Full text of DOI:10.1177/1094428102005002003

OSaRnGchAeNz,IVZiAsTwIeOsvNaAraLnR/ ETeSmEApoRraClHSeMpaErTatHioOnDS
The Effects of Temporal Separation on
the Relations Between Self-Reported
Work Stressors and Strains
JUAN I. SANCHEZ
CHOCKALINGAM VISWESVARAN
Florida International University
The assumption that a longitudinal design reduces the size of the relationships be-
tween work stressors and strains was tested. A comparison of cross-sectional and
longitudinal correlations suggested that temporal separation of stressors and
strains reduced correlations between role ambiguity and primarily attitudinal
strains. Cross-sectional correlations between role conflict and strains, as well as
correlations involving health-related strains, were practically unaffected by switch-
ing to a longitudinal design. A theory-driven approach to the adoption of temporal
separation of stressor and strain measures in organizational stress research is
outlined.
The empirical association between employee stressors and strains is extensively docu-
mented (Fisher & Gitelson, 1983; Jackson & Schuler, 1985). However, most of these
studies have relied on cross-sectional designs and mono-method self-reports of both
stressors and strains (Jackson & Schuler, 1985, p. 27). This overreliance on all-self-
report, cross-sectional designs is not surprising, since sources of data other than self-
reports are often unavailable or difficult to obtain in organizational research, espe-
cially when designs require filling out a survey at more than one point in time (Avolio,
Yammarino, & Bass, 1991; Campbell & Fiske, 1959; Gupta & Beehr, 1982; Schmitt &
Stults, 1986).
Ganster, Fusilier, and Mayes (1986, p. 104) noted that the widespread practice of
gathering self-reports of both stressors and strains on the same questionnaire adminis-
tration can exacerbate common method variance, thereby rendering flawed accounts
of the true association between these variables. However, stressors and to some extent
strains represent perceptions of an objective reality rather than the reality itself. Thus,
it is unknown whether the reduced correlations that may be observed in longitudinal
designs are a better or worse representation of the “true” stressor-strain relationship.
One may even argue that the ability to make comparative judgments among items
increases the validity of self-reports, and that such ability is hampered by a separation
of the assessments over time. In fact, it has been argued that method variance can
sometimes attenuate rather than increase correlations (Williams & Brown, 1994).
Organizational Research Methods, Vol. 5 No. 2, April 2002 173-183
© 2002 Sage Publications
173
Downloaded from orm.sagepub.com at Kungl Tekniska Hogskolan / Royal Institute of Technology on March 8, 2015

174 ORGANIZATIONAL RESEARCH METHODS
Spector and Brannick (1995) specifically noted that if method components are
uncorrelated (between measures of two traits), method variance will function as error
variance, thereby attenuating discriminant (i.e., multitrait) correlations.
Despite the philosophical and the measurement-related difficulties inherent in
ascertaining the “true” stressor-strain relationship, it is still important to determine
whether and when longitudinal designs provide estimates of stressor-strain relations
that differ from those provided by cross-sectional designs. There may be time varia-
tions in the onset of different stressors, and also differences in the onset of the effects of
the same stressor on different strains. Ignoring such time variations may lead investi-
gators to miss important stressor-strain relationships, and organizations to draw the
wrong conclusions about the effectiveness of stress-related interventions. Unfortu-
nately, organizational researchers often leave the decisions about when to measure
variables to convenience or tradition (Mitchell & James, 2001). The aim of the study
described herein was to evaluate the extent to which stressor-strain correlations are
changed by the employment of a longitudinal design. A comparison of synchronous
and cross-lagged or panel correlations (Cook & Campbell, 1979, pp. 309-321; Kenny,
1975) allowed us to test this research question.
In the work stress domain, the extent to which employee reports of theoretically dis-
tinct stressors and strains share common method variance has been debated (Spector,
1987; Spector, Van Katwyk, Brannick, & Chen, 1997; Williams, Cote, & Buckley,
1989). A rather prolific ramification of this debate has questioned whether a third vari-
able, namely negative affectivity, might bias relationships between self-reported
stressors and strains (Brief, Burke, George, Robinson, & Webster, 1988; Burke, Brief,
& George, 1993; Chen & Spector, 1991; Parkes, 1990; Schaubroeck, Ganster, & Fox,
1992; Spector, Zapf, Chen, & Frese, 2000; Watson, Pennebaker, & Folger, 1986; Wil-
liams, Gavin, & Williams, 1996). This debate is still alive, with one side arguing that
negative affectivity has a biasing effect on the stressor-strain relation whereas the other
argues for a substantive role of negative affectivity.
Common method bias, however, can exercise its effects in ways other than through
third variables such as negative affectivity. Specifically, the employment of the same
method of data collection for both measures at the same time is likely to exacerbate the
effects of artifacts such as consistency of responses, priming, and even attribution
effects. For instance, having just reported a relatively high level of role conflict may
compel one to remain consistent and therefore report a similarly high level of work
anxiety on the same survey. In addition, the widely held belief that stressful work trig-
gers health problems may lead employees to attribute any health problems they might
be experiencing to their work, therefore reporting higher levels of work stressors than
they would report had they not been questioned about health problems first.
The suggestion that a longitudinal design may be less vulnerable to common
method variance than the widely employed cross-sectional design is not new (Ganster
et al., 1986; Jackson, 1983; Schaubroeck et al., 1992; Spector, Dwyer, & Jex, 1988). In
fact, temporal separation in the data collection of stressor and strain measures has been
employed as a means to attenuate common method variance (Ganster et al., 1986;
Podsakoff & Organ, 1986; Schaubroeck et al., 1992). However, Ganster et al. (1986)
indicated that the presumably attenuating effects of temporal separation on correla-
tions have not been directly tested.
Alternatively, one can argue that the effects of a stressor at one point in time have a
spiraling effect over time. If so, stressor-strain correlations measured with a time lag
Downloaded from orm.sagepub.com at Kungl Tekniska Hogskolan / Royal Institute of Technology on March 8, 2015

Sanchez, Viswesvaran / Temporal Separation 175
could conceivably be higher than cross-sectional correlations. In fact, Mitchell and
James (2001) outlined several possible effects for time on an effect size. Unfortu-
nately, the extant literature on stressor-strain relationships does not provide much
guidance here. Prior studies have been either longitudinal or cross-sectional, and have
rarely combined the two approaches (i.e., administering stressor and strain measures
at more than one point in time).
More than 15 years have past since Ganster et al. (1986) noted that prior research
offered no empirical documentation of whether temporal separation of scale adminis-
trations mitigates common method variance. We reviewed the recent literature in
search of possible new evidence. A number of studies employing temporal separation
of stressors and strains were identified. However, some of these studies reported only
aggregated correlations (Abramis, 1994; Stewart & Barling, 1996), only cross-sec-
tional correlations (Schaubroeck, Ganster, Sime, & Ditman, 1993), only correlations
of measures gathered after an intervention (Moss & Lawrence, 1997), or only correla-
tions of stressors gathered at time 2 with strains gathered at time 1 (Spector, Chen, &
O’Connell, 2000). Nevertheless, two studies were deemed to have comparable cross-
sectional and longitudinal correlations that could shed light on the focal issue (Agho,
Mueller, & Price, 1993; Moyle, 1998). A closer examination of these two studies sup-
ported the assumption that cross-sectional correlations between role ambiguity and
job satisfaction are higher than their longitudinal counterparts (Agho et al., 1993,
reported r = .49 vs. .34, and Moyle, 1998, reported r = .37 vs. .28); in contrast, the lon-
gitudinal correlation between role conflict and job satisfaction was higher than its
cross-sectional counterpart (Agho et al., 1993, reported r = .34 vs. .30).
In summary, artifacts associated with common method variance such as temporary
mood states, priming, consistency, and attribution can be exacerbated by cross-sec-
tional designs. Although some authors have temporally separated the collection of
stressors and strains in an attempt to reduce common method variance, there is a pau-
city of tests of this presumption in the literature. Moreover, the scarce results available
are mixed and seemingly inconclusive. The potential for a spiraling effect of stressors
over time, which may increase longitudinal correlations, should not be ignored. Our
objective here was to explicitly compare longitudinal and cross-sectional stressor-
strain correlations.
Method
Participants
Twelve teams of three to five survey administrators were recruited, trained on sur-
vey administration, and instructed to locate individuals who were currently employed
in a metropolitan area in the southeastern United States. A total of 213 individuals con-
tacted by the survey teams chose to participate in the study. Subjects who provided
largely incomplete surveys were discarded, resulting in a final sample of 166 individu-
als. The sample was 45% female, 2.5% Asian, 6% Black, 49% Hispanic, and 27%
White, with the remainder choosing not to report their ethnicity. This ethnic composi-
tion resembled the latest figures for the area’s population (Fiedler, 1992). The modal
age was represented by the option termed “between 30 and 39 years old.” A total of 62
different occupations were represented: secretary (13 incumbents), manager (26),
Downloaded from orm.sagepub.com at Kungl Tekniska Hogskolan / Royal Institute of Technology on March 8, 2015

176 ORGANIZATIONAL RESEARCH METHODS
administrative officer (12), salesperson (6), account executive (4), and real estate agent
(5) were most numerous.
Approximately 60 days after the first administration, the same subjects were asked
to complete a second questionnaire. A total of 143 respondents, or 86% of those whose
data were obtained in the first administration, responded to this follow-up survey. This
sample size seemed to provide adequate statistical power, as it exceeded the size rec-
ommended by Cohen (1992, p. 158) for detecting small to medium effect correlations
with a statistical power of .80. Tests of differences in proportions for gender and eth-
nicity revealed no demographic differences between the gender and ethnic composi-
tion of respondents and nonrespondents to the second administration of surveys.
Procedure
Respondents were asked to provide the last four digits of their social security num-
ber so that their answers to both questionnaires could be matched. Otherwise, ques-
tionnaires were anonymous and confidentiality was assured.
Measures
Role ambiguity and role conflict were measured on the first survey using the Rizzo,
House, and Lirtzman (1970) scales (for additional psychometric information on these
scales, see House, Schuler, & Levanoni, 1983; Schuler, Aldag, & Brief, 1977). Sample
items are “Clear, planned goals and objectives exist for my job” and “I receive incom-
patible requests from two or more people” for role ambiguity and role conflict,
respectively.
Two types of strains were measured on both survey administrations: health out-
comes and work-related attitudes. Note that work-related attitudes such as poor orga-
nizational commitment and job dissatisfaction have been conceptually defined as
strains because they capture employee affect (Beehr, 1985). Items for each strain scale
were added to derive total scores. The following scales, listed in this same order, were
included in the questionnaires. Anxiety was measured with the State Anxiety Scale
from the State-Trait Anxiety Inventory (Spielberger, Gorsuch, & Lushene, 1970). Par-
ticipants were instructed to indicate the manner in which they had usually felt in the
past 30 days. A sample item is “I feel calm.” Five-point scales ranging from not at all
(1) to very often (5) were employed. Our symptoms scale included 15 items dealing
with physical symptoms selected from similar scales used by Chen and Spector
(1991). These items covered stomach problems (“I have been bothered by acid indi-
gestion or heartburn”), weakness and fatigue (“I have felt weak all over”), shortness of
breath (“I have experienced rapid heart beating when not exercising”), and loss of
sleep and nightmares (“I have had trouble sleeping”). Respondents were instructed to
indicate the extent to which they had felt each symptom in the past 30 days on a scale
ranging from never (1) to very often (5). Organizational commitment was measured
with the 9-item scale developed by Cook and Wall (1980) (e.g., “I am proud to be able
to tell people who it is I work for”). Job satisfaction was gauged with the three-item
scale employed in the Michigan Organizational Assessment Questionnaire (Cammann,
Fichman, Jenkins, & Klesh, 1979) (e.g., “All in all, I am satisfied with my job”). Work
tension was assessed with the seven-item subscale developed by House and Rizzo
(1972) (e.g., “I work under a great deal of tension”). Possible answers to the latter three
Downloaded from orm.sagepub.com at Kungl Tekniska Hogskolan / Royal Institute of Technology on March 8, 2015

Sanchez, Viswesvaran / Temporal Separation 177
scales ranged from I strongly disagree (1) to I strongly agree (5). Depression was
gauged with the 15-item Beck Depression Inventory (Beck & Steer, 1987). Respon-
dents chose one of four possible options describing the way they had been feeling the
week in which they answered the survey. A sample item ranged from I do not feel par-
ticularly guilty (1) to I feel guilty all the time (4). Sex, ethnicity, number of dependents,
age, and occupation were gathered on both surveys.
Results
Descriptive Statistics
Means, standard deviations, Cronbach’s alpha reliabilities, and range of scores
for the different measures at both Time 1 and Time 2 are provided in Table 1. Also
included is the available range on all scales. The observed range for all variables, with
the exception of the Beck Depression Inventory, covered nearly all the available range.
Scores on the Beck Depression Inventory did not span over high values. This result
might be explained by the nature of the inventory, which is aimed at clinically
depressed individuals.
Stressor-Stressor and Strain-Strain Intercorrelations
Role conflict and role ambiguity were significantly correlated (r = .33, p < .05).
Strainintercorrelations atTime 1 and Time 2 , together with test-retest correlations, are
presented in Table 2. Correlation ranges and correlation patterns appeared comparable
to those previously reported in the literature (Brief et al., 1988; Chen & Spector, 1991;
Jackson & Schuler, 1985). Test-retest correlations were similar in magnitude across
strains, ranging from .42 to .58.
Correlations Between Stressors and Strains
Table 3 presents the correlations (both observed and corrected for attenuation)
between stressors and strains gathered at concomitant and separate points in time. A
series of t tests for dependent correlations were computed to ascertain statistically sig-
nificant differences between the cross-sectional and longitudinal stressor-strain corre-
lations (McNemar, 1969, p. 158).
A comparison of the observed correlations suggested that temporal separation of
stressor and strain measures resulted in generally lower correlations than synchronous
administration of scales when the relationships between role ambiguity and work-
related attitudinal strains were considered. With the exception of the role conflict-
work tension relationship, temporal separation of stressors and strains did not result in
statistically significant attenuation of the correlations between stressors and health-
related strains.
Changes in the relations between variables over time constitute a violation of the
quasi-stationarity assumption that should hold when comparing synchronous and
panel correlations (Kenny, 1975). To check this assumption, we performed a LISREL
VII test of the equality of covariance structures between strain intercorrelations at both
Time 1 and Time 2 (Jöreskog & Sörbom, 1985, p. V-9). This test supported the quasi-
stationarity of our panel data because the same underlying covariance structure was
Downloaded from orm.sagepub.com at Kungl Tekniska Hogskolan / Royal Institute of Technology on March 8, 2015

Table 1
Descriptive Statistics and Internal Consistency Reliabilities (N = 143)
Time 1
Time 2
Available
Range
Actual
Range
Variable
M
SD
M
SD
Range
Role ambiguity
Role conflict
State anxiety
Symptoms
Commitment
Job satisfaction
Work tension
Depression
5-25
8-40
20-100
15-75
9-45
3-15
7-35
15-60
18.24
21.19
56.91
34.33
32.31
11.61
17.89
20.80
3.76
6.63
13.10
10.40
6.38
2.68
6.58
5.85
8-25
8-39
23-90
15-67
9-44
3-15
7-31
15-47
.77
.86
.93
.89
.76
.82
.88
.88
N/A
N/A
N/A
N/A
N/A
N/A
22-92
15-59
12-45
3-15
N/A
N/A
.93
.89
.83
.79
.87
.87
53.01
30.47
31.50
11.27
18.27
19.42
12.82
9.46
6.39
2.74
6.09
4.77
7-34
15-37

Sanchez, Viswesvaran / Temporal Separation 179
Table 2
Pearson Zero-Order Correlations Among Strains (N = 143)
Variable
1
2
3
4
5
6
1. State anxiety
2. Symptoms
(.56**)
.65**
.53**
(.46**) –.07
–.23**
–.22**
–.16
.37**
.45**
.55**
.45**
3. Commitment
4. Job satisfaction
5. Work tension
6. Depression
–.42**
–.30**
.37**
–.30**
–.17*
.40**
.61**
(.56**)
.76**
(.42**) –.32**
–.36**
–.26**
–.32**
–.18*
–.18*
.33**
.69**
–.38**
–.42**
(.58**)
.35**
.66**
(.47**)
Note.Correlations above the diagonal represent Time 1, whereas those below the diagonal repre-
sent Time 2. Test-retest correlations are presented in parentheses on the diagonal.
*p < .05. **p < .01.
Table 3
Synchronous Versus Panel Stressor-Strain Correlations
Dc
td
rxy
rxy
Dg
a
b
e
f
Strain/Stressor
rxy
rxy
State anxiety
Role ambiguity
Role conflict
Symptoms
.23** .23** .00
0.00
.12
.27
.08
.35
0.12
.08
.04
.07
.08
.01
Role ambiguity
Role conflict
Commitment
Role ambiguity
Role conflict
Job satisfaction
Role ambiguity
Role conflict
Work tension
Role ambiguity
Role conflict
Depression
.03
.16*
.10
.14
.07
–.02
0.66
0.19
.04
.18
.17
.22
.13
.04
–.59** –.35** –.24
–.31** –.30** –.01
3.01** –.78
0.19 –.38
–.53
–.43
–.25
.05
–.56** –.35** –.21
–.26** –.29** .03
2.50** –.71
0.21
–.61
–.48
–.10
.17
–.31
.29** .23** –.06
.50** .33** –.17
0.56
2.06*
.35
.58
.34
.47
–.01
–.11
Role ambiguity
Role conflict
.38** .22** –.16
1.64
0.50
.46
.18
.37
.33
–.09
.15
.16*
.21*
.05
a. Synchronous zero-order correlations.
b. Panel zero-order correlations.
c. D = difference synchronous-panel correlations.
d. t test for dependent correlations with 140 degrees of freedom.
e. Synchronous zero-order correlations corrected for attenuation using cross-sectionally com-
puted alpha reliabilities.
f.Panel zero-order correlations corrected for attenuation using Time 1 alpha reliabilities for stress-
ors and test-retest reliabilities for strains.
g. D = difference corrected synchronous-corrected panel correlations.
*p < .05. **p < .01.
shared by survey administrations at the two points in time (χ2(15) = 14.23, p > .50,
goodness of fit index = .98). Thus, differences between synchronous and panel corre-
lations did not seem attributable to changes over time in the relationships between
variables.
Downloaded from orm.sagepub.com at Kungl Tekniska Hogskolan / Royal Institute of Technology on March 8, 2015

180 ORGANIZATIONAL RESEARCH METHODS
Discussion
This study tested the widely held assumption that stressor-strain relations would be
reduced by a longitudinal as opposed to a cross-sectional design. Indeed, it has been
suggested that temporal separation in the collection of self-reported stressors and
strains is a useful procedural aid to mitigate artifacts associated with common method
variance such as priming, consistency, and attribution bias (Ganster et al., 1986;
Podsakoff & Organ, 1986). Our data, however, hinted at a more nuanced approach to
the use of temporal separation in correlational studies of stressor-strain relationships.
Specifically, temporal separation of scales might reduce correlations involving attitu-
dinal strains such as organizational commitment and job satisfaction, but not those
involving health-related strains such as physical symptoms and anxiety.
When deciding whether to separate scales over time, investigators should consider
the stability of stressors and strains as measured by their test-retest correlations. Insta-
bility is a form of unreliability likely to hamper the ability to identify valid relation-
ships over time. Alternate measures of role stressors with measurement properties
superior to the widely used Rizzo et al. (1970) scales may prove more stable and
should be incorporated in future research (Breaugh & Colihan, 1994). However, the
reduction in panel correlations for organizational commitment and job satisfaction
was hardly explained by change/instability in these variables over time, since their
test-retest correlations were similar in magnitude to those of all other variables. In fact,
the same pattern of differences between cross-sectional and longitudinal correlations
was observed when comparing the correlations that were corrected for attenuation
using test-retest strain correlations.
Together with prior studies (Agho et al., 1993; Moyle, 1998), our results suggest
that the correlations involving role ambiguity might be more likely to change due to the
use of a longitudinal design compared to those involving role conflict (the average
absolute change in magnitude from cross-sectional to panel correlations was .12 and
.05 for role ambiguity and role conflict, respectively). Such difference might be attrib-
utableinpart toroleconflict being more stable over time than role ambiguity; however,
it is possible that the onset of role conflict effects on strains takes longer than the onset
of role ambiguity effects. Using the terms employed by Mitchell and James (2001),
role conflict might require a longer equilibration period to reach its new equilibrium
condition compared to role ambiguity.
Note that Lazarus (1966) maintained that stress is triggered by a perceived mis-
match between environmental demands and coping responses. It may take a few trials
and some time for individuals to perceive that they lack the resources to cope with role
conflict, whereas the uncertainty provoked by role ambiguity may more immediately
elicit the feelings of vulnerability that activate the stress process. This phenomenon
may also explain the more pronounced reductions observed among panel correlations
involving attitudinal strains such as job dissatisfaction and organizational commit-
ment. That is, changes in work-related attitudes may represent more immediate reac-
tions thanchangesinhealth-related strainssuch as symptomsand anxiety, whose onset
may require a prolonged exposure to the role stressors.
In essence, measurement stability over time may not matter when the causal lag
between stressors and strains is not adequately captured by the timing of the measure-
ment. For example, despite a perfect test-retest reliability, role conflict effects on
health symptoms may not be apparent after a 1-day test-retest period.
Downloaded from orm.sagepub.com at Kungl Tekniska Hogskolan / Royal Institute of Technology on March 8, 2015

Sanchez, Viswesvaran / Temporal Separation 181
Taken together, our results challenge the assumption that temporal separation in the
data collection of self-reported stressors and strains guarantees a reduction in their
observed relationships. Note that the superiority of longitudinal designs in other areas,
such as their ability to ascertain cause and effect, has also been challenged (Rogosa,
1980). Indeed, even though longitudinal designs adhere to the positivist view of causa-
tion in that the cause precedes the effect in time, temporal precedence is a necessary
but not a sufficient condition for causation (Cook & Campbell, 1979, pp. 10-36).
We argue that time should be a variable of interest in work stress theories and that
measurement should proceed according to theory-driven time specifications. We have
outlined here how stress theory (e.g., Lazarus, 1966) may help formulate propositions
concerning the timing associated with the lag that takes different stressors to provoke
measurable changes in strains.
Future research employing repeated-measures designs that manipulate the time lag
between stressors and strains is needed for the development of time-sensitive theories
of work stressor effects on strains (Kenny, 1975; Mitchell & James, 2001; Rogosa,
1980). Time series designs that collect data at more than two points in time seem most
capable of answering questions about the dynamics of stressor-strain relationships
because different stressor-strain combinations may differ not only in the onset of
strains but also in their duration and their rate of change. However, an examination of
both within- and between-employee variation in stressor-strain relationships over time
is necessary. Time-embedded data of this sort can be analyzed using hierarchical linear
models (Hofman, 1997). Having multiple measures of multiple employees should fur-
ther our understanding of whether stressor-strain relationships evolve in a linear or a
nonlinear manner, whether these longitudinal trajectories differ across employees, and
what types of third variables account for such individual differences.
References
Abramis, D. J. (1994). Relationships of job stressors to job performance: Linear or an inverted-
U? Psychological Reports, 75, 547-558.
Agho, A. O., Mueller, C. W., & Price, J. L. (1993). Determinants of employee job satisfaction:
An empirical test of a causal model. Human Relations, 46, 1007-1027.
Avolio, B. J., Yammarino, F. J., & Bass, B. M. (1991). Identifying common methods variance
with data collected from a single source: An unresolved sticky issue. Journal of Manage-
ment, 17, 571-587.
Beck, A. T., & Steer, R. A. (1987). Beck Depression Inventory manual. San Antonio, TX: Psy-
chological Corporation.
Beehr, T. A. (1985). The role of social support in coping with organizational stress. In T. A.
Beehr & R. S. Bhagat (Eds.), Human stress and cognition in organizations (pp. 375-398).
New York: John Wiley.
Breaugh, J. A., & Colihan, J. P. (1994). Measuring facets of job ambiguity: Construct validity
evidence. Journal of Applied Psychology, 79, 191-202.
Brief, A. P., Burke, M. J., George, J. M., Robinson, B. S., & Webster, J. (1988). Should negative
affectivity remain an unmeasured variable in the study of job stress? Journal of Applied
Psychology, 73, 193-198.
Burke, M. J., Brief, A. P., & George, J. M. (1993). The role of negative affectivity in understand-
ing relations between self-reports of stressors and strains: A comment on the applied psy-
chology literature. Journal of Applied Psychology, 78, 402-412.
Downloaded from orm.sagepub.com at Kungl Tekniska Hogskolan / Royal Institute of Technology on March 8, 2015

182 ORGANIZATIONAL RESEARCH METHODS
Cammann, C., Fichman, M., Jenkins, D., & Klesh, J. (1979). The Michigan Organizational As-
sessment Questionnaire. Unpublished manuscript, University of Michigan.
Campbell, D. T., & Fiske, D. W. (1959). Convergent and discriminant validation by the
multitrait-multimethod matrix. Psychological Bulletin, 56, 81-105.
Chen, P. Y., & Spector, P. E. (1991). Negative affectivity as the underlying cause of correlations
between stressors and strains. Journal of Applied Psychology, 76, 398-407.
Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155-159.
Cook, J., & Wall, T. D. (1980). New work attitude measures of trust, organizational commitment
and personal need for non-fulfillment. Journal of Occupational Psychology, 53, 39-52.
Cook, T. D., & Campbell, D. T. (1979). Quasi-experimentation: Design and analysis issues for
field settings. Boston: Houghton Mifflin.
Fiedler, T. (1992, April 19). South Florida’s new identity: It’s almost America. The Miami Her-
ald, p. 4C.
Fisher, C. D., & Gitelson, R. (1983). A meta-analysis of the correlates of role conflict and ambi-
guity. Journal of Applied Psychology, 68, 320-333.
Ganster, D. C., Fusilier, M. R., & Mayes, B. T. (1986). Role of social support in the experience of
stress at work. Journal of Applied Psychology, 71, 102-110.
Gupta, N., & Beehr, T. A. (1982). A test of the correspondence between self-reports and alterna-
tive data sources about work organizations. Journal of Vocational Behavior, 20, 1-13.
Hofmann, D. A. (1997). An overview of the logic and rationale of hierarchical linear models.
Journal of Management, 23,723-744.
House, R. J., & Rizzo, J. R. (1972). Role conflict and ambiguity as critical variables in a model
oforganizational behavior. OrganizationalBehavior andHuman Performance, 7, 467-505.
House, R. J., Schuler, R. S., & Levanoni, E. (1983). Role conflict and ambiguity scales: Reality
or artifacts? Journal of Applied Psychology, 68, 334-337.
Jackson, S. E. (1983). Participation in decision-making as a strategy for reducing job-related
strain. Journal of Applied Psychology, 68, 3-19.
Jackson, S. E., & Schuler, R. S. (1985). A meta-analysis and conceptual critique of research on
role ambiguity and role conflict in work settings. Organizational Behavior and Human De-
cision Processes, 36, 16-78.
Jöreskog, K. G., & Sörbom, D. (1985). LISREL VI: Analysis of linear structural relationships by
the method of maximum likelihood. Mooresville, IN: Scientific Software.
Kenny, D. A. (1975). Cross-lagged panel correlation: A test of spuriousness. Psychological Bul-
letin, 82, 887-903.
Lazarus, R. S. (1966). Psychologicalstress andthe coping process. NewYork:McGraw-Hill.
McNemar, Q. (1969). Psychological statistics (4th ed.). New York: John Wiley.
Mitchell, T. R., & James, L. R. (2001). Building better theory: Time and the specification of
when things happen. Academy of Management Review, 26, 530-547.
Moss, S. E., & Lawrence, K. G. (1997). The effects of priming on the self-reporting of perceived
stressors and strains. Journal of Organizational Behavior, 18, 393-403.
Moyle, P. (1998). Longitudinal influences of managerial support on employee well-being. Work
and Stress, 12(1), 29-49.
Parkes, K. R. (1990). Coping, negative affectivity, and the work environment: Additive and in-
teractive predictors of mental health. Journal of Applied Psychology, 75, 399-409.
Podsakoff, P. M., & Organ, D. W. (1986). Self-reports in organizational research: Problems and
prospects. Journal of Management, 12, 531-544.
Rizzo, J. R., House, R. J., & Lirtzman, S. I. (1970). Role conflict and ambiguity in complex orga-
nizations. Administrative Science Quarterly, 15, 150-163.
Rogosa, D. (1980). A critique of cross-lagged correlation. PsychologicalBulletin, 88, 245-258.
Schaubroeck, J., Ganster, D. C., & Fox, M. L. (1992). Dispositional affect and work-related
stress. Journal of Applied Psychology, 77, 322-335.
Downloaded from orm.sagepub.com at Kungl Tekniska Hogskolan / Royal Institute of Technology on March 8, 2015

Sanchez, Viswesvaran / Temporal Separation 183
Schaubroeck, J., Ganster, D. C., Sime, W. E., & Ditman, D. (1993). A field experiment testing
supervisory role clarification. Personnel Psychology, 46, 1-25.
Schmitt, N., & Stults, D. M. (1986). Methodology review: Analysis of multitrait-multimethod
matrices. Applied Psychological Measurement, 10, 1-22.
Schuler, R. S., Aldag, R. J., & Brief, A. P. (1977). Role conflict and ambiguity: A scale analysis.
Organizational Behavior and Human Performance, 20, 111-128.
Spector, P. E. (1987). Method variance as an artifact in self-reported affect and perceptions at
work: Myth or significant problem? Journal of Applied Psychology, 72, 438-443.
Spector, P. E., & Brannick, M. T. (1995). The nature and effects of method variance in organiza-
tional research. International Review of Industrial and Organizational Psychology, 10,
249-274.
Spector, P. E., Chen, P. Y., & O’Connell, B. J. (2000). A longitudinal study of relations between
job stressors and job strains while controlling for prior negative affectivity and strains.
Journal of Applied Psychology, 85, 211-218.
Spector, P. E., Dwyer, D. J., & Jex, S. M. (1988). Relation of job stressors to affective, health, and
performance outcomes: A comparison of multiple data sources. Journal of Applied Psy-
chology, 73, 11-19.
Spector, P. E., Van Katwyk, P. T., Brannick, M. T., & Chen, P. Y. (1997). When two factors don’t
reflect two constructs: How item characteristics can produce artifactual factors. Journal of
Management, 23(5), 59-78.
Spector, P. E., Zapf, D., Chen, P. Y., & Frese, M. (2000). Why negative affectivity should not be
controlled in job stress research: Don’t throw out the baby with the bath water. Journal of
Organizational Behavior, 21(1), 79-80.
Spielberger, C. D., Gorsuch, R. L., & Lushene, R. E. (1970). Manual for the State-Trait Anxiety
Inventory. Palo Alto, CA: Consulting Psychologists Press.
Stewart, W., & Barling, J. (1996). Daily work stress, mood and interpersonal performance: A
traditional model. Work and Stress, 10, 336-351.
Watson, D., Pennebaker, J. W., & Folger, R. (1986). Beyond negative affectivity: Measuring
stress and satisfaction in the workplace. Journal of Organizational Behavior Management,
8, 141-157.
Williams, L. J., & Brown, B. K. (1994). Method variance in organizational behavior and human
resources research: Effects on correlations, path coefficients and hypothesis testing. Orga-
nizational Behavior and Human Decision Processes, 57, 185-209.
Williams, L. J., Cote, J. A., & Buckley, M. R. (1989). Lack of method variance in self-reported
affect and perceptions at work: Reality or artifact? Journal of Applied Psychology, 74, 462-
468.
Williams, L. J., Gavin, M. B., & Williams, M. L. (1996). Measurement and nonmeasurement
processes with negative affectivity and employee attitudes. Journal of Applied Psychology,
81, 88-101.
Juan I. Sanchez is an associate professor of management and international business at Florida Interna-
tional University. His research interests include strategic work analysis and the effects of work-related per-
ceptions on performance and well-being.
Chockalingam (Vish) Viswesvaran (Ph.D., University of Iowa) is an associate professor of pyschology at
Florida International University. His research interests include personnel selection, performance ap-
praisal, personality assessments in the workplace, and meta-analysis.
Downloaded from orm.sagepub.com at Kungl Tekniska Hogskolan / Royal Institute of Technology on March 8, 2015