deposits (CDs) or Guaranteed Investment Certificates (GICs)
5. None of the above

103

 

D. lnmmet Skills

Please mswer the follovn'ng questions which test your lntemal and computer skills. You should use the
Internet to help you answer the questions below. Please type your responses in the blank cells to the left of
fermions questions

 

'13 1. Search for the term sleep disorder using the Google search engine. List the
first 5 web pages that come up from your search, NOT including the
sponsored links.

 

2. How many miles are needed for a standard award roundtrip First Class
Ticket on Northwest Airlines traveling from New York City to San Francisco?

 

3. I want to order a pizza from Bocce Club Pizzeria near Buffalo. NY. What
are the addresses and phone numbers of their two Buffalo-area locations?

 

4. As construction on I-96 continues, travelers face many detours and lane

2 changes. Go to the Michigan Department of Transportation website. What
does MDOT plan to do for westbound traffic traveling on I-96 at the 36" Street
interchange in Kent County when construction begins there in 2006?

 

 

 
  

5. Save this application to your computer. Use your last name as the file name (e. 9., Smith. doc). Send
this at alication as an attachment to info ud amicresearcho-rouco

 

 

Submitting this application via email to Dynamic Research Group will serve as your electronic
signature. I attest that I am 18 years of age or older, all information included in this
application is accurate, and all responses are truthful as of the date of this application.

104

APPENDIX C

DYNAMIC Rita/slalohurfj GROUP
T arms of Employment and Employee Expectations

As an employee of Dynamic Research Group (DRG), this document explains the
terms of your employment and some of our organization’s expectations.

Terms of Employment

You will be hired as a short term employee of DRG as part of our Internet Consulting
Project. You will be an employee of DRG for one 3-hour time period in the coming
weeks. During this time, you will be asked to work independently on an intemet research
project, write a report about your findings, and answer several surveys about your
experiences. You will not be expected to do anything beyond what is outlined for you
here.

The company who has hired DRG for this project is interested in the quality of your work
product, your reactions to the work, and how you feel about your experience. Employees
are expected to arrive to their work sessions on time and remain at the session for the
entire 3-hour period. We are interested in your individual work projects so we
discourage any conversation between employees druing the session. At the conclusion of
your work session, you will be paid $30.

Employee Expectations

F ormal-external

To ensure that employees meet our work standards, we will be monitoring your
workplace behavior. Our code of conduct states that employees must be on task, work
efficiently, and conduct themselves in a professional manner at all times. We will be
monitoring your Internet activity, tracking the web pages you look at, and checking in
with you throughout your work session. You will be provided with these expectations
again in writing when you arrive at your work session.

F ormal-internal

To ensure that employees meet our work standards, you will be responsible for keeping a
work diary. This diary should be a list of all of your activities during your work session
today. Our code of conduct states that employees must be on task, work efficiently, and
conduct themselves in a professional manner at all times. You will be expected to
commit to making a list of the web pages you visit and of all work-related behaviors that
you do. You will be provided with these expectations again in writing when you arrive at
your work session.

105

Informal-external

To ensure that employees meet our work standards, the administrator will be walking
through the room to ensure that you are on task, working efficiently, and conducting
yourself in a professional manner at all times. Ethical conduct is a defining characteristic
of all DRG employees. Everyone is expected to conduct themselves appropriately and in
the past, our employees have had a strong desire to comply with these principles.

Informal-internal

To ensure that employees meet our work standards, we trust that you will commit to our
values of keeping on task, working efficiency, conducting yourself in a professional
manner at all times. Ethical conduct is a defining characteristic of all DRG employees.
Everyone is expected to conduct themselves appropriately and in the past, our employees
have had a strong desire to comply with these principles.

106

APPENDIX D

:e' :1.” ."7 W:
DYNAMIC RESEAlfiCgl GROUP
Internet Research Project
Dynamic Research Group’s client is a new organization whose founder is a graduate of
Michigan State University. This entrepreneur is interested in pursuing some new ideas
related to outsourcing. As you may know, companies have been increasingly focused on
cutting costs and identifying new ways of getting work done. This can include outsourcing.
Outsourcing involves contracting for cheaper labor than what it would cost the organization
while still trying to preserve the quality of work done. Our client is interested in outsourcing
a component of his work process — namely research -— to less costly, skilled individuals
instead of having full time employees on staff whose salary and benefits costs are much more
expensive. Our client would like to know if college students and community members are
capable of doing this work. Because the founder is a Michigan State graduate, we are
beginning this work at MSU but will eventually expand our efforts to other colleges and
communities.

Your Task

 

The websites listed above are the only websites that you need to visit today. Feel free to
make notes about your research as you are working on the paper provided. Please work
individually on your research and refrain from discussing your work with others in the room
today. If you do have any questions, please let the administrator know.

107

Outcomes

DRG’s client is interested in the nature and content of the reports that are generated from this
session. They will be reviewing the content and making an evaluation on whether or not
individuals from your demographic are able to create quality reports in a 3-hour period versus
having their own employees take company time to complete the same job. Therefore, we
expect that you will create the highest quality report possible in the time permitted.

Report Expectations

This report will have three sections: opening statement, research evaluation, and closing
statement. We would like you to begin your report by writing an opening statement or
paragraph that summarizes what you will be discussing in the text. In the next section of
your report, provide a short summary of the investment information that you gather from
each of the websites. Next, compare and contrast the investment information that you gather
from the websites listed above. Use this opportunity to critically evaluate the information. In
your final paragraph, use your best judgment from your own experience and the information
you have been evaluating to make a recommendation to the client as to which of the websites
listed (if any) you would suggest as quality resources. You should focus your work efforts
on the written portion of your report, but if you feel that a table or figure would add to the
quality of your final document, please insert these graphics at the end of your document.
This element is not required but could benefit our client.

The final element of your report will be to evaluate the prospects with investing in a
particular company, based on your research and what you have learned. You can keep your
analysis at a fairly high level and include your thoughts at the end of your written report.
This section does not need to exceed I page.

Report Format

Your report should be near 4 pages in length. When you are ready to begin writing your
report, create a Microsoft Word document. You should write this document as if it will be
included in a report to be sent to top management. Use proper grammar, spelling, and
language. Be sure to save this document as you are working. You have been assigned an
employee number, which is posted on top of your computer work station. Save this
document as YourEmployeeNumber.doc, so for example AlOO8.doc or C3389.doc on the
floppy disk provided. You must write your report using your own words. DO NOT cut and
paste information from the websites directly into your report. Because the expectation is that
you will synthesize the information and make your own critical evaluations, we expect that
you will not plagiarize any material. Please use Times New Roman lZ-point font, double
space, l-l.25” normal margins. Enter your employee number on the first line of your
document and then begin your report.

Additional Information

I You will be able to take two 5 minute breaks during the session today. The breaks will
be announced by the session administrator. When it is time for your break, we would like
for you to get up, stretch your legs, and leave the room.

I You will have until 15 minutes prior to the end of today’s session to finish your report.
The last 15 minutes of the session will be used to answer some surveys and distribute

payment.

108

APPENDIX E

Manipulation Check

Do the following statements correspond to the type of organizational system used in your
work session today? Respond using a 1-5 scale, with l=strongly disagree and 5=strongly

agree-

1. Dynamic Research Group monitored my workplace behavior, Internet activity, and
tracked the web pages I looked at.

2. At the start of today’s session, I was given a written, formal document explaining
how I was expected to keep on task, work efficiently, and conduct myself in a
professional manner at all times.

3. Dynamic Research Group told me that they trusted I would commit to their values
and comply with their principles.

4. I was asked to keep a work diary, including a list of the web pages I visited and all the
work-related behaviors I did.

5. The administrator at the front of the room walked through the room to ensure that I
was on task, working efficiently, and conducting myself in a professional manner at
all times.

6. At the start of today’s session, I was told in an unwritten, informal manner about the
ethical conduct expected from all Dynamic Research Group employees.

7. I knew that I, and not my company Dynamic Research Group, was charged with
monitoring my own behavior.

8. I knew that my company Dynamic Research Group was charged with monitoring my
behavior.

Justice

(Modified from Colquitt, J. A. (2001). On the dimensionality of organizational justice: A
construct validation of a measure. Journal of Applied Psychology, 86, 386-400.)

Please answer the following items using a 1-5 scale, with l=strongly disagree and
5=strongly agree.

Procedural Justice
The following items refer to the procedures used in today’s work session.

I have been able to express my views and feelings during today’s procedures.
I have had influence over the procedures used in today’s work session.
Today’s procedures have been applied consistently.

Today’s procedures have been free of bias.

Today’s procedures have been based on accurate information.

I have been able to appeal the use of the procedures.

Today’s procedures have upheld ethical and moral standards.

NP‘MPP’NF

109

Distributive Justice
The following items refer to your work behavior today.

1. My pay reflects the effort I have put into my work.

2. My pay is apprOpriate for the work I have completed.

3. My pay reflects what I have contributed to the organization.
4. My pay is justified given my performance.

Interpersonal Justice
The following items refer to your reactions towards Dynamic Research Group.

1. Dynamic Research Group has treated me in a polite manner.

2. Dynamic Research Group has treated me with dignity.

3. Dynamic Research Group has treated me with respect.

4. Dynamic Research Group has refrained from improper remarks or comments.

Trust

(Modified from Robinson, S. L., & Rousseau, D. M. (1994). Violating the psychological
contract: Not the exception but the norm. Journal of Organizational Behavior, 15, 245-
259.)

The following questions relate to how much you believe your employer trusts you.
Please answer using a 1-5 scale, with 1=strongly disagree and 5=strongly agree.

I am not sure Dynamic Research Group fully trusts me.

I am open and upfront with Dynamic Research Group.

Dynamic Research Group believes I have high integrity.

In general, Dynamic Research Group believes my motives and intentions are good.
I am not always honest and truthful.

I don’t think Dynamic Research Group treats me fairly.

I can expect Dynamic Research Group to treat me in a consistent and predictable
fashion.

.‘IQ‘P'PP’I‘N‘

Attitudes about the Internet

(Modified from Lavoie, J. A. A., & Pychyl, T. A. (2001). Cyberslacking and the
procrastination superhighway: A web-based survey of online procrastination, attitudes,
and emotion. Social Science Computer Review, 19, 431-444.)

Please answer the following questions about your attitudes towards the Internet. Use a 5
point rating scale from 1=strongly disagree to 5=strongly agree.

1. I feel the Internet makes me more productive.

2. I ofien use the Internet to relieve stress.
3. I think that the Internet is a useful tool for research.

110

4.
5.

People rely on the Internet for information more often than they should.
The Internet is a way for me to stay connected to the outside world.

Attitudes about Outsourcing

Please answer the following questions about your attitudes about outsourcing. Use a 5
point rating scale from 1=strongly disagree to 5=strongly agree.

1. Individuals who have outsourced jobs are not as competent as in-house employees.

2. Outsourcing takes away jobs from people who deserve them.

3. Outsourcing is an effective way for companies to save money and become more
efficient.

4. I think that companies should only consider the bottom line when deciding to
outsource jobs.

5. Only jobs which are not core to a company’s business should be outsourced.

Withholding Eflort

(Modified from Kidwell, R. E., & Robie, C. (2003). Withholding effort in organizations:
Toward development and validation of a measure. Journal of Business and Psychology,
17(4), 537-561.)

1. I gave less than 100% effort on my work today.

2. I took longer breaks than I should have today.

3. I put in less effort in my work than I know I could today.
4. While at work today I was daydreaming.

5. While at work today I pretended to be busy.
Demographics

1. Age (in years)

2. Gender

a Male
b. Female

3. Total Years of Work Experience (in years)

4. Background
a. Asian/Pacific Islander
b. Black/African American
c. Hispanic
d. Native American/American Indian
e. White
f. Other

111

Additional Interview Items used during Pilot Testing Phase

I. What did you think about today’s experience? What are your reactions?

2. Did any part of today’s session raise your suspicions about why you were here today?
3. Were you convinced that you were an employee of this consulting organization?

4. What could be done to increase the realism of this experience for future participants?
Of the manipulations? Of the location? Of the task itself? Of the questionnaires following

the work?

5. Did you feel free to do what you wanted to get the Internet research project done? Did
you go outside the bounds of what was established for you?

6. How did you react to the type of organizational control that was used in this session?
Did you take it seriously, respond naturally, disregard it, etc.?

7. Were there other behaviors that you did that we did not look at? Either positive or
negative?

8. What were the norms that guided your behavior?

112

APPENDIX F

Dynamic Research Group Debriefing F orm

In the wake of several corporate scandals such as Enron, Worldcom, and Adelphia, organizations are
increasingly concerned with employees following proper standards of conduct. The work that youjust
participated in is part of a psychology research activity designed to understand how control systems designed to
curb negative employee behavior may inadvertently result in fewer positive employee behaviors as well.
Dynamic Research Group is a fictional company that was created for the purposes of this research. Control
systems are any tool that the organization uses to bring out desirable behavior in their employees, whether a
monitoring system, ethics program, or norms that encourage people to do the right thing. These systems are
typically designed to limit negative behavior in the organization (e.g., employee thefl or withholding effort), but
an unintended consequence is that employees may react negatively and are no longer as willing to go above or
beyond what is required of them. We are interested in your justice and trust reactions to the type of control that
was being used and then will link your reactions to your behavior. We measured negative behavior by the length
of your break and the webpages you visited that were outside the authorized list given to you by the session
administrator. We will also look at your positive contributions by whether or not you went above and beyond in
your written report or if you answered the additional questionnaires given to you at the end of the session. We
hope that this research will benefit organizations that are looking to understand employees’ behavior and
maximize employees’ positive contributions while minimizing negative behavior.

We wanted you to feel as though you were an employee of a real organization so that you could feel
free to engage in actual work behavior. Had we used a true laboratory experiment, we felt it would have created
a situation that was too artificial. We wanted to avoid a situation where you felt as if you were being watched as
part of an experiment and then refrain from your natural behavior. Creating a situation where you felt as though
you were an employee allows us to more accurately capture our variables of actual positive and negative
behavior in a work setting.

If you have any questions about this research or the research process, please let the session
administrator know. If any questions or concerns arise after you leave the session today, please contact Dr. Dan
Ilgen, Department of Psychology, 340A Psychology Building, East Lansing, MI 48824, 517.353.4873, email:
ilgen(a2msu.edu or Jaclyn Nowakowski, Department of Psychology, 346 Psychology Building, East Lansing, MI
48824, 517.355.2171, email: nowakol 8@msu.edu. You may also contact Michigan State University’s
Committee on Research Involving Human Subjects (anonymously if you wish) at 202 Olds Hall, East Lansing,
M148824, 517.355.2l80, email: ucrihs@msu.edu. This project has received their approval.

We assure you that none of the data that we have gathered here can be tied to your identity. All of your
responses will be given a number that we will use for our research purposes. While we do have your name on
file, we will never associate your name with your data alter the process of data collection has been completed.
Your responses will remain confidential and we would like to use your data for our research. You may decline
without any penalty. Even though the company titled Dynamic Research Group is a fictional company that was
created for the purpose of this project, all participants will receive the $30 as advertised for your time and work
effort.

Please mark the box and indicate your permission to allow the researchers to use your data for

research purposes only. Evegyone is asked to keep the details of this activity confidential dge t2 the
gnaitive nature of the information.

D The researchers MAY use my data for research purposes.
D The researchers MAY NOT use my data for research purposes.
Employee:

Printed Name
Signed Name

 

 

Administrator:
Printed Name Jaclyn Nowakowski
Signed Name Date

 

113

.E. V a a... ”3. V a s A u 5:333 usgm d u :82 £on 2w com .mmm n 2 632

 

 

odd Nod- Nod- Ed- Ed 2d 2d. 2d- 2858526 30525 8:8 329:8 .x. .2
mod Ed- mod odd- Ed- Ed 2d. vod- 285.3336 {03:88 8:8 329:8 o\.. .2
mod Ed Nod- Ed cod mdd- Ed- Ed- $82.35 $236 team-m .3
odd- cod odd Ed- odd Ed- Ed- ...m fid- 5m5. omen toned .m g
odd odd mod- :cmd Nod Ed said Ed 3263 woman does—2 dick-:0: 53852 .Q
*smmd 32d Ed- immd odd ifd ~_d Ed Bum? 3:303 down—2 xcoB-eo: H3252 .2
immd Immd Ed 8d. mod- mod Nod- mod Ame» u g e: u 8 8:303 area-:0: an; d—
.13 d inmd mod- mod- Ed- Ed mod an _d 3303 $me Eugen 8on c3852 d
.3..de Ed- Ed- 2 d 9d Ed- Ed @863 82303 30526 «98 58:32 .w
*m _ d- mod- mod- Ed Edd Ed- Ame» u _ o: n 8 8:303 Eocene «55 55 H
86. 3o- 85 33- 2.? caused seas a 335 23 .0
odd Ed .3..de mod 905335 x3 to bus .m
Ed am _ d id 305335 83295 5953 :30 .v
d _ .0. ix _ d- 3835:: x35 Sic mo Swami .m
tad 92:55 N as; do ewes. .N
-- massed a seas do 533 ._

m n o w v m m _ o3mtm>

 

95.3:me 938.:me hmumhau3=8m§ Eetfimheuafis

O EDZmEA<

114

 

 

taco So .26 36. So- So- 8.? 25:8is Eofié can... sages as .2
:3 tone :5. So. 8d- :3. saméoeasc eczema 9:3 2238 so .2
to; So So- so So can case .3
:3 8.0- :.o- 36. swag 88 :25 .2
in; its .23 Bass mama case eases. 88:2 .2
tame a _ .o cases 8:303 @852 e228 H3.52 .:
.c _ .o 8:303 seats: ES .2
2 E 2 S : 2 o afifi>

 

2333mm 82%me hmumhaeneaamfie meetsmteoamcé

3.283 c 528%..

115

APPENDIX H

List of Hypotheses and Analyses

 

Hypothesis Analysis Supported?

H1: Control systems with an internal locus of HLM No
responsibility will result in more positive

procedural justice reactions compared to

systems with an external locus of

responsibility.

H2: Formal control systems will result in HLM No
more positive distributive justice reactions
compared to informal control systems.

H3: External control systems will result in HLM No
less positive interpersonal justice reactions
compared to internal control systems.

H4: Control systems with an internal locus of HLM No
responsibility will result in more positive

trust reactions compared to systems with an

external locus of responsibility.

H5: The more positive the justice reactions, Multiple No
the greater the frequency of constructive regression

deviance.

H6: The more negative the justice reactions, Multiple Partial
the greater the frequency of destructive regression

deviance.

H7: The more positive the trust reactions, the Regression No
greater the frequency of constructive

deviance.

H8: The more positive the trust reactions, the Regression Yes

lower the frequency of destructive deviance.

116

Appendix H (cont’d.)

List of Hypotheses and Analyses

 

 

Hypothesis Analysis Supported?
H9a: Conscientiousness will moderate the Moderated 9a — Yes
relationship between employee trust reactions hierarchical 9b - No
and destructive deviance, such that when regression 9c — No
conscientiousness is low, the relationship will 9d — No

be stronger.

H9b: Conscientiousness will moderate the
relationship between employee justice
reactions and destructive deviance, such that
when conscientiousness is low, the
relationship will be stronger.

H9c: Conscientiousness will moderate the
relationship between employee trust reactions
and constructive deviance, such that when
conscientiousness is low, the relationship will
be stronger.

H9d: Conscientiousness will moderate the
relationship between employee justice
reactions and constructive deviance, such that
when conscientiousness is low, the
relationship will be stronger.

H10a: Proactive personality will moderate the Moderated 10a - No
relationship between employee trust reactions hierarchical 10b - No
and constructive deviance, such that when regression

proactivity is high, the relationship will be

weaker.

H10b: Proactive personality will moderate
the relationship between employee justice
reactions and constructive deviance, such that
when proactivity is high, the relationship will
be weaker.

117

Appendix H (cont’d.)

List of Hypotheses and Analyses

 

 

Hypothesis Analysis Supported?
H1 Ia: Formal control systems are more likely HLM 11a — No
to limit destructive deviance than informal 11b — No

control system, and

H1 1b: The relationship between the formality
of the control system and the frequency of
destructive deviance will be mediated by
justice reactions.

H12a: Formal control systems will be more HLM 12a — No
likely to restrict constructive deviance as 12b — No
compared to informal control systems, and

H12b: The relationship between the formality
of control and the frequency of constructive
deviance will be mediated by justice

reactions.

Hl3a: External control systems will be more HLM 13a — No
likely to restrict destructive deviance as 13b — No
compared to internal control systems, and 13c — No

The relationship between locus of
responsibility and the frequency of
destructive deviance will be mediated by
(H13b) justice and (Hl3c) trust reactions.

H14a: External control systems will be more HLM 14a — Yes
likely to restrict constructive deviance as 14b — No
compared to internal control systems, and 14c — No

The relationship between locus of

responsibility and the frequency of

constructive deviance will be mediated by
fl-I 14b) justice and (H14c) trust reactions.

118

REFERENCES

Adam, A. M., & Rachman-Moore, D. (2004). The methods used to implement an ethical
code of conduct and employee attitudes. Journal of Business Ethics, 54, 225-244.

Adams, J. S. (1965). Inequity in social exchange. In L. Berkowitz (Ed), Advances in
experimental social psychology (Vol. 2, pp. 267-299). New York: Academic.

Aiello, J. R., & Kolb, K. J. (1995). Electronic performance monitoring and social context:
Impact on productivity and stress. Journal of Applied Psychology, 80(3), 339-353.

Alder, G. S., & Ambrose, M. L. (2005). An examination of the effect of computerized
performance monitoring feedback on monitoring fairness, performance, and
satisfaction. Organizational Behavior and Human Decision Processes, 97, I61-
I77.

Alge, B. J. (2001). Effects of computer surveillance on perceptions of privacy and
procedural justice. Journal of Applied Psychology, 86(4), 797-804.

Ambrose, M. L., & Alder, G. S. (2000). Designing, implementing, and utilizing
computerized performance monitoring: Enhancing organizational justice. In G. R.
Ferris (Ed.), Research in personnel and human resources management (V 01. 18,
pp. 187-219). Stamford, CT: JAI Press.

Ambrose, M. L., Seabright, M. A., & Schminke, M. (2002). Sabotage in the workplace:
The role of organizational injustice. Organizational Behavior and Human
Decision Processes, 89(1), 947-965.

Barker, J. R. (1993). Tightening the iron cage: Concertive control in self-managing
teams. Administrative Science Quarterly, 38, 408-437.

Barley, S. R., & Kunda, G. (1992). Design and devotion: Surges of rational and
normative ideologies of control in managerial discourse. Administrative Science
Quarterly, 37(3), 363-399.

Barrick, M. R., & Mount, M. K. (1991). The Big Five personality dimensions and job
performance: A meta-analysis. Personnel Psychology, 44(1), 1-26.

Bateman, T. S., & Crant, J. M. (1993). The proactive component of organizational
behavior: A measure and correlates. Journal of Organizational Behavior, 14(2),
103-1 18.

Bell, N. E., & Staw, B. M. (1989). People as sculptors versus sculpture: The role of
personality and personal control in organizations. In M. B. Arthur, D. T. Hall &
B. S. Lawrence (Eds), The handbook of career theory (pp. 232-251). Cambridge,
England: Cambridge University Press.

119

Bies, R. J ., & Moag, J. F. (1986). Interactional justice: Communication criteria of
fairness. In R. J. Lewicki, B. H. Sheppard & M. H. Bazerman (Eds), Research on
negotiations in organizations (Vol. I, pp. 43-55). Greenwich, CT: JAI Press.

Bolino, M. C. (1999). Citizenship and impression management: Good soldiers or good
actors? Academy of Management Review, 24(1), 82-98.

Bonke, J. (2005). Paid work and unpaid work: Diary information versus questionnaire
information. Social Indicators Research, 70, 349-368.

Cadrain, D. (2005). GPS on rise; workers' complaints may follow. HRMagazine, 50(4),
32, 40.

Cardinal, L. B., Sitkin, S. B., & Long, C. P. (2004). Balancing and rebalancing in the
creation and evolution of organizational control. Organization Science, 15(4),
411-431.

Carlson, R. G. (1987). The organizational control system. Journal of Property
Management, 52(3), 6.

Chan, D. (2006). Interactive effects of situational judgment effectiveness and proactive
personality on work perceptions and work outcomes. Journal of Applied
Psychology, 91(2), 475-481.

Clark, K., & James, K. (1999). Justice and positive and negative creativity. Creativity
Research Journal. 12(4), 311-320.

Colbert, A. E., Mount, M. K., Harter, J. K., Witt, L. A., & Barrick, M. R. (2004).
Interactive effects of personality and perceptions of the work situation on
workplace deviance. Journal of Applied Psychology, 89(4), 599-609.

Colquitt, J. A. (2001). On the dimensionality of organizational justice: A construct
validation of a measure. Journal of Applied Psychology, 86(3), 386—400.

Conlon, D. E., Meyer, C. J ., & Nowakowski, J. M. (2005). How does organizational
justice affect performance, withdrawal, and counterproductive behavior? In J. A.
Colquitt & J. Greenberg (Eds), Handbook of organizational justice (pp. 301-
327). Mahwah, NJ: Lawrence Erlbaum Associates.

Cortina, L. M., & Magley, V. J. (2003). Raising voice, risking retaliation: Events
following interpersonal mistreatment in the workplace. Journal of Occupational
Health Psychology. 8(4), 247-265.

Crossen, B. R. (1993). Managing employee unethical behavior without invading
individual privacy. Journal of Business and Psychology, 8(2), 227-243.

120

dc Cremer, D., Snyder, M., & Dewitte, S. (2001). "The less I trust, the less I contribute (or
not)?’ The effects of trust, accountability and self-monitoring in social dilemmas.
European Journal of Social Psychology, 31(1), 93-107.

Douthitt, E. A., & Aiello, J. R. (2001). The role of participation and control in the effects
of computer monitoring on fairness perceptions, task satisfaction, and
performance. Journal of Applied Psychology, 86(5), 867-874.

El-Alayli, A., & Messe, L. A. (2004). Reactions toward an unexpected or
counternormative favor-giver: Does it matter if we think we can reciprocate?
Journal of Experimental Social Psychology, 40(5), 633-641.

Enzle, M. E., & Anderson, S. C. (1993). Surveillant intentions and intrinsic motivation.
Journal of Personality and Social Psychology, 64(2), 25 7-266.

Ezzamel, M., & Willmott, H. (1998). Accounting for teamwork: A critical study of
group-based systems of organizational control. Administrative Science Quarterly,
43, 358-3 96.

Federal Bureau of Investigation. (2004). Workplace violence: Issues in response.
Quantico, VA: National Center for the Analysis of Violent Crime.

Fox, 8., Spector, P. E., & Miles, D. (2001). Counterproductive work behavior (CWB) in
response to job stressors and organizational justice: Some mediator and moderator
tests for autonomy and emotions. Journal of Vocational Behavior, 59(3), 291-309.

Frenkel, S., & Orlitzky, M. (2005). Organizational trustworthiness and workplace labor
productivity: Testing a new theory. Asia Pacific Journal of Human Resources,
43(1), 34-51.

Frink, D. D., & Klimoski, R. J. (1998). Toward a theory of accountability in
organizations and human resource management. In G. R. Ferris (Ed.), Research in
personnel and human resources management (Vol. 16, pp. 1-51). Stamford, CT:
JAI Press.

Gabarro, J. J ., & Athos, J. (1976). Interpersonal relations and communications. New
York: Prentice Hall.

Galbraith, J. R. (1977). Organization design. Reading, MA: Addison-Wesley.

Gilliland, S. W. (1994). Effects of a procedural and distributive justice on reactions to a
selection system. Journal of Applied Psychology, 79, 691-701.

Glassman, C. A. (2005). The Sarbanes-Oxley Act of 2002: A culmination of corporate
reform initiatives by the Bush administration, the SEC and Congress.
Washington, DC: Securities and Exchange Commission.

121

Goldberg, L. R. (1999). A broad-bandwidth, public-domain, personality inventory
measuring the lower-level facets of several five-factor models. In I. Mervielde, I.
Deary, F. D. Fruyt & F. Ostendorf (Eds), Personality psychology in Europe (V 01.
7, pp. 7-28). Tilburg, the Netherlands: Tilburg University Press.

Greenberg, J. (1993). The social side of fairness: Interpersonal and informational classes
of organizational justice. In R. Cropanzano (Ed.), Justice in the workplace:
Approaching fairness in human resource management (pp. 79-103). Hillsdale, NJ:
Erlbaum.

Greenberg, J. (2002). Who stole the money, and when? Individual and situational
determinants of employee theft. Organizational Behavior and Human Decision
Processes, 89, 985- 1 003.

Greenberg, L., & Barling, J. (1999). Predicting employee aggression against coworkers,
subordinates and supervisors: The roles of person behaviors and perceived
workplace factors. Journal of Organizational Behavior, 20(6), 897-913.

Heckert, A., & Heckert, D. M. (2004). Using an integrated typology of deviance to
analyze ten common norms of the US. middle class. Sociological Quarterly,
45(2), 209.

Hewitt, P. J. (1983). Performance evaluation: How to make it easier on you - and your
employee. Super Vision, 45(6), 14-18.

Hoffman, D. A. (1997). An overview of the logic and rationale for hierarchical linear
models. Journal of Management, 23(6), 723-744.

Hoffman, D. A., & Gavin, M. B. (1998). Centering decisions in hiearchical linear
modeling: Implications for research in organizations. Journal of Management,
24(5), 623-641.

Hoover, E. (2002, May 3). Honor for honor's sake? Retrieved June 6, 2005, from
http://chronicle.com/free/v48/i34/34a03501.htrn.

Hultman, K. (2004). Let’s wipe out systemic mistrust. Organization Development
Journal, 22( l ), 1 02-106.

Judge, T. A., Erez, A., Bono, J. E., & Thoresen, C. J. (2003). The core self-evaluations
scale: Development of a measure. Personnel Psychology, 56(2), 303-331.

Kidwell, R. E., & Bennett, N. (1994). Electronic surveillance as employee control: A
procedural justice interpretation. Journal of High Technology Management
Research, 5(1), 39-57.

Kidwell, R. E., & Robie, C. (2003). Withholding effort in organizations: Toward
development and validation of a measure. Journal of Business and Psychology.
17(4), 537-561.

122

Kramer, R. M., & Goldman, L. (1995). Helping the group or helping yourself? Social
motives and group identity in resource dilemmas. In D. A. Schroeder (Ed.), Social
dilemmas: Perspectives on individuals and groups (pp. 49-67). London: Praeger.

Lavoie, J. A. A., & Pychyl, T. A. (2001). Cyberslacking and the procrastination
superhighway: A web-based survey of online procrastination, attitudes, and
emotion. Social Science Computer Review, 19(4), 431-444.

Lawler, E. E., & Rhode, J. G. (1976). Information and control in organizations. Pacific
Palisades, CA: Goodyear Publishing Company, Inc.

LePine, J. A., Erez, A., & Johnson, D. E. (2002). The nature and dimensionality of
organizational citizenship behavior: A critical review and meta-analysis. Journal
of Applied Psychology, 87(1), 52-65.

Leventhal, G. S. (1976). Fairness in social relationships. In J. W. Thibaut, J. T. Spence &
R. C. Carson (Eds), Contemporary topics in social psychology (pp. 212-239).
Morristown, NJ: General Learning.

Little, H. T., Magnet, N. R., & Welker, R. B. (2002). The fairness of formal budgetary
procedures and their enactment. Group and Organization Management, 27(2),
209-225.

Litzky, B. E., Eddleston, K. A., & Kidder, D. L. (2006). The good, the bad, and the
misguided: How managers inadvertently encourage deviant behaviors. Academy
of Management Perspectives, 20(1), 91-103.

Malhotra, D., & Mumighan, J. K. (2002). The effects of contracts on interpersonal trust.
Administrative Science Quarterly, 47(3), 534-559.

Mayer, R. C., Davis, J. H., & Schoorman, F. D. (1995). An integrative model of
organizational trust. Academy of Management Review, 20(3), 709-734.

Mesch, D. J ., & Dalton, D. R. (1992). Unexpected consequences of improving workplace
justice: A six-year time series assessment. Academy of Management Journal,
35(5), 1099-1114.

Niehoff, B. P., & Moorman, R. H. (1993). Justice as a mediator of the relationship
between methods of monitoring and organizational citizenship behavior. Academy
of Management Journal. 36(3), 527-556.

O'Brien, G. E. (1984). Locus of control, work, and retirement. In H. M. Lefcourt (Ed.),
Research with the locus of control construct: Extensions and limitations (V 01. 3,
pp. 7-68). New York: Academic Press.

Ones, D. S., & Viswesvaran, C. (2001). Integrity tests and other criterion-focused
occupational personality scales (COPS) used in personnel selection. International
Journal of Selection and Assessment, 9(1-2), 31-39.

123

O'Reilly, A. P. (1973). Skill requirements: Supervisor - subordinate conflict. Personnel
Psychology, 26(1). 75-80.

Organ, D. W. (1988). Organizational citizenship behavior: The good soldier syndrome.
Lexington, MA: Lexington Books.

Organ, D. W., & Ryan, K. (1995). A meta-analytic review of attitudinal and dispositional
predictors of organizational citizenship behavior. Personnel Psychology, 48(4),
775-802.

Ouchi, W. G. (1977). The relationship between organizational structure and
organizational control. Administrative Science Quarterly, 22, 95-113.

Ouchi, W. G. (1979). A conceptual framework for the design of organizational control
mechanisms. Management Science, 25, 833-848.

Ouchi, W. G. (1980). Markets, bureaucracies, and clans. Administrative Science
Quarterly, 25(1), 129-141.

Ouchi, W. G., & Johnson, J. B. (1978). Types of organizational control and their
relationship to emotional well being. Administrative Science Quarterly, 23, 293-
317.

Ouchi, W. G., & Maguire, M. A. (1975). Organizational control: Two functions.
Administrative Science Quarterly, 20(4), 559-569.

Pearce, C. L., & Herbik, P. A. (2004). Citizenship behavior at the team level of analysis:
The effects of team leadership, team commitment, perceived team support, and
team size. Journal of Social Psychology, 144(3), 293-310.

Pennock, G. ( l 930). Industrial research at Hawthorne. Personnel Journal, 8, 296-313.

Podsakoff, P. M., MacKenzie, S. B., Paine, J. B., & Bachrach, D. G. (2000).
Organizational citizenship behaviors: A critical review of the theoretical and
empirical literature and suggestions for future research. Journal of Management,
26(3), 513-563.

Rabinowitz, A. M. (1996). Rebuilding public confidence in auditors and organizational
controls. The CPA Journal, 66(1), 30.

Raudenbush, S., Bryk, T., & Congdon, R. (2000). HLM 6 Hierarchical Linear and
Nonlinear Modeling (Version 6): Scientific Software International.

Robinson, S. L., & Bennett, R. J. (1995). A typology of deviant workplace behaviors: A
multi-dimensional scaling study. Academy of Management Journal, 38, 555-572.

Robinson, S. L., & Bennett, R. J. (1997). Workplace deviance: Its definition, its
manifestations, and its causes. Research on Negotiation in Organizations, 6, 3-27.

124

Robinson, S. L., & Rousseau, D. M. (1994). Violating the psychological contract: Not the
exception but the norm. Journal of Organizational Behavior, I 5(3), 245-259.

Roth, N. L., Sitkin, S. B., & House, A. (1994). Stigma as a determinant of legalization. In
S. B. Sitkin & R. J. Bies (Eds.), The legalistic organization (pp. 137-168).
Thousand Oaks, CA: Sage.

Rousseau, D. M., Sitkin, S. B., Burt, R. 8., & Camerer, C. (1998). Not so different after
all: A cross-discipline view of trust. Academy of Management Review, 23(3), 393-
404.

Schroth, H. A., & Pradhan Shah, P. (2000). Procedures: Do we really want to know
them? An examination of the effects of procedural justice on self-esteem. Journal
of Applied Psychology, 85(3), 462-471.

Seibert, S. E., Crant, J. M., & Kraimer, M. L. (1999). Proactive personality and career
success. Journal of Applied Psychology, 84(3), 416-427.

Sewell, G. (1998). The discipline of teams: The control of team-based industrial work
through electronic and peer surveillance. Administrative Science Quarterly, 43,
397-428.

Shapiro, D. L., & Brett, J. M. (1993). Comparing three processes underlying judgments
of procedural justice: A field study of mediation and arbitration. Journal of
Personality and Social Psychology, 65, 1167-1177.

Shapiro, D. L... & Brett, J. M. (2005). What is the role of control in organizational justice?
In J. Greenberg & J. A. Colquitt (Eds), Handbook of Organizational Justice.
Mahwah, NJ: Lawrence Erlbaum Associates.

Sitkin, S. B., & George, E. (2005). Managerial trust-building through the use of
legitimating formal and informal control mechanisms. International Sociology,
20(3), 307-338.

Smith, E. D. (2005, May 30). The company is watching Abon Beacon Journal, pp. D1,
D3.

Snow, D. L., Swan, S. C., Raghavan, C., Connell, C. M., & Klein, 1. (2003). The
relationship of work stressors, coping and social support to psychological
symptoms among female secretarial employees. Work and Stress, 1 7(3), 241-263.

Spreitzer, G. M., & Sonenshein, S. (2004). Toward the construct definition of positive
deviance. American Behavioral Scientist, 47(6), 828—847.

Stanton, J. M. (2000). Reactions to employee performance monitoring: Framework,
review, and research directions. Human Performance, 13(1), 85-113.

125

Stanton, J. M., & Weiss, E. M. (2000). Electronic monitoring in their own words: an
exploratory study of employees' experiences with new types of surveillance.
Computers in Human Behavior, 16, 423-440.

Stevens, J. M., Steensma, H. K., Harrison, D. A., & Cochran, P. L. (2005). Symbolic or
substantive document? The influence of ethics codes on financial executives'
decisions. Strategic Management Journal, 26, 181-195.

Strutton, D., Pelton, L. E., & Lumpkin, J. R. (1993). The relationship between
psychological climate and salesperson-sales manager trust in sales organizations.
Journal of Personal Selling and Sales Management, 13(4), 1-14.

Swanson, S. (2002). Employers take a closer look. Information Week, 897, 40-41.

Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistics. Boston: Allyn
and Bacon.

Tabak, F., & Smith, W. P. (2005). Privacy and electronic monitoring in the workplace: A
model of managerial cognition and relational trust development. Employee
Responsibilities and Rights Journal, 17(3), 173-189.

Thibaut, J. W., & Walker, L. (1975). Procedural justice: A psychological perspective.
Hillsdale, NJ: Erlbaum.

Trevino, L. K., & Weaver, G. R. (2001). Organizational justice and ethics program
"follow-through": Influences on employees' harmful and helpful behavior.
Business Ethics Quarterly, 11(4), 651-671 .

Tyler, T. R. (1987). Conditions leading to value-expressive effects in judgments of
procedural justice: A test of four models. Journal of Personality and Social
Psychology, 52, 333-344.

Van Dyne, L., Graham, J. W., & Dienesch, R. M. (1994). Organizational citizenship
behavior: Construct redefinition, measurement, and validation. Academy of
Management Journal, 37(4), 765-802.

Voyer, J. J. (1994). Coercive organizational politics and organizational outcomes: An
interpretive study. Organization Science, 5(1), 72.

Warren, D. E. (2003). Constructive and destructive deviance in organizations. Academy
of Management Review, 28(4), 622-632.

Weaver, G. R., Trevino, L. K., & Cochran, P. L. (1999). Corporate ethics programs as
control systems: Influences of executive commitment and environmental factors.
Academy of Management Journal, 42(1), 41-57.

126

   

l11111111111111111711