. mm?

$8.? 1!
:1;
I

.x‘r D. 943.
4025.}.“233fi I. i .
213., l. i fiww. . ;.

V . 1... I." 5.. 4:. :41

a...“ «e, .t. .cflhtya.

Nguyh‘fivnfimmuwm .
. i (tutu. I." .m

..\lo\~|.

iat‘n-fii I
. «5.. 2:.1 3.

All v\0 .

l... ‘9 15"}. 11v
5 Vlwar. II A
i .q 2;. rolxvi.

~. 2.: . 5 5.

’05 ‘
T J 1.
‘ .. y).

 

 

 

was * UBRARY
n . Michigan State
J35 University

 

 

This is to certify that the
dissertation entitled

SILENCE OR VOICE? USING FACEWORK AND
COMMUNICATION APPREHENSION TO EXPLAIN
EMPLOYEE RESPONSES TO AUTONOMY AND
COMPETENCE FACE THREATS POSED BY NEGATIVE
FEEDBACK

presented by

CATHERINE YORK KINGSLEY WESTERMAN

has been accepted towards futfillment
of the requirements for the

PhD. degree in Communication

 

 

Kim”! (is LL’F/Jc-W’Ii 141/
Major Professor’s Signature

ma’PN’v 0%, «200 z?

 

Date

MSU is an affinnative-action, equal-opportunity employer

 

PLACE IN 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

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K‘IProj/AccsPres/CIRC/DateDue.indd

 

SILENCE OR VOICE?
USING FACEWORK AND COMMUNICATION APPREHENSION TO EXPLAIN
EMPLOYEE RESPONSES TO AUTONOMY AND COMPETENCE FACE THREATS
POSED BY NEGATIVE FEEDBACK

By

Catherine York Kingsley Westerman

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Department of Communication

2008

ABSTRACT
SILENCE OR VOICE: USING FACEWORK AND COMMUNICATION
APPREHENSION TO EXPLAIN EMPLOYEE RESPONSES TO AUTONOMY AND
COMPETENCE FACE THREATS POSED BY NEGATIVE FEEDBACK
By
Catherine York Kingsley Westerman

This study examined the type of communication responses employees choose
after receiving potentially face-threatening negative feedback from their supervisors. The
facework framework suggests that the addition of preventative facework to negative
feedback messages may reduce threats to autonomy and competence face. Different
levels of face threat may lead respondents to choose voice or silence response modes and
to be defensively or prosocially motivated following reception of negative feedback.
Communication apprehension (CA) is also examined as a factor in employees’ choice of
voice or silence.

This study provided a test of eight negative feedback messages that differed in
level and type of face threat, their predicted effects on perceptions of face threat, and
selection of voice or silence responses as a function of both face threat and CA. Voice
responses may be more functional than silence responses because they allow employees
to share ideas and clarify any misunderstandings about feedback they receive.
Hagedoorn, Van Yperen, Van De Vliert, and Buunk (1999) suggested reporting of
problems as opposed to silence responses can have two positive outcomes: it may result

in an early alert for the organization that procedures or practices may not be working and

it may also reduce distress for employees experiencing the problem. A new individual

difference variable known as facework strategy preference was also tested for its effect
on perceptions of face threat and selection of voice or silence responses.

This study (N = 443) employed a 2 (type of face threat: autonomy, competence) x
4 (level of threat: threat only, threat with tact, threat with approbation, threat with tact and
approbation) design with CA as a co-variate to test these ideas. Results indicated that
approbation reduced perceptions of both autonomy and competence threat, whereas tact
was not effective in reducing the perception of either type of threat. Thus, approbation
was the most effective facework message type. As employees felt more threatened (either
autonomy or competence) they were less likely to use prosocial voice and more likely to
use defensive silence as a response. As a general rule, higher perceived threat was
negatively associated with the functional response of prosocial voice; as threat decreased,
prosocial voice responses increased. Prosocial voice was also negatively associated with
the dysfunctional responses of defensive silence and defensive voice.

The association between defensive voice and autonomy threat was negative and
significant for high CAs but nonsignificant for low CAs, as was the correlation between
prosocial voice and competence threat. The correlation between defensive silence and
competence threat was positive and significant for high CAs and nonsignificant for low
CAs. The preference for direct facework was negatively associated with both defensive
voice and silence and positively associated with prosocial voice. The preference for
indirect facework was positively associated with both defensive voice and silence but not
correlated with prosocial voice. Implications of these findings, limitations, and future

directions are discussed.

ACKNOWLEDGEMENTS

Sandi, I could not have completed this dissertation without you. Working with
you gave me new confidence in my abilities as a scholar. It was a lucky turn of events
that allowed me to work with you and enjoy the advantage of your expertise. I hope you
know how much I appreciated (and still appreciate) your guidance and encouragement
throughout the process. I hope our relationship continues for a long time.

Dave, I think you are the only one who knows exactly how difficult this was for
me. Apparently some other people saw it on my face, but they weren’t there when I spent
every hour of the day writing or when we left Christmas with your family early so I could
get home to work or all those other times I won’t mention. You gave me support, comfort
and encouragement when I needed it most. THANK YOU for everything you did and
said to help me achieve this.

Mom and Dad, you provided me with the emotional support and listening ear(s) I
needed to get through. I know it wasn’t the easiest job to listen to me for a while there,
but it really made a big difference. Thanks for your support and thanks for always
thinking I am the best. Unconditional love is a great feeling.

One other person really made the difference: Brandon Vanderheide. Without you,
I would not have my data!! Thank you, thank you! I also owe thanks to others who
provided emotional support, a place to vent, and a shoulder to cry on when things were

really going to h***. You know who you are—thank you.

iv

TABLE OF CONTENTS

LIST OF TABLES ........................................................................................................... VII
LIST OF FIGURES ....................................................................................................... VIII
LITERATURE REVIEW ................................................................................................... 1
FACE ................................................................................................................................ 4
Autonomy. .................................................................................................................. 5
Competence ................................................................................................................. 9
FEEDBACK ..................................................................................................................... 10
Control Perspective. .................................................................................................. 12
Operant Conditioning Perspective. ........................................................................... 13
Social Cognitive Perspective. ................................................................................... 14
Face Perspective ........................................................................................................ 15
Feedback Valence. .................................................................................................... 16
Face threats and facework ......................................................................................... 19
Preventative Facework. ............................................................................................. 2O
Facework Strategies. ................................................................................................. 21
Voice and Silence. .................................................................................................... 22
Communication Apprehension. ................................................................................ 29

F acework Strategy Preference. ................................................................................. 34
Supervisor support & feedback severity. .................................................................. 35
METHOD ......................................................................................................................... 38
OVERVIEW ..................................................................................................................... 38
Participants ................................................................................................................ 38
Design. ...................................................................................................................... 41
Procedure. ................................................................................................................. 42
Feedback Messages ................................................................................................... 43
MEASUREMENT .............................................................................................................. 45
PRCA-Org ................................................................................................................. 46

F acework strategy preference. .................................................................................. 47
Supervisor supportiveness. ....................................................................................... 47
Feedback severity ...................................................................................................... 48
Realism. .................................................................................................................... 48
Face threat. ................................................................................................................ 49
Voice. ........................................................................................................................ 49
Silence. ...................................................................................................................... 50
Feedback sign ............................................................................................................ 50
Work information & Demographics. ........................................................................ 50
Para-ANALYSIS ............................................................................................................... 51
Work Information & Demographics. ........................................................................ 51
Feedback Experience & Realism Check. .................................................................. 51
Manipulation Check. ................................................................................................. 53

RESULTS ......................................................................................................................... 54

OVERVIEW ..................................................................................................................... 54
PRETEST ......................................................................................................................... 54
MAIN DATA ANALYSIS .................................................................................................. 55
POST HOC ANALYSES ...................................................................................................... 62
Facework Strategy Preference. ................................................................................. 62
Voice/Silence Regressions. ....................................................................................... 63
SUMMARY OF FINDINGS ................................................................................................. 69
DISCUSSION ................................................................................................................... 73
SUMMARY ...................................................................................................................... 73
EXPLANATION/IMPLICATIONS OF FINDINGS .................................................................... 73
Face Threat ................................................................................................................ 73
Defensive Voice. ....................................................................................................... 77
Prosocial Voice. ........................................................................................................ 81
Defensive Silence ...................................................................................................... 84
Prosocial Silence. ...................................................................................................... 86
IMPLICATIONS FOR PRACTITIONERS ............................................................................... 86
IMPLICATIONS FOR RESEARCHERS ................................................................................. 87
LIMITATIONS .................................................................................................................. 88
FUTURE DIRECTIONS ...................................................................................................... 90
Message issues. ......................................................................................................... 90
Individual difference variables. ................................................................................ 91
Supervisor-subordinate interaction. .......................................................................... 92
Commtmication responses. ....................................................................................... 92
CONCLUSION .................................................................................................................. 93
EN DNOTES ..................................................................................................................... 96
APPENDICES ................................................................................................................ 117
APPENDIX A. MESSAGE COMPONENTS ........................................................................ l 17
APPENDIX B. MESSAGE CONDITIONS ........................................................................... 1 18
APPENDIX C. MEASUREMENT ...................................................................................... 1 19
REFERENCES ............................................................................................................... 120

vi

LIST OF TABLES

TABLE 1. DEFINITIONS AND EXAMPLES OF FACE-SAVING MESSAGES FOR

SOLIDARITY, APPROBATION, AND TACT (LIM & BOWERS, 1991) .................... 97
TABLE 2. VOICE/SILENCE TABLE (ADAPTED FROM VAN DYNE, ANG, &

BOTERO, 2003) ............................................................................................................... 99
TABLE 3 . SCALE ITEMS & FACTOR LOADINGS .................................................. 100

TABLE 4. ONE-WAY AN OVA RESULTS FOR CONDITION ON CONTINUOUS
CONTROL VARIABLES. ............................................................................................. 107

TABLE 6. PERCEPTION OF AUTONOMY AND COMPETENCE FACE THREAT
BY CONDITION. ........................................................................................................... 108

TABLE 7. ZERO-ORDER CORRELATIONS AMONG TRAIT CA, FACE THREAT,
VOICE/SILENCE, AND DIRECT & INDIRECT F ACEWORK STRATEGY

PREFERENCE. .............................................................................................................. 109
TABLE 8. CORRELATIONS AMONG VARIABLES FOR LOW CA QUARTILE

SPLIT. ............................................................................................................................. 111
TABLE 9. CORRELATIONS AMONG VARIABLES FOR HIGH CA QUARTILE

SPLIT. ............................................................................................................................. 112
TABLE 10. HIERARCI-IICAL REGRESSION ANALYSES RESULTS. .................... 113
TABLE 11. HIERARCHICAL REGRESSION ANALYSES RESULTS. .................... 114

vii

LIST OF FIGURES

FIGURE 1. SIMPLE SLOPES FOR THE INTERACTION OF CA AND
COMPETENCE THREAT ON DEFENSIVE VOICE. ................................................... 64

FIGURE 2. SIMPLE SLOPES FOR THE INTERACTION OF CA AND
COMPETENCE THREAT ON PROSOCIAL VOICE. ................................................... 66

FIGURE 3. SIMPLE SLOPES FOR THE INTERACTION OF AUTONOMY AND
COMPETENCE THREAT ON PROSOCIAL VOICE. ................................................... 67

FIGURE 4. SIMPLE SLOPES FOR THE INTERACTION BETWEEN CA AND
COMPETENCE THREAT ON DEFENSIVE SILENCE. ............................................... 68

FIGURE 5. SIMPLE SLOPES FOR THE INTERACTION BETWEEN
APPROBATION-NO APPROBATION AND DIRECT FACEWORK PREFERENCE
ON DEFENSIVE SILENCE ............................................................................................. 69

viii

Literature Review

Researchers and practitioners alike agree that performance feedback is important
to organizations as it helps to improve performance and keep employees satisfied in their
jobs. However, a meta-analysis by Kluger and DeNisi (1996) suggested all feedback may
not have the desired effect of improving performance. They found that one third of the
time, performance decreased after feedback delivery. This finding highlights the need for
further research into receiver responses and the elements of performance feedback that
render it ineffective one-third of the time. This study will investigate some factors which
may influence employees’ propensity to communicate further with their supervisors
about negative feedback they receive, including face threat, communication apprehension
(CA), direct facework preference, indirect facework preference, supervisor
supportiveness, and perceptions of feedback severity.

Although the feedback a supervisor may give to an employee is important, it is
also important for the employee to be able to communicate with the supervisor about the
feedback. In order to improve their performance, employees Should be able to Share ideas
and questions with their supervisors after receiving feedback. Informal feedback in
particular may serve the function of opening a dialogue between supervisors and
employees because it can occur more frequently and with less formality than performance
appraisals. Informal feedback is defined here as unscheduled, interactive, involving a
supervisor-employee pairing, informal language and speech register, and having an
unarranged agenda (based on informal communication definition; Kraut, Fish, Root, &
Chalfonte, 1993). Here, the focus is on informal feedback rather than performance

appraisals for two reasons: I) a paucity of research exists on day-to-day communication

interactions in organizations (Bartoo & Sias, 2004) and 2) informal feedback can provide
information soon after problematic behavior, as recommended by previous researchers
(Anderson, 1987; Ilgen, Fisher, & Taylor, 1979; Lindsley, Brass, & Thomas, 1995).
Despite the potential for informal feedback to be a constructive tool, employees may not
feel comfortable communicating with their supervisors because performance feedback
can threaten the autonomy and competence face of the receiver (i.e., employee).

Face is “the public self-image that every member wants to claim for himself”
(Brown & Levinson, 1987, p. 61). Autonomy face is threatened by messages that impose
on the employee, such as criticism or orders, whereas competence face is threatened by
messages that imply the employee is not a capable individual, such as ridiculing
messages. If either type of face is threatened, employees may not feel comfortable
voicing ideas or questions about the feedback they received. The ability to use voice after
receiving feedback is important because voice is one way employees can assert their
freedom to participate by contributing to their organization (Ellis & Van Dyne, in press).

Additionally, when employees perceive a threat, they may be motivated to behave
in specific ways. If the face threat presented by feedback is perceived as severe,
employees may be motivated to protect themselves, or to behave defensively. For
example, they may attempt to place blame on others or to divert attention from problems
brought up in the feedback. Using voice behaviors such as criticisms and complaints is
one way employees seek to self-protect when they are mistreated (Ellis & Van Dyne, in
press). These types of behaviors may be employees’ responses to face-threatening
feedback and can prove damaging for both the employee and the organization. However,

if the threat is perceived as minimal, employees may be motivated to help the

organization, or to behave prosocially. For example, they may offer solutions to the
problems brought up in the feedback. In other. words, more constructive behavior may
result from diminished face threats in negative feedback messages.

Preventative facework provides a way to mitigate threat thus allowing employees
to communicate with their supervisors in desirable ways (Lim & Bowers, 1991). The use
of tact or approbation messages in concert with feedback known to be threatening can
help to diminish employees’ perceptions of face threat while still getting the information
across. With the reduction of face threat, employees should feel more comfortable
voicing ideas, concerns, or questions about the feedback to their supervisors, thus
facilitating the process of improving performance cooperatively.

In addition to face threat, a second concern is that some employees may be
apprehensive about communicating in general, whereas others may prefer commrmicating
regardless of the threat. Communication apprehension (CA) is “an individual’s level of
fear or anxiety associated with either real or anticipated communication with another
person or persons” (McCroskey, 1977a, p. 78). Because CA is an individual
characteristic, it is expected to affect employees’ communication with everyone they
encounter, including their supervisors. Pinder and Harlos (2001) suggested that receiver
CA may affect tendencies toward voice and silence. Those with high CA may find it
preferable to remain silent, whereas those with low CA may be more willing to Speak up
in these situations. Learning more about the effects of CA on employees’ attempts to
communicate with their supervisors can be useful in helping organizations deal more

effectively with the varying needs and abilities of their employees.

The combination of the level of face threat and employees’ level of trait CA will
determine whether they feel comfortable voicing their ideas and questions about feedback
they receive and their motives for choosing to use voice or to remain Silent. This study
will examine how the type and level of face threat in feedback messages, along with
employee CA, affects employees’ motives and communication responses (i.e., silence or
voice). A review of the face and feedback literatures will be provided, along with
treatments of autonomy, competence, silence, voice, and CA. Hypotheses and research
questions are posed, and the method, results, and discussion of the findings are presented.
Face

F ace is defined as the public display of one’s self (Holtgraves, Eck, & Lasky,
1997) or “the public self-image that every member wants to claim for himself” (Brown &
Levinson, 1987, p. 61 ). Each person has an image he/She wants to present in a given
setting and these images may differ across settings. For example, a CEO may have a
more relaxed “face” to be displayed at social events with clients and a more formal
“face” to be displayed when working with employees. In order to maintain these different
faces people must choose to enact behaviors that are consistent with the image they want
to project. Not only do people behave in a way consistent with their desired image, they
also help others in a situation to maintain their desired images. This is because each
individual’s face depends not only on their own behavior but also on others’ ability to
maintain their face in the interaction. Cooperative face maintenance in an interaction
helps maintain social order. (Cupach & Metts, 1994; Goffman, 1955).

Face is threatened when an individual’s face wants or needs are violated. These

face wants include the want to be included, the want to have one’s abilities respected, and

the want to be free of others’ impositions (Brown & Levinson, 1987; Lim & Bowers,
1991) and are commonly referred to as positive face (i.e., inclusion and respect) and
negative face (i.e., freedom from imposition). Ting-Toomey (1988) suggested that face
may be based on concern for one’s own face (self-face concern), for the other’s face
(other-face concern), or for both one’s own and the other’S face (mutual-face concern).
This study examines self-face concern in feedback receivers.

Negative feedback in the workplace is likely to threaten two specific face needs:
autonomy and competence needs. When a supervisor gives an employee advice, this may
threaten the employee’s need to be flee from imposition (i.e., autonomy), because the
supervisor is imposing his/her view on the employee. In the same way, other kinds of
messages such as reminding or threatening employees may also threaten their sense of
autonomy in the workplace because the supervisor is imposing his/her will on the
employee. In the case of criticism, the supervisor presents a face threat to the employee’s
need to have his/her abilities respected (i.e., competence). It is also likely that receiving
negative feedback in general presents a threat to competence, as the simple delivery of
negative feedback can suggest that the employee is not seen as performing competently.
Further details on autonomy and competence concerns in the workplace are presented
next.

Autonomy. Lim and Bowers (1991) defined the autonomy need as “the want not to
be imposed upon” (p. 420). The need for autonomy is a fundamental psychological need
(Deci & Ryan, 2000) and is also known as negative face (Brown & Levinson, 1987; Lim
& Bowers, 1991). Employees’ need for autonomy can be threatened by negative feedback

because negative feedback imposes supervisors’ requirements thus constraining

employees’ choice in how to behave. Feedback generally includes information meant to
inform an individual about his/her performance on the job (Geddes, 1993). Thus,
imposition presented by performance feedback involves a supervisor commenting on
how, when, where, or how fast employees should perform the required tasks for their jobs
which limits their sense of free choice. Lim and Bowers (1991) suggested that the
messages most threatening to autonomy are those giving orders or invoking obligation to
a group or organization. Because part of a supervisor’s job is to tell employees what to do
and how to do it, employees frequently deal with this kind of imposition. This loss of
autonomy can be hard on employees, supervisors, and the organization as a whole. For
example, Ellis and Van Dyne (in press) suggested that negative feelings may result fiom
job creep, which is the gradual inclusion of extra-role behaviors into job expectations,
and other types of workplace mistreatment because these impositions affect employees’
sense of freedom and control (i.e., their autonomy).

Previous studies of feedback have looked at autonomy in terms of participation in
performance appraisals. Participation in performance appraisals indicates that employees
are allowed some independence in how they are evaluated. In general, this type of
participation has been associated with positive outcomes. As Fedor, Eder, and Buckley
(1989) stated, “Inviting the subordinate to participate and share control over the
conversation [i.e., performance appraisal] has positive effects on subordinate feedback
reactions” (p. 400). For example, Giles and Mossholder (1990) factor analyzed
employees’ responses to items regarding their supervisors’ behavior during performance
appraisals. One of the major factors extracted was participation of the employee. The

participation factor was positively correlated with job satisfaction, satisfaction with the

supervisor, satisfaction with the performance appraisal session, and satisfaction with the
system. Thus, having the freedom to participate in the performance appraisal session led
to many positive outcomes both for the employee and the organization. With a sample of
nurses, Burke, Weitzel, and Weir (1978) found that the opportunity to present ideas and
feelings (i.e., participation) during an appraisal interview was positively associated with
perceptions of the supervisor as helpful and constructive, the perception that job
problems were cleared up, and the perception that goal-setting was completed. Burke and
Wilcox (1969) also found that for phone operators (information, toll, and assistance
operators), the opportunity to present ideas and feelings was positively associated with
perceptions of the supervisor as helpful and constructive, the perception that job
problems were cleared up and the perception that goal-setting was completed. In sum, the
ability to participate in the performance appraisal process has generally produced positive
outcomes.

Previous feedback studies, such as those discussed above, focused generally on
what would happen if participation was allowed or encouraged in the actual process of
performance appraisals. The same benefits and positive effects Should accrue in the
informal feedback situation as well. Benefits should result if employees are allowed
participation in feedback interactions just as they do when employees participate in
performance appraisals. If, conversely, the fieedom to participate is threatened,
employees might be expected to have negative responses. This question has not yet been
investigated. Specifically, employee responses to threats to their autonomy face have not
been examined. This unknown factor presents an important avenue for study because of

the potential for negative effects of autonomy-threatening feedback messages and

positive effects of autonomy-preserving messages for both the employee and the
organization. Some of the possible effects are discussed next.

The general importance of autonomy to employees is likely to affect how they
might respond to threats to their autonomy. Hodson (1991) found that employees most
frequently cited infringements on autonomy as the reason they were not enthusiastic
about their work. The “infi-ingements” mentioned included things such as overly strict
supervising. Hodson also observed that workers found ways to create autonomy if they
did not feel the organization granted it to them. For example, in order to regain some
autonomy employees may do things that are detrimental to the organization, such as
avoid work or take long breaks. Hodson’s discussion of autonomy was based largely on
observation. Other studies present further evidence on the positive value of feelings of
autonomy in the workplace.

In one study (Lewis & Hayward, 2003), both autonomy and motivation to learn
were increased by affording students the opportunity to choose the type of assignment
they would complete for a class requirement. Parker, Axtell, and Turner (2001) found
positive correlations between job autonomy and both organizational commitment and
supportive supervision. In addition, supportive supervision was also positively correlated
with communication quality. In essence, Parker et al. found a link between supportive
supervisors, high communication quality, and high job autonomy. This suggests the
potential for a well-phrased feedback message to have a positive effect on autonomy
needs. In sum, threats to autonomy can be damaging not only to the individual, but also
to the organization because of their potential to elicit negative behaviors, whereas

allowing individuals to perceive that they have autonomy can have positive results,

particularly in terms of the employee-supervisor relationship. Another important concern
for employees is the need to be seen as competent.

Competence. Lim and Bowers (1991) defined the competence need as “the want
that one’s abilities be respected” (p. 420). Competence itself can be defined as “a
condition or quality of effectiveness, ability, sufficiency, or success” (Elliot & Dweck,
2005, p. 5). Like autonomy, competence is also viewed as a fundamental psychological
need (Deci & Ryan, 2000). Meyer (1975) noted the importance of feeling competent at a
job: “If a person identifies at all with his work and takes pride in what he is doing, he
should think he iS good at it—‘well above average,’ if you will” (p. 45).

Few studies have examined competence or similar constructs in conjunction with
feedback. Some have postulated potential relationships, but not tested them empirically.
Ashford and Cummings (1983) suggested that the motive to learn about one’s own
competence is an important motivator behind feedback-seeking behavior. Geddes (1993)
suggested that performance feedback valence and sensitivity may affect perceptions of
competence, and more specifically that feedback that threatens self perceived competence
Should lead to unfavorable responses from employees such as lowered performance.

Although few feedback studies have empirically examined the variable of
competence, it has been associated with desirable organizational outcomes such as job
satisfaction. Sense of competency was a strong positive predictor of job satisfaction in a
study by James, Treadway, Conner, and Hochwarter (2005). Gottfiedson and Holland
(1990) also found that the ability to use skills at which one was competent (i.e., skills that

were already learned from a previous job or “things I do best”) correlated positively with

expected job satisfaction. Thus, feeling competent in a job is important for job
satisfaction.

While previous studies have examined competence and perceptions of
competence in various ways, the need or desire to be seen as competent has been
overlooked in the feedback literature. The difference between perceptions of one’s own
competence and the need for competence described by face scholars is that being
competent and/or thinking you are competent is different from wanting to be seen as
competent. Lim and Bowers (1991) defined the need for competence as a need to be
respected by others in terms of “intelligence, appearance, and general ability to cope with
the world” (p. 420). Keeping up one’s competence face depends on assessments of
others’ impressions of one’s own competence, or ability level. To date, no previous
studies address this particular need in the context of performance feedback.

Feedback

Feedback can be defined as information received by an individual about his/her
past behavior (Ammons, 1956; Annett, 1969). According to Boume (1966), feedback
generally provides information about the correctness, accuracy, or adequacy of the
behavior. It can be descriptive, prescriptive, evaluative, or comparative (Haeggberg,
2000). Prescriptive feedback provides information on what should be done next;
descriptive feedback, also known as velocity feedback (Kozlowski, Toney, Mullins,
Weissbein, Brown, & Bell, 2001 ), provides information on what the individual has
completed; and comparative feedback, also known as normative feedback (Kozlowski et
al., 2001), compares performance with that of another individual. Evaluative feedback,

also known as labeled feedback (Kozlowski et al., 2001), is feedback that provides

10

information about whether performance has met an unreferenced standard by putting a
positive or negative label on the performance. This type of feedback may include phrases
such as “good enou ” or “not good enough” while not mentioning the standard of
comparison.

The current study focuses on feedback that is evaluative in nature, and that comes
from an external party (i.e., a supervisor). Evaluative feedback is critical to study because
of its corrective potential, and the supervisor as an external party represents the source
most likely to garner cooperation fi'om the employee (Ilgen et al., 1979). In addition, this
study is focused on informal feedback, defined here as unscheduled, interactive,
involving a supervisor—employee pairing, informal language and speech register, and
having an unarranged agenda (based on informal communication definition; Kraut et al.,
1993). Informal feedback is important to examine for two reasons. First, there is a lack of
research on informal feedback. Bartoo and Sias (2004) pointed out that research on day-
to—day communication interactions in organizations has not been conducted and that this
type of work could “provide important insights into a variety of workplace issues and
help explain variation in employee communication behavior” (p. 16). Second, some
scholars (Anderson, 1987; Ilgen et a1, 1979; Lindsley et al., 1995) suggest that feedback
may be “a dish best served hot;” that feedback Should be given soon after any behavior
that needs to be corrected. Lindsley et al. noted “When feedback is delayed... actors may
continue to employ inappropriate strategies or effort” (p. 653). Most formal
organizational feedback typically occurs in performance appraisals, which often occur

too late to be maximally corrective. That is, informal feedback can be useful for

11

correcting behavior immediately, so studying it can help organizations with their
feedback processes.

Employees may also benefit firm the opportunity to improve their work in a more
timely fashion than annually. Informal feedback may open up a dialogue between
supervisors and employees about both problems and solutions as they develop. This may
provide employees with early warning for behaviors which otherwise might lead to more
serious job consequences (i.e., discipline or firing). A “feedback dialogue” may also
allow employees the opportunity to participate in deciding how best to accomplish
individual and organizational goals on a day-to-day basis.

Feedback in organizations is a widely studied topic, evidenced by several studies
aimed at synthesizing the feedback literature (Balcazar, Hopkins, & Suarez, 1986;
Clement & Frandsen, 1976; Gardiner, 1971; Kluger & DeNisi, 1996). As noted by both
Clement and Frandsen (1976) and Gardiner (1971), early feedback research did not use
theory to drive predictions. Despite this, theoretical viewpoints of feedback are evident in
some feedback studies (e.g., Kluger & DeNisi, 1996; Rhodes & Frandsen, 1975; Scheidel
& Crowell, 1966; Stoltz & Tannenbaum, 1963; Stone & Stone, 1985). The main
theoretical viewpoints used to study feedback include control theory (Carver & Scheier,
1981), Operant conditioning (Skinner, 1953), and social learning theory (Bandura &
Cervone, 1983). Further elaboration on these viewpoints follows.

Control Perspective. Some researchers take a control theory or cybernetic
perspective whereby individuals attempt to reconcile discrepancies in “feedback loops.”
Control theory (Carver & Scheier, 1982) focuses on negative feedback loops, which help

a given system reduce discrepancies between the current state and the desired (goal) state.

12

Similarly, in their synthesis of feedback studies in human communication, Clement and
Frandsen (1976) seem to endorse a cybernetic or “system” perspective. They defined
feedback as a process of mutual control whereby each communicator influences and is
influenced by the other. Defined this way, performance feedback focuses on both sender
and receiver and how they influence each other by making a discrepancy evident; this is
the central defining feature of feedback using a control theory perspective. In terms of
performance feedback, control theory is most useful when studying performance
appraisals because prior to appraisals, most supervisors set expectations or goals to be
met. The “discrepancy” focus of control theory makes sense in the context of an
evaluation of whether or not goals and expectations have been met (i.e., a performance
appraisal). However, the current study is focused on informal feedback, and this type of
feedback does not involve a supervisor reporting on clear-cut goals and expectations.
Operant Conditioning Perspective. A second perspective on feedback is based in
Simple Operant conditioning (Skinner, 1953). From this viewpoint, feedback is seen as an
action-reaction sequence whereby individuals cease or repeat behavior based on rewards
and punishments (Clement & Frandsen, 1976). The main principle behind this idea is
Thomdike’s (1927) law of effect, which suggests people will repeat or increase behaviors
associated with positive affect or cognition, and decrease or entirely cease behaviors
associated with negative affect or cognition. As Haeggberg (2000) stated, feedback may
influence performance through its “behavioral reward properties” (p. 53). This viewpoint
suggests that the association of correct behaviors with positive feedback and incorrect
behaviors with negative feedback will naturally cause productive behaviors to be

repeated or increased, and cause non-productive behaviors to be decreased or stopped

13

altogether. The law of effect has been criticized for its Simplicity (Kluger & DeNisi,
1996), which renders it unable to explain some complex results (e.g., positive feedback
can lead to relaxed effort). The findings presented in Kluger and DeNisi’S (1996) meta-
analysis Showed decreases in performance can occur after receiving either positive or
negative feedback. In other words, although the “punishment” of negative feedback
caused performance to decrease, the “rewar ” of positive feedback did not cause people
to increase their performance. Though the current study focuses only on negative
feedback, Kluger and DeNisi’s study shows why Operant conditioning and the law of
effect, while parsimonious, are too simple to explain the aftermath of feedback.

Social Cognitive Perspective. Still other researchers (e. g., Haeggberg, 2000;
Podsakoff & Farh, 1989; Wang & Netemeyer, 2002) view feedback with a social learning
theory perspective. Social learning theory (SLT; or social cognitive theory, SCT;
Bandura, 1991) suggests that feedback provides individuals with information that will
boost or lower their self-efficacy. Self-efficacy is a personal judgment of how well one
can perform the necessary behaviors for situations that may arise in one’s life (Bandura,
2006). Different sources of information can strengthen or weaken self-efficacy. For
example, success is expected to strengthen self-efficacy unless too much success occurs
too easily, in which case people are easily discouraged when confronted with failure.
SLT focuses on feedback obtained through observing others’ behavior. This does not fit
well with study of performance feedback because performance feedback is intentional
communication delivered directly by one party to another. Rather than the employee
watching others and learning from their successes and failures, performance feedback

focuses on intentional, dyadic communication of areas that one party believes the other

14

party, the employee, needs to improve. SLT may be applicable in future studies as a
companion to direct dyadic communication, but is not ideal for examining and making
predictions about communication responses.

These three theories provide useful ways of looking at feedback from different
perspectives. However, what they do not provide is a focus on the communication aspects
of negative feedback. None of the three theories discussed above, and ofien invoked in
feedback studies, examines message characteristics of feedback as a way of explaining
responses. Face (Goffrnan, 1955) affords a framework for examining feedback messages
closely and separating the effects of different message characteristics on employee
responses to feedback. Thus, this is a new theoretical lens through which to view the
provision of negative feedback and the resulting outcomes.

Face Perspective. This study presents a new perspective on feedback, using face
as a framework to guide discussion of message effects on receivers. Receivers’ need to
maintain face in front of a supervisor has not previously been examined as a determinant
of responses to feedback. Based on feedback studies of similar messages, face can
provide additional explanation for receiver responses. Past studies have found that using
different types of delivery for feedback messages has had positive effects on various
responses. For example, Steelrnan and Rutkowski (2004) found that delivery done in a
considerate way resulted in a stronger motivation to act on the feedback content than did
delivery done in an inconsiderate way. Baron (1988, 1990) conducted multiple studies
showing that the use of constructive rather than destructive (i.e., “negative feedback that
is harsh in tone, nonspecific in nature, and focused on internal causes of substandard

performance,” Baron, 1988, p. 199) criticism resulted in reduced anger and tension for

15

feedback receivers. The responses in these studies could be attributed to inconsiderate
and destructive feedback being face threatening, whereas constructive and considerate
feedback was less face threatening. Carson and Cupach (2000) also found that more
severe reproaches (e.g., rebukes) were perceived by receivers as more face threatening
whereas less severe reproaches (e.g., hinting) were perceived as less face threatening.

When individuals receive feedback messages delivered in a way that mitigates
potential face threat, they tend to respond more positively (e. g., less negative emotions,
more motivation to perform better) than when the feedback messages are delivered
without mitigation. The fiamework of face threats and facework messages provided by
the face literature (see Brown & Levinson, 1987; Lim & Bowers, 1991) can help to
organize and direct the study of mitigated delivery of feedback messages and receiver
responses to these messages. These studies (Baron, 1988, 1990; Steelrnan & Rutkowski,
2004; Tata, 2002) provide some empirical evidence that face may be a useful way to
predict how people will respond to negative feedback.

The scope of this particular study is limited to learning more about the receiver
perspective and how receivers will respond to different ways of delivering feedback.
Receivers’ perceptions influence whether or not and how they choose to implement
feedback; because of this, studying receiver responses is critical (Balcazar, Hopkins, &
Suarez, 1985; Prue & F airbank, 1981). Receiver responses may be determined in part by
valence of the feedback message.

Feedback Valence. An important issue in feedback research is the definition of
positive and negative feedback. Positive feedback is defined as information about

meeting or exceeding expected performance, whereas negative feedback is defined as

16

information about performing below expectations. Feedback can include things like
positive and negative behaviors or emotions. Much performance feedback research
follows a similar definition, including an evaluative component and a positive/negative
valence component (Geddes, 1993; Geddes & Baron, 1997; Jacobs, Jacobs, Feldman, &
Cavior, 1973).

The separation of feedback into “positive” and “negative” is used conventionally
by many feedback researchers (Geddes & Baron, 1997; Jacobs et al., 1973; Kluger,
Lewinsohn, & Aiello, 1994; Stoltz & Tannenbaum, 1963). However, this is not just a
facetious division. Various researchers have completed empirical research showing
positive and negative feedback are separate dimensions of feedback and that individuals
respond differently to positive and negative feedback. Geddes (1993) factor analyzed
feedback and found valence (i.e., positive/negative) to be “an easily identifiable
dimension” (p. 204). Geddes and Linnehan (1996) also completed a feedback factor
analysis and found separate dimensions for positive and negative feedback. In their
review and summary of feedback literature, Ilgen et al. (1979) identified valence, or sign,
as one of three dimensions of feedback that affect feedback perceptions. Consistent with
Ilgen et al. (1979), Geddes defined valence “as the message Sign, or simply, positive
versus negative feedback” (p. 204). For ease of understanding and consistency with
previous definitions of positive and negative feedback (Geddes & Baron, 1997;
Haeggberg, 2000; Jacobs et al., 1973; Kluger et al., 1994; Stoltz & Tannenbaum, 1963),
the term “negative feedback” will be used here to refer to instances where the feedback
conveys negative information (e. g., about mistakes made, low scores, low production,

and other types of low performance) and the term “positive feedback” will be used to

17

refer to instances where the feedback imparts positive information (e.g., about a job well
done, high scores, high production, and other types of high performance).

Previous research indicates that negative feedback generates more difficulties
than positive feedback, and this is one reason why it is important to study. People
generally prefer positive to negative feedback (Moreland & Sweeney, 1984). Jablin
(1978) found that subordinates most preferred to receive a confirming (positive in both
content and relational levels) response from their supervisors after (hypothetically)
delivering an unfavorable message. In addition, those receiving negative feedback may
deny or distort the feedback (Ilgen et al., 1979), but receive and recall positive feedback
easily. A study by Lundgren and Rudawsky (1998) found the degree of negativity of
feedback to be positively associated with both negative affect and likelihood of rejecting
the feedback. Jacobs et al. (1973) found negative feedback was consistently rated as less
credible, less desirable, and as having less impact on the receiver than positive feedback.
Halperin, Snyder, Shenkel, and Houson (1976) found that negative feedback was only
accepted fi‘om a high status source, while positive feedback was accepted from any
source. Performance outcomes also differed after receiving positive versus negative
feedback. Two studies (Shrauger & Rosenberg, 1970; Stoltz & Tannenbaum, 1963)
indicated that positive feedback led to improved performance, whereas negative feedback
led to decreased performance. Another study showed that negative feedback led to
significantly lower perceptions of efficacy than positive feedback (Anderson & Rodin,
1989). In addition, delivering negative feedback to employees is not an enjoyable task for

superiors conducting performance appraisals (Antonioni, 1994; London, 2003; Villanova,

18

Bemardin, Dahmus, and Sims, 1993). Because of these difficulties generally associated
with negative feedback, the main focus of this paper will be on negative feedback.

A second reason to study negative feedback is the potential for it to result in
positive outcomes. For example, Podsakoff and Farh (1989) found that negative feedback
made people increase their goals and perform better. In their study, negative feedback
groups outperformed the control group, while positive feedback groups performed
similarly to the control group. Jacobs et al. (1973) found that negative feedback that was
focused on specific behaviors was viewed as having more impact on the receiver than
negative feedback focused on emotions of the observer (e.g., you made me feel angry) or
a combination of behavioral and emotional negative feedback. Anderson and Rodin
(1989) found that when given self-determination cues, participants receiving mild
negative feedback had increased intrinsic motivation. The apt use of negative feedback
can have positive results for both individuals and organizations. As previously
mentioned, one factor that may affect employees’ responses to negative feedback is the
need to maintain face.

Face threats and facework Some situations are inherently face-threatening
because they require the exchange of face-threatening information. Negative feedback
delivery is one of these situations. Expressions regularly used in feedback (e.g., criticism,
advice, reminders, threats, reproaches) may threaten positive and/or negative face.
Previous studies of these types of messages have used different labels for face-
threatening and face-saving messages, such as reproach (Carson & Cupach, 2000) or

destructive/constructive criticism (Baron, 1988, 1990).

19

When the delivery of potentially face-threatening information is required, people
will likely attempt to tone down the face threat while delivering the necessary message
(Lim & Bowers, 1991 ). One potential way to alleviate face threats posed by negative
feedback is to use preventative facework (Cupach & Metts, 1994).

Preventative F acework. Preventative facework is defined as using message
features in anticipation of presenting a face threat, and consists of strategies, which are
general categories of facework designated to address specific types of face needs (i.e.,
autonomy, competence, inclusion), and messages, which are the specific messages used
to carry out a given strategy.

Some types of messages have been successful (Morand, 2000) in softening the
impact of face-threatening acts (F TAs; Brown & Levinson, 1987), including 1) being
conventionally indirect, or asking whether the individual is able to comply, 2) minimizing
imposition, or using phrases that minimize the request, 3) hedging, or using words such
as “perhaps” which soften the request, 4) using formal word choices such as “Sir” to
show deference and 5) using past tense, creating distance in time by saying things like “I
had been wondering if I could. . .”.

Previous studies have examined the use of facework in teacher-student
relationships (Kerssen-Griep, 2001; Kerssen-Griep, Hess, & Trees, 2003), interpersonal
relationships (Carson & Cupach, 2002), influence Situations (Wilson, Aleman, &
Leatham, 1998; Wilson & Kunkel, 2000), and survey items (Holtgraves et al., 1997).
Some findings from those studies indicate that preventative facework messages may have
positive effects. For example, students who received their teachers’ messages with

facework perceived them more positively than messages without facework (Kerssen-

20

Griep, 2001). In a study of surveys, Holtgraves et al. found respondents to be more
honest when answering questions containing facework messages than questions without
facework messages.

Jablin (1978) conducted a study asking subordinates what type of message was
appropriate for a supervisor to use in response to an unfavorable message from a
subordinate. For employees in an “open” communication climate, disconfirming (i.e.,
irrelevant) responses were considered inappropriate, as were repudiating (i.e., negative on
relational and content levels) responses. J ablin also found that subordinates preferred to
receive confirming (i.e., positive on relational and content levels) and disagreeing (i.e.,
positive on relational, negative on content level) responses than acceding (i.e., negative
on relational, positive on content level), repudiating, and disconfirrning responses. In
general, there was a preference for messages with positive relational content, or messages
that allow receivers to maintain face.

Other findings showed that certain types of messages might be more threatening
than others. Wilson and Kunkel (2000) examined the differences between giving advice
and asking for a favor. They found that favor-asking was more threatening to negative
face (i.e., autonomy) than advice-giving and giving advice was more threatening to
positive face (i.e., competence and inclusion combined) than asking for a favor. In
addition, dispositional complaints were found to pose more of a face threat than relational
or behavioral/physical appearance complaints (Carson & Cupach, 2002).

F acework Strategies. According to Lim and Bowers (1991), choice of a facework
strategy is dependent on the type of face that may be threatened. Their framework

suggests that there are three strategies, solidarity, approbation, and tact, which address

21

the three face needs. All strategies and messages are from Lim and Bowers (1991); see
Table l for examples of each message type within strategy. Solidarity addresses threats to
the need for inclusion (i.e., fellowship face). The strategy of solidarity may be carried out
using one of a number of solidarity messages, including fi'iendship reaffirmation,
cooperation, empathy, social acknowledgement, and agreement. Approbation addresses
threats to the need to be seen as competent (i.e., respect for abilities). An approbation
strategy may be carried out using one of a number of messages, including admiration,
support, contradiction, suggestion, and being diminutive about the problem. The third
strategy, tact, addresses the need for autonomy (i.e., freedom from imposition). This
strategy may be carried out using one of a number of tact messages, including imposition
sharing, experimenting, and unconventional indirectness. All three face needs are
fundamental psychological needs (Deci & Ryan, 2000). However, the needs for
autonomy and competence are vital for recipients of feedback by supervisors in the
organizational context (for autonomy, see Burke et al., 1978; Burke & Wilcox, 1969;
Fedor et al., 1989; Hodson, 1991; Parker et al., 2001; for competence, see Geddes, 1993;
James et al., 2005; Meyer, 1975), so the current study focuses on those two needs. When
autonomy and competence face needs are threatened by negative feedback, employees
may respond differently depending on the strength of the threat. Their responses may
involve the use of voice or silence as a way of self-protecting or helping others.
Discussion of voice and silence is provided next to explain how and with what motives
receivers may respond.

Voice and Silence. After receiving negative feedback, employees will respond to

their supervisor in some manner. Two ways they may do this are by voicing ideas or

22

opinions about the feedback (i.e., voice), or by withholding ideas or opinions about the
feedback (i.e., silence). Silence and voice were conceptualized by Van Dyne, Ang, and
Botero (2003) as two separate constructs rather than as opposite poles of one construct.
They define silence as “intentionally withholding ideas,” and voice as “expressing ideas.”

Van Dyne et al. (2003) argued that voice and Silence are not just the presence or
absence of speaking, but involve “the actor’s motivation to withhold versus express ideas,
information, and opinions about work-related improvements” (Van Dyne et al., 2003, p.
1360). Specifically, Van Dyne et al. discussed three motives for employees to be silent or
to use voice, the types of which include prosocial, acquiescent, and defensive voice and
silence. Each of the three motives can result in either voice or silence.

Prosocial motives are based on wanting to help and making a constructive
contribution, acquiescent motives are based on resignation or feeling unable to make a
difference, and defensive motives stern fiom feeling afiaid and personally at risk.
Acquiescent and defensive motives are based in the work of Morrison and Milliken
(2000), who suggested that silence is manifested in organizations because of 1) fear of
negative repercussions or 2) feeling that speaking up about an issue would not make a
difference.

Related to the concepts of silence and voice is Rusbult, Farrell, Rogers, and
Mainous’ (1988) typology of responses to dissatisfaction on the job. This typology
includes four categories: exit, voice, loyalty, and neglect, and is commonly referred to as
“EVLN.” Exit responses consist of “leaving an organization by quitting, transferring,
searching for a different job, or thinking about quitting” (Rusbult et al., 1988, p. 601),

whereas voice “describes actively and constructively trying to improve conditions

23

through discussing problems with a supervisor or co-workers, taking action to solve
problems, suggesting solutions, seeking help from an outside agency like a union, or
whistle-blowing” (p. 601). Loyalty is “passively but optimistically waiting for conditions
to improve—giving public and private support to the organization, waiting and hoping
for improvement, or practicing good citizenship” (Rusbult et al., 1988, p. 601) and
neglect is “passively allowing conditions to deteriorate through reduced interest or effort,
chronic lateness or absences, using company time for personal business or increased error
rate” (Rusbult et al., 1988, p. 601).

The ‘Woice” concept discussed by Van Dyne et al. (2003) builds on the concepts
identified in EVLN research (Rusbult et al., 1988) by identifying added motives. Rusbult
et al. focus on voice as an active and constructive way of helping solve organizational
problems. This viewpoint assumes the employee is always acting constructively and in
the organization’s best interests when using voice. Van Dyne et al. enrich the concept by
suggesting that voice may be motivated in a variety of ways. More specifically, the use of
voice may be motivated not only constructively (prosocial motive), but also by giving up
(acquiescence motive) or saving oneself (defensive motive). For example, an employee
may be motivated to speak up in order to defend his or her performance by pointing out
positive actions he or she has taken. An employee may also choose to speak up to express
agreement with a decision made in the organization not because he/she thinks it will help
or make a difference but because he or she believes giving that particular opinion is
expected. By considering these additional motives, Van Dyne et al. added another

dimension to the study of voice, and to the study of silence, which will be discussed next.

24

Pinder and Harlos (2001) differentiated silence into “Acquiescent” and
“Quiescent” silence, connoting Silence based on passive, accepting resignation
(acquiescent) or fear of both speaking up about a problem and potential consequences of
the change which might result from speaking up (quiescent). In terms of surface
outcomes, these ideas are conceptually similar to Rusbult et al.’s (1988) loyalty, wherein
a worker waits and trusts that something will change to improve the organization’s
situation. That is, the worker stays with the organization and does not use voice. Van
Dyne et al. (2003) discussed a prosocial motive, which may be the most similar to loyalty
because of a more optimistic focus. That is, Van Dyne et al. suggested workers may be
motivated to remain silent for prosocial reasons, or because of a desire to help others
(e.g., the organization or coworkers). Even this prosocial motive is conceptually more
active than the loyalty motive in that prosocial silence involves actively choosing to
withhold information or opinions whereas loyalty focuses on passively waiting and
hoping something will change. The concept of loyalty also focuses solely on optimism as
a motivator and lacks the consideration of fear and/or resignation evident in Pinder and
Harlos’ discussion as well as Van Dyne et al.’s conceptualization of silence. Whereas
Van Dyne et al. do consider a somewhat optimistic focus in prosocial motives, both
conceptualizations focus on fear and self-protection (defensive motive) and resignation
(acquiescence motive) as motivations for silence.

Each of the three motives discussed by Van Dyne et al. (2003) can manifest in
silence or voice, yielding Six communication types including prosocial Silence, prosocial
voice, acquiescent silence, acquiescent voice, defensive silence, and defensive voice (See

Table 2 for a more detailed explanation). Considering the type of feedback studied in the

25

current research effort (evaluative, informal), defensive and prosocial motives are most
relevant. Employees may feel acquiescent afier receiving negative feedback, but given
the parameters of this study (i.e., delimited to a specific feedback incident), it is unlikely
that acquiescence would have time to develop sufficiently to motivate the use of
acquiescent voice or silence. Rather, a single informal incident of negative feedback
should result in more defensive feelings if it is perceived as face threatening or in more
prosocial feelings if it is sufficiently cushioned with preventative facework. Because of
this, for the purpose of face predictions, only defensive and prosocial motives will be
discussed. First, more explanation of the motives, voice and Silence follows.

Prosocial silence involves “intentional and proactive behavior that is primarily
focused on others” (Van Dyne et al, 2003, p. 1368). This is characterized by an employee
with the knowledge of his/her options who decides not to share information based on the
desire to protect others (e.g., fellow coworkers, the organization). Defensive silence is
“intentional and proactive behavior intended to protect the self from external threats”
(Van Dyne et al., p. 1367). This is characterized by an employee who does not share
ideas because he/She is afraid of the personal consequences that may result from speaking
up.

Prosocial voice involves purposeful communication behavior intended to help
others. An employee using prosocial voice would be motivated to help others and would
do so by giving solutions to problems that will help the organization or group. Defensive
voice involves purposeful behavior motivated by self-protection. An employee using

defensive voice might not take personal responsibility for things, attribute outcomes to

26

external factors, shift attention and blame to others, emphasize positive features and
divert attention from the problem and engage in other similar behaviors.

In terms of motives, it is probable that employees experiencing a face threat
would respond defensively. In his essay on defensive communication, Gibb (1961)
elaborated on defensive-supportive communication pairings. A relevant pairing is
evaluation (defensive) and description (supportive). Based on this, evaluative feedback
alone without facework may tend to make receivers feel defensive, and negative
evaluation in particular is likely to evoke defensive receiver responses. Kluger et al.,
(1994) reported on the use of negative feedback as a prompt to elicit negative mood, also
indicating defensiveness may be likely. As previously mentioned, other studies (e. g.,
Baron, 1988, 1990) have shown that people react negatively (e.g., with anger and tension)
to more threatening messages and more positively to less threatening messages. If, as
expected, the addition of facework indeed renders messages less threatening, then
messages without facework would be expected to evoke more negative responses (e. g.,
defensiveness), whereas messages with facework would be expected to evoke more
positive responses (e.g., prosocialness).

Lim and Bowers’ (1991) ideas about face threat and proper strategies to address
each type of face threat can shed light on this discussion. As previously mentioned, they
suggested that a specific strategy addresses each type of face threat most effectively. A
tact strategy should address threats to autonomy most effectively, and an approbation
strategy should address threats to competence most effectively. Lim and Bowers also
examined individuals’ natural attempts at facework (without manipulating the type used)

and found that use of multiple strategies was common. They commented that “the oo-

27

occurrence of different types of facework is not only possible but sometimes required of a
socially competent person” (Lim and Bowers, p. 448). This may indicate that use of
multiple strategies may be more effective than one strategy alone, however well-targeted
it is. Based on these as-yet untested expectations and the expectation of certain responses
to more- or less-threatening messages, different predictions can be made for responses to
autonomy and competence threats.

In general, it is expected that the threat-only condition for both autonomy and
competence will evoke the most defensiveness. However, the level of defensiveness is
expected to decrease differently for each type of threat. It is also expected that the threat
plus tact and approbation condition (i.e., the multiple-strategy condition) for both
autonomy and competence will evoke the most prosocialness. The level of prosocialness
is also expected to vary differently for each type of threat. Thus, the following hypotheses
are proposed.

H1: For autonomy threats, the level of threat will descend in the following order

(fi'om most to least threatening): threat only, threat with approbation, threat with

tact, threat with tact and approbation.

H2: For autonomy threats, defensive voice and silence will be positively

associated with the level of threat posed by the message.

H3: For autonomy threats, prosocial voice and Silence will be negatively

associated with the level of threat posed by the message.

Threats to competence should descend in a slightly different order, according to
Lim and Bowers’ (1991) suggested threat-strategy matches. Based on this ordering, the

relationship between level of threat and both defensiveness and prosocialness Should be

28

different for competence than for autonomy. The following hypotheses are proposed for
competence threats.

H4: For competence threats, the level of threat will descend in the following order

(from most to least threatening): threat only, threat with tact, threat with

approbation, threat with tact and approbation.

H5: For competence threats, defensive voice and silence will be positively

associated with the level of threat posed by the message.

H6: For competence threats, prosocial voice and silence will be negatively

associated with the level of threat posed by the message.

Communication Apprehension An additional factor that may affect receiver
responses to negative feedback is their level of CA. CA is defined as “an individual’s
level of fear or anxiety associated with either real or anticipated communication with
another person or persons” (McCroskey, 1977a, p. 78). McCroskey (1997a) pointed out
that CA has generally been studied as a trait-like variable, but also has been examined as
a state variable. He suggested that CA should be considered as if it were on a continuum
where individuals’ level of CA could be affected by both personal and situational
characteristics. The current study focuses on CA mainly stemming from personality
rather than the situation. It is important to note that studies have shown both personality-
and situation-based CA are amenable to training efforts (Ayres, Hopf, & Will, 2000;
Dwyer, 2000).

Learning more about how individuals’ levels of CA may affect their behavior in
an organizational context can be useful because of the associations between CA levels

and some important organizational variables. For example, Winiecki and Ayres (1999)

29

found that CA was negatively correlated with position in the company and individuals’
salaries. Those high in CA tended to have lower positions and salaries than those low in
CA. In addition, high CAS have been Shown to have lower job satisfaction (Harville,
1992), perhaps due to their difficulty developing and maintaining relationships. Harville
also found that even people with low CA were not highly satisfied with their jobs when
there were low communication requirements, indicating more knowledge of both high
and low CAs would be helpful.

The diagnosis of problems associated with high and/or low CA can be useful to
organizations because of the treatability of both state and trait CA. In a Solomon four-
group study, Ayres et al. (2000) found that both state and trait CA were reduced using a
systematic desensitization intervention (i.e., systematic exposure to uncomfortable
situations over time to decrease the reaction to those Situations). They suggested similar
studies be undertaken to determine the effectiveness of other CA treatments. Also, Dwyer
(2000) found that trait CA was reduced by both a traditional intervention (i.e., a skills-
based public speech class) and a program based on tailoring individual treatments for CA
to individual personality dimensions. Helping high CA individuals become more
comfortable with communicating through some type of intervention may, in turn, lead to
positive individual and organizational outcomes.

Few recent studies have examined CA in the organizational context, although
previous findings indicate that CA could affect individuals and organizations alike. The
propensity for high CA individuals to avoid communication could be damaging in a
feedback Situation where clarification and mutual understanding are important. More

specifically, it may be problematic if high CA individuals tend to use Silence to respond

30

to feedback when it is important to talk with their supervisors to understand the feedback
better and to be able to implement it. If individuals with high CA tend not to use voice, it
will be very difficult to have mutual understanding between a supervisor and an
employee. The responses of low CA individuals to negative feedback may also provide
useful information on how to go about delivering feedback. If individuals with low CA
tend to respond to threatening feedback with silence as well, then threat becomes a much
more important variable. That is, if face threats are so powerful they elicit Silence from
people who like communicating in general, then perceived threats are wide-ranging
problem for organizations. If low CAs respond with voice even when there is high face
threat, then perhaps threat will be less of a concern because it does not seem to alter their
natural pattern (i.e., tending to communicate).

One of the basic theoretical assumptions behind CA (McCroskey, 1977a) is that
those high in CA will withdraw or avoid communication if possible. In a meta-analysis of
CA studies, Allen and Bourhis (1996) found that indeed, “CA is associated with lower
levels of communication behavior (as measured in quality or quantity)” (p. 222). In other
words, the findings across studies Show that individuals with high CA did not
communicate as well or as much as those with low CA. Ayres, Keereetaweep, Chen, &
Edwards (1998) found that low CA’s perform better in interviews than high CA’S. High
CA’S “tried not to think about the impending interview, did not prepare as thoroughly,
and believed that they would not perform well compared to low CA’S” (Ayres et al.,
1998, p. 432). This study also showed that high CA’S talked less and asked fewer
questions than low CA’s. In sum, high CA individuals tend to avoid communication and

communicate less than low CA3.

31

Pinder and Harlos (2001) suggested that individuals low in CA will be more
likely to choose voice, whereas individuals high in CA will be more likely to choose
silence in response to injustice in the workplace, such as negative feedback, which is
likely to be perceived as unjust. When high CAS experience a high level of threat,
because they are experiencing a strong personal threat to either their autonomy or their
competence, their response will be motivated by self-protection, and thus will be
defensive. Also, higher CA’S Should be more likely to choose silence over voice because
of the anxiety and other negative emotions they associate with communicating.

H7: For participants high in CA, there will be a significant positive relationship

between face threat and defensive silence.

Because lower CA individuals do not experience the levels of anxiety and other
negative emotions associated with communicating that higher CAs do, they may be more
able to strategize regarding the use of voice and silence. Ellis and Van Dyne (in press)
suggest that voice could be a way for employees to restore their fi'eedom (i.e., autonomy)
by choosing to make a contribution. Following with this suggestion, lower CA employees
are expected to use voice when their autonomy is most threatened, as a way of regaining
some autonomy. In addition, low CA’S were found to prefer (Daly & McCroskey, 1975)
and be most satisfied when in jobs requiring high levels of communication (Harville,
1992). In general, they like to communicate. Lower CA’S may also find that using voice
in the face of a competence threat presents them with an opportunity to Show off or
highlight their abilities. This may be a way of disproving the negative feedback, and a
way of focusing attention away from their seeming incompetence and onto their areas of

excellence. Also, Similarly to the higher CA’S, the lower CA’S will react to strong threats

32

with efforts to self-protect. In other words, they will be defensively motivated. Thus, the
following hypotheses are proposed across autonomy and competence threat.

H8: For low CA participants, there should be a significant positive relationship

between face threat and defensive voice.

If the negative feedback message is perceived as posing little threat, then the
motive may change from defensive to prosocial because there iS no reason to defend if
little threat exists. Higher CA’S Should still choose silence because of their aversion to
communication in general. However, if the personal threat is greatly reduced, their
motive will change from self-protection to other-protection. For example, they may avoid
mentioning problems in others’ performance to their supervisor rather than avoiding
mentioning their own performance problems. They may also focus on choosing voice or
silence as a way of protecting the organization (e. g., not bringing up problems, or
pointing out problems that could easily be fixed) and thus be prosocially motivated. Thus,
the less threatened high CAS feel, the more likely they will be to choose a prosocial
response.

H9: For participants high in CA, there will be a significant negative relationship

between face threat and prosocial silence.

One study of CA individuals (Harville, 1992) indicated low CA’S were less
satisfied with jobs requiring little communication and were most satisfied with jobs that
required a lot of communication. This may indicate they prefer to verbalize regardless of
the situation. However, in the conditions perceived as less threatening, lower CA’S may

be motivated by prosocial rather than defensive motives. The less threatened they feel,

33

the more likely low CAs will be to choose a prosocial response, and because they are low
CAS, it is likely to be a voice response. The following hypotheses are posed.

H10: For low CA participants, there Should be a significant negative relationship

between face threat and prosocial voice.

F acework Strategy Preference. Concerns about individual differences in
responses to feedback led to the inclusion of a variable of “facework strategy preference.”
This was defined as individuals’ preference for directness/indirectness in delivery of
information (which encompassed both information about performance and general
impressions). The conceptualization and measurement for this were adapted from Merkin
(2006), who measured direct and indirect strategic responses to face threatening
situations. Merkin focused on the use of these strategies in face threatening situations.
Direct responses were “straightforward and candid” (Merkin, p. 142), whereas indirect
responses were “roundabout” (Merkin, p. 142) and divergent. This study frames the
concept in a receiver preference mode. That is, the consideration here was to ascertain
how direct and/or indirect individuals would prefer those delivering face-threatening
messages to be. No other research exists on this concept considered in this way, so
research questions pertinent to other variables of interest are posed.

RQl: How will individual facework strategy preference affect individuals’

perceptions of the level of autonomy and competence face threat?

RQ2: How will individual facework strategy preference affect individuals’ choice

of voice or silence strategy in response to feedback?

Facework strategy preference may have some effect on individuals’ choice of

voice or Silence Simply because they may choose strategies to use themselves that are

34

similar to or different from their own preferences. In addition, some other factors that
may affect individuals’ choice of voice or silence responses might include how
supportive they perceive their supervisor to be (Taylor & Bowers, 1972) and how serious
they perceive the problem posed by the feedback they receive to be.

Supervisor support & feedback severity. Supervisor support is defined as “the
degree to which supervisors value their [employee] contributions and care about their
well-being” (Rhoades & Eisenberger, 2002, p. 700). This could clearly have an effect on
employee voice and silence responses to negative feedback. For example, the perception
that one’s supervisor is very supportive may increase the likelihood of the choice of a
voice response and decrease the tendency for silence responses. Studies of supervisor
support have generally found that perceptions of high supervisor support are positively
associated with positive organizational variables such as personal accomplishment (Cook,
Banks, & Turner, 1993), perceptions of organizational support and subsequently lower
turnover (Shanock & Eisenberger, 2006; Eisenberger, Stinglhamber, Vandenbergl,
Sucharski, & Rhoades, 2002), and job satisfaction (Steinhardt, Dolbier, Gottlieb, &
McCalister, 2003). Jablin (1981) also linked high supervisor supportiveness with more
openness of communication with and higher satisfaction with supervisors. Shanock and
Eisenberger (2006) found that as perceptions of supervisor support increased, employees’
perceptions of support from the organization, and their in-role (i.e., required tasks) and
extra-role (i.e., non-required additional tasks) performance all increased. That is, when
employees felt supported by their supervisor and their organization, they were more
likely to choose to take on extra tasks not required by their job descriptions. Assuming

taking on these extra tasks indicates a propensity of employees who feel supported to go

35

out of their way to help the organization, one active way of “going the extra mile” to help
their organizations might be for them to use prosocial voice. So, this could indicate an
association between perceived support and an increased tendency to use prosocial voice
responses in an effort to help the organization.

Perceptions of supervisor support have also been negatively associated with
undesirable constructs such as depression and hostility (J ones-J ohnson & Johnson, 2001)
and job stress (Steinhardt et al., 2003). The association of supervisor support with
depression and hostility (J ones-Johnson & Johnson) could lead employees to withdraw
and remain silent when perceived supervisor support is low.

In addition to supervisor support, the receiver’s perception of the seriousness of
the feedback may have an effect on how he or She responds. There is no previous
research on receiver perceptions of severity of a feedback message. However, intuitively,
employees might respond differently based on their perception of the severity of the
problem. For example, employees may be more likely to use voice defensively if they
think the problem is very severe, whereas they may speak up with prosocial motives if
they think the problem is less serious for them (i.e., they will not be fired because of it).
Similarly, employees who think the feedback they received indicates a very serious
problem might tend to keep quiet for defensive reasons. If the problem doesn’t seem very
severe, employees rrright keep their comments quiet in an effort not to upset their
manager or hurt the organization by bringing up unnecessary (i.e., small or
inconsequential) problems. In sum, it is unclear exactly how severity of the feedback will
affect responses, so a research question is posed. This research question addresses both

supervisor support and severity of the feedback.

36

RQ3: Will perceived supervisor support or severity of feedback predict

employees’ tendencies to use voice and Silence?

37

Method

Overview

Working adults (N = 443) were solicited to participate in an experimental survey
containing feedback messages delivered with differing levels of face threat. The design
incorporated eight conditions, including four message conditions (face threat message,
face threat message with tact, face threat message with approbation, and face threat
message with approbation and tact). These four message conditions were completely
crossed with two types of face threat: autonomy threat and competence threat. First, each
participant completed a facework preference scale, perceived supervisor supportiveness
scale and the PRCA-Org (Scott, McCroskey, & Sheahan, 1978), and provided detailed
information on their job and their demographic information. Following this, they read one
of the eight message manipulations. After reading the message, participants were asked to
write out how they would respond to the situation, then they completed scales measuring
feedback severity, autonomy face threat, competence face threat, defensive voice,
prosocial voice, defensive Silence, prosocial silence, vignette realism, and past feedback
experience.

Participants. A sample of 443 working adults was obtained. The sample consisted
of 45.1 % males and 52.4 % females (2.5 % did not give their sex) recruited from a
variety of organizations. Participants were solicited through a snowball-type sampling
method. Students in communication classes at a Mid-Atlantic university were asked to
recruit non-student adults who had worked for a minimum of one year at their current
job. Students received class credit for recruiting a maximum of two participants each

(most were allowed to recruit only one participant for credit). The students were given a

38

letter explaining the research and the procedure required to give to participants they
recruited. This method of data collection was chosen because the purpose of the current
research was to study individual responses to the negative feedback Situation rather than
to learn about the context of one specific organization. By gathering a broad sample and
using random assignment to conditions, the current study focuses on how varying types
of individuals respond to this situation rather than how an organizational context affects
individuals who reside within a specific organization.

The participants had a mean age of41.82 years, with the youngest at 19 years old
and the oldest at 66 (12 participants did not give their age). The sample consisted of 92.8
% Caucasians, 1.4 % Hispanics, 1.1 % Afiican Americans, 1.1 % Asian Americans, 0.5
% Native Americans, and 0.7 % of other ethnicities (2.5 % did not indicate an ethnicity).
Participants were asked to indicate their highest level of education attained. The sample
was fairly well-educated, with 0.5 % indicating they had some high school education,
13.5 % indicating they had a high school degree, 17.2 % had some college, 7.7 % were
currently attending college, 5.9 % had an associate’s degree, 33.2 % had a bachelor’s
degree, 0.9 % were currently attending a graduate school, 14.4 % indicated they had a
graduate degree, and 4.1 % indicated “other” rather than choose one of the categories (2.7
% did not indicate their education level). In sum, over 50% had at least an associate’s
degree.

One point one percent of the participants worked at a publisher/newspaper, 9 %
worked at a manufacturing company, 16.3 % worked at an analytical organization
(analyzing/utilizing information; included banking, law, insurance, real estate-associated

organizations), 22.6 % worked at a technical organization (performing technical tasks;

39

included all health-related organizations), 23 % worked at a service organization
(hospitality/restaurant), 7.9 % worked in education-related organizations, and 17.34 %
chose “other” as the type of organization (2.7 % did not indicate an organization type).
Participants were also asked to indicate their Specific job type. Of 428 participants who
indicated a job type, 0.5 % were nonfarrn laborers, 0.5 % were transport workers, 1.2 %
were factory workers, 10.3 % were Skilled laborers, 12.1 % were
clerical/secretarial/support staff, 12.6 % were in sales, 50.9 % were
professional/managerial workers, and 11.9 % indicated they were in other job types.
Participants indicated they had worked a mean of 10.1 years with their current
employer, and that they worked a mean of 42.9 hours per week. Most of the sample
reported receiving feedback with a frequency of at least once a year (91.8 %), whereas
4.3 % reported receiving feedback with some other fi'equency, and 3.8 % did not indicate
their fiequency of feedback. More information on the participants’ experience with
feedback is reported below with the realism check information. Participants reported
having 63.2 % male supervisors, 30.5 % female supervisors, and 6.3 % did not report
their supervisor’s sex. Participants were asked to indicate their organizational status on an
organizational triangle (displayed in Appendix C), which had the numbers one through
five listed from the top of the triangle to the bottom (similar to measurement in Mahoney,
1979). A few of the participants indicated they were at the highest level of the
organization (8.4 %), followed by 16.3 % at the second highest level, 34.1 % at the third
highest level, 23.0 % at the second-to-lowest level, and finally 12.9 % reported being at
the lowest level of the organization. When asked if they currently supervise others, 54 %

indicated that they did supervise others at the time they responded to the survey, whereas

40

43.3 % indicated they did not supervise others (2.7 % did not respond to the question).
Those who indicated they did supervise others were asked how many they supervise. The
number supervised ranged fiom one to 300 employees, with a mean of 20.53 employees
supervised. About 8 % supervised 50 or more employees, and about 92 % supervised
fewer than 50 employees.

Design. The design was a 2 (threat: autonomy, competence) x 4 (type of
facework: none, tact, approbation, tact & approbation) fully crossed design. Threats to
autonomy and competence were manipulated and measured. In addition, facework
strategy preference, supervisor supportiveness, message realism, past feedback
experience, trait CA, feedback severity, defensive voice, prosocial voice, defensive
silence, and prosocial silence were measured.

When conducting research which asks participants to imagine some situation and
their responses to it, one way to improve responses is by creating research “stories” that
include the participant as a main character and portray situations familiar to the
participant (Greenberg & Eskew, 1993). Because of this, the vignettes used asked
participants to imagine themselves as part of the story. Participants were asked to think of
their real job and real supervisor (direct report) before reading the feedback message.
They read a brief introduction describing a situation in which they received feedback
from their real boss on their real job. The introduction reads as follows, “Imagine your
supervisor stops by your work area to talk with you about some recent work you had
performed. In his/her estimation there is a problem in that you are not performing as well
as you should given your ability. He/she provides you with a brief and general statement

and you are left to interpret what was meant.” After reading these instructions,

41

participants read one of eight different feedback messages ostensibly delivered by their
boss.

Procedure. Participants completed the survey online but were offered the option
of procuring a paper copy if they desired. The online survey was offered for the
convenience of the participants and to allow for participation of individuals who
otherwise might not be able to participate because of physical distance from where the
study was being conducted. The paper survey was also offered to ensure that no one was
excluded based on lack of intemet access and also for individuals concerned about
surveillance at work. No participants requested a paper copy. Participants who completed
the web survey were directed to a webpage which displayed the IRB consent form for the
study and a link to the survey at the bottom of the page. The survey link contained a
randomizer which randomly assigned participants to one of the eight conditions in the
study. For logistical reasons, there were eight separate survey versions which were
identical except for the message manipulation. When clicked, the randomizer chose one
of the eight at random and linked to the front page for that survey, thus providing random
assignment to conditions. At the end of the survey, participants were sent to yet another
separate webpage to provide their name and daytime phone number so they could be
contacted if necessary but still retain anonymity on their survey responses. No names
were associated with particular responses.

The survey consisted of initial scales for CA and facework strategy preferences,
along with demographics, work information, and a scale for perceived supervisor
supportiveness, followed by the message manipulation (see Appendix B for complete

messages). After reading the message, participants responded on post-message measures

42

including scales for feedback severity, autonomy threat, competence threat, defensive
voice, prosocial voice, defensive Silence, prosocial silence, vignette realism, and past
feedback experience. See Appendix C for all scales and items.

Feedback Messages. The feedback messages are included in Appendix A. These
messages contain either an autonomy or a competence face threat. Each face threat
message consists of a face threat alone or the face threat plus one of three types of
preventative facework: tact, approbation, or a combination of tact and approbation. The
feedback messages are comprised of combinations of autonomy or competence threat and
one of three types of facework.

The feedback messages were created using combinations of different statements
that were added to a base message. The individual statements were developed through a
brainstorming session with four individuals who were informed of the definitions of the
different types of face threat and facework. These individuals worked alone to come up
with potential statements, and then shared their statements and generated more ideas
based on those statements. Then, the statements generated through brainstorming were
combined by the researcher to create the messages used in this study. More description of
the messages is provided below.

Threats to autonomy specifically focus on the want not to be imposed upon. The
autonomy message threatens this face need in three important ways: by insinuating a loss
of independence, invoking obligation to the organization, and by giving an order. Lim
and Bowers (1991) suggested these are the three statements most threatening to
autonomy needs. Threats to competence focus on the want to be seen as a capable

individual; in this case it is the need to be seen as capable of doing an acceptable job. The

43

competence message threatens this face need in three ways also: through ridicule,
focused disapproval, and blunt disapproval of performance.

In order to combat autonomy face threats, previous research claims that tact
should be used (Lim & Bowers, 1991). Tact messages are designed to provide the
receiver with a boost to autonomy. This is manipulated by the use of statements sharing
responsibility for the problem and incurring debt (to the sender), creating shared
responsibility for the need to redo the work, and putting the supervisor in a position of
owing the employee. To combat competence face threats, previous research claims that
approbation Should be used (Lim & Bowers, 1991). Approbation messages are designed
to Show the sender believes in the receiver’s capabilities. The message with approbation
consists of statements showing support and admiration for the employee, showing that
the employee’s ability is indeed respected. Finally, the message with both tact and
approbation includes all elements fiom both the tact and approbation messages. The
actual messages used in the study included eight different combinations: autonomy threat,
autonomy threat with tact, autonomy threat with approbation, autonomy threat with
tact/approbation, competence threat, competence threat with tact, competence threat with
approbation, and competence threat with tact/approbation. The components of these
messages are in Appendix A and the messages as displayed in each of the eight
conditions are included in Appendix B.

A pretest was conducted to ensure the competence threat manipulation was
effective. The autonomy threat manipulation had already been tested and found effective
in a prior study (Westerman & Park, 2007). Results of the pretest are reported at the

beginning of the results section.

Measurement

Participants completed scales measuring CA, facework strategy preference,
supervisor supportiveness, feedback severity, autonomy face threat, competence face
threat, defensive voice, prosocial voice, defensive silence, prosocial silence, vignette
realism, and past feedback experience. The CA, facework strategy preference, supervisor
supportiveness, message realism, and past feedback experience scales used Likert-type
scaling with 5-point scales. This scaling was based on previous successful use of these
scales with 5-point scaling (see Westerrnan, Park, & Lee, 2007; Bartoo & Sias, 2004;
Harville, 1992). Feedback severity was measured using semantic differentials. Both face
threat scales and all four voice/silence scales used a seven-point Likert-type format. Each
scale is described individually in more detail below. Confirmatory Factor Analysis (CFA)
was performed on all scales to test for unidirnensionality. This process involved checking
for differences between observed inter-item correlations and expected inter-item
correlations. First, expected inter-item correlations were calculated as products of factor
loadings for individual items. Confidence intervals were also calculated for these
expected inter-item correlations. Then, the observed correlations were examined to
determine whether they fell outside of the calculated confidence intervals. A small
number of these observed correlations can be expected to fall outside of the confidence
intervals by chance. If more observed correlations than expected by chance fell outside of
the confidence intervals for a given scale, items with multiple problematic correlations
were dropped from the analysis until a urridimensional solution was found. Retained
items and their factor loadings can be seen in Table 3. Although some items have factor

loadings below .30, these items were not removed because they did not present large

45

errors that needed to be removed from the factor analyses according to the requirements
described above. That is, although their factor loadings were low, these items presented
fewer errors than would be expected by chance. Reliabilities were then calculated based
on the resulting unidirnensional scales.

PRCA-Org. The PRCA-Org scale was designed to measure CA in an
organizational setting. The measure was originally created by Scott et al. (1978), who
used 30 items from three existing scales (PRCA, Verbal Reticence, and Unwillingness to
Communicate) and 20 new items designed to tap into CA displayed in organizations.
Their results showed the new scale (PRCA-Org) was Similar to other measures of trait
CA. The correlation between the PRCA and this new scale was .90. Scott et al. concluded
that the PRCA-Org would provide increased face validity for measuring CA when
dealing with organizational samples. Subsequently, this scale has been used in a variety
of organizational studies (e.g., Bartoo & Sias, 2004; Harville, 1992; Winiecki & Ayres,
1999). It consists of twenty Likert-type items asking about communication in
organizational settings (i.e., People can usually count on me to keep a conversation
going,” and “I enjoy talking to my supervisor”) Scott et al. (1978) found a reliability
(Cronbach’s a) of .91. Thirteen of the twenty items were chosen for face validity and to
reduce the amber of items participants were required to answer. The items were
answered on a five-point Likert-type scale, with 1 being strongly agree and 5 being
strongly disagree. Thus, higher numbers on this scale would indicate higher trait CA.
CFA was performed on the scale, and four items were dropped because too many large
errors were found initially between observed and expected correlations and these four

items had problematic correlations with multiple other items. Nine items were retained in

46

a unidimensional scale that had a reliability (Cronbach’s a) of .86. Dropped and retained
items are indicated in Appendix C.

F acework strategy preference. Facework strategy preference is individuals’
preference for directness/indirectness of delivery of information, and encompassed both
performance and general thoughts about the person. The items for this were adapted from
Merkin (2006), who measured direct and indirect facework strategies in a cross-cultural
setting. The adapted items measured the preference for more direct delivery and for more
indirect delivery. Sample “direct” items include “If something is wrong with my
performance, I would want to be told directly,” and “I like my supervisor to express
his/her thoughts on my ability forthrightly.” Sample “indirect” items include, “I want my
supervisor to express any issues with my performance indirectly,” and “I want others to
deliver their reactions to my performance sensitively.” Factor analysis was performed
separately on each scale. The direct scale that resulted from CFA was one 5-item
unidimensional factor that had a reliability (Cronbach’s o.) of .87. All items were retained
for this scale. The indirect scale was four items that formed a unidimensional factor that
had a reliability (Cronbach’s a) of .66. One item was dropped from the indirect scale
because of problematic correlations with the other items.

Supervisor supportiveness. The supervisor supportiveness scale (Taylor &
Bowers, 1972) was designed to measure participants’ perceptions of how supportive their
supervisor was of them. The scale included three items, such as “My supervisor is willing
to listen to my problems.” Participants responded to these items on a 5-point Likert-type

scale with 1 equal to strongly agree and 5 equal to strongly disagree. CFA was performed

47

and resulted in all three items being retained in a unidimensional factor with a reliability
(Cronbach’s a) of .93.

Feedback severity. This scale measured participants’ perceptions of how serious a
problem was posed by the feedback they received. This scale was created by the
researcher and consisted of four semantic differential items, such as “Trivial..Crucial.”
CFA was performed on the four items and all were retained in a unidimensional factor
with a reliability (Cronbach’s a) of .80

Realism. Participants’ perceptions of the realism of the situation presented in the
survey may depend on their beliefs about whether or not they would ever receive the type
of feedback presented in the survey and whether or not the situation described seems like
something that would really happen in their work life. Both were measured as part of the
manipulation check. These scales were measured on a 5-point Likert-type scale (1 -
strongly disagree to 5-Strongly agree).

Previous feedback experience was measured with four items created by
Westerman and Park (2007). They found a reliability (Cronbach’s a) of .86. These items
include samples such as “I consider this type of feedback normal,” and “I would never
expect to hear this kind of message.” CFA revealed the four items formed a
unidimensional factor that had a reliability (Cronbach’s a) of .80.

Realism of the situation (i.e., vignette realism) was measured with a five-item
measure previously used by Westerman and Park (2007). This scale includes items such
as, “I can imagine being in a Situation like this one,” and “This Situation could happen, or

has happened.” Westerman and Park found this scale had a reliability (Cronbach’s a) of

48

.84. CF A conducted for this study retained all five items as one unidimensional factor
with a reliability (Cronbach’s a) of .75.

Face threat. The current study focused on autonomy face (synonymous with
negative face) and competence face (combined with solidarity, makes up positive face).
For autonomy threats, Westerman and Park (2007) added items to Carson and Cupach’s
(2000) original scale and found their version had a reliability (Cronbach’s a) of .90. The
seven-item scale fi'om Westerman and Park was used to measure autonomy threat. For
this study, CFA was conducted and the seven items were retained in a unidimensional
factor that had a reliability (Cronbach’s o.) of .87.

Threats to competence were measured with two modified items from Carson and
Cupach’s (2000) face threat scale and six original items from a scale previously used by
Westerman and Mieksztyn (2007). The items include, “My boss’s statements showed
he/ she thought I was incompetent at the task,” and “My boss’s statements pointed out
he/she thought I failed miserably.” Westerman and Mieksztyn found this scale to be
unidimensional with a (Cronbach’s a) of .89. CFA was conducted on the current data set
and all eight items were retained in a unidimensional factor with a reliability (Cronbach’s
a) of .89. Both face threat scales are Included in Appendix C.l

Voice. Scales for voice were created using a combination of items adapted from
Van Dyne et al. (2003) and items written Specifically for the feedback Situation to
increase face validity of the scale. The scale for prosocial voice includes 5 items adapted
from Van Dyne et al. and 4 items written by the researcher. The scale for defensive voice
consists of five items adapted from Van Dyne et al. Responses were given on a 7-point

Likert-type scale. These scales were tested for unidirnensionality using CFA. For

49

defensive voice, all five items were retained in a unidimensional factor, and had a
reliability (Cronbach’s a) of .87. For prosocial voice, all nine items were retained in a
unidimensional factor with a reliability (Cronbach’s a) of .93. The scales are included in
Appendix C.

Silence. Similar to the measurement for voice, the silence scales are a
combination of items adapted from Van Dyne et al. (2003) and items written by the
researcher to increase face validity for this study. The prosocial silence scale consists of 5
items adapted fiom Van Dyne et al. and 5 items written by the researcher. The defensive
silence scale includes 5 items adapted from Van Dyne et al. and 4 items written by the
researcher. Responses were given on a 7 -point Likert-type scale. These scales were tested
for unidimensionality using CFA. The defensive silence scale retained nine items in a
unidimensional factor with a reliability (Cronbach’s o.) of .89. For the prosocial silence
scale, four items were dropped in order to form a unidimensional factor with the
remaining six items. This scale had a reliability (Cronbach’s a) of .60.2 Both scales are
included in Appendix C.

Feedback sign. The Sign of the feedback was also checked with a three-item
semantic differential manipulation check adapted from Haeggburg (2001) and one
additional item written by the researcher. These items were tested for unidimensionality
using CFA. Three of the four items formed a unidimensional factor, including poor-good
and low-high. These three items had a reliability (Cronbach’s o.) of .82. The items are
included in Appendix C.

Work information & Demographics. Working adults were asked for more detailed

information on their work situation and some of their own demographics. Some of this

50

information was included as control variables, including organization Size, supervisor’s
sex, hours worked per week, whether or not they supervise others, type of job performed,
organizational status, and type of organization. Participants were also asked to provide
their tenure at the organization, as it may have affected their perceptions of competence
and autonomy threats. In addition, they were asked to share how often they receive
feedback, and whether they receive written or verbal feedback, or written and verbal
feedback. This information helped in establishing the ability of the sample to comment on
feedback experiences. Participants were also asked to share their own age, their sex, their
level of education, and their ethnicity.
Pre-analysis

Work Information & Demographics. To determine whether any control variables
had an effect on the results, tests were conducted on the participant’s age, sex, ethnicity,
and education level, their job and organization type, their organizational status, whether
they were a supervisor or not, their tenure, hours per week, the sex of their supervisor,
and the size of the organization. Table 4 shows the results of a one-way AN OVA with
condition (eight message conditions) as the independent variable and each of the
continuous variables as the dependent variable, and Table 5 shows the results of chi-
square tests for the effects of study condition on the categorical control variables. None
of the control variables were associated significantly with the eight conditions.

Feedback Experience & Realism Check. Two types of realism were tested to
determine participant perceptions of each. These included previous feedback experience

and vignette realism. Each provided information about a different aspect of the feedback

51

experience of the participant and each participant’s perceptions of the hypothetical
Situation.

First, previous feedback experience can give some insight into participants’
previous experience with receiving feedback. This was measured on a five-point Likert-
type scale with one indicating strong disagreement and five indicating strong agreement.
A one-sample t-test (with 3, or the midpoint, as the test value) was performed to
determine whether participants had previously received the kind of feedback delivered in
the study. Participants (M = 2.70, SD = 0.87) had not previously experienced this type of
feedback, t (422) = -7.04, p < .001, n = .11. The interpretation of the items for this scale
could have led people to indicate they had not experienced this type of feedback even if
they had received negative feedback previously. All four items in some way refer to “this
type” of feedback, which could indicate negative feedback, or feedback phrased in this
certain way, or literally this message itself. There is no way to know exactly how
participants interpreted these items. Future studies should use more specific items to
determine exactly what type of feedback participants have previously experienced.

In addition, responses to questions about how often and through which channel(s)
participants received feedback can give some insight into their previous experience with
feedback. Participants reported variously receiving feedback daily (18.3 %), weekly (35
%), monthly (22.1 %), bi-annually (10.3 %), and annually (9.9 %). Four and a half
percent reported receiving feedback at other levels of frequency, such as “as needed,” or
“only when you would do something wrong.” In addition, participants reported how the
feedback was delivered. Two hundred and thirty two participants (52.4 %) reported

receiving feedback verbally, twenty (4.5 %) received written feedback, and one hundred

52

seventy-two (38.8 %) received both verbal and written feedback, whereas one individual
(0.2 %) reported receiving no feedback. Eighteen participants did not respond to the
question about feedback delivery.

Realism of the situation was a direct measure of whether or not participants saw
the described Situation as realistic. This was measured using a Likert-type scale of one to
five, with one being strongly disagree and five being strongly agree. A one-sample t-test
(with 3 as the test value) indicated that participants found the scenario believable (M =
3.27, SD = 0.71), t (422) = 7.73,p < .001, n = .12.

Manipulation Check. Whether participants indeed saw the feedback message they
received as negative was of crucial importance. This was measured with three semantic
differential items on a seven-point scale, with lower values indicating more negative
evaluations. A one-sample t-test with the midpoint (4) as the test value revealed the
feedback was seen as more negative than positive (M = 3.72, SD = 1.54), t (425) = -3.76,
p < .001, n = .03. Based on this, it was concluded that manipulation of the negative

valence in these messages was effective.

53

Results

Overview

A pretest was run prior to the main data collection to determine the effectiveness
of the competence message manipulation. The autonomy message manipulation was
previously tested in Westerman and Park (2007). Following the results of the pretest, the
main data was collected and analyzed. These results follow those of the pretest.
Pretest

The autonomy face threat messages were pretested for their manipulation effects
in a prior study by Westerman and Park (2007). They were shown to differ sufficiently in
their effects on perceptions of autonomy face threat. The competence face threat
messages were pretested specifically as a precursor to this study. All four competence
face threat messages were presented to participants as repeated measures and
counterbalanced. Participants rated the level of face threat after reading each message in
turn. A rating of five was most threatening and a rating of one was least threatening. A
repeated subjects one-way AN OVA revealed the four types were perceived Significantly
different in terms of competence threat, F (1, 31) = 22.03, p < .001. Post-hoe paired t-
tests revealed the mean ratings for the threat-only condition (M = 3.84, SD = 0.66)
differed significantly hour each of the other three messages: approbation (M = 2.75, SD =
0.85), t (31) = -9.53,p < .001, tact (M= 2.92, SD = 0.65), t (31) = 8.12, t < .001 and
combined tact and approbation (M = 2.68, SD = 0.79), t (31) = -10.09, p < .001. As
expected (and intended), the threat level of the messages descended in the following

order (by mean): threat-only was most threatening, followed by tact, followed by

54

approbation, and finally the least threatening message was the combined tact and
approbation message.

Perceptions of face threat were also tested for order effects. The only difference
found was that the threat-only message was perceived as less threatening when presented
first as opposed to any other position (2“, 3rd, 4‘”). As the messages were not presented in
repeated subjects form in the main study, this provides evidence that testing each
message alone provides the most stringent conditions for finding significant results, as
participants’ views of the threat-only message were not potentially inflated by reading
any of the other messages first.

Main Data Analysis

Hypothesis one, which predicted threats would descend in a certain order for
autonomy, was tested with a one—way ANOVA. Hypothesis one was not supported. This
AN OVA revealed that there were not significant differences in the ratings of autonomy
threat due to experimental condition, F (7,421) = .68, p > .05, n2 = .01. Means and
standard deviations for this analysis are provided in Table 6. A Tukey post hoc analysis
revealed no Significant differences between any pairing of two conditions out of the eight.
Subsequently, a planned contrast was run to compare the ratings of autonomy face threat
in the autonomy face threat conditions (all four) to the ratings of autonomy face threat in
the competence face threat conditions (all four). This contrast revealed that the two
groups (autonomy versus competence conditions) were not significantly different from
each other, t (421) = -.09, p = .93, n2 = .0001 . Finally, because the approbation conditions
(across autonomy & competence conditions) appeared to elicit the lowest perception of

autonomy threat, a planned contrast was run to compare the two approbation conditions

55

(i.e., autonomy threat with approbation & competence threat with approbation) against
the other six conditions (i.e., autonomy threat only, with tact, with tact & approbation,
competence threat only, with tact, with tact & approbation). This contrast was significant,
t (421) = 1.94, p < .05, n2 = .01, indicating the two approbation conditions were perceived
as significantly less threatening to participants’ autonomy than the other six conditions
combined.

Hypothesis four, which suggested threats would descend in a certain order for
competence, was tested with a one-way AN OVA. Hypothesis four was not supported.
This analysis Showed significant differences in ratings of competence threat due to
experimental condition, F (7, 420) = 4.31, p < .001, n2 = .07. Means and standard
deviations are provided in Table 6. A Tukey post hoc analysis revealed significant
differences between the competence threat condition and each of the following
conditions: autonomy with tact, autonomy with approbation, autonomy with tact and
approbation, and competence with approbation. These differences are represented in
Table 6 with subscripts. A planned contrast of the autonomy threat conditions and the
competence threat conditions was run to determine whether the two groups were
different. The competence threat ratings were Significantly different fiem the autonomy
threat conditions (all four) versus the competence threat conditions (all four), t (420) = -
3.27, p < .001, n2 = .02. A planned contrast was run to compare among the four
competence threat conditions. This contrast revealed that the two conditions containing
approbation (i.e., approbation and tact & approbation conditions) were significantly

different fi'om the other two conditions, t (206) = -2.73, p < .001, n2 = .03. That is, the

56

approbation conditions were significantly less threatening to participants’ competence
than those without approbation.

Autonomy threat ratings were affected by both types of threat (i.e., autonomy and
competence), but competence threat ratings were only affected by competence threat.
Overall, approbation messages resulted in the least perceived autonomy or competence
threat, whereas tact messages did not reduce either type of threat. Because of this
assessment of the differences across conditions, additional analyses were undertaken to
determine if these two groups (i.e., approbation-only message conditions vs. all other
message conditions) differed significantly in terms of the major variables in the study. T-
tests were run to compare the two approbation-only Conditions with the other six message
conditions on CA, defensive voice, prosocial voice, defensive Silence, direct facework
preference, indirect facework preference, supervisor supportiveness, and feedback
severity. Of these variables, only one variable measured before the message induction,
indirect facework preference, differed significantly between the approbation (M = 2.91 ,
SD = .63) and the non-approbation (M = 2.71, SD = .67) condition groups, t (434) = -
2.80, p < .01, n2 = .02. All pre-analysis variables, including participants’ age, sex,
ethnicity, and education level, their job and organization type, their organizational status,
whether they were a supervisor or not, their tenure, hours per week, the sex of their
supervisor, and the size of the organization, were also submitted to t-tests using the same
groupings and none were significantly different across the two groups. This lack of
differences led to the use of perceived threat rather than comparison of participants based

on condition for the remaining analyses.

57

Hypotheses two and five, which predicted a positive correlation between
defensive responses, both in terms of voice and silence, and the level of autonomy threat
and competence threat, respectively, were tested by correlating each autonomy threat and
competence threat with defensive voice and silence. Correlations are reported in Table 7.
Autonomy face threat was significantly and positively correlated with defensive silence, r
(420) = .10, p < .05, r2 = .01, and negatively, but not Significantly correlated with
defensive voice, r (419) = -.05, p > .05, r2 = .003. Hypothesis two is thus partially
supported. Competence face threat was positively and significantly correlated with
defensive silence, r (419) = .19, p < .01, r2 = .04, but not significantly correlated with
defensive voice, r (418) = .07, p > .05, r2 = .005. Hypothesis five is also partially
supported. Overall, these findings indicate that as face threats of either type increase, the
propensity to use defensive silence, but not defensive voice, increases.

Hypotheses three and six, which predicted a negative correlation between
prosocial voice and silence and the level of autonomy and competence threat,
respectively, were tested with correlations. The correlations are reported in Table 7.
Autonomy face threat was significantly and negatively correlated with prosocial voice, r
(419) = -.19, p < .01, r2 = .04. Thus, hypothesis three is partially supported. Competence
face threat was significantly and negatively correlated with prosocial voice, r (418) = -
.23, p < .01 , r2 = .05, providing partial support for hypothesis six. AS noted above,
prosocial Silence was not tested due to measurement concerns. These findings indicate a
greater propensity to choose prosocial voice when threat is perceived to be low.

Hypotheses seven and eight, which stated that high CAS would have a positive

relationship between face threat and defensive silence, and that low CAS would have a

58

positive relationship between face threat and defensive voice, were tested with
correlations using sub-samples of the highest and lowest CA individuals for both
autonomy and competence face threat. The upper (n = 103, M = 3.21, SD = 0.39) and
lower (n = 119, M = 1.56, SD = 0.27) quartiles of CA were used for this analysis.
Complete correlation tables for the high and low quartiles of CA are provided in Tables 6
and 7.

Hypothesis seven predicted a positive relationship between face threat and
defensive Silence for those high in CA. The upper quartile of CA was analyzed, and a
nonsignificant relationship was found between autonomy threat and defensive silence, r
(99) = .11, n.s., r2 = .01 Also for those high in CA, a positive relationship was found
between competence threat and defensive silence, r (99) = .27, p < .01, r2 = .07.
Hypothesis seven was partially supported because of the predicted positive relationship
between autonomy threat and defensive silence.

Hypothesis eight predicted a positive relationship between face threat and
defensive voice for those low in CA. The lower quartile of CA was analyzed, and a
nonsignificant relationship was found between autonomy threat and defensive voice, r
(113) = -.10, n.s., r2 = .01, and between competence threat and defensive voice, r (112) =
-.09, r2 = .01. Hypothesis eight was not supported, as no significant findings were
evident.

Hypotheses nine and ten, which stated that for those high in CA, there would be a
negative relationship between face threat and prosocial Silence, and that for those with
low CA, there would be a negative relationship between face threat and prosocial voice,

were also tested using the upper and lower quartiles of CA.

59

Hypothesis nine predicted a negative relationship between face threat and
prosocial silence for those high in CA. This hypothesis could not be tested because of the
measurement problems with prosocial silence. Hypothesis ten predicted a significant
negative relationship between face threat and prosocial voice for those low in CA. The
lower quartile of CA was analyzed, and a significant negative relationship was found
between autonomy threat and prosocial voice, r (113) = -.19, p < .05, r2 = .04. A
nonsignificant relationship was found between competence threat and prosocial voice, r
(112) = -.10, n.s., r2 = .01. Thus, hypothesis ten was partially supported because of the
negative correlation between autonomy threat and prosocial voice.

In sum, for low CAS, the association between autonomy threat and prosocial voice
was significant and negative, but the association between competence threat and
prosocial voice was nonsignificant. For high CAS, the correlation between defensive
silence and competence threat was positive and significant, but autonomy threat and
defensive Silence were not significantly related.

Research questions one and two, which asked how facework strategy preferences
would affect perceptions of autonomy and competence face threats and the choice of
voice or Silence, were answered using correlations between preference for indirect and
preference for direct strategies with each type of facework (RQl) and direct and indirect
individual facework strategy with each type of voice/Silence response (RQ2). Prosocial
silence was excluded from the analyses because of the measurement issues discussed
earlier. To answer RQl , the preference for direct facework strategy was not significantly
correlated with either autonomy, r (429) = -.02, p > .05, r2 = .001, or competence face

threat, r (428) = -.07, p > .05, r2 = .005. This indicates that there was no relationship

60

between the preference for direct delivery of feedback and perceptions of either type of
face threat. The preference for indirect facework strategy was significantly and positively
correlated with both autonomy, r (429) = .10, p < .05, r2 = .01, and competence face
threat , r (428) = .12, p < .05, r2 = .01, indicating that the more strongly people preferred
indirect delivery of feedback, the more they felt both their autonomy and competence
were threatened by the feedback message.

To answer RQ2, the preference for direct facework strategies was negatively and
significantly correlated with defensive voice, r (419) = .35, p < .01, r2 = .12, and
defensive Silence, r (420) = -.36, p < .01, r2 = .13. Those who preferred direct facework
strategies were less likely to use either defensive response (i.e., voice or silence). The
preference for direct facework strategies was positively correlated with prosocial voice, r
(419) = .32, p < .01, r2 = .10. That is, individuals who preferred direct strategies for
feedback delivery were also more likely to use prosocial voice as a response to the
feedback in an effort to help others or their organization. The preference for indirect
facework strategies was positively correlated with both defensive voice, r (419) = .20, p <
.01, r2 = .04, and defensive Silence, r (420) = .21, p < .01, r2 = .04. Those who preferred
an indirect strategy for feedback delivery were more likely to choose a defensive strategy,
including both voice and silence. The preference for indirect facework strategies was not
significantly correlated with prosocial voice, r (419) = -.06, p > .05, r2 = .004. In other
words, there was no relationship between the preference for indirect strategies and the
tendency to choose to respond prosocially with voice. Those who preferred more direct
delivery were less likely to use defensive response strategies (voice or silence), and did

not seems to feel threat more or less strongly because of the preference. Those who

61

preferred more indirect delivery were more likely to respond defensively (voice or
silence) and felt threat more strongly the more they preferred indirect delivery.

To test research question three, three regressions were rim with the criterion
variables of defensive voice, prosocial voice, and defensive silence, respectively, and
predictor variables including perceived supervisor support and severity of the feedback.
Supervisor support and feedback severity were centered before entering them into the
equation (of, Cohen, Cohen, West, & Aiken, 2003), and a product term for the two
variables was included to test for interaction effects.

The regression for defensive voice showed that perceived feedback severity (b = -
0.109, p < .05) had a significant negative effect on the tendency to use defensive voice,
but supervisor supportiveness (b = -0.072, p =.144) did not, and the interaction between
the two (b = -0.034, p = .488) was also not significant. The more severe the feedback was
perceived to be, the less likely participants were to use defensive voice. For prosocial
voice, supervisor supportiveness (b = 5.362, p < .001) was the only Significant predictor,
indicating that as supervisor supportiveness increased, so did prosocial voice. Finally, for
defensive silence, there were no significant predictors. All regression results are reported
in Table 10.

Post hoc analyses

F acework Strategy Preference. An unexpected finding emerged from the
exploratory variables of preference for indirect facework strategies and preference for
direct facework strategies. There was a significant, negative correlation between the two,
r (435) = -.36, p < .01, r2 = .13, indicating that they were measuring separate preferences.

These individual differences have not previously been tapped; this provides an indicator

62

that they are separate preferences rather than two ends of a continuum of preference. In
addition, the direct strategy preference was negatively correlated with CA, r (43 5) = -.36,
p < .01, r2 = .13, indicating that those with lower CA, or those more comfortable
communicating in general, had a preference for more direct strategies. Preference for
indirect strategies was correlated positively with CA, r (435) = .27, p < .01, r2 = .07,
indicating that those with higher CA, or those generally uncomfortable communicating,
preferred more indirect strategies.

Voice/Silence Regressions. Although the hypotheses were answered using
correlations, in order to determine the variables which had the most effect on voice and
silence, an additional set of regressions were rtm. Each regression had predictors of CA,
autonomy face threat, competence face threat, direct facework preference, and indirect
facework preference. The criterion variables were defensive voice, prosocial voice, and
defensive silence, respectively. The predictors were centered, then entered into the
equation (of, Cohen, Cohen, West, & Aiken, 2003), and product terms were included to
test for second—order effects. Results for these regressions are reported in Table 11.

For defensive voice, autonomy face threat (b = -0.19) and direct facework
preference (b = -0.27) were Significant negative predictors and competence face threat (b
= 0.16), and CA (b = 0.15) were significant positive predictors. Thus, the more autonomy
threat and the more an individual preferred direct facework, the less likely they were to
use defensive voice. The more competence face threat and the higher their CA level, the
more likely they were to use defensive voice.

This regression had one significant second-order effect; CA interacted with

competence threat to affect defensive voice. This interaction showed the positive

63

relationship between perceived competence threat and defensive voice was strengthened
by the individuals’ CA level, B = 0.163, p < .05. The Significant interaction indicates that
all three slopes are significantly different from each other. However, for individuals with
CA scores at 1 standard deviation below the mean (i.e., low CA), the simple slope was
positive but not significantly different from zero (b = 0.048, p > .25). The higher
individuals’ level of CA, the more the Simple Slope increased (b = 0.188, p > .05 for
moderate CA; b = .328, p < .005 for high CA), but only the high CA slope was
Significantly different from zero. This means that when CA is one standard deviation
above the mean and competence threat is high, the use of defensive voice increased

significantly. The simple Slopes are shown in Figure 1.

....... Mean CA j

Defensive Voice
.5

 

— - Low CA

Competence Threat

 

Figure 1. Simple Slopes for the interaction of CA and competence threat on defensive
voice.

For prosocial voice, CA (b = -0.19) and competence face threat (b = -0.14) were
significant negative predictors and direct facework preference (b = .30) and indirect

facework preference (b = .13) were significant positive predictors. This indicates that

64

when CA and competence face threat were higher, prosocial voice decreased. Prosocial
voice increased with more preference for direct and indirect facework. There were two
significant second order effects. CA interacted with competence threat, and autonomy
and competence threat interacted with each other to affect prosocial voice.

First, CA and competence threat interacted to affect prosocial voice. This showed
the negative relationship between perceived competence threat and prosocial voice was
strengthened by the individuals’ CA level, B = -0.17, p < .05. The significant interaction
indicates that all three slopes are significantly different from each other. The higher CA
was, the stronger the negative relationship between prosocial voice and competence
threat. For individuals with CA scores at 1 standard deviation below the mean (i.e., low
CA), the simple slope was negative but not significantly different from zero (b = -.019, p
> .40). The higher individuals’ level of CA, the stronger the negative Simple slope
became (b = -.l43, p > .05 for moderate CA; b = -.266, p < .01 for high CA), but only the
high CA slope was significantly different from zero. That is, when CA is one standard
deviation above the mean and competence threat is high, the use of prosocial voice

decreased significantly. The simple slopes are shown in Figure 2.

65

 

 

 

%’ 4.5

>

5 4 HIGHCA

g 3.5 ....... MEANCA l
3 ' ---LOWCA l
2.5 - V

N

Competence Threat ;

 

Figure 2. Simple Slopes for the interaction of CA and competence threat on prosocial
voice.

A second Significant interaction was revealed between the two types of threat
(autonomy & competence) on prosocial voice. The negative relationship between
perceived competence threat and the use of prosocial voice becomes more pronounced as
perceived autonomy threat decreases, b = 0.102, p < .05. The significant interaction
indicates that all three slopes are different from each other. In addition, the Simple Slope
for individuals with scores of perceived autonomy threat at 1 standard deviation below
the mean was significantly different from zero, b = -.217, p < .01. AS autonomy threat
increased, the relationship between prosocial voice and competence threat weakened (b =
-.l43, p < .05 for the mean of autonomy threat, and b = -.069, p > .10 for ISD above the
mean). When autonomy face threat is one standard deviation below the mean or when
autonomy face threat is at the mean, and competence threat is high, there is a Significant

decrease in prosocial voice. A graph of the slopes is provided in Figure 3.

66

 

 

 

 

, 6
i
l 5.8
i
5.6 -
0
g 5.4 - “‘
2 M‘ ‘ HIGH AUTONOMY
g 5.2 I \ THREAT
g 5‘ cu; ------- MEANAUTONOMY
THR AT
I 4.8 E
l - - LOWAUTONOMY
i ‘35 THREAT
4.4
Competence Threat
l

 

Figure 3. Simple slopes for the interaction of autonomy and competence threat on
prosocial voice.

For defensive silence, direct facework preference (b = -0.28) was a significant
negative predictor, and CA (b = 0.16) and competence face threat (b = 0.16) were
Significant positive predictors. The more individuals preferred direct facework, the less
they used defensive silence, and the higher their CA and the competence threat, the more
they used defensive Silence. There were two significant second-order effects, with CA
and competence threat interacting and the approbation-no approbation variable
interacting with direct facework preference to affect defensive Silence.

First, the CA-competence threat interaction showed the positive relationship
between perceived competence threat and defensive voice was strengthened by the
individuals’ CA level, B = 0.167, p < .05. The significant interaction indicated that all
three slopes are Significantly different from each other. However, for individuals with CA

scores at 1 standard deviation below the mean (i.e., low CA), the Simple slope was

67

positive but not significantly different from zero (b = 0.017, p > .40). The higher
individuals’ level of CA, the more the simple slope increased (b = 0.139, p > .05 for
moderate CA; b = .261, p < .01 for high CA), and both the moderate and the high CA
Slopes were significantly different fi'om zero. This means that when CA is one standard
deviation above the mean and competence threat is high, there is a Significant increase in

defensive silence. The Simple slopes for this interaction are Shown in Figure 4.

 

—-—HIGH CA

   

....... MEAN CA i

Defensive Silence
5

 

—- - LOW CA

Competence Threat

 

Figure 4. Simple Slopes for the interaction between CA and competence threat on
defensive silence.

Second, the approbation-no approbation interaction showed that for those in the
no-approbation conditions, the higher their preference for direct facework, the less likely
they would be to choose to use defensive silence. Those in the approbation message
conditions did not display a difference in their choice of using defensive silence
depending on their preference for direct facework. Thus direct facework preference was
more influential in the choice to use defensive silence when the feedback message

contained no approbation.

68

 

     

 

4
3.5

3 "T'f‘"
2.5 ----Approbation

2 NoApprobation
1.5 .

3.4 4.04 4.68

Direct Facework Preference l

 

Figure 5. Simple slopes for the interaction between approbation-no approbation and
direct facework preference on defensive silence.
Summary of findings

For autonomy threat ratings, there were no Si gnificant differences among
individual conditions. Further investigation showed the two approbation-only conditions
were significantly lower than the other six conditions in terms of ratings of autonomy
threat. For competence threat ratings, the four competence conditions were Significantly
different fiom the four autonomy conditions. Among the four competence conditions, the
two conditions containing an approbation message (i.e., approbation & combined tact &
approbation) were perceived as significantly less threatening than the competence threat-
only and competence with tact conditions.

Defensive voice was not significantly correlated with autonomy and competence
face threat ratings. Prosocial voice was Significantly and negatively correlated with both
autonomy and competence face threat ratings. Defensive silence was significantly and

positively correlated with autonomy and competence face threat ratings. Prosocial silence

69

was not significantly correlated with either type of threat, but the measurement of this
construct needs to be improved before definitive statements are made.

For high CA participants, defensive silence was positively correlated with
competence threat, but not significantly correlated with autonomy threat. For low CA
participants, defensive voice was not Si gnificantly correlated with either autonomy or
competence threat, and autonomy threat was negatively and significantly correlated with
prosocial voice.

Perceived supervisor supportiveness was a Significant positive predictor of
tendency to use prosocial voice, and feedback severity was a negative predictor of using
defensive voice. In other words, when supervisor supportiveness was high, prosocial
voice was likely to be used. When feedback severity was low, defensive voice was more
likely to be used.

The preference for direct facework strategies was not correlated with ratings of
autonomy or competence face threat. However, the preference for indirect facework
strategies was significantly and positively correlated with ratings of both types of face
threat. The more participants preferred indirect delivery, the stronger they perceived both
types of threats to be.

Preferences for direct facework strategies negatively and significantly correlated
with both defensive silence and defensive voice and positively and significantly
correlated with prosocial voice. Those who had a high preference for direct strategies
were less likely to respond defensively and more likely to respond with prosocial voice.
There was not a significant correlation between direct facework strategy preference and

prosocial silence. The preference for indirect facework strategies was positively and

70

Significantly correlated with defensive Silence and defensive voice but not Significantly
correlated with either prosocial voice or prosocial Silence. Participants who preferred
indirect strategies more were more likely to use defensive strategies in response to
threats.

The preference for indirect and direct facework strategies were negatively and
significantly correlated with each other, indicating two separate variables rather than two
ends of a continuum. In addition, the direct strategy preference correlated negatively with
trait CA and preference for indirect strategies correlated positively with trait CA.

Autonomy face threat and direct facework preference were significant negative
predictors of defensive voice, whereas CA, competence face threat, and indirect facework
preference were significant positive predictors of defensive voice. In addition, the
positive relationship between competence threat and defensive voice was strengthened by
higher levels of CA, such that there was a more pronounced difference in people with
different levels of CA in their choice to use defensive voice when competence threat was
high than when it was low.

Competence face threat and CA were significant negative predictors of prosocial
voice, whereas direct and indirect facework preference were significant positive
predictors of prosocial voice. In addition, the negative relationship between prosocial
voice and competence threat was strengthened by higher levels of CA so that there was a
more pronounced difference in people with different levels of CA in their choice to use
prosocial voice when competence threat was high than when it was low. A second
interaction revealed that the negative relationship between prosocial voice and

competence threat was weakened by higher autonomy threat, such that when competence

71

threat was low, autonomy threat had a stronger effect on the choice of prosocial voice
than when competent threat was high.

Direct facework preference was a Significant negative predictor of defensive
Silence, whereas CA and competence face threat were significant positive predictors of
defensive silence. Also, the positive relationship between defensive silence and
competence threat was strengthened by higher CA such that the difference among
different people with different levels of CA in their choice to use defensive Silence was

greater when competence threat was high than when it was low.

72

Discussion

Summary

This study examined employee responses to different types of negative feedback
messages. Both autonomy and competence threats and their corresponding facework
messages (i.e., tact & approbation) were examined for their effects on employees’
tendencies to respond with defensive voice, prosocial voice, defensive silence, or
prosocial silence. CA was examined for its potential to affect people’s responses; those
with low CA were expected to communicate more fi'eely whereas those with high CA
were expected to be more inhibited in their communication. Two new individual
difference variables were studied for their relationship with the voice and silence
responses: the preference for direct facework strategies and the preference for indirect
facework strategies. Supervisor supportiveness was also investigated as a potential
contributor to employees’ choice of response to negative feedback. It should be noted that
whereas numerous statistically significant findings were presented in the results section
of this study, some of the findings are not large in terms of effect sizes. In light of this
and because this is an initial study, the conclusions drawn below should be considered
tentative.
Explanation/Implications of findings

Face Threat. The message manipulations were not completely effective in
manipulating the two types of threats separately. Two potential explanations for this will
be discussed. One explanation is that the messages worked, but the two types of threat

were somehow entangled in a way unforeseen previously. A second explanation is that

73

there was a problem with the Specific phrasing of the messages (i.e., an artifact) which
caused systematic variance in participant responses.

First, if it is assumed that the messages were effective, but that threats to
autonomy and competence are not necessarily completely separate, then the first problem
is that autonomy threat ratings did not differ for either autonomy versus competence-
threatening conditions. This means the perception of an autonomy threat was evoked by
an actual autonomy threat and by a competence threat. Why might this be? Consider that
people who are not competent are also not given freedom or responsibility (i.e.,
autonomy). This is particularly true in the workplace, where an incompetent performance
by any individual can cause problems for all. Thus, in organizations, those considered
incompetent are watched closely and “kept on a Short leash.” Translated in terms of
feedback for an average employee, if a negative feedback message, such as the one in this
study, indicates that an individual is not seen as competent, could it not be assumed by
the receiver that he/ she was also experiencing a threat to his/her autonomy? That is,
perhaps implying that one is not competent also implies a reduction in fieedom. LMX
theory (Scandura & Graen, 1984) suggests that individuals in high-LMX relationships
with their supervisor are seen as trustworthy, competent, and are given extra
responsibility and freedom because of the relationship. The implication of a lack of
competence may also imply that the employee is entering a lower status relationship
where he/she will not be trusted with freedom or responsibility. This could result in the
lack of difference in autonomy face threat ratings across the two types of conditions

(autonomy & competence). That is, autonomy threat messages threaten autonomy

74

because they were designed to do so, and competence threat messages threaten autonomy
because it is assumed that a loss of competence also means a loss of autonomy.

Despite the lack of difference in autonomy ratings, the competence ratings did
differ depending on the type of threat. This could be because although a competence
threat may imply a loss of autonomy, an autonomy threat may not imply that one is seen
as incompetent. Employees could easily come up with explanations for why, despite their
competence at their job, their supervisor might threaten to take away their freedom. It
could be that one’s supervisor is power-hungry, bossy, a micromanager, or a control
freak. Looking at the actual autonomy threat message, an employee retort along these
lines iS quite clear (Supervisor: “This is not how I would have done it.” Employee: “It
may not be how you would have done it, but that doesn’t mean it isn’t right”). Also, a
threat to autonomy could be seen as rote, or part of the supervisor’s legitimate power
(French & Raven, 1959); thus it would not be threatening to one’s sense of competence
because it is Simply part of the supervisor’s job. When a supervisor threatens autonomy,
it does not inherently imply the employee is incompetent, or implicate that a loss of
competence is imminent as a result of the loss of autonomy. These explanations leave
room for rating competence threat differently when an autonomy versus a competence
threat message is posed.

A second explanation for the failure of the message manipulations to elicit the
expected threat is that there was something in the specific messages that caused
participants to respond contrarily to how they were expected to respond. A close
examination of the messages reveals that the competence threat messages started with

“You screwed this up” and the autonomy threat messages did not start with the same

75

personal and accusatory tone. In fact, the autonomy threat was centered on the sender by
saying “This is not the way that I would have done it.” The difference in the tone of the
messages could have caused a difference in how individuals responded to them. More
specifically, because the autonomy threat message started out by focusing on the
supervisor’s preference, the employee may have been more inclined or more able to
blame the problem on the supervisor. However, because the competence threat message
focuses on “You,” it very clearly indicates that the problem is indeed the employee’s
fault, so avoiding that threat by blaming the supervisor may not have been possible or
may have been more difficult than for the autonomy threat message.

AS for threat levels of the different messages, approbation was clearly the strategy
that reduced face threat the most, across all conditions and for both types of threat. The
autonomy threat ratings were the lowest for the approbation-only condition across
conditions for both types of threat. The competence threat ratings were the lowest for the
approbation-only condition and approbation plus tact conditions across conditions for
both types of threat. In other words, no matter what type of threat was provoked, the
message most effective in reducing that threat was the one that made clear that even
though there was a problem with the receiver’s performance, the supervisor still believed
the person had the ability to do the job as desired. This indicates Lim and Bowers (1991)
may be correct about approbation addressing competence threats, but it may reach further
than competence threats alone.

Tact, on the other hand, was not a highly effective facework tactic in terms of
threat reduction. Ratings of autonomy and competence for the tact message conditions

did not differ fiom the threat-only messages for autonomy or competence. One element

76

of the tact message was to share some of the responsibility for the problem. It could be
that sharing responsibility would hold more sway if done in front of a third party, such as
the supervisor’s boss, or if the supervisor indicates he/ she would let his/her supervisor
know the responsibility for the error was shared. However, these findings indicate Lim
and Bowers’ (1991) assumption that tact addresses autonomy threats may not hold for
this sample and perhaps for other organizational workers.

In sum, Lim and Bowers’ (1991) approbation appears to be the more effective
feedback strategy of the two tested here. This study provides some evidence supporting
the idea that approbation addresses a competence threat. However, the tact strategy does
not appear to have an effect on autonomy threats or competence threats, and even appears
to hinder the effectiveness of approbation messages, as autonomy threat ratings were
higher for the combined conditions than for the approbation conditions alone.

These findings indicate the need for further research into the effectiveness of the
different types of facework that address different types of face threat. Research should
address the separation of autonomy and competence face threats; specifically of interest
is whether they can be separated in an organizational context. Practitioners can take a
clear and safe message fiom the findings: use approbation in negative feedback
messages! Whereas this is only one study, so the findings Should be interpreted with
caution, it seems unlikely that the use of approbation messages could be harmful to
employees as they provide generally confirming and supportive messages. Thus,
practitioners can feel relatively secure in implementing the use of this type of message.

Defensive Voice. The tendency to use defensive voice did not appear to be related

to ratings of either autonomy or competence threat. That is, these results indicate that if

77

people are going to use defensive voice, they will choose that response regardless of how
threatening or nonthreatening a feedback message may be. It could be that some people
tend to respond defensively in general.

However, defensive voice was positively related to CA. This indicates that those
who are highly apprehensive about communicating in general have a greater tendency to
use defensive voice, regardless of the situation. Because of their anxiety about
communicating itself, high CAS may have felt defensive simply because they were asked
to communicate in response to a negative feedback message. This could create higher
tendencies to use defensive voice in response to negative feedback. It could also be that
because high CAS generally avoid communicating, they lack skill in communicating and
may not realize that using defensive voice might have negative outcomes for them.

Taken together, these findings have implications for both researchers and
practitioners. First, the findings provide an additional confirmation for CA researchers by
indicating yet another negative association in organizations for those with high CA.
Previous research indicated those high in CA are at a disadvantage in various areas. For
example, high CAS have been demonstrated to have lower job satisfaction (Harville,
1992) and to perform more poorly in interviews than low CAS (Ayres et al., 1998). A
tendency to use defensive voice could be one reason why it is difficult for those high in
trait CA to obtain jobs through interviewing and to attain job satisfaction.

Practitioners may want to consider the level of CA of current employees to
determine the best course of action when dealing with different individuals. Previous CA
research (Ayres et al., 2000; Dwyer, 2000) indicates that training can reduce the level of

both state and trait CA for some individuals, so organizations may want to consider

78

implementing some CA training to help employees who are high in CA deal with
receiving and responding to negative feedback. Ayres et al.’s study would suggest
systematic desensitization would be effective in reducing trait CA. Dwyer would suggest
a traditional skills-based training program would work, or training based on personality
dimensions could be effective. Alternatively, organizations might provide training for
managers in dealing with defensive voice responses when they arise.

In addition to the correlational findings for CA and defensive voice, another effect
emerged which indicated that when competence threat was low, the choice of defensive
voice was similar for individuals low, moderate, and high in CA, but when competence
threat was high, high CAS were most likely to use defensive voice, followed by moderate
CAS and low CAS. In other words, the positive relationship between perceived
competence threat and defensive voice became more pronounced for individuals with
high CA. Previous studies (MacIntyre, 1994; McCroskey, 1977b) suggest that CA is
negatively correlated with self-esteem; that is, high CA individuals tend to have lower
self-esteem. This may be one reason why high CAS seem to be more strongly affected by
competence threat than do low CAS. The high CAS reacted very strongly by choosing
defensiveness more when threat was high, even choosing to use a response that would be
considered out of character for them normally—voice. Perhaps the low self-esteem of
high CA individuals caused them to feel even more defensive when their competence was
threatened, and to take the extreme measure of using voice, albeit defensive voice.

Defensive voice was negatively related to prosocial voice, indicating those who
choose to use defensive voice will be less likely to use prosocial voice. This may be an

important finding for organizations to consider, as prosocial voice is probably the most

79

functional response to negative feedback of the four responses examined here. Hagedoom
et al. (1999) suggest voice is a more functional response than silence, as it can provide an
early alert for the organization of problems that may exist and may also relieve
employees’ distress over the problem. Hirschman (1970) would probably consider exit to
be an alternate and somewhat functional type of voice, because by exiting, people are
making a statement and attempting to have an effect on the organization. However, exit
was not examined here. Of the two voice responses used in this study, prosocial voice is
preferable to defensive voice because it is constructive, making an effort to help either
the organization or a fellow coworker, whereas defensive voice is more selfishly
motivated and may prove damaging to the organization and the individual.

Defensive voice was also negatively related to direct facework preference and
positively related to indirect facework preference. A higher preference for direct
facework indicated less use of defensive voice, and a higher preference for indirect
delivery indicated more use of defensive voice. This may suggest that people who want
information delivered in a clear, uncensored way are less likely to respond in a defensive
manner, whereas those who prefer information delivered in a more sensitive manner are
more likely to respond defensively. Organizations may want to consider determining their
employees’ preferences and training managers in delivery styles that accommodate those
preferences.

Feedback severity was found to be a negative predictor of defensive voice. The
more serious the feedback was perceived to be, the less likely individuals were to use
defensive voice. This can be explained by simple fear of repercussions. Perhaps using

defensive voice is perceived as implicating oneself in the problem at hand. When the

80

problem is not very serious, the feedback may seem unjust, spurring more use of
defensive voice to protest the unfairness of the negative feedback and explain away the
problem behavior(s). Baron (1990) found that destructive criticism was viewed as less
fair than constructive criticism, indicating a view of negative feedback as unjust is likely.
Baron (1988) also suggested that destructive criticism, which is “negative feedback that
is harsh in tone, nonspecific in nature, and focused on internal causes of substandard
performance” (p. 199), can lead to emotional responses such as anger or resentment, and
render the receiver more contentious. In particular, “nonspecific” feedback may be likely
to be viewed as unjustified and may Spur a defensive response.

Prosocial Voice. The tendency to use prosocial voice was negatively correlated
with ratings of both autonomy and competence face threat. This indicates those who
perceived a greater threat were less likely to attempt to help the organization or their
coworkers by using prosocial voice. This further indicates that the use of facework
messages to soften negative feedback may help in leading to the positive and functional
response of prosocial voice. The importance of delivering negative feedback (i.e., to
improve performance and help employees avoid consequences) combined with the
knowledge that reduced threat leads to a more functional response, Should lead
organizations to believe that reducing threat in feedback messages is a valuable pursuit.
Managers should be taught how to deliver negative feedback in less threatening ways in
order to elicit more desirable responses (i.e., prosocial voice). For researchers, this
indicates future study of facework strategies and their use in organizations is worthwhile.

Tendency to use prosocial voice was also negatively correlated with CA,

indicating that lower apprehension to communicating meant participants were more likely

81

to try to use voice to assist either their organization or a coworker. That is, those who are
generally more comfortable communicating are more likely to use prosocial voice. For
practitioners, these results strengthen the need for treatments to reduce CA for employees
high in CA and for training for managers in dealing with different levels of CA in their
employees. For researchers, this provides more evidence that low CA can be an
advantageous characteristic in employees.

An effect emerged for prosocial voice similar to that for defensive voice. The
level of CA strengthened the negative relationship between perceived competence threat
and prosocial voice so that those higher in CA were the least likely to choose prosocial
voice when competence threat was high, followed by those with a moderate level of CA
and finally those with a low level of CA. The low CAS were just as likely to choose
prosocial voice whether they perceived a competence threat or not. This finding points to
high CA individuals being more sensitive about competence than low CA individuals. As
noted above, this could be a result of the negative correlation found to exist between CA
and self-esteem in other studies (MacIntyre, 1994; McCroskey, 1977b). That is, higher
CAS reacted more strongly to higher competence threat by purposely not using prosocial
voice, whereas low CAS did not evidence much of a difference in their use of prosocial
voice based on the level of perceived competence threat.

The level of perceived autonomy threat also had an effect on the relationship
between perceived competence threat and the use of prosocial voice. The presence of
autonomy threat weakened the relationship of competence threat with prosocial voice.
When autonomy threat was low, there was a strong negative relationship between

competence threat and the use of prosocial voice, but when autonomy threat was high, the

82

negative relationship between competence threat and prosocial voice was not as strong.
This indicates the perception of each type of threat did not operate independently,
although only one type of threat was presented at a time in the messages. This could
mean that competence threat has a stronger effect if autonomy threat is low—that is, an
autonomy threat is more important than a competence threat. Because this interaction is
relatively isolated in that it did not reach significance in defensive voice or defensive
Silence, further investigation Should be conducted to determine if this relationship will
hold in different studies.

Prosocial voice was also positively associated with direct facework preference but
not associated with indirect facework preference. Those individuals who prefer
information to be delivered directly tend to be more likely to use prosocial voice. The
preference for receiving information directly may be indicative of individuals’ preferred
style of delivery for when they themselves send messages. That is, perhaps participants’
stated preferences for receiving information are the same as their preferred methods of
sending information. If this is the case, then those who prefer direct delivery may
naturally feel comfortable delivering “advice” to a supervisor who has just delivered
negative feedback. Perhaps they expect the supervisor to be prepared for the same
delivery style in retum—if the negative feedback is delivered bluntly and directly by the
supervisor, then the employee can also deliver helpful (but also potentially negative)
feedback to the supervisor in a blunt manner. Jablin (1978) found that employees
predicted that subordinates would respond to supervisors with the same message they
originally sent. That is, subordinates expected they would respond to supervisors’

repudiation message most often with repudiation, and they would respond to supervisors’

83

confirmation message with confirmation messages of their own. That these employees
would choose the same message type as their supervisors may also indicate they would
choose a similar delivery style to respond to their supervisors.

Supervisor supportiveness was found to be a positive predictor of prosocial voice.
Feeling that their supervisor was supportive led employees to want to help the
organization and/or their coworkers in an active way. This seems a natural finding for
supervisor supportiveness, given the association between perceived supervisor support
and perceived organizational support found by Rhoades and Eisenberger (2002). If an
employee feels supported by his/her supervisor and organization, he/She will want to help
them.

Defensive Silence. The tendency to use defensive silence was positively correlated
with ratings of both autonomy and competence face threat. This makes sense intuitively,
as those who felt more threatened tended to be more defensively Silent. One way of
dealing with negative feedback might be to avoid giving more evidence of one’s own
problematic performance and to avoid appearing to agree with the negative feedback by
remaining silent. Defensive silence could also be a coping mechanism participants
wished to employ while absorbing the negative feedback they had just received.
Defensive silence may be seen as functional for the organization in as much as it is
similar to Rusbult et al.’s (1988) loyalty, which is defined as passively and optimistically
staying with the organization while waiting for the situation to improve. For example,
defensive silence may be functional if it keeps a valuable employee from leaving the
organization and the problem can be fixed through another route. However, the types of

Silence (defensive and prosocial) used in this study are not really the same as loyalty,

84

especially with the motivations attached to them. Further discussion of loyalty related to
prosocial silence is provided in the limitations section.

Defensive Silence was also positively related to CA. This finding parallels the
finding for defensive voice. Perhaps high CA individuals are also generally more nervous
and thus tend to be more defensive on the whole because of their anxiety about
communicating. To these individuals, defensive motives would be natural because they
are receivers in a communication transaction, and silence would be natural because of
their apprehension about commrmicating further.

The relationship between defensive silence and competence threat was different
for individuals with different levels of CA. Those with high CA responded to higher
competence threat with more defensive silence, whereas those with low CA did not
change their defensive silence responses based on competence threat. This finding
parallels the finding for defensive voice. Assuming that high CA individuals generally
have low self-esteem (see MacIntyre, 1994; McCroskey, 1977b), it may also be true that
high CA individuals responded with more defensive silence to competence threats
because of their low self-esteem, whereas low CAS, who presumably did not have self-
esteem problems, were not affected by competence threat in terms of their choice to use
defensive silence.

Defensive silence, like defensive voice, was negatively correlated with direct
facework preference and positively correlated with indirect facework preference. The
tendency to be defensive seems to hold across voice and Silence in terms of these
relationships. In general, then, those who prefer direct delivery tend to be less defensive

overall, and those who prefer indirect delivery tend to be more defensive overall. This

85

could be because people who like to hear information directly are more prepared for
unsettling feedback and are thus less likely to react defensively than those who prefer
more sensitivity in message delivery.

Prosocial Silence. Unfortunately, the measurement for prosocial silence in the
present study was not sufficiently reliable to draw conclusions about this particular
variable. Prosocial Silence presents a tricky measurement situation, as the researcher must
attempt to measure what is essentially a non-response—being silent. Being silent to help
others may be particularly difficult for participants to understand unless they possess
some information which needs to be kept secret. Employees without “secret” information
may remain Silent simply because they are hoping things will get better, as Rusbult et al.
(1988) suggested with their loyalty construct, or because they think speaking up will not
make a difference (i.e., acquiescence). It may be worth measuring a silence construct
motivated by optimism in addition to prosocial, defensive, or acquiescent motives (the
latter of which was not measured in this study). Creative thinking may be needed to
discover a way of measuring this tendency accurately and getting at the reasons why
individuals remain silent in organizations.

Implications for Practitioners

Practitioners can reap useful information fiom the current study. First, because
CA appears to be important for certain outcomes in organizations, such as
communication responses, practitioners may want to consider providing training for their
managerial staff in how to deal with employees with different levels of CA. It may also
be worthwhile to consider providing training to employees with high CA in an effort to

help them be more comfortable communicating. This training could take the form of

86

systematic desensitization (per Ayres et al., 2000) or Skills-based training (per Dwyer,
2000.

Second, organizations may want to consider determining their employees’
preferences for direct and indirect facework and helping managers to develop delivery
styles that meet those needs for their employees. Finally, the use of approbation message
to reduce threat seems to be a good idea for organizations to implement. Face threat
seems to lead to undesirable outcomes, so reducing threat using approbation messages is
a good option for practitioners in organizations to consider.

Implications for Researchers

In general, researchers can take a few key points away from the current study.
First, CA seems an important variable to consider in organizations. Employees’ CA
seems to make a difference in how they respond to negative feedback. Also, the findings
here corroborate other findings that high CA can be problematic in organizations. Thus,
further investigation of CA is important and worthwhile.

Second, face threat was shown to have an effect on employee responses, and at
least one type of facework was shown to be effective in reducing the perception of face
threat. As Lim and Bowers’ (1991) ideas have not been investigated in other studies,
these findings provide a first step toward determining whether their thinking was accurate
and/or how it should be adjusted in the future. The current findings indicate there are
nuances to investigate in the study of face threat and preventative facework.

Finally, the voice and Silence results also provide evidence that the area of voice
and silence is worth investigating further. The problematic measurement of prosocial

voice indicates there is room for improvement, and the interesting findings for defensive

87

voice, prosocial voice, and defensive silence suggest future studies will be enlightening
and useful in developing theory about communication responses in organizational
settings.

Limitations

A few limitations are evident after completion of the study, including problems
with the induction of threat, the measurement of prosocial silence, the sampling method,
the use of self-report and the use of responses based on intentions. First, the threat
manipulations did not work exactly as planned. Although the threat messages had been
used in previous studies and pretested, they were not perceived as expected. Fortunately,
measures of perceived threat were able to stand in for the manipulated conditions in the
analyses. Before another study is done, these messages will need to be adjusted and
pretested together to ensure they are sufficiently different from each other. Alternatively,
methodology could be changed to ask for reports of feedback messages rather than
manipulating the type of message participants receive.

A second issue is that of the measurement of prosocial silence. This issue can be
considered on a conceptual or a measurement level. For this study, measurement was the
problem. Asking people whether they would keep quiet about some unknown company
or coworker secret may not be effective unless individuals actually have secrets to keep.
In future studies it may be useful to ask participants whether they possess information
that could be damaging to the company or to a coworker as a way of differentiating
meaningful from meaningless answers. Alternatively, adjusting the writing of the items to
focus less on keeping information quiet may also improve the measurement of prosocial

silence.

88

Future researchers in the area may want to consider a more macro issue: whether
prosocial silence is problematic at a conceptual level. Is there a better way of conceiving
of prosocial silence? The examples and items from Van Dyne et al. (2003) focus on
keeping private information quiet. Perhaps there is a conceptualization that would apply
across all organizational citizens rather than only to those with privileged information. It
may be that the conceptualization of prosocial silence should be more similar to Rusbult
et al’s (1988) loyalty, comprising more focus on Optimism and hope for the future of the
organization as reasons for keeping quiet. This conceptualization could get at the motive
of wanting to help, but have a wider focus than just keeping secrets.

The sampling method used here may have been problematic. Although the data
were tested for systematic differences across conditions due to a variety of demographic
and work variables (e. g., supervisor gender, tenure at the organization), and no
differences were found, it is possible there were other unknown variables creating
unmeasured systematic variance. This problem should be remedied in the future by
collecting a sample from one organization for comparison with the current sample.

Using self-report scales can be considered a limitation. Self-report based on a
manipulated message can only approximate how people would respond if put in a similar
situation in real life. This problem was combated by using a sample of working adults
who would be likely to have had feedback experiences similar to the one described in the
survey. Observing individuals’ responses to real feedback messages in their real
workplaces fi'om their real supervisors may yield richer, and potentially different
findings. However, using working participants should yield a fairly accurate picture of

how people might respond in a feedback situation.

89

Related to the issue of self-report is the issue of intentions. This study did not
measure behavior, but measured intentions to behave in certain ways. The Theory of
Reasoned Action (TRA, Ajzen & Fishbein, 1980) suggests that behavioral intentions lead
to behavior consistent with the intentions. Kim and Hunter (1993) conducted a meta-
analysis Showing a correlation of .82 between behavioral intentions and behavior,
indicating that intentions are indeed likely to lead to the corresponding behavior. Thus,
intentions are not a perfect way, but a reasonably accurate way of studying constructs of
interest.

Future Directions

The current study provides rich potential for future studies in the area of feedback
in organizations. Feedback is crucial to the success of both organizations and their
employees because it allows both parties the benefits of improved performance and helps
both to avoid negative consequences of low performance. This future research can be
grouped into four areas: message issues, individual difference variables, supervisor-
subordinate interaction, and communication responses.

Message issues. The order of delivery might make a difference (Asch, 1946;
Rosnow, 1966). It could be that the messages were viewed similarly because the threat
came before facework in each message. This would explain the lack of significant
differences among autonomy conditions for ratings of autonomy face threat. That is,
people may have been so offended by the first statement that the facework statements,
which came afterward, did not have a chance to “soften the blow.” Delivering the face-
threatening statement first, followed by a face-saving message may have a different effect

than delivering the face-saving message first, followed by a face-threatening message.

90

Conventional wisdom advocates the “sandwich” method (i.e., positive-negative-positive;
Dohrenwend, 2002); perhaps putting the face threat between two face-saving statements
may be the most effective. This should be empirically investigated in future studies.

Second, future research should also investigate the effectiveness of various face-
saving tactics on the delivery of negative information. Delivery of negative information
can extend beyond negative feedback to other contexts, such as the delivery of terminal
disease diagnoses in hospitals. This research should investigate Lim and Bowers’ (1991)
claims regarding which types of facework address which types of face threats to
determine whether the claims are valid. The current study provides some evidence to
answer this question, but more should be gathered in future studies. In addition, future
investigations of the types of messages which are most successful (approbation in
particular) may be informed by the social support literature.

Future research may also attempt to address the issues of message “range.” That
is, the current study only investigated one specific message, whereas there is a range of
messages that might be considered threatening to either autonomy or competence threat,
and a range of messages that might provide relief from each type of threat. A different
methodology may be required; asking people in organizations what type of messages they
receive could help to create a sense of what the range of potential feedback messages in
real organizations may be. Future studies could then test this “bank” of messages,
enabling researchers to provide practitioners with a broader range to use in feedback
delivery.

Individual difi’erence variables. Future studies should investigate individual

difference variables as a way of explaining differences in employee responses to

91

messages they receive at work. This study found participants had different levels of
preference for direct and indirect facework, and these corresponded differently with their
responses to negative feedback. CA also related differently to the types of voice and
Silence. These are just a few of a variety of individual difference variables that could
affect responses to negative feedback in organizations. For example, leader-member
exchange (LMX), need for feedback (Miller, Johnson, Hart, & Peterson, 1999), self-
esteem and the big five personality traits could all have effects on employee voice/silence
responses. These individual difference variables Should be tested in future studies.

Supervisor-subordinate interaction. Future studies might also investigate other
aspects of the interaction between supervisors and subordinates. The current study
focuses on the subordinate as a receiver and what the receiver does after receiving a
feedback message. A second question to consider in terms of supervisor-subordinate
interactions is whether the findings of this study would hold if feedback was being
delivered upward—if the negative feedback message was going from an employee to a
supervisor. More Specifically, what types of voice/ silence responses would supervisors
choose to use in response to negative feedback? The combination of the reactions from
both the subordinates and the supervisors could identify destructive communication
patterns in feedback interactions and eventually help to defuse those situations by
learning what types of messages lead to what types of responses. Supervisor
supportiveness may become an even more important variable in future studies of this
interaction.

Communication responses. Finally, future investigation into voice and Silence

responses is important. The current results are just a first look into these responses.

92

Future studies should investigate what types of messages elicit which types of voice and
silence and should also investigate consequents of the use of different types of voice and
Silence. Acquiescent silence should also be included in future studies when relevant.
Voice and silence responses could also be investigated in organizational relationships
other than the supervisor-subordinate relationship. For example, some doctor-patient
relationships may require the sharing of negative information both ways: the doctor may
have to deliver a serious diagnosis, or the patient may have to explain his or her
dissatisfaction with the doctor’s behavior. Patient responses in particular could be
affected by their ability to change health care providers (HCPS) and the type of
relationship they have with their HCP. More specifically, patients may choose not to
speak up because they do not have other options for HCPS or because they do not have a
close relationship with their HCP. Learning more about the propensity to use voice and/or
silence responses in the doctor-patient relationship could help doctors and hospitals
improve their procedures and may help patients be able to express their concerns more
freely and clearly.

In addition to communication responses, it may also be informative to consider
emotional responses as they are related to communication responses. Considering that
some responses to negative feedback may be irrational, it may be useful to consider
action-emotion patterns which may be activated in response to negative feedback. These
could provide richer explanation for the communication responses to negative feedback.
Conclusion

This study investigated the effects of different facework messages on employee

responses to negative feedback that could induce autonomy or competence threat. In

93

addition, employees’ level of CA was studied as a potential factor in the choice of voice
and silence responses to feedback. Direct and indirect facework preference were created
as new variables which affected how people respond to negative feedback delivered with
different levels of face threat.

This study provided a test of eight different feedback messages, some more and
some less threatening to autonomy and competence face. An approbation strategy was
found to be the most effective preventative facework strategy for reducing perceptions of
both autonomy and competence threat. A tact strategy was not effective in reducing either
type of face threat, contrary to previous thinking in the area of face. The most functional
employee response, prosocial voice, was found to be negatively related to perceptions of
autonomy and competence threat, indicating that more threat would lead to fewer
prosocial voice responses. Both types of threat were also positively correlated with
defensive silence, indicating employees who perceived they were threatened were likely
to choose to use defensive silence as a response. Finally, CA moderated the effect of
competence threat on prosocial voice, defensive voice, and defensive silence. The higher
an individual’s CA levels, the more they responded to threat by using less prosocial voice
and by using more defensive silence and defensive voice.

Using approbation messages is recommended as an immediate solution to the
threats posed by negative feedback, until further research can be conducted. Practitioners
are advised to be aware of the effects of threat and CA on how employees respond to
negative feedback, and to consider implementing training to help deal with high CA

levels and defensive voice responses from employees. Finally, future research should be

94

conducted into facework strategies and their effectiveness, and employee voice and

silence responses, as these are two areas currently under-researched by researchers.

95

Endnotes
1 Because the autonomy and competence face threat scales were highly correlated with
each other (r = .72, p < .01), CFA was run with the scales combined as one scale and as
one scale with two factors (to check for second-order turidimensionality). The one-scale
CF A revealed that 30% of the observed correlations were outside of 95 % confidence
intervals for the predicted correlations. The second-order unidimensional CFA revealed
that 50% of the observed correlations were outside of 95 % confidence intervals for the
predicted correlations. Thus, the scales were kept separate for later analyses.
2 Measurement difficulties with prosocial silence precluded running analyses using this
variable. The analyses would have little meaning unless validity and reliability of the
measure are acceptable, and they are not in this case. This issue is discussed fmther in the

limitations section.

96

Table 1. Definitions and Examples of face-saving messages for solidarity, approbation,

and tact (Lim & Bowers, 1991).

 

 

 

 

 

 

 

 

 

 

Definition Sample Item

Solidarity

Friendship reaffirmation To express intimate “I know I can trust you.”
emotions toward the other

Cooperation To emphasize the necessity “We have to work on the
to cooperate with each other problem together.”

Empathy To Show understanding of “I understand what you are
the other’s emotional state trying to say.”

Social acknowledgement To appreciate work-related “You are a good
aspects of the other colleague.”

Agreement To agree or seek agreement “I had a Similar thought.”

Approbation

Admiration To approve of the other “You did a very good job.”
without any reservation

Support To approve of some other “You have some good

aspects with minimization

of the problem

ideas, but you need to
support them with

evidence.”

 

97

Table 1, cont.

 

 

 

 

 

 

 

Approbation (cont)

Contradiction To approve of some other “I know you put a lot of
aspects without time and effort into it, but it
understating the problem just doesn’t have the focus

you need.”

Suggestion To suggest ways to make “You need to add more
the performance even better evidence to strengthen your

arguments.”

Being diminutive To trivialize the “I know that I’m asking a
problematic area lot, but it needs a little more

research.”

Tact

Imposition Sharing To Share the responsibility “Could we sit together and

work on the ways to
improve it?”

Experimenting To explore the possibility “Would you be willing to

for the other to volunteer

work on the paper again?”

 

Unconventional Indirectness

Not to state the imposition

explicitly

“I think you are the best

candidate to write the group

paper.”

 

98

Table 2. Voice/Silence Table (Adapted from Van Dyne, Ang, & Botero, 2003)

 

Employee Silence.

Intentionally withholding
ideas, information, and
opinions related to negative
feedback

Employee Voice.

Intentionally expressing
ideas, information, and
opinions related to negative
feedback

 

Disengaged behavior
Based on resignation,
feeling unable to make a
difference

Acquiescent silence.

Examples:

Withholding ideas based on
resignation

Keeping opinions to self
due to low self-efficacy to
make a difference

Acquiescent voice.

Examples:

Expressing supportive ideas
based on resignation
Agreeing with the group
due to low self-efficacy to
make a difference

 

Defensive silence.

Defensive voice.

 

 

 

 

. . Examples: Examples:
gitgrgfgge :figglfiaid Withholding information on Expressing ideas that shift
and personally at risk problems based on fear attention elsewhere based
Omitting facts to protect the on fear
self Proposing ideas that focus
on other to protect the self
Prosocial silence. Prosocial voice.

. . Examples: Examples:
Emigeclbtsgebrzfizxm Withholding confidential Expressing solutions to
feeling cooperative and information based on problems based on
altruistic cooperation cooperation

Protecting proprietary Suggesting constructive
knowledge to benefit the ideas for change to benefit
organization the organization

 

99

 

Table 3. Scale Items & Factor Loadings

 

 

 

 

PRCA-Org 20 (5-pt. Likert-type scale) Factor
Loadings
I feel self-conscious when I am called upon to answer a question or give an .55
opinion. (recode)
I am basically an outgoing person. .57
When I’m with other people, I often have difficulty thinking of the right .66
things to talk about. (recode)
I’m afraid to speak up in conversations. (recode) .65
I look forward to an opportunity to speak in public. .50
In most Situations, I generally know what to say to people. .56
I talk less because I’m shy. (recode) .81
Conversing with people who hold positions of authority causes me to be .67
fearful and tense. (recode)
I consider myself to be the Silent type. (recode) .78
Previous Feedback Experience (5 pt. Likert-type scale)
I consider this type of feedback normal. .77
I would be Shocked to receive this feedback.(recode) .54
I would never expect to hear this kind of message.(recode) .88
I don’t know what it would be like to hear this kind of message. (recode) .65

 

100

Table 3, cont.

 

Scenario Realism (5 pt. Likert-type scale)

 

 

 

 

 

I didn’t have any problem with the realism of this situation. .48
It was difficult to make myself feel that this Situation was real. (recode) .57
This Situation could happen, or has happened. .76
A situation like this could develop in real life. .50
I can imagine being in a situation like this one. .75
Autonomy Face Threat @receded by “My boss ’s statements 7 pt. Likert-

type scale)

Made me uncomfortable. .60
Required me to do things exactly his/her way. .61
Did not allow me to choose how I do my job. .74
Made me feel I owed something back. .64
Constrained my choices. .84
Took away some of my independence. .83
Could make me look bad in the eyes of others. .66
Competence Face Threat (preceded by “My boss’s statements 7 pt. Likert-

Ope scale)

Showed he/she thought I was incompetent at the task .86
Indicated he/she thought I did not know what I was doing .88

101

Table 3, cont.

 

 

 

 

 

 

Pointed out he/she thought I failed miserably .79
Showed he/ she expected a lot better work to complete the task well. .55
Insulted my abilities .83
Questioned my competence+ .89
Allowed me to determine appropriate solutions.+ (recode) -. 14
Indicated doubt in my competence“ .79
+Item from Carson & Cupach (2002)

Defensive Voice (7 pt. Likert-tjpe scale)

I would not express anything except agreement with my supervisor, based on

fear. .79
I would express ideas that Shift attention to others, because I would be afraid. .82
I would provide explanations that focus the discussion on others in order to

protect myself. .72
I would go along and communicate support for my supervisor, based on self-

protection .67
I would express agreement with my supervisor, because I would be

motivated by fear. .81
Prosocial Voice (7 pt. Likert-type scale)

I would try to offer suggestions that would help other people at my company. .70
I would try to offer suggestions that would help my company. .83

102

Table 3, cont.

 

I would Share my opinions in an effort to help the company change for the

 

 

better. .81
I would want to help my organization, so I would make an effort to

contribute solutions to problems in my company. .80
I would give solutions to problems with the cooperative motive of benefiting

my organization. .84
I would develop and make recommendations concerning the feedback. .71
I would communicate my opinions about the feedback even if my supervisor .57
disagrees.

I would Speak up with ideas for new projects that might benefit the

organization. .84
I would suggest ideas for change, based on constructive concern for my

organization. .80
Defensive Silence (7 pt. Likert-type scale)

I would withhold ideas of mine because I’d be afraid something bad would

happen if I was to share them. .63
I would leave out information that would make me look bad. .61
If, based on this feedback, I thought my job or reputation was in danger, I

wouldn’t mention negative information about myself. .37
I would withhold ideas of mine because I wouldn’t want others to take

advantage of me. .64

103

Table 3, cont.

 

I would not speak up and suggest ideas for changing my performance, based

 

 

on fear. .80
I would withhold relevant information due to fear. .79
I would omit pertinent facts in order to protect myself. .67
I would avoid expressing ideas for improvements, based on wanting to

protect myself. .83
I would withhold my solutions to problems because I would be afraid. .84
Prosocial Silence (7 pt. Likert—type scale)

If I knew someone else wasn’t doing so well, I wouldn’t bring it up to my

supervisor. .47
I would keep quiet rather than “rock the boat” by bringing up ideas that will

cause problems for my supervisor. .62
I wouldn’t mention negative information about others at work to avoid

looking like a tattletale. (RC) .49
I would withhold thoughts about my performance, based on COOperation. .53
I would withstand pressure from others to discuss the feedback I received. .26
I would refuse to divulge information that might harm the organization. .35

 

104

Table 3, cont.

 

Feedback sign manipulation check (adapted fiom Haeggburg, 2001;

semantic dijfcrential scaling fiom 1-7)

 

 

 

 

 

 

 

 

Good-Poor .65
Low-High .83
Below average-Above average“ .88
* Additional item written by researcher

Severity (Semantic difizrential scale fiom 1-7)

Trivial-Crucial .66
Significant-Insignificant (recode) .79
Weak-Strong .63
Major-Minor (recode) .78
Supervisor Supportiveness (Taylor & Bowers, I 972)

My supervisor is friendly and easy to approach. .85
When I talk with my supervisor, he or she pays attention to what I’m saying. .93
My supervisor is willing to listen to my problems. .92
F acework Strategy Preferences (adapted from Merkin, 2006)

Direct

If something is wrong with my performance, I would want to be told directly. .75
I like my supervisor to express his/her thoughts on my ability forthrightly. .82

105

Table 3, cont.

 

 

 

I want others to share their real thoughts about my performance fiankly. .82
I want others to say what they think with no reservations. .71
I want others to acknowledge what they see as the truth straight to my face. .74
Indirect

I want others to express themselves in question form (e.g., asking if I need

some help doing my tasks). .53
I want others to express their thoughts on my ability indirectly. .50
I want others to deliver their reactions to my performance sensitively. .56
I want others to respond to my mistakes or successes in a subtle way. .70

 

106

Table 4. One-Way AN OVA results for condition on continuous control variables.

 

 

df F
Participant’s Age 7, 423 1.04
Organizational Status 7, 411 0.73
Supervisor Status 7, 423 0.86
Tenure 7, 419 1.98
Hours per week 7, 420 0.53

 

Note. * p < .05, ** p < .01, *** p < .001

Table 5. Chi-square results for condition on nominal control variables

 

 

df ,8
Participant’s Sex 7 8.41
Ethnicity 35 39.97
Education Level 56 61.01
Job Type 49 45.82
Organization Type 42 39.92
Supervisor Sex 7 4.27
Organization Size 35 21.06

 

Note. * p < .05, ** p < .01, *** p < .001

107

Table 6. Perception of autonomy and competence face threat by condition.

 

 

 

Condition N M SD
Autonomy Autonomy Threat 54 4.38 1.32
face threat
Autonomy with Tact 53 4.30 1.18
Autonomy with Approbation 56 4.08 1.26
Autonomy with Tact & 55 4.24 1.12
Approbation
Competence Threat 49 4.42 1.09
Competence with Tact 53 4.32 1.12
Competence with 53 4.05 1.09
Approbation
Competence with Tact & 56 4.25 1.11
Approbation
Competence Competence Threat 49 4.4721de 1.22
face threat
Competence with Tact 53 4.12 1.28
Competence with 53 3-71d 1.20
Approbation
Competence with Tact & 55 3.95 1.18
Approbation
Autonomy Threat 54 4.07 1.40
Autonomy with Tact 53 3-70a 1.09
Autonomy with Approbation 56 3-51b 1.06
Autonomy with Tact & 55 3.47c 1.02
Approbation

 

Note. Significantly different means, according to Tukey tests, are marked with pairs of
matching letters in subscript.

108

Table 7. Zero-order correlations among trait CA, face threat, voice/silence, and direct &

indirect facework strategy preference.

 

PRCA- Autonomy Competence Defensive Prosocial

 

 

Org face threat face threat voice voice
PRCA-Org
Autonomy
facethreat .17**
Competence
facethreat .21" .72“
Defensive
voice .27** -.05 .07
Prosocial
voice -.30** -.19** -.23** -.46**
Defensive
silence .30** .10* .19" .74** -.52**
Prosocial
silence .13“ .05 .02 .24M -.10*
Direct
facework -.36** -.02 -.07 -.35** .32**
preference
Indirect
facework .27” .10* .12* .20** -.06
preference
M 2.29 4.25 3.87 3.06 5.20
SD 0.64 1.16 1.21 1.23 1.04

 

Note: N’s range from 418-435. ** Correlation is significant at the 0.01 level (2-tailed).
* Correlation is significant at the 0.05 level (2-tailed).

109

Table 7, cont.

 

Direct Indirect
Defensive Prosocial facework facework
silence silence preference preference

 

PRCA-Org

Autonomy
face threat

Competence
face threat

Defensive
voice

Prosocial
voice

Defensive
silence

Prosocial
silence .47**

Direct

facework -.36** -.09

preference

Indirect

facework .21 ** .05 -.36**

preference

 

M 3.22 4.10 4.04 2.76

SD 1.06 0.86 0.63 0.66

 

Note: N’s range fi'om 418-435. ** Correlation is significant at the 0.01 level (2-tailed).
* Correlation is Significant at the 0.05 level (2-tailed).

110

Table 8. Correlations among variables for low CA quartile split (11 = 119, M = 1.56, SD =

 

 

.27).
Autonomy Competence Defensive Prosocial Defensive Prosocial
face threat face threat voice voice silence silence

Autonomy

face threat

Competence 6 08"

face threat '

Difensw" -.096 -.089

vorce

P’S’S‘flal -.187* -.103 -.468**

vorce

Pefens‘v" .081 .020 .690" -.408**

Silence

P.’°S°°‘al -.084 -.111 225* -.056 .473"

Silence

 

Note. ** p < 0.01 (2-tailed), * p < 0.05 (2-tailed).

111

Table 9. Correlations among variables for high CA quartile split (11 = 103, M = 3.21, SD

 

 

= .39).
Autonomy Competence Defensive Prosocial Defensive Prosocial
face threat face threat voice voice silence silence

Autonomy

face threat

Competence 810**

face threat '

Difens’ve -.205* .006

vorce

”950““ —.210* -.328** -.390**

vorce

Pefenswe .108 .268" .649" -.512**

Silence

Pios‘f‘al .126 .119 .311** -.105 .528"

Silence

 

Note. ** p < 0.01 (2-tailed), * p < 0.05 (2-tailed).

112

Table 10. Hierarchical Regression Analyses Results.

 

 

Defensive Voice b so. t
Feedback Severity -0.109 0.048 -2.222*
Supervisor Supportiveness -0.072 0.070 -1.464
F B Severity x Supv. -0.034 0.055 -0.694
Support

F(2, 410) = 3.927, p < .05, adjusted R2 = .014
Fchange (1, 409) = 0.481, p = .49, Ram, = .001

 

 

Prosocial Voice b so. t
Feedback Severity 0.002 0.039 -0.033
Supervisor Supportiveness 0.257 0.057 5362*"
FB Severity x Supv. 0.054 0.046 1.131
Support

F(2, 410) = 14.559, p < .001, adjusted R2 = .062
Fchmg, (l, 409) = 1.278, p = .259, Rich...“ = .003

 

 

Defensive Silence b so. t
Feedback Severity -0.034 0.041 -0.682
Supervisor Supportiveness -0.075 0.060 -1 .5 l 9
F B Severity x Supv. -0.015 0.048 -0.312
Support

F(2,411)=1.514,p = .221, adjusted R2 = .002
Fchmg, (l, 410) = 0.097, p = .755, Ram, = .000

 

Note. ***p < .001, Mp < .01, *p < .05

113

Table 11. Hierarchical Regression Analyses Results.

 

 

Defensive Voice B so. t
Appmbatlo“ vs' No 0.055 0.131 1.188
Approbatlon
CA 0.156 0.094 3.131"
Autonomy Face Threat -0.194 0.068 -3.028*"'
Competence Face Threat 0.148 0.066 2288*
Direct Facework Pref. -0.261 0.101 -5.098***
Indirect Facework Pref. 0.070 0.093 1.397
AppxPRCA -0.348 0.225 -1.635
Appruto -0.050 0.158 -0.398
AppxComp 0.022 0.158 0.162
Apprirect -0.138 0.243 -1.265
Appxlndirect -0.068 0.235 -.0623
PRCAXAuto -0.160 0.129 -1.924
PRCAxComp 0.180 0.1 10 2.163“
PRCAxDirect 0.049 0.163 0.91 1
PRCAxIndirect 0.101 0.147 1.756
AutoxComp 0.01 1 0.037 0.224
AutoxDirect 0.03 1 0. 127 0.407
Autoxlndirect 0.050 0.1 14 0.661
Comprirect 0.030 0.133 0.394
CompxIndirect -0.030 0.1 13 -0.396
Directxlndirect —0.034 0. 1 52 -0.641

F(6,411)=14.121,p<.01,adjustedR2 = .159
Am. (15, 396) = 1.138, p = .320, R2...“ = .034

 

114

Table 11, cont.

 

Prosocial Voice

Approbation vs. No
Approbation

CA

Autonomy Face Threat
Competence Face Threat
Direct F acework Pref.
Indirect Facework Pref.
AppxPRCA
Appruto

AppxComp
Apprirect
Appxlndirect
PRCAxAuto
PRCAxComp
PRCAxDirect
PRCAxIndirect
AutoxComp
AutoxDirect
AutoxIndirect
Comprirect
CompxIndirect
Directxlndirect

F(6,411)=16.621,p < .001, adjustedR2 = .195
Fang. (15, 396) = 1.503, p = .101, Ram. = .043

 

B so. T
0028 0.109 -0.607
-0. 195 0.078 -3.990***
-0.078 0.057 -1.236
-0.136 0.055 -2.l3l*
0.292 0.084 5.791***
0.122 0.078 2.464*
0.070 0.186 0.702
0.061 0.130 0.497
-0.072 0.131 -0.544
0.069 0.200 0.650
0.210 0.194 1.954
0.079 0.106 0.965
-0.167 0.091 -2.049*
0.018 0.134 0.354
0.052 0.121 0.930
0.098 0.030 2.022*
0.102 0.105 1.343
0.018 0.094 0.244
0.006 0.109 0.083
0.066 0.093 0.894
0.019 0.125 0.369

 

115

Table 11, cont.

 

 

Defensive Silence B s.e. T
APPmbatlo“ vs: No 0.013 0.111 0.278
Approbation
CA 0.165 0.080 3.353“
Autonomy Face Threat -0.044 0.058 -0.700
Competence Face Threat 0.161 0.056 2516*
Direct Facework Pref. 0.278 0.083 ~5.546***
Indirect Facework Pref. 0.050 0.078 1.019
AppxPRCA -0.168 0.191 -1.654
Appruto -0.064 0.134 -0.509
AppxComp 0.087 0.134 0.647
Apprirect -0.236 0.205 -2.137*
Appxlndirect -0.078 0.199 -0.721
PRCAxAuto -0.158 0.109 -1.911
PRCAxComp 0.183 0.094 2.222*
PRCAxDirect -0.071 0.133 -1.364
PRCAxIndirect -0.062 0.125 -1.078
AutoxComp 0.058 0.03 1 l .179
AutoxDirect -0.03 1 0. 103 —0.395
AutoxIndirect 0.080 0.095 1.081
Comprirect 0.085 0.1 13 1.073
CompxIndirect -0.024 0.096 -0.321
Directxlndirect -0.039 0.120 -0.745

F(6, 412) = 15.847, p < .001, adjusted R2 = .188
chg, (15, 397) = 1.056, p = .396, Fang, = .031

 

Note. ***p < .001, ”p < .01, *p < .05

116

APPENDICES

Appendix A. Message Components

Autonomy threat
Loss of independence / obligation / order

 

This is not how I would have done it. / You have an obligation to the organization to
roduce quality work. / Keep working until your work is up to my standards.

 

 

 

Tact Message
Imposition sharing/debt incurrence

 

What was expected of you may not have been clear, though. / I’d greatly appreciate it if
you would keep putting in good effort.

 

 

 

Competence threat
Ridicule/ Comparative disapproval/Blunt disapproval

 

You screwed this job up. / Your performance is not up to standard, and / this job needs to
be redone.

 

 

 

Approbation Message
Support/admiration

 

I But, I want you to know you’re on the right track and your work has potential.

 

Combined Tact & Approbation message

 

What was expected of you may not have been clear, though. You’re on the right track and
your work has potential. I’d greatly appreciate it if you would keep putting in good effort.

 

 

 

117

Appendix B. Message Conditions

Imagine your supervisor stops by your work area to talk with you about some recent
work you had performed. In his/her estimation there is a problem because you did not
perform a certain task required of you as well as you should. He/she provides you with a
brief and general statement about that task and you are left to decide what to do next.

Autonomy threat

 

This is not how I would have done it. You have an obligation to the organization to
produce quality work. Keep working until your work is up to my standards.

 

 

 

Autonomy threat with tact

 

This is not how I would have done it. You have an obligation to the organization to
produce quality work. What was expected of you may not have been clear, though. I’d
greatly appreciate it if you would keep putting in good effort. Keep working until your
work is up to my standards.

 

 

 

Autonomy threat with approbation

 

This is not how I would have done it. You have an obligation to the organization to
produce quality work. But, I want you to know you’re on the right track and your work
has potential. Keep working until your work is up to my standards.

 

 

 

Autonomy threat with tact and approbation

 

This is not how I would have done it. You have an obligation to the organization to
produce quality work. What was expected of you may not have been clear, though.
You’re on the right track and your work has potential. I’d greatly appreciate it if you
would keep putting in good effort. Keep working until your work is up to my standards.

 

 

 

 

Competence threat

 

You screwed this job up. Your performance is not up to par, and the job needs to be
redone.

 

 

 

Competence threat with tact

 

You screwed this job up. Your performance is not up to par, and the job needs to be
redone. What was expected of you may not have been clear though. I'd greatly appreciate
it if you would keep putting in good effort.

 

 

 

Competence threat with approbation

 

You screwed this job up. Your performance is not up to par, and the job needs to be
redone. But, I want you to know you're on the right track and your work has potential.

 

 

 

Commtence threat with tact and approbation

 

You screwed this job up. Your performance is not up to par, and the job needs to be
redone. What was expected of you may not have been clear, though. You're on the right
track and your work has potential. I'd greatly appreciate it if you would keep putting in
good effort.

 

 

 

118

Appendix C. Measurement
Organizational status measurement

Your status in the organization: (please circle the appropriate number)

 

119

REFERENCES

Allen, M., & Bourhis, J. (1996). The relationship of communication apprehension to
communication behavior: A meta-analysis. Communication Quarterly, 44, 214-
226.

Ammons, R. B. (1956). Effects of knowledge of performance: A survey and tentative
theoretical formulation. Journal of General Psychology, 54, 279-299.

Anderson, J. R. (1987). Skill acquisition: Compilation of weak-method problem
solutions. Psychological Review, 94, 192-210.

Anderson, S., & Rodin, J. (1989). IS bad news always bad?: Cue and feedback effects on
intrinsic motivation. Journal of Applied Social Psychology, 19, 449-467.

Annett, J. (1969). Feedback and human behavior. Baltimore, MD: Penguin Books.

Antonioni, D. (1994). The effects of feedback accountability on upward appraisal ratings.
Personnel Psychology, 4 7, 349-356.

Asch, S. E. (1946) Forming impressions of personality, Journal of Abnormal and Social
Psychology, 41, 258-290.

Ashford, S. J ., & Cummings, L. L. (1983). Feedback as an individual resource: Personal
strategies of creating information. Organizational Behavior and Human
Performance, 32, 370-398.

Ayres, J ., Hopf, T., & Will, A. (2000). Are reductions in CA an experimental artifact? A
solomon four-group answer. Communication Quarterly, 48, 19-26.

Ayres, J ., Keereetaweep, T., Chen, P. E., & Edwards, P. A. (1998). Communication
apprehension and employment interviews. Communication Education, 4 7, 1-17.

Azjen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social behavior.
Englewood Cliffs, NJ: Prentice-Hall.

Balcazar, R, Hopkins, B. L., & Suarez, Y. (1985). A critical, objective review of
performance feedback. Journal of Organizational Behavior Management, 7, 65-
89.

Bandura, A. (1991). Social cognitive theory of self-regulation. Organizational Behavior
and Human Decision Processes, 50, 248-287.

Bandura, A. (2006). Guide for creating self-efficacy scales. In F. Pajares & T. C. Urdan
(Eds), Self-efi‘icacy beliefs of adolescents. Greenwich, CT: Information Age
Publishing.

120

Bandura, A., & Cervone, D. (1983). Self-evaluative and self-efficacy mechanisms
governing the motivational effects of goal systems. Journal of Personality and
Social Psychology, 45, 1017-1028.

Baron, R. A. (1988). Negative effects of destructive criticism: Impact on conflict, self-
efficacy, and task performance. Journal of Applied Psychology, 73, 199-207.

Baron, R. A. (1990). Countering the effects of destructive criticism: The relative efficacy
of four interventions. Journal of Applied Psychology, 75, 23 5-245.

Bartoo, H. & Sias, P. M. (2004). When enough is too much: Communication

apprehension and employee information experiences. Communication Quarterly,
52, 15-26.

Boume, L. E. (1966). Comments on professor I. M. Bilodeau's paper. In E. A. Bilodeau
(Ed.), Acquisitions of skill. New York: Academic Press.

Brown, P. & Levinson, SC. (1987). Politeness: Some universals in language usage. New
York, NY: Cambridge University Press.

Burke, R. J ., Weitzel, W., and Weir, T. (1978). Characteristics of effective employee
performance review and development interviews: Replication and extension.
Personnel Psychology, 31, 903-919.

Burke, R. J ., & Wilcox, D. S. (1969). Characteristics of effective employee performance
review and development interviews. Personnel Psychology, 22, 291 -305.

Carson, C. L., & Cupach, W. R. (2000). Facing corrections in the workplace: The
influence of perceived face threat on the consequences of managerial reproaches.
Journal of Applied Communication Research, 28, 215-234.

Carson, C. L., & Cupach, W. R. (2002). Characteristics and consequences of
interpersonal complaints associated with perceived face threat. Journal of Social
and Personal Relationships, 19, 443-462.

Carver, C. S., & Scheier, M. F. (1982). Control theory: A useful conceptual fiamework
for personality-social, clinical, and health psychology, Psychological Bulletin, 92,
1 1 1-135.

Clement, D. A. & F randsen, K. D. (1976). On conceptual and empirical treatments of
feedback in human communication. Communication Monographs, 43, 11-28.

Cohen, J ., Cohen, R, West, S. G., & Aiken, L. S. (2003). Applied multiple
regression/correlation analysis for the behavioral sciences (3rd ed.). Mahwah, NJ:
LEA.

Cook, 3., Banks, 8., & Turner, R. (1993). The effects of work environment on burnout in
the newsroom. Newspaper Research Journal, I 4, 123—136.

121

Cupach W. R., & Metts, S. (1994). Facework. Thousand Oaks, CA: Sage.

Daly, J. A., & McCroskey, J. C. (1975). Occupational desirability and choice as a
function of communication apprehension. Journal of Counseling Psychology, 22,
309-313.

Deci, E. L. & Ryan, R. M. (2000). The “what” and “why” of goal pursuits: Human needs
and the self-detennination of behavior. Psychological Inquiry, 11, 227-268.

Dohrenwend, A. (2002). Serving up the feedback sandwich. Family Practice
Management, 9, 43-46.

Dwyer, K. K. (2000). The multidimensional model: Teaching students to self-manage
high communication apprehension by self-selecting treatments. Communication
Education, 49, 72-81.

Eisenberger, R., Stinglhamber, F., Vandenbergl, C., Sucharski, I., L., & Rhoades, L.
(2002). Perceived supervisor support: Contributions to perceived organizational
support and employee retention. Journal of Applied Psychology, 87, 565-573.

Elliot, A. J ., & Dweck, C. S. (2005). Competence and motivation: Competence as the
core of achievement motivation. In A. J. Elliot and C. S. Dweck (Eds), Handbook
of Competence and Motivation (pp. 3-14). New York, NY: Guilford Press.

Ellis, J. B., & Van Dyne, L. (in press). Voice and silence as observer reactions to
defensive voice: Predictions based on communication competence theory. In J.
Greenberg & M. S. Edwards (Eds), Voice and Silence in Organizations. Oxford,
UK: Elsevier.

Fedor, D. F., Eder, R. W., Buckley, M. R. (1989). The contributory effects of supervisor
intentions on subordinate feedback responses. Organizational Behavior and
Human Decision Processes, 44, 396-414.

French, Jr., J. R. P., & Raven, B. (1959). The bases of social power. In D. Cartwright
(Ed), Studies in Social Power. Ann Arbor: University of Michigan.

Gardiner, J. C. (1971). A synthesis of experimental studies of Speech communication
feedback. The Journal of Communication, 21, 17-35.

Geddes, D. (1993). Examining the dimensionality of performance feedback messages:
Source and recipient perceptions of influence attempts. Communication Studies,
44, 200-215.

Geddes, D., & Baron, R. A. (1997). Workplace aggression as a consequence of negative
performance feedback. Management Communication Quarterly, 10, 433-454.

Geddes, D., & Linnehan, F. (1996). Exploring the dimensionality of positive and negative
performance feedback. Communication Quarterly, 44, 326-344.

122

Gibb, J. R. (1961). Defensive communication. Journal of Communication, 1 I, 141-148.

Giles, W. F., & Mossholder, K. W. (1990). Employee reactions to contextual and session
components of performance appraisal. Journal of Applied Psychology, 75 , 371-
377.

Goffrnan, E. (1955). On facework: An analysis of ritual elements in social interaction.
Psychiatry, 18, 319-345.

Gottfredson, G. D., & Holland, J. L. (1990). A longitudinal test of the influence of
congruence: Job satisfaction, competence utilization, and counterproductive
behavior. Journal of Counseling Psychology, 3 7, 389-398.

Greenberg, J. & Eskew, D. E. (1993). The role of role playing in organizational research.
Journal of Management, 19, 221-241.

Halperin, K., Snyder, C. R., Shenkel, R. J ., & Houston, B. K. (1976). Effect of source
status and message favorability on acceptance of personality feedback. Journal of
Applied Psychology, 61, 85-88.

Haeggberg, D. (2000). The influence of goal orientation on individuals’ cognitive,
affective, and behavioral reactions to different types of performance feedback.
Unpublished doctoral dissertation, Ohio University, Athens.

Hagedoom, M., Van Yperen, N. W., Van De Vliert, E., & Buunk, B. P. (1999).
Employees’ reactions to problematic events: A circumplex structure of five
categories of responses, and the role of job satisfaction. Journal of Organizational

Behavior, 20,309—321.

Harville, D. L. (1992). Person/job fit model of communication apprehension in
organizations. Management Communication Quarterly, 6, 150-165.

Hirschman, A. O. (1970). Exit, voice, and loyalty. Cambridge, MA: Harvard University
Press.

Hodson, R. (1991). The active worker: Compliance and autonomy at the workplace.
Journal of Contemporary Ethnography, 20, 47-78.

Holtgraves, T., Eck, J ., & Lasky, B. (1997). Face management, question wording, and
social desirability. Journal of Applied Social Psychology, 27, 1650-1671 .

Ilgen, D. R., Fisher, C., & Taylor, MS. (1979). Consequences of individual feedback on
behavior in organizations. Journal of Applied Psychology, 64, 349-371.

J ablin, F. (1978). Message-response and “openness” in superior-subordinate
communication. Communication Yearbook, 2, 294-309.

123

J ablin, F. (1981). Subordinate’s sex and superior-subordinate status differentiation as
moderators of the Pelz_ effect. Communication Yearbook, 4, 349-366.

Jacobs, M., Jacobs, A., Feldman, G., & Cavior, N. (1973). Feedback II--the “credibility
gap”: Delivery of positive and negative and emotional and behavioral feedback in
groups. Journal of Consulting and Clinical Psychology, 41, 215-223.

James, M., Treadway, D. C., Conner, D., & Hochwarter, W. A. (2003). The interactive
effects of positive affect and a sense of competency on the politics perceptions-
job satisfaction relationship. Journal of Management, 29, 551-567.

J ones-Johnson, G. & Johnson, W. R. (2001). Subjective underemployment and
psychosocial stress: The role of perceived social and supervisor support. The
Journal of Social Psychology, 132, 11-21.

Kerssen-Griep, J. (2001). Teacher communication activities relevant to student
motivation: Classroom facework and instructional communication competence.
Communication Education, 50, 256-273.

Kerssen-Griep, J ., Hess, J. A., & Trees, A. R. (2003). Sustaining the desire to learn:
Dimensions of perceived instructional facework related to student involvement
and motivation to learn. Western Journal of Communication, 67, 357-381.

Kim, M., & Hunter, J. E. (1993). Relationships among attitudes, behavioral intentions,

and behavior: A meta-analysis of past research, part 2. Communication Research,
20, 331-364.

Kluger, A.N., & Denisi, A. (1996). The effects of feedback interventions on performance:

A historical review, a meta-analysis, and a preliminary feedback intervention
theory. Psychological Bulletin, 119, 254-284.

Kluger, A. N., Lewinsohn, S. & Aiello, J. R. (1994). The influence of feedback on mood:
Linear effects on pleasantness and curvilinear effects on arousal. Organizational
Behavior and Human Decision Processes, 60, 276-299.

Kozlowski, S. W. J., Toney, R. J ., Mullins, M. E., Weissbein, D. A., Brown, K. G., &
Bell, B. S. (2001). Developing adaptability: A theory for the design of integrated-
embedded training systems. In E. Salas (Ed.), Advances in human performance
and cognitive engineering research (V 01. 1, pp. 59-123). Amsterdam: JAI/Elsevier
Science.

Kraut, R. E., Fish, R. 8., Root, R. W., & Chalfonte, B. L. (1993). Information
communication in organizations: Form, function and technology. In R. M.
Backer, (Ed.), Readings in Groupware and Computer-supported Cooperative
Work, (pp. 287-314). Morgan Kaufrnann, San Mateo, CA.

124

Lewis, L. K. & Hayward, P. A. (2003). Choice-based learning: Student reactions in an
undergraduate organizational communication course. Communication Education,
52, 148-156.

Lim, T., & Bowers, J. W. (1991). Facework: Solidarity, approbation, and tact. Human
Communication Research, I 7, 415-450.

Lindsley, D. H., Brass, D. J ., & Thomas, J. B. (1995). Efficacy-performance spirals: A
multi-level perspective. The Academy of Management Review, 20, 645-678.

London, M. (2003). Job feedback: Giving, seeking, and using feedback for performance
improvement. Mahwah, NJ: Lawrence Erlbaum.

Lundgren, D. C., & Rudawsky, D. J. (1998). Female and male college students’
responses to negative feedback from parents and peers. Sex Roles, 39, 409-429.

Mahoney, T. A. (1979). Organizational hierarchy and position worth. The Academy of
Management Journal, 22, 726-737.

MacIntyre, P. D. (1994). Variables underlying willingness to communicate: A causal
analysis. Communication Research Reports, 1 I , 135-142.

McCroskey, J. (1977a). Studies of the relationship between communication apprehension
and self-esteem. Human Communication Research, 3, 269-277.

McCroskey, J. (1977b). Communication apprehension: A summary of recent theory and
research. Human Communication Research, 4, 78-96.

Merkin, R. S. (2006). Uncertainty avoidance and facework: A test of the Hofstede model.
International Journal of Intercultural Relations, 30, 213-228.

Meyer, H. H. (1975). The pay-for-performance dilemma. Organizational Dynamics, 3,
39-50.

Miller, V. M., Johnson, J. R., Hart, 2., & Peterson, D. (1999). A test of antecedents and
outcomes of employee role negotiation ability. Journal of Applied
Communication Research, 27, 24-48.

Morand, D. A. (2000). Language and power: An empirical analysis of linguistic strategies
used in superior-subordinate communication. Journal of Organizational
Behavior, 21, 235-248.

Moreland, R. L., & Sweeney, P. D. (1984). Self-expectancies and reactions to evaluations
of personal performance. Journal of Personality, 52, 156-176.

Morrison, E. W., & Milliken, F. J. (2000). Organizational silence: A barrier to change
and development in a pluralistic world. The Academy of Management Review, 25,
706-725.

125

Parker, S. K., Axtell, C. M., & Turner, N. (2001). Designing a safer workplace:
Importance of j ob autonomy, communication quality, and supportive supervisors.
Journal of Occupational Health Psychology, 6, 211-228.

Pinder, C. C., & Harlos, K. P. (2001). Employee silence: Quiescence and acquiescence as
responses to perceived injustice. Research in Personnel and Human Resources
Management, 20, 331-369.

Podsakoff, P. M., & F arh, J. (1989). Effects of feedback Sign and credibility on goal
setting and task performance. Organizational Behavior and Human Decision
Processes, 44, 45-67.

Prue, D. M. & F airbank, J. A. (1981). Performance feedback in organizational behavior
management: A review. Journal of Organizational Behavior Management, 3, l-
16.

Rhoades, L. & Eisenberger, R. (2002). Perceived organizational support: A review of the
literature. Journal of Applied Psychology, 87, 698-714.

Rhodes, S. C., & Frandsen, K. D. (1975). Some effects on the instruction in feedback
utilization on the fluency of college students' Speech. Speech Monographs, 42, 83-
89.

Rosnow, R. (1966). Whatever happened to the "Law of Primacy." Journal of
Communication, 1 6, 10-3 1 .

Rusbult, C. E., Farrell, D., Rogers, G., & Mainous III, A. G. (1988). Impact of exchange
variables on exit, voice, loyalty, and neglect: An integrative model of responses to
declining job satisfaction. Academy of Management Journal, 31, 599-627.

Scandura, T. A., & Graen, G. B. (1984). Moderating effects of initial leader-member
exchange status on the effects of a leadership intervention. Journal of Applied
Psychology, 69, 428-436.

Scheidel, T. & Crowell, L. (1966). Feedback in small group communication. Quarterly
Journal of Speech, 52, 273-278.

Scott, M. D., McCroskey, J. C., & Sheahan, M. E. (1978). Measuring communication
apprehension. Journal of Communication, 28, 104-111.

Shanock, L. R. & Eisenberger, R. (2006). When supervisors feel supported: Relationships
with subordinates’ perceived supervisor support, perceived organizational
support, and performance. Journal of Applied Psychology, 91, 689-695.

Shrauger, J .S., & Rosenberg, SE. (1970). Self-esteem and the effects of success and
failure feedback on performance. Journal of Personality, 38, 404-417.

Skinner, B. F. (1953). Science and human behavior. New York: McMillan.

126

Steelman, L, A., & Rutkowski, K. A. (2004). Moderators of employee reactions to
negative feedback. Journal of Managerial Psychology, 19, 6-18.

Steinhardt, M., Dolbier, C. L., Gottlieb, N. H., & McCalister, K. T. (2003). The
relationship between hardiness, supervisor support, group cohesion, and job stress
as predictors of job satisfaction. American Journal of Health Promotion, 1 7, 382-
389.

Stoltz, W. S., & Tannenbaum, P. H. (1963). Effects of feedback on oral encoding
behavior. Language and Speech, 6, 218-228.

Stone, D. L. & Stone, E. F. (1985). The effects of feedback consistency and feedback
favorability on self-perceived task competence and perceived feedback accuracy.
Organizational Behavior and Human Decision Processes, 36, 167-185.

Tata, J. (2002). The influence of managerial accounts on employees’ reactions to
negative feedback. Group & Organization Management, 27, 480-503.

Taylor, F. C., & Bowers, D. G. (1972). Survey of organizations: A machine scored
standardized questionnaire instrument. Ann Arbor, MI: Institute for Social
Research, University of Michigan.

Thorndike, E. L. (1927). The law of effect. The American Journal of Psychology, 3.9,
212-222.

Ting-Toomey, S. (1988). Intercultural conflict styles: A face negotiation theory. In Y.
Kim & W. Gudykunst (Eds), Theories in intercultural communication (pp. 213-
235). Newbury Park, CA: Sage.

Van Dyne, L., Ang, S., & Botero, I. C. (2003). Conceptualizing employee silence and
employee voice as multidimensional constructs. Journal of Management Studies,
40, 1359-1392.

Villanova, P., Bemardin, H. J ., Dahmus, S. A., and Sims, R. L. (1993). Rater leniency
and performance appraisal discomfort. Educational and Psychological
Measurement, 53, 789-799.

Wang, G. and Netemeyer, R. G. (2002). The effects of job autonomy, customer
demandingness, and trait competitiveness on salesperson learning, self-efficacy,
and performance. Academy of Marketing Science Journal, 30, 217-228.

Westerman, C. Y. K. & Mieksztyn, J. (2007). Part-time workers and organizational
justice. Unpublished manuscript, Michigan State University.

Westerman, C. Y. K. & Park, H. S. (2007). Can supervisor facework strategies buffer
negative feedback to improve receiver performance? Unpublished manuscript,
Michigan State University.

127

Westerman, C. Y. K., Park, H. S., & Lee, H. (2007). A test of equity theory in
multidimensional friendships: A comparison of the United States and Korea.
Journal of Communication, 57, 576-598.

Wilson, S. R., Aleman, C. G., & Leatham, G. B. (1998). Identity implications of
influence goals: A revised analysis of face-threatening acts and application to

seeking compliance with same-sex fiiends. Human Communication Research, 26,
64-96.

Wilson, S. R. & Kunkel, A. W. (2000). Identity implications of influence goals:
Similarities in perceived face threats and facework across sex and close
relationships. Journal of Language and Social Psychology, 19, 195-221.

Winiecki, K. L., & Ayres, J. (1999). Communication apprehension and receiver
apprehension in the workplace. Communication Quarterly, 4 7, 430-440.

128

   

lljlljljiijjljjjljjljl