JOURNAL OF EXPERIMENTAL SOCIAL PSYCHOLOGY 26, 389-407 (l!

i@@

When It Is Especially Important to Explain Why: Factors

Affecting the Relationship between Managers’
Explanations of a Layoff and Survivors’
Reactions to the Layoff

JOEL BROCKNER

Columbia University

ROCKI LEE DEWITT

Penn State University

STEVEN GROVER

Indiana University

AND

THOMAS REED

Texas A&M

Received April 21, 1989

Recent research has shown that subordinates are generally more accepting of
their managers’ resource allocation decisions to the extent that managers offer
clear explanations (or accounts) of the reasons underlying their decisions. This
study sought to delineate the conditions under which the clarity of managers’
accounts is especially related to the favorability of subordinates’ reactions. We
hypothesized that when subordinates’ need for information is high-i.e., when
they are relatively uncertain about why the resources were allocated in a par-
ticular way, and when they attach greater importance to the resource allocation
decisions-they will be greatly influenced by the presence or absence of clear
managerial accounts. These hypotheses were tested in the context of a field study

The authors thank Bob Bies and Rob Folger for their comments on an earlier version
of the manuscript. Please address correspondence to Joel Brockner, Graduate School of
Business, Columbia University, New York, NY 10027.

389
0022-1031/90 $3.00
Copyright Q 1990 by Academic Press, Inc.
All rights of reproduction in any form reserved.
390 BROCKNER ET AL.

exploring survivors’ reactions to the layoffs of co-workers. Survivors indicated

how clearly the reasons for the layoff had been explained to them by management;
they also reported how much their organizational commitment, work effort. and
turnover intention had changed, relative to the prelayoff period. In general, the
positive relationship between the clarity of managers’ explanations and the fa-
vorability of survivors’ reactions was especially pronounced under conditions of
high uncertainty and high importance. The theoretical and practical implications
of these results, and the limitations of the study, are discussed. Q 1990 Academic
Press. Inc.

Recent research has shown that individuals are much more accepting
of unfavorable resource allocation decisions if they have been given clear
and adequate explanations of the bases underlying those decisions (see
Bies, 1987). This finding is of interest to social psychologists working
from a number of different theoretical perspectives. For example, in the
past 15 years (Thibaut & Walker, 1975), justice scholars have become
especially interested in the relationship between procedural and distrib-
utive justice. One key finding is that individuals’ negative behavioral and
attitudinal reactions to resource allocations perceived as distributively
unjust can be significantly reduced when the procedures accompanying
those decisions-such as the provision of an explanation of the reasons
underlying the resource allocation decisions-were deemed to be fair
(e.g., Greenberg, 1987; Rousseau & Anton, 1988; Tyler, 1990). Moreover,
attribution theorists (e.g., Kelley, 1967) and students of social interaction
(e.g., Goffman, 1959) have long been interested in the interpersonal
consequences of individuals’ explaining why they acted in a particular
way.
The following studies have shown that individuals are more accepting
of undesirable resource allocations that have been accompanied by clear
and adequate explanations. Bies and Shapiro (1987) explored subjects’
reactions to a manager who appeared to have stolen ideas from his
subordinate. Half of them were provided a causal explanation suggesting
that external factors had influenced the manager’s actions, whereas half
were not. As predicted, subjects’ feelings of injustice were much lower,
and their evaluation of the manager was much higher in the former than
the latter condition. In a second study, Bies and Shapiro explored in-
dividuals’ reactions to an inequitable sales contract or budget allocation.
Once again, those who had received a causal account suggesting that
factors external to the decision maker led to the allocation decision were
much less disapproving of the action and the decision maker.
Folger and his colleagues have obtained conceptually analogous re-
sults. In one study, subjects’ reactions to an inequitable rule change were
assessed (Folger, Rosenfield, & Robinson, 1983). Half were provided
with an adequate explanation of the reasons for the rule change, whereas
half were not. As expected, participants’ feelings of resentment were
considerably lessened when they had been provided an adequate account
 WHEN TO EXPLAIN WHY 391

of why the rule change occurred, particularly when the new rules caused
them to receive lower rewards than they would have received under the
old rules. In a second study (Folger & Martin, 1986), the experimenter’s
actions caused subjects to be denied a favorable outcome. In one con-
dition subjects received good reasons for the experimenter’s actions
whereas in the other they did not. Once again, subjects had much more
favorable reactions (to the experimenter) when they had been provided
a reasonable explanation of the experimenter’s actions. Taken together,
the results of these and other studies (see Bies, Shapiro, & Cummings,
1988; Weiner, Amirkhan, Folkes, & Verette, 1987) suggest that, “a causal
account that claims mitigating circumstances can reduce the perception
of injustice, dampen the feelings of moral outrage, and lessen the erosion
of support for the harmdoer relative to a similar situation in which no
causal account is provided” (Bies, 1987, p. 300).’
Unfortunately, much less empirical research has helped to explain why
the presence of clear and adequate explanations has generally positive
interpersonal consequences. One way to address this matter is by de-
lineating the conditions under which clear and adequate explanations
have favorable effects; put differently, it is important to identify the
factors that moderate the impact of clear and adequate explanations on
interpersonal relations.
It seems likely that the presence or absence of a clear and adequate
explanation will matter more when the recipient of that information is
highly motivated to understand why the allocating party acted as he/she
did. Basic cognitive and social psychological theory and research have
explored the antecedents of individuals’ information-seeking tendencies
(e.g., Berlyne, 1960; Lanzetta & Driscoll, 1968). Individuals who are
motivated to understand why resources were allocated in a particular
way are engaging in a form of information-seeking. Therefore, theory
and research on the antecedents of information search are relevant to
the present analysis in that attributional instigation-the onset of people
asking why-is one form of information search.
Research on information search in general and attributional instigation
in particular have shown that two broad categories of factors affect when
people ask why: (1) uncertainty about why resources were allocated in
a particular way, and (2) the importance of knowing why resources were
allocated in a particular way. Although these two categories may not be
completely independent, they are conceptually somewhat different from
one another. Uncertainty refers to individuals’ inability to make sense
of why resources were allocated in a given way, whereas importance
refers to their motivation to understand the basis of the resource allo-
cation decision.
’ The terms “account,” “causal account.” and “explanation” are being used inter-
changeably in this article.
392 BROCKNER ET AL.

The present study sought to identify when the clarity of managers’

explanations of their organization’s actions is more vs. less related to
the favorability of subordinates’ reactions to the organization’s actions.
The context was a field study exploring the effects of job layoffs on the
work behaviors and attitudes of the employees who were not laid off:
the survivors (e.g., Brockner, 1988). One independent variable was the
extent to which survivors felt that the reasons underlying the layoffs had
been clearly explained to them by management. The chief dependent
variables were survivors’ retrospective reports of how their organiza-
tional commitment, work effort, and turnover intention had changed,
relative to the prelayoff period. We also assessed the factors hypothe-
sized to moderate the relationship between the clarity of management’s
explanation and survivors’ reactions. We now turn to a more complete
discussion of the moderator variables.
The two broad categories of moderator variables-survivors’ uncer-
tainty about, and the importance of knowing, why the layoffs occurred-
each include numerous factors. More specifically, survivors may be un-
certain because the layoff was an unusual event, one that violated em-
ployees’ expectations in light of the typical corporate culture. Further-
more, uncertainty may be aroused to the extent that the process and/or
outcome of the layoff is of questionable legitimacy. If the management
of the layoff is of questionable legitimacy, survivors may become un-
certain about the organization’s management style in the aftermath of
the layoff; moreover, this uncertainty may generalize to other issues of
considerable importance to surviving employees, such as their career
path and the nature of their work. Put differently, layoff management
of questionable legitimacy should make survivors hungry for information
that will enable them to interpret the meaning of the layoff. One poten-
tially relevant source of information is why the layoffs occurred in the
first place.
Five factors believed to reflect uncertainty were included as moderator
variables in the present study. The first one to be described tapped
uncertainty as determined by the perceived unusualness of the layoff,
whereas the remaining four were hypothesized to reflect the questionable
legitimacy of the layoff.
Factors Affecting Uncertainty
We hypothesized that the presence or absence of clear managerial
explanations of the layoffs will have a stronger effect on survivors’ work
behaviors and attitudes when survivors’ level of uncertainty is high rather
than low. The uncertainty variables included:
(I) Unusualness of the layoff. To the extent that the layoff is perceived
to be unusual-i.e., incongruent with the prior management style or
organizational culture-survivors are likely to be uncertain about why
 WHEN TO EXPLAIN WHY 393

the layoff occurred. Layoffs perceived to be unusual violate survivors’

expectations about how the organization is likely to act; as attribution
research has shown (Pyszczynski & Greenberg, 1981; Wong & Weiner,
1981), it is especially in the face of unexpected outcomes that individuals
seek to understand the causes of those outcomes. Survivors should have
a relatively high need to know why layoffs deemed to be unusual oc-
curred; therefore, the presence or absence of clear explanations from
management should be more likely to have a significant impact on sur-
vivors’ reactions when the layoffs are perceived to be unusual.
(2) Avoidability of the layoff. To the extent that the layoff is perceived
to be avoidable, survivors may question its legitimacy. The perception
of avoidability implies that the layoff did not have to happen, which
could make survivors uncertain about why it did. More generally, sur-
vivors also may begin to wonder whether the organization will take other
actions of questionable legitimacy in the future.
(3) Lack of clarity of the decisional basis to keep certain employees
and lay off others. Once an organization implements layoffs, it has a
number of options in deciding whom to keep versus lay off (e.g., sen-
iority, merit). If survivors are unclear about the decision basis that was
used, they may question the legitimacy of the layoff management process;
moreover, their lack of clarity about the decisional basis may suggest
that they are also uncertain about whether the organization will act
legitimately in the ways it handles its future affairs.
(4) Unfairness of the decisional basis to keep certain employees and
lay off others. It is one thing for the decisional basis to be clear; it is
quite another, however, for it to be fair (although the two certainly can
be related). If the decisional basis is clear, then management probably
has done an effective job in communicating the content of the decision
basis. However, the content itself may or may not be fair in the minds
of survivors. If the decisional basis is deemed to be unfair, survivors
are likely to be uncertain about the legitimacy of the layoff, and, more
generally, whether management can be trusted to act fairly in the future.
(5) Perceived adequacy of organizational caretaking in the event of
future [Link] vary considerably in the amount of care-
taking that they provide for the layoff victims. Some offer generous
severance packages and/or considerable provisions to help those laid
off find new jobs, whereas others do not. Suppose that survivors believed
that the layoff victims were not well taken care of during the layoff
process, and that by extension, the survivors would not be well provided
for as well. If so, survivors are more likely to question the fairness or
legitimacy of the layoff (Brockner, Grover, Reed, Dewitt, & O’Malley,
1987).
In summary, it was predicted that survivors would report having re-
acted more negatively-i.e., their organizational commitment and work
394 BROCKNER ET AL,

effort will be lower and their turnover intention higher-to the extent
that they did not feel that the reasons for the layoff had been clearly
explained. Moreover, these predicted effects should be stronger when:
(1) the layoff was viewed to be more unusual, (2) the layoff was believed
to be more avoidable, (3) the decisional basis was seen as more unclear,
(4) the decisional basis was perceived to be more unfair, and (5) the
perceived adequacy of organizational caretaking in the event of future
layoffs was low.
Factors Affecting Importance
The presence or absence of clear managerial explanations of the layoff
should be more strongly related to survivors’ reactions to the extent that
survivors attach personal significance to the layoff. Two variables related
to personal significance were explored in the present study:
(1) Perceived likelihood of additional layoffs. If survivors believe that
layoffs are likely to recur in the near future, then they should be especially
motivated to understand why the layoffs occurred the first time. Such
reasons might provide clues about whether they will be personally af-
fected by future cutbacks.
(2) Prior attachment to the layoff victims. The discussion of the pre-
vious factor implies that survivors will view layoffs as more personally
significant if the layoff threatens their sense of economic well-being. We
also believe that interpersonal variables affect the personal significance
that survivors ascribe to the layoff (e.g., Brockner et al., 1987). Specif-
ically, to the extent that survivors felt attached to at least some of the
layoff victims prior to the layoff--e.g., suppose that survivors worked
closely with the layoff victims for many years, or socialized with them
during the lunch hour or outside of the work setting-they are more
likely to feel personally affected by layoff, in turn, such survivors will
be more motivated to know why the layoffs occurred.
We predict that survivors’ perceived likelihood of layoffs in the near
future, and their prior attachment to the layoff victims will moderate the
relationship between the perceived clarity of management’s explanations
of the layoff and survivors’ reactions to the layoff. Although survivors
generally are expected to react more negatively when managerial ac-
counts are unclear rather than clear, this should be especially true when:
(1) the perceived likelihood of layoffs in the near future is relatively high,
and (2) survivors felt relatively attached to the layoff victims prior to
the layoff.
METHOD
Participants
Participants were drawn from a chain of small retail stores throughout the United States.
Three hundred stores were randomly selected from the store population (N = 773) for
 WHEN TO EXPLAIN WHY 395

participation in the study. Questionnaires were sent to 1602 individuals. The 597 respond-
ents-reflecting a response rate of 37.3%-were predominantly female (91%). white
(83.5%), and married (70%). Average age was 37.06 years (SD = 8.59). They were either
sales clerks (74%) or store managers and had a mean tenure of 4.13 years (SD = 2.3)
with the company. Given that many stores in the chain had been closed during the previous
12-month period (in which thousands of employees had lost their jobs), all employees in
the stores that remained open could be considered survivors of the layoff.

Procedure
Each store in which at least one person returned the questionnaire employed 2-11
employees (M = 5.8, SD = 1.47). To heighten control of the survey administration process,
we contacted each district manager by telephone to determine the exact number of em-
ployees at each store: this practice ensured that the proper number of surveys were sent
to each store. To maintain anonymity, subjects were asked not to write their names
anywhere on the survey.
The store manager of each targeted store received a packet that included a letter from
the researchers urging participation, a letter from a top manager in the organization also
requesting compliance, one survey for each employee and the store manager, and stamped
envelopes addressed to the university address of the principal researcher. Subjects were
instructed to return their completed surveys directly to the principal researcher, not to
their superiors in the organization.
All of the sampled stores were sent a follow-up letter several weeks after the surveys
were mailed that urged participation from those who had not already returned question-
naires. Not all subjects completed all of the measures: hence, the number of observations
for each measure was a bit lower than the size of the entire sample.

Dependent Variables
For each dependent variable respondents were asked to report to what degree each item
applied to them before relative to after the layoffs. Specifically, subjects were instructed
to indicate their opinion “now in comparison to how you felt one month prior to first
hearing that there would be layoffs.” Eleven-point scales were used, the endpoints of
which were labeled “applied to me more before the layoff than now” and “applies to me
more now than before the layoff.” The midpoint was labeled “applies to me the same.”
Measures included:
(I) Organiza~iuna/ commifment. The scale was an 18-item inventory developed by Schwy-
hart and Smith (1972), and measures attributes such as company identification and pride.
Sample items include “I like working for this company,” and “I am proud to tell my
friends I work for this company,” (coefficient alpha = .95)
(2) Work effort. We constructed a 3-item index, including such statements as, “I try to
work as hard as possible” (coefficient alpha = .87).
(3) Turnover intention. The single item measure which we developed was, “I have every
intention of continuing to work in this organization-rather than a different organization
or job-for the foreseeable future.” (Note that this variable was reverse-scored.)

Independent Variables
All of the independent variables were measured with the use of 7-point scales.
Managerial accounts of the layoff. Two items measured the clarity of managerial ex-
planations or accounts: “How clearly were the reasons for the layoff explained to you by
(top management) (your supervisor)?” Scale endpoints were “not at all clearly” (1) and
“very clearly” (7). The two measures were highly related, r(584) = .69. p < .OOl, and
summed into an index.
396 BROCKNER ET AL.

Unusualness of the layoff. The term “organizational culture” was defined for all subjects
as “rules, procedures, norms, and values for conducting business.” Three highly related
items (coefficient alpha = 32) tapped unusualness, including, “In light of the usual culture
at (name of organization), how unusual was it for there to be layoffs?” Endpoints were
“not at all unusual” (1) and “very unusual” (7). The three items were summed into an
index,
Avoidubiliry. The single item measure was, “How avoidable were the layoffs” Endpoints
were “not at all avoidable” (1) and “very avoidable” (7).
Unclarify of decision rule; fairness of decision rule. Subjects were asked to rate how
important various criteria were in management’s decision to close certain stores and keep
others open (e.g., productivity, sales volume, seniority, politics, etc.). The lack of clarity
item was “No clear decision criteria (I do not know how it was decided that certain stores
would be closed and others kept open).” Endpoints were “not at all” (1) and “very much”
(7).
The fairness measure was, “How fair were the criteria used to decide which stores
would be closed and which would remain open?” Endpoints were “not at all fair” (I) and
“very fair” (7).
Adequacy of organizational caretaking. Subjects had just completed several measures
asking them to rate how well they believed the organization had provided for the layoff
victims, (i.e., in the form of severance pay and assistance in finding a job elsewhere).
Next, subjects were asked: “Suppose you were to be laid off. If so, how well do you
think the assistance that management offered the laid-off people would provide for your
needs?” Endpoints were “not at all” (1) and “very much” (7).
Likelihood of fufure layoffi. The single item measure was “To what extent do you
believe more layoffs in (the organization) are likely to occur in the near future?” Endpoints
were “not at all likely” (1) and “very likely” (7).
Prior attachment lo layoff vicfims. Subjects rated the extent to which they agreed with
the following two statements: I had a close (working) (personal) relationship with at least
some of the laid off people. The two items were highly related, r(579) = .79, p < .OOl,
and summed into an index.

RESULTS
Intercorrelations between all of the independent and dependent vari-
ables are presented in Table 1. Several aspects are worth noting. First,
four of the five uncertainty items (avoidability, unclarity of the decision
rule, fairness of the decision rule, and adequacy of caretaking) tended
to be significantly related to each other. The absolute value of the average
correlation among these four measures was .22, p < .OOl; moreover,
five out of the six possible correlations were significant (r at least .21).
Second, the unusualness measure was uncorrelated with these (and all
other) independent variables, suggesting that different types of uncer-
tainty existed. Third, the two measures of importance (likelihood of
future layoffs and prior attachment) were only weakly related (r = .09),
suggesting that different types of importance existed. Fourth, the un-
certainty and importance items generally were not significantly correlated
with each other (only 2 out of the 10 possible relationships were signif-
icant). Fifth, the dependent variables were significantly correlated (ab-
solute values of the correlations ranging from .38 to .50); however the
 TABLE I
SUMMARY STATISTICS FOR AND INTERCORRELATIONS BETWEEN MAIN VARIABLES
-~ -
Independent variables Dependent variables

1. Managerial accounts of layoffs (M = 7.05: SD = 3.86) 9. Organizational commitment (M = 88.95; SD = 34.56)

2. Unusualness of layoffs (M = 15.59; SD = 4.96) 10. Work effort (M = 20.30; SD = 5.77)
3. Avoidability (M = 4.50; SD = 1.73) 11. Turnover intention CM = 6.03: SD = 2.74)
4. Unclarity of decision rule (M = 3.72; SD = 2.08)
5. Fairness of decision rule (M = 3.92: SD = 1.68)
6. Likelihood of future layoffs (M = 6.21: SD = 1.29)
7. Adequacy of caretaking (M = 2.89: SD = 1.76)
8. Prior attachment to victims (M = 8.43; SD = 4.55)

I 2 3 4 5 6 7 8 9
.___
1
2 .oo
3 - .29*** .06
4 - ,27*‘* - .Ol .22***
5 .42*** .oo - .36*** - .21***
6 - .07 .04 - .03 .04 - .14***
7 .25*“* - .05 - .24*** - .03 .25*** - .18***
8 -.06 .oo .04 - .02 -.08 .09* -.02
9 .31*** -.I]* - .26*** - .04 .33*** - .25*** .37*** - .17***
10 lo* .w -.06 .04 - .16** .09* lo*** - .04 .38***
II -.10* .03 .02 - .03 -.12** .21*** - .14*** .07 - .49*** - .50***
Note. For all measures except turnover intention. higher scores reflect greater presence of the attribute.
* p < .05.
** p < .Ol.
*** p < ,001.
398 BROCKNER ET AL.

dependent variables were not so highly related with each other to suggest
that they were all measuring the same underlying construct.
The central hypotheses were that the accounts variable would be pos-
itively correlated with change in organizational commitment and work
effort, and inversely related to change in turnover intention. Moreover,
these predicted effects were supposed to be moderated by the various
uncertainty and importance factors. Of greatest interest, therefore, were
the main effects for the accounts variable, and the interaction effects
between the accounts variable and the moderator variables.
Main effects. Several findings suggested that survivors reacted more
negatively to the extent that they perceived management’s causal ex-
planations of the layoff to be unclear. First, the correlations reported in
Table 1 demonstrated that the relationships between managerial accounts
and each of the three dependent variables was significant, particularly
in the case of organizational commitment.
Second, we performed a multiple regression in which the accounts
variable and each of the seven moderator variables were entered si-
multaneously. The accounts variable yielded a significant effect on the
measure of organizational commitment, F(1, 500) = 10.38, /3 = 1.28,
p < .Ol; as expected, the greater the clarity of the accounts of the layoff,
the more positive was their change in organizational commitment. The
accounts variable was not related, however, to work effort or turnover
intention in these regression analyses.
Interaction effects. We performed a series of multiple regressions, in
which the three dependent variables were analyzed as a function of the
accounts factor and the seven moderator variables. The accounts variable
was paired separately with each of the moderator variables. More spe-
cifically, the main effect for the accounts variable, the main effect for
the moderator variable, and the interaction between the two were entered
simultaneously. This procedure was followed 21 times, allowing for all
combinations of the accounts variable with each of the seven moderator
variables (across the three dependent variables).
Of greatest importance in these regression analyses are the interaction
effects. As can be seen in Table 2, of the 21 hypothesis tests, 12 were
significant (p < .05 at least), and one was marginally significant. More-
over, the sign of the beta weights reveal that all of these effects were
in the predicted direction. Thus, the relationship between clarity of ac-
counts and positive change in organizational commitment was more pro-
nounced to the extent that: (1) the layoff was seen as avoidable, (2) the
decision rule was perceived to be unfair, (3) layoffs were viewed as likely
to recur, (4) caretaking in the event of additional layoffs was judged to
be inadequate, and (5) survivors felt attached to at least some of the
layoff victims prior to the layoff.
The relationship between accounts clarity and positive change in work
 TABLE 2
BETA WEIGHTS AND p VALUES FOR INTERACTIONS EFFECTS INVOLVING THE ACCOUNTS VARIABLE

Organizational Work Turnover

commitment effort intention

beta P bega P beta

Accounts x avoidability 0.46 .02 0.04 .I.5 0.03 .04 2

Accounts x unusualness 0.03 .65 0.03 .03
Accounts x lack of clarity 0.24 .I8 0.06 .04 0.01 .Ol E
0.03 .02
Accounts x fairness - 0.48 .03 -0.07 .07 -0.04 .03 f-
Accounts x future likeli- 2
hood of layoffs 0.93 .Ol 0.02 .67 0.03 .19
Accounts x adequacy of
caretaking -0.55 .Ol - 0.05 .18 -0.04 .02
Accounts x prior
attachment 0.25 .Ol 0.01 .53 0.01 .23

Note. All beta weights are unstandardized.

400 BROCKNER ET AL.

effort was greater to the extent that: (1) the layoff was rated as unusual,
(2) the decision rule was perceived to be unclear, and (3) the decision
rule was seen as unfair (p < .07). Finally, the negative relationship
between accounts clarity and turnover intention was more marked to
the extent that: (1) the layoff was avoidable, (2) the layoff was unusual,
(3) the decision rule was unclear, (4) the decision rule was unfair, and
(5) organizational caretaking in the event of future layoffs was inadequate.
Participants in this study were of two organizational ranks: store man-
agers (26%) or sales clerks (74%). To evaluate the possible role of this
variable we performed several types of analyses. First, we repeated all
of the tests of the interaction effects reported in Table 2, controlling for
rank. More specifically, we entered simultaneously the main effect for
the accounts variable, the main effect for the moderator variable. the
interaction between the two, and the main effect for rank (coded as a
dummy variable). The results were virtually identical to those reported
in Table 2. Of the 12 significant interaction effects (p < .05 or better)
in Table 2, 11 remained significant. The lone exception was the Accounts
x Avoidability interaction on the organizational commitment measure,
in which the effect was marginally significant @ < .06). Moreover, all
of these effects were in the predicted direction. On the work effort
measure the significance level of the Accounts x Fairness interaction-
previously .07-slipped to .14. However, the p value of the Accounts
x Adequacy of Caretaking interaction-previously .18-rose to .08.
Second, we reanalyzed the results in Table 2 to determine whether
the findings were more strongly observed with one group rather than the
other. Specifically, we entered simultaneously the main effect for the
accounts variable, the main effect for the moderator variable, the main
effect for rank, all two-way interactions among these three variables,
and the triple interaction. As in the analyses reported in Table 2, this
procedure was performed 21 times. Of greatest interest is the triple
jnteraction effect. Only one was significant at the .05 level (Accounts x
Future Likelihood of Layoffs x Rank on the measure of turnover in-
tention), and two were marginally @ < .08) significant (Accounts x
Adequacy of Caretaking x Rank on the measure of organizational com-
mitment and Accounts x Culture x Rank on work effort). These triple
interaction effects suggested that the two-way interaction effects de-
scribed in Table 2 were more pronounced among the sales clerks than
the store managers. However, the most accurate summary of these tests
for triple interaction effects is that the two groups-sales clerks and store
managers-did not significantly differ from one another in the extent to
which they yielded the results described in Table 2.
Recall that several of the moderator variables were singie-item mea-
sures with unknown reliabilities. Moreover. several of them correlated
with each other, suggesting that some of the significant interaction effects
 WHEN TO EXPLAIN WHY 401

in Table 2 simply were repeating each other. Therefore, we performed

a second series of multiple regression analyses. First, based upon the
results of the correlation matrix (see Table l), we combined the uncer-
tainty items that correlated significantly with one another into an index.
Specifically, the avoidability, lack of clarity, fairness, and adequacy of
caretaking items were summed into an index (after the last two terms
were reverse scored). The other three moderator variables-unusualness,
likelihood of future layoffs, and prior attachment-were generally un-
correlated with all other moderator variables.
Consequently, we simultaneously entered the four-item uncertainty
index, each of the other three moderator variables, and the accounts
variable as independent variables in three separate multiple regressions,
aimed to predict each of the three dependent variables (organizational
commitment, work effort, and turnover intention). First, the main effects
were entered simultaneously. The accounts variable was significant in
the case of organizational commitment (p < .OOl), but not in the instance
of work effort and turnover intention (both p values > .15). Then the
main effects and the two-way interactions of the accounts variable with
each of the uncertainty index, unusualness, likelihood of future layoffs,
and prior attachment factors were entered simultaneously.
The interaction effect results are summarized in Table 3. As in the
separate analyses presented in Table 2, many of the interaction effects
(6 out of 12) are significant (at least at the .05 level); one other is mar-
ginally significant. All are in the predicted direction. Moreover, the re-
sults representing the five interaction terms that were not significant all
were in the predicted directions (i.e., the sign of the beta weights was
always positive).’

DISCUSSION
In summary, the results generally confirmed predictions. Two broad
categories of factors moderated the positive relationship between the
clarity of managerial accounts and the favorability of survivors’ reactions:
(1) survivors’ uncertainty about why the layoff occurred (e.g., whether
the layoff was inconsistent with the corporate culture), and (2) the im-
portance of layoff-associated outcomes (e.g., whether layoffs are likely

’ A factor analysis of the seven moderator variables also suggested the appropriateness
of the four moderator variables employed in the analyses in Table 3. More specifically,
neither the unusualness nor the prior attachment variable loaded onto any factors in two-
factor and three-factor models. Hence, we dropped these two terms and then performed
a factor analysis of the remaining five variables. In the emerging two-factor model, all of
the measures except likelihood of future layoffs loaded more heavily on the first than the
second factor (minimum loading of .31, maximum of 66). The likelihood measure was the
only one to load substantially t.61) on the second factor. We thank Lori Bartner for her
assistance in performing the factor analysis.
 TABLE 3
MULTIPLE REGRESSION RESULTS BASED UPON ACCOUNTS, UNCERTAINTY INDEX, LIKELIHOOD OF FUTURE LAYOFFS, UNUSUALNESS,
AND PRIOR ATTACHMENT

Organizational Work Turnover

commitment effort intention

beta beta beta m

P P P
Accounts 1.36 .ool 0.10 .I9 0.04 .26 E
Uncertainty index -2.01 .ool -0.09 .17 -0.03 .37 2
Layoff likelihood - 4.55 ,001 -0.26 .I9 -0.36 a01 E
Unusualness -0.58 .04 -0.01 .88
-0.01 .58 2
Prior attachment - 1.02 .OOl - 0.04 .44 - 0.03 .20
Accounts x uncertainty 0.16 .03 0.03 .03 0.02 .Ol F
Accounts x layoff likelihood 0.65 .02 0.01 .89 0.01 .58
Accounts x unusualness 0.04 51 0.02 .09 0.01 .Ol
Accounts x prior attachment 0.20 .02 0.01 .40 0.01 .24
Overall R' = .24, p < .OOl Overall R' = .04. p < .02 Overall R' = .07,p < 301

Note. Beta and p values for main effects are based upon the absence of interaction effects from the regression equation; beta and p values for
interaction effects are based upon the simultaneous entry of all main and interaction effects. All beta weights are unstandardized.
 WHEN TO EXPLAIN WHY 403

to occur again). Taken together, the results suggest that managerial ac-
counts bear a stronger relationship to survivors’ reactions when survivors
are relatively uncertain about why the layoffs occurred, and when sur-
vivors attach greater significance to the outcomes of the layoff.
Theoretical Implications
Managerial accounts. Recent research has shown that employees are
much more accepting of undesirable resource allocations to the extent
that they have received adequate explanations of the reasons for the
resource allocations (Bies et al., 1988; Rousseau & Anton, 1988). The
present findings may help explain why adequate managerial accounts
soften the blow of unfavorable resource allocations, by delineating some
conditions under which accounts are more vs. less likely to relate to
employee reactions. We hypothesized that when individuals’ needs for
information are high-either because they lack the ability to answer
questions on their own, or because the resource allocation decision holds
important outcomes for them-they are particularly likely to search their
environment. When that environment satisfies (fails to satisfy) their in-
formation needs, i.e., when management offers (withholds) clear expla-
nations of why the layoffs occurred, their reactions are more positive
(negative). Furthermore, when employees’ information needs are rela-
tively low, the presence or absence of clear managerial accounts is much
less strongly related to employee reactions.
The present findings suggest that employees may be uncertain about
several aspects of their work environment; moreover, they may attach
importance to events in their work setting for more than one reason.
Thus, our participants may have been uncertain about why the layoff
occurred in the first place, especially if they viewed it as unusual in light
of the existing corporate culture. In addition, they could have been
uncertain about the legitimacy of the layoff management process, par-
ticularly if they believed that the layoff could have been avoided, and/or
if they believed that the decision rule used was unclear or unfair. The
results of the correlation matrix and factor analysis, moreover, suggested
that these are two largely unrelated types of uncertainty. However, each
is likely to put surviving employees in need of information that would
help them interpret the meaning of the layoff; one such source of in-
formation is why the layoffs occurred.
Similarly, the importance of knowing why the layoffs occurred may
stem from a number of largely unrelated sources. The present findings
suggest that survivors’ beliefs about the likelihood of additional layoffs,
as well as their prior attachment to the layoff victims are two such
determinants of perceived importance. Each is likely to motivate sur-
vivors to know why the layoffs occurred. More generally, the present
findings identify the general categories of factors-uncertainty and im-
404 BROCKNERETAL.

portance-that are likely to moderate the effects of accounts on indi-

viduals’ reactions to resource allocation decisions. A challenge for future
research is to delineate the (likely myriad of) factors that affect uncer-
tainty and importance.
Attribution theory. Attribution theorists posit that individuals make
sense of their worlds through the causal explanations they make for their
own and others’ behavior (Kelley, 1973). Therefore, an important task
is to specify the conditions under which people instigate the attribution
process, i.e., when is it that individuals ask why? Several studies have
shown that the attribution process is invoked in the face of negative
and/or unexpected outcomes (Pyszczynski & Greenberg, 1981; Wong &
Weiner, 1981). The present results are consistent with those found earlier.
For example, survivors are likely to view the layoff as more negative
when it is expected to recur, and when they felt attached to the layoff
victims. Moreover, the layoff is likely to be seen as more unexpected
or unusual to the extent that it violated the corporate culture. The pros-
pect that individuals are most likely to initiate the attribution process in
response to negative and/or unexpected outcomes implies the following:
that in the face of negative and/or unexpected outcomes individuals are
especially apt to be influenced by the presence or absence of information
that will help them understand why such outcomes occurred. Of course,
such results are precisely those that were obtained in the present study.
Slarvivors’ reactions. Given that layoffs are a pervasive organizational
cost-cutting strategy (Greenhalgh, Lawrence, & Sutton, 1988), it is both
practically and theoretically important to discover the effects of layoffs
on survivors. Previous research has shown that survivors’ reactions to
layoffs are quite variable; Brockner (1988), for example has shown that
survivors may work harder, less hard, or no differently, depending upon
certain situational and dispositional factors. Theory and research needs
to explain the variability in survivors’ reactions. The present and past
research suggest that survivors are apt to react more negatively to the
extent that they believe that the layoff was managed unfairly. Numerous
factors may affect survivors’ perceptions of how fairly the layoff was
managed, including whether and how well the organization offered pro-
visions to the layoff victims (e.g., severance pay, outplacement coun-
seling; Brockner et al., 1987), and, in the present study, how clearly
management explained the reasons for the layoff. Moreover, the present
study helped identify some conditions under which the fairness of the
layoff management, i.e., the extent to which the reasons for the layoff
were clearly explained, was more versus less likely to be associated with
survivors’ reactions.
Practical Implications
Theory and research which identify factors related to survivors’ re-
actions may help managers handle layoffs in ways that elicit the most
 WHEN TO EXPLAIN WHY 405

positive (or least negative) reactions among the remaining workforce.

Let us assume that the presence of clear managerial accounts of the
layoff not only predicted, but also causally affected survivors’ reactions
to be more favorable. If so, then it is important for downsizing orga-
nizations to ensure that its remaining workforce is clearly told why the
organization has chosen to lay off employees. Moreover, our findings
help specify when it is especially important for management to provide
clear explanations of the layoff, e.g., when the layoff is an unexpected
event, and/or when the layoff has important outcomes for the survivors.
A key practical matter, of course, is how management may offer such
accounts so as to produce the most beneficial survivor reactions. What
exactly should management say? Who should do the explaining? When
should the accounts be offered? These are some of the difficult questions
that must be addressed before the present theorizing can be translated
into actual managerial practice.

Limitations/Future Research
The present study suffers from a number of shortcomings, each of
which needs to be addressed in further research. First, the design of the
study was correlational, making it more difficult for us to conclude that
the independent variables had a causal impact on the dependent variables.
It would be useful to manipulate the independent variables in a future
research study-be it in the laboratory or the field-so that their causal
impact could be evaluated more fully.
Second, the construct validity of the moderator variables are open to
question. We hypothesized that in the face of uncertainty-i.e., about
why the layoffs occurred, or uncertainty concerning the legitimacy of
how the layoffs were managed-and in the face of high importance lay-
offs, survivors would be especially influenced by the presence or absence
of clear managerial accounts for the layoffs. However, uncertainty and
importance were not measured directly. Instead, some of their anteced-
ents were. It is likely that the independent variables were at least some-
what valid, in that uncertainty and importance were operationalized in
different ways. Nevertheless, it is not entirely clear that our interpretation
of the constructs underlying the moderator variables was the most
appropriate.
Third, future research needs to explore the boundary conditions for
the positive relationship between managerial accounts and employee re-
actions. In the present study we have suggested that it is the clarity of
managerial explanations that leads to more accepting reactions on the
part of subordinates. Previous research has shown that it is the adequacy
of managerial explanations that facilitate subordinate acceptance (Bies
et al., 1988). Disentangling these two attributes of explanations may be
somewhat problematic, as it has been shown that the correlation between
406 BROCKNER ET AL.

the perceived clarity and adequacy of explanations is quite high (i.e.,

Konovsky and Folger, 1989, obtained a correlation of .82 in their field
study of layoff victims).
Fourth, although the present findings suggest that managerial accounts
(and the moderator variables) account for some of the variation in sur-
vivors’ reactions, we do not mean to imply that other factors are not
equally (if not more) important determinants of survivors’ reactions. In
fact, many of the significant interaction effects reported in Table 2 ac-
counted for a statistically significant, albeit small, amount of variance.
Thus, it is important to delineate other factors that explain the variation
in survivors’ reaction. For example, Brockner and Greenberg (1990) have
suggested that survivors’ reactions are mediated by their perceptions of
how fairly the layoff was managed, on both distributive and procedural
grounds. Numerous factors are likely to influence perceived fairness,
including the presence of managerial explanations of the reasons for the
layoff (Rousseau & Anton, 1988), the amount of caretaking that the
organization provides to the layoff victims (Brockner et al., 1987), the
degree of advanced warning that the organization offers, and the extent
to which survivors blame the environment (rather than the organization)
for the occurrence of the layoff. Moreover, factors other than perceived
justice are likely to mediate the effects of layoffs on survivors. For
instance, the extent to which the layoff: (I) threatens survivors’ job
security, and (2) leads to a change in survivors’ work responsibilities or
conditions should also influence survivors’ reactions to the layoff.
In conclusion, the present study builds bridges between several areas
of inquiry in social and organizational psychology. Based on fundamental
principles of information search and attributional instigation, we delin-
eated some of the conditions under which managerial accounts are more
versus less strongly associated with individuals’ work attitudes and be-
haviors. Furthermore, the present findings add to the growing literature
on survivors’ reactions (Brockner, 1988). Finally, the discussion of the
limitations of the study identifies ways in which the present study may
be refined or extended on both methodological and conceptual grounds.

REFERENCES
Berlyne, D. E. (1960). Conflicr, arousal. rind curiosity. New York: McGraw-Hill.
Bies, R. J. (1987). The predicament of injustice: The management of moral outrage. In
L. L. Cummings & B. M. Staw (Eds.), Research in orgnniznriond behavior (Vol. 9,
pp. 289-319). Greenwich, CT: JAI Press.
Bies, R. J.. & Shapiro, D. L. (1987). Interactional fairness judgments: The influence of
causal accounts. Social Justice Research. 1, 199-218.
Bies, R. J.. Shapiro, D. L., & Cummings, L. L. (1988). Causal accounts and managing
organizational conflict. Communication Research. 15, 381-399.
Brockner. J. (1988). The effects of work layoffs on survivors: Research. theory, and
 WHEN TO EXPLAIN WHY 407

practice. In B. M. Staw & L. L. Cummings (Eds.), Research in organizational behavior

(Vol. 10, pp. 213-255). Greenwich, CT: JAI Press.
Brockner, J., & Greenberg, J. (1990). The impact of layoffs on survivors: An organizational
justice perspective. In J. Carroll (Ed.), Applied social psychology and organizational
settings (pp. 45-75).Hillsdale, NJ: Erlbaum.
Brockner, J., Grover, S., Reed, T., Dewitt, R., & O’Malley, M. (1987). Survivors’ reactions
to layoffs: We get by with a little help for our friends. Administrative Science Quarterly,
32, 526-541.
Folger, R., & Martin, C. (1986). Relative deprivation and referent cognitions: Distributive
and procedural justice effects. Journal of Experimental Social Psychology, 22, 531-
546.
Folger, R.. Rosenfield, D., & Robinson, T. (1983). Relative deprivation and procedural
justifications. Journal of Personality and Social Psychology, 45, 268-273.
Goffman, E. (1959). Thepresentation of selfin everyday lifp. Garden City, NY: Doubleday,
Anchor Books.
Greenberg, J. (1987). A taxonomy of organizational justice theories. Academy of Man-
agement Review, 12, 9-22.
Greenhalgh, L., Lawrence, A. T., & Sutton, R. 1. (1988). Determinants of work force
reduction strategies in declining organizations. Academy of Management Review, 13,
241-254.
Kelley, H. H. (1967). Attribution theory in social psychology. In D. Levine (Ed.), Nebraska
symposium on motivation (Vol. 15, pp. 192-238). Lincoln: University of Nebraska
Press.
Kelley. H. H. (1973). The process of causal attribution. American Psychologist, 28, 107-
128.
Konovsky. M., & Folger, R. (1989). The effectiveness of procedural justice, social ac-
counts, and adequacy of explanation in mitigating the consequences of involuntan,
job loss. Manuscript under editorial review.
Lanzetta, J. T., & Driscoll, J. M. (1968). Effects of uncertainty and importance on infor-
mation search in decision making. Journal of Personality and Social Psychology, 10,
479-486.
Pyszcznski, T. A., & Greenberg, J. (1981). Role of disconfirmed expectancies in the
instigation of attributional processing. JournaL of Personality and Social Psychology,
40, 31-38.
Rousseau, D. M., & Anton, R. J. (1988). Fairness and implied contract obligations in
termination: A policy capturing study. Human Performance, 1, 273-289.
Schwyhart, W. R., & Smith, P. C. (1972). Factors in the job involvement of middle
managers. Journal of Applied Psychology, 56, 227-233.
Thibaut, I., & Walker, L. (1975). Procedural justice: A psychological analysis. Hillsdale,
NJ: Erlbaum.
Tyler, T. R. (1990). Why people obey the law. New Haven: Yale University Press.
Weiner, B., Amirkhan, J., Folkes, V. S., & Verette, J. A. (1987). An attributional analysis
of excuse-giving: Studies of a naive theory of emotion. Journal of Persona&y and
Social Psychology, 52, 3 16-324.
Wong, P. T. P., & Weiner, B. (1981). When people ask “why” questions. and the heuristics
of attributional search. Journal of Personality and Social Psychology, 40, 650-663.