An Empirical Investigation of Job Stress

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 24

ORGANIZATIONAL BEHAVIOR AND HUMAN PERFORMANCE 27, 279--302 (1981)

An Empirical Investigation of Job Stress, Social


Support, Service Length, and Job Strain
GARY B L A U

University o f Cincinnati

Although different approaches to job stress exist, the person-environment


fit approach seems to have the widest current acceptance. The most well-
developed p e r s o n - e n v i r o n m e n t ( P - E ) fit formulation of job stress has been
produced by French and his associates at the University of Michigan. The
present study evaluated several relationships within the French job stress
model, using bus operators from a midwestern transit authority as the sample.
Results provided only limited support for the model, and length of service was
found to be an important variable to consider in future research. The useful-
ness of measuring job stress as a discrepancy score and issues regarding the
construct validity of P - E fit job stress are also discussed.

The general concept of stress has generated a great deal of research and
many definitions have been offered. Illustrations of the diversity of these
definitions include: stress as a situational condition (e.g., Hill, 1958);
stress as a reaction (e.g., Janis, 1958); stress as a general state type of
phenomena (e.g., Selye, 1956); and stress as a term for designating a
broad area of study (e.g., McLean, 1974). The unsettled general
conceptualization of stress has made its imprint upon efforts to define job
or occupational stress.
After taking a backseat to such work issues as quality of work life and
unemployment for many years, job stress has recently come into its own
as a prominent work-related research topic (e.g., Cooper & Payne, 1978).
Part of the reason for the emergence of job stress is that no panacea has
been found for eliminating negative effects in a work situation. In addi-
tion, job stress readily allows for the inclusion of a variable set generally
absent from organizational research--physiological measures.
Most approaches dealing with occupational stress in the work setting
(e.g., Gross, 1970; Landy & Trumbo, 1976) have involved listings of
various sources of stress, e.g., task (speed/load demands), role (conflict/
ambiguity), physical environment (noise, temperature), and social
environment (interpersonal discord). H o w e v e r , as McGrath (1970)
pointed out, several methodological weaknesses are apparent in such an
The author wishes to acknowledge the helpful technical comments of Dr. Ralph Kater-
berg and the typing expertise of Julie Menchen, Administrative Assistant, both in the De-
partment of Management, University of Cincinnati. Please address all correspondence to:
Gary Blau, Department of Management, University of Cincinnati, Mail Location #20, Cin-
cinnati, OH 45221.

279
0030-5073/81/020279-24502.00/0
Copyright© 1981by AcademicPress, Inc.
All fightsof reproductionin any form reserved.
280 GARY BLAU

approach to stress, including: (1) no accepted units of measurement which


can serve as a common metric for stresses of different origins; (2) ordinal
measurements for many kinds of stress (e.g., isolation); and (3) difficulty
in accounting for individual differences in perceptions of stress. In an
attempt to alleviate these problems and to enhance the conceptual useful-
ness of the term, McGrath (1970) offered the following definition of stress:
" A perceived substantial imbalance between demand and response capa-
bility, under conditions where failure to meet demands has important
perceived consequences."
Stress is here defined in terms of a relationship between the person and
the environment. Thus, where either an environmental (job) demand ex-
ceeds a person's response capability (overload), or the person's capabil-
ities exceed the environmental demand (underload), the resulting misfit
represents stress. This person-environment ( P - E ) fit approach to stress
allows for increased measuring precision (through commensurate P and E
measures), and gives full recognition to individual differences.
Person-environment fit formulations of job stress have become increas-
ingly popular for organizational researchers, e.g., Beehr and Newman
(1978) and Schuler (1979). However, the most well-developed P - E fit job
stress model has been produced by French and his associates at the Uni-
versity of Michigan.

THE FRENCH MODEL RELATING JOB STRESS TO HEALTH


Over the course of many years, French and his associates (e.g., French
& Kahn, 1962; French & Caplan, 1972; French, Rogers, & Cobb, 1974)
have developed a general model of relationships between job stress and
health. Within this model, two perceived person-environment ( P - E ) fit
definitions of job stress are considered. One kind of fit incorporates
McGrath's (1970) definition of stress and examines the extent to which a
person's skills/abilities match the demands of his/her job. The second kind
of fit measures the extent to which the job environment provides supplies
to meet an individual's needs (i.e., stress equals n e e d - s u p p l y misfit).
The French model of job stress is conceptually similar to the Minnesota
model of work adjustment (Dawis, England, & Lofquist, 1968). Work
adjustment is defined by the degree of correspondence (fit) between: (1)
abilities of the individual and ability requirements of the job, and (2) needs
of the individual and reinforcement of those needs by the work environ-
ment. However, each model utilizes different measures and has a differ-
ent focus of prediction; vocational selection and tenure are predicted by
the Minnesota work adjustment model vs worker health by the French job
stress model.
The foundation of the French job stress model (French & Kahn, 1962) is
based upon the theoretical work of Lewin (1938) and Murray (1938).
EMPIRICAL INVESTIGATION 281

Lewin' s tenet that behavior is a function of the person and the environ-
ment provided the necessary idea of an organism-situation interaction.
Murray (1938) developed a person-environment process model based
upon individual need frustration or gratification. Individual need frustra-
tion resulted from the environment (press) not satisfying the individual's
needs, while need gratification was due to proper environmental supplies.
In addition, Murray (1938) proposed that individuals and environments be
classified along similar or commensurate dimensions (in terms of needs),
and he distinguished between the environment as it is (a press) and the
environment as it is perceived to be (/3 press).
A schematic representation of person-environment fit within the
French job stress model (French, Note 1) is shown in Fig. 1. The three top
boxes of variables (A, B, and C) refer to the environment, while the
bottom three boxes (D, E, and F) refer to characteristics of the person.
The arrows represent hypothetical causal relations running from cause to
effect. An arrow terminating on a second arrow represents a conditioning
or moderating effect (e.g., social support moderates the relationship be-
tween P - E fil job stress and strains). Boxes A and B distinguish between
the objective environment (a press) and the subjective environment (/3
press). Box C, labeled social support, involves measuring the amount of
perceived instrumental and/or socioemotional help which the individual
receives from others. Boxes D and E distinguish between the objective
person and the subjective person (i.e., self-concept).
An individual's perceived motives (needs) for working are related to
his/her perceived job supplies and perceived abilities are related to his/her
perceived job demands. Job demands reflect j o b requirements (e.g., typ-

A B c
OBJECTIVESOCIAL I I SUBJECTWESOCIAL ' ' J
ENVIRONMENT j I ENVIRONMENT ggIO,CT,VE
hersSUpervi
WorkSOC'
S°ra,
SUPPORT
ALl
Occupation
Family Situation
~
I
I
Perceived Supplies for Motives
J Perceived Job Demands
I I People at Home

G
sUgJECTIVE'PERSON.
I ENVIRONMENTPIT
Motives for Working vs Supplies
SBi
aAIv,"o;o,I _ s"oE2TtITZ;SI2iots
I for These Motives Physiological Morbidity
J Job Demands vs Abilities to Psychologica,I.I M..... y
Meet These Demands

D
I OBJECTIVEPERSONI SUBJECTIVEPERSON
~Perceived Motives tar Working I
| Perceived Abilities J
FIG. 1. A general model of the effects of the social environment on psychological strain
and health (adapted from French, 1976).
282 GARY BLAU

ing speed, market knowledge), while abilities include education, training,


etc. The broken line linking the person and the environment (see Fig. 1)
represents the "goodness of fit" between the person and his/her environ-
ment. If the motives (needs) in the person and corresponding perceived
job supplies for these motives (or perceived abilities-job demands) can
be measured on similar dimensions, then a discrepancy score yielding a
quantitative measure of the goodness of fit between the person and envi-
ronment can be calculated. According to the model, this discrepancy
score indicates job stress, and the greater the discrepancy the greater the
job stress. Thus, j o b stress is conceived as the sustained motive arousal
which occurs when the job environment either does not provide what the
individual seeks, or provides too much.
A central hypothesis in the French model (see Fig. 1) states that job
stress (subjective P - E fit) is a major cause of psychological, physiologi-
cal, and behavioral strains. According to the model, job dissatisfaction is
an example of a psychological strain and poor job performance is an
example of a behavioral strain. Examples of a physiological strain would
include: high blood pressure, abnormal heart rate, and high cholesterol
level. Such a predicted relationship between job stress and job strain is
based upon the assumption that sustained motive arousal will produce
response deviations. According to French (Note 1), "strain refers to any
deviation of normal responses in a person." These job strains are in turn
hypothesized to affect various measures of health and illness (e.g., heart
disease). Other researchers (e.g., Matteson & Ivancevich, 1979) have
developed a similar stress-strain relationship conceptualization.
Very little research has been done relating a P - E fit definition of job
stress to behavioral strains such as poor job performance. Yet, we know
from laboratory research (e.g., Yerkes & Dodson, 1908; McClelland &
Apicella, 1947; and Finkelman & Glass, 1970) that stress of a more general
nature causes performance deterioration at some point. One study which
investigated the relationship between a P - E fit definition of job stress and
performance (Andrews & Farris, 1972) found mixed support for the
French model. In a panel study of time pressure and performance for
scientists, results showed that when experienced time pressure was opti-
mal (good motive-supply fit) innovation and productivity, two subjec-
tively judged job performance measures, were highest. Judged usefulness
of the scientist's work, however, was equally good when time pressure
was either optironl or too much. It appears thus far that no study has used
hard performance data in investigating the relationship between P - E fit
job stress and job strain. Such a study is needed to provide a more objec-
tive test of the hypothesized relationship.
Most research studying the relationship between job stress and job
dissatisfaction (a psychological strain) has involved working with role
stresses (e.g., role conflict; role ambiguity). The prevalent finding has
EMPIRICAL INVESTIGATION 283

been that role conflict and role ambiguity are negatively related to job
satisfaction (e.g., Kahn, Wolfe, Quinn, Snoek,& Rosenthal, 1964; House
& Rizzo, 1972; Johnson & Stinson, 1975; LaRocco & Jones, 1978). In
studies specifically using a P - E fit (need-supply) definition of job stress,
French (1974) and Caplan, Cobb, French, Harrison, and Pinneau (1975)
found that misfits in terms of responsibility for people and role ambiguity
were positively related to job dissatisfaction. In addition, using a P - E fit
job stress measure based upon demand-ability (education required vs
education achieved), Coburn (1975) found that workers with a poor P - E
fit (education too high or too low) tended to be more dissatisfied with their
jobs. With regard to research on the stress-physiological-strain relation-
ship, an extensive series of correlational analyses between various P - E
fit job stresses (e.g., quantitative work load, responsibility for people, job
complexity, role ambiguity) and physiological measures (e.g., heart rate,
blood pressure, cholesterol, uric acid, thyroid hormones) produced very
few significant results (Caplan et al., 1975).
Another important concept within the French model is the perceived
environmental concept of social support (Caplan et al., 1975). Social sup-
port grew out of the earlier Ohio State (e.g., Fleishman, 1957; Halpin &
Winer, 1957) and Michigan (e.g., Katz, Maccoby, & Morse, 1950) leader-
ship studies which investigated several ideas, including: (a) the degree to
which supervisor behavior could be characteriized as considerate or
employee oriented, and (b) the relationship of perceived supervisor
considerate behavior to subordinate job satisfaction. Whereas considera-
tion was only concerned with the psychological-emotional support given
by a supervisor to his/her subordinates, social support has also included
an instrumental support dimension (e.g., helping a person to complete a
task) and looked at co-worker and off-job (e.g., spouse, friends) influ-
ences.
The French model predicts that social support will be negatively related
to job strains. Regarding the relationship between social support and job
dissatisfaction, research strongly indicates that supervisor social support
is negatively related to job dissatisfaction (e.g., Caplan et al., 1975; House
& Wells, Note 2; LaRocco & Jones, 1978). In reference to the relation-
ships of co-worker and off-job social support to job dissatisfaction, signi-
ficant but less impressive negative relationships were also found (e.g.,
Caplan et al., 1975; House & Wells, Note 2). House and Wells (Note 2)
found that only supervisor social support was negatively related to
physiological strains (e.g., ulcers, itching/rash). It appears that no study
has investigated the relationship between social support and poor job
performance (a behavioral strain). This relationship needs to be examined
and, according to the French model, a negative correlation should be
found.
The French model predicts that social support plays two roles in rela-
284 GARY BLAU

tion to job stress and job strain: (a) additive and (b) interactive or mod-
erating. The additive role holds that social support can have a significant
additive effect upon reducing job strain beyond what happens when job
stress is reduced. Regression research studies by Pinneau (Note 3) and
LaRocco and Jones (1978) dealing with supervisory and co-worker social
support in relation to job dissatisfaction have verified this position. How-
ever, neither of the above-mentioned regression research studies con-
trolled for potentially influential variables such as length of service.
The moderating role predicts that social support will buffer the job-
stress-job-strain relationship. As perceived level of social support in-
creases, the relationship between job stress and job strain should diminish
(i.e., people with high social support will show a lower stress-strain
relationship). Empirical research investigating social support as a moder-
ator has generated controversy, since some research supports the mod-
erator position (e.g., House, McMichael, Kaplan, & Wells, Note 4; Cobb,
1976; House & Wells, Note 2) and some does not (e.g., Pinneau, Note 3;
LaRocco & Jones, 1978). Additional research is needed to resolve this
controversy.
Regarding the variable length of service, Argyris (1957) has suggested
that with considerable job longevity and stability, various psychological
defenses (e.g., rationalization) could allow individuals to adopt to their
job environments by modifying their needs or expectations. Caplan et al.
(1975) found that length of service was negatively related to the P - E fit
job stresses of job complexity and responsibility for people. In a similar
theme, Hulin and Smith (1965) showed that length of service was posi-
tively related to job satisfaction as a result of the individual's ability to
adjust his/her expectations as to what the job environment was likely to
provide. These research results suggest that length of service is an im-
portant variable to consider in studying job stress and job strain and that
further exploration is needed.
In this study, a P - E fit job stress measure (in terms of need-supply)
was developed and several hypothesized relationships within the French
model were investigated. First, the hypothesized relationships between
job stress and two job strains (job dissatisfaction and ineffective job
performance) were explored. Then, the hypotheses of social support
playing additive and/or moderating roles in relation to job stress and job
strain were tested. Finally, this study explored the relationship of length
of service to job stress, job strain, and social support.
METHOD
Sample
The sample consisted of 166 bus operators across four bus stations
within a midwestern transit authority. Information from transit authority
EMPIRICAL INVESTIGATION 285

records (in terms of sex and race) revealed that the sample taken was
representative of the total bus operator population. Sixty-three percent of
the sample had less than 5-years' length of service as a bus operator.

Procedure
Sample participants were contacted and informed of the study at their
respective bus stations. This contact was carried out primarily through
posted memos which identified the researcher as being from a nearby
university, and indicated that management and union approved of the
study. The memos also indicated that the participants would be asked to
fill out a questionnaire in which individual answers would be kept confi-
dentiN.
The researcher administered the questionnaire at each of the four bus
stations. Respondents completed the questionnaire immediately at their
station and returned it to the researcher. Subjects were asked to print
their names with the understanding that this would also be kept confiden-
tial. A cutoff was established that bus operators must have had at least 3
months of experience in order to be fully exposed to their jobs and its
stressful/nonstressful aspects. Job performance and length of service data
were collected by using company records for those bus operators willing
to put down their names. One hundred twenty-eight bus operators (78% of
166) were willing to give their names, and data were collected for 123 bus
operators (four records could not be found and one could not be used due
to questionable data). Union and management permission were obtained
prior to collecting any job performance/length of service data.
Instruments
A pretested questionnaire contained the job stress, social support, and
job satisfaction measures.
Stress. Questionnaire items measuring job stress were based upon what
48 interviewed bus operators perceived to be "upsetting" about their
jobs. To measure job stress as a discrepancy score, bus operators were
asked to first rate the extent to which an item was present on their bus
run/job (E), and then to rate the amount of that item which they would
need in order to say that they had a good job (P), i.e., a comfortable job
where things were going smoothly. Based upon their interview answers,
18 situations were identified and broken into P and E terms. The mag-
nitude of job stress across P - E items was calculated by adding up the
response discrepancies. On 15 of the 18 situations, only discrepancies in
the direction of P > E (the individual's specific need was greater than the
environmental supply for that need) were considered as stressful. Thus,
where a numerical comparison on an item yielded either P -- E or E > P
(very rare), there was assumed to be no stress. This operationalization
286 GARY BLAU

utilized the kind of needs (see Appendix, first 15 items) where an over-
supply would not be likely to generate motive tension (House, 1972). The
remaining 3 items all involved serious on-the-job problems (see Appendix,
last 3 items). Only where the numerical relationship between P and E
items in these three situations was E > P (the environment supplies more
problems than the operator needs to have a good job) was this counted as
stress. Support for this interpretation came from pretest and interview
data, and from the common-sense notion that few people would want to
be involved in serious problems where they were physically threatened.
Responses to both the E and P items were marked on identical scales from
1 (none) to 7 (all the time). Since the 18 items were multidimensional, a
factor analysis was done upon the stress measure using the principal-
factor method, followed by varimax rotation.
Social support. The social support measure consisted of 12 items taken
from Job Demands and Worker Health (Caplan et al., 1975). For this
study, each bus operator indicated how supportive (instrumentally and
socioemotionally) three different groups of people were to him/her: the
bus district superintendent, fellow operators, and significant off-job
people (e.g., spouse, friends, relatives). The rating scale was enlarged
from the original 4-point scale to a 7-point scale for this study (1 = not at
all; 7 = all the time).
Job satisfaction~dissatisfaction. The job satisfaction measure consisted
of 50 items taken from the 100 item long-form Minnesota Satisfaction
Questionnaire (MSQ) developed by Weiss, Dawis, England, and Lofquist
(1967). Time constraints and questionnaire length necessitated the cut-
back. Forty-five out of the fifty items were selected in order to maintain
adequate three-item scale reliabilities for 15 particularly relevant facet
scales (e.g., ability utilization, compensation, recognition, social service,
working conditions). The remaining five items were selected to allow the
short-form MSQ (20 items) to be embedded within the overall job satis-
faction measure. Each item was rated on a 5-point scale ranging from 1
(very satisfied) to 5 (very dissatisfied).
Job performance. A point scale for classifying ineffective/effectivejob
performance data was devised by the author with the help of transit au-
thority personnel. The point scale included the following types of
performance incidents already recorded in bus operator files (the numbers
in parentheses give the point values assigned for each incident): charge-
able accident (10 pts.); suspension (7 pts./incident, 1 pt./day); unexcused
absence (4 pts.); miss or coming late (4 pts.); supervisory reports (2 pts,/
violation, - 2 pts./commendation); and public relations reports (1 pt./
violation, - 2 pts./commendation). In order to control for the effect of
length of service upon number of performance points accumulated, the
EMPIRICAL INVESTIGATION 287

job performance measure used was a points/month average based upon


each bus operator's job performance and months of service.

Analyses
The potential moderating effect of social support (supervisory, co-
worker, and off-job) between job stress and job strain was measured by
moderated regression. According to the moderated regression procedure
(Zedeck, 1971), the multiple correlation coefficients produced by equa-
tions combining length of service, stress factors, and social support mea-
sures in an additive manner were compared to coefficients produced by
prediction equations containing the a b o v e - m e n t i o n e d terms and
multiplicative interaction terms (stress factor × social support measure).
For an interaction to occur, there must be a significant incremental
change in R 2 after the main effects (additive model) are accounted for.

RESULTS
Measures
A factor analysis of the 18 job stress items followed by varimax rotation
yielded three interpretable stress factors (see Appendix): Factor I - -
Passenger/intracompany (P/I) stress (9 items); Factor II--Scheduling/
assistance (S/A) stress (6 items); and Factor III--Physical danger (PD)
stress (3 items). These three stress factors accounted for 55% of the total
variance. This factor structure was chosen based upon the relative size of
eigenvalues (see Appendix). The internal consistency reliabilities of each
factor were adequate (Factor I = .77, Factor II = .77, and Factor I I I =
.65). Furthermore, a factor analysis of the individual P and E components
comprising the job stress measure yielded a comparable factor structure.
The internal consistency reliabilities found for each type of social sup-
port were: bus district superintendent (.89); fellow operator (.72); and
off-job (.88). The internal consistency reliability of the job satisfaction
measure was .95.
Main Data
Simple Pearson product correlations and partial correlations controlling
for length of service among simple correlation variables are reported in
Table 1. The partial correlations appear in parentheses below the simple
correlations and are written as rp. The abnormal length of service standard
deviation value is a result of the skewness of the distribution. Many initial
zero-order correlations provide support for the predicted relationships
between: job stress and job strains (ineffective job performance, job
dissatisfaction), social support (especially supervisor) and job strains, and
t-,#
TABLE 1 O0
PEARSON PRODUCT--MOMENT CORRELATIONS AMONG THE STRESS FACTORS, SOCIAL SUPPORT, LENGTH OF SERVICE, JOB PERFORMANCE, AND
JOB SATISFACTION MEASURES, a AND CORRESPONDING PARTIAL CORRELATIONS CONTROLLING FOR LENGTH OF SERVICE (IN PARENTHESES) b

Variable 1 2 3 4 5 6 7 8 9 Mean SD

1. Physical danger
stress factor 4.09 3.00
2. Passenger/intracompany
stress factor .08 -- 22.85 9.19
3. Scheduling/assistance
stress factor .19"* .58** -- 13.77 6.35
4. Superintendent
social support -.04 -.49** -.34** -- 13.59 5.30
(-.13) (-.46**) (-,36**)
5. Co-worker social
support .15" -.19"* -.18"* .35** -- 16.19 3.81
t"
(.12) (-.15") (-.11) >
6. Off-job social
support - . 15" -.08 - . 13 .25** .29** -- 21.53 5.55
(-.13) (-.02) (-.11)
7. Length of service .38** -.23** -.01 .31"* .12 .05 -- 83.54 84.89
8. Ineffective job
performance .14 .20** .07 - . 17" - . 11 .08 -.53"* -- 4.24 3.21
(.15") (.10) (.08) (-.01) (-.06) (,13)
9. Job dissatisfaction .10 .50** .39** -.56** -.28** - . 16" -.22"* .04 -- 144.93 27.55
(. 14) (.45**) (.38**) (-.55**) (-.29**) (-.09)

" The N for any specific correlation varies from 123 to 166.
~' The N for any specific partial correlation varies between 106 and 123,
* p < .05.
** p < .01.
EMPIRICAL INVESTIGATION 289

social support and job stress. Partialing out length of service from these
relationships reduces several relationships considerably (e.g., supervisor
social support and ineffective job performance, passenger/intracompany
stress and ineffective job performance). However, such a partialing out is
necessary in order to provide for a less contaminated test of the model.
Yet even under this more rigorous test, most of the initially significant
relationships remain significant.
Results provide only limited support for the French model's hypothe-
sized relationships between job stress and job strains. Although the re-
lationship between job stress and psychological strain (job dissatisfac-
tion) was confirmed (rp = .45, rp = .38), the results failed to support a
job stress-behavioral-strain (ineffective job performance) relationship.
Only the relationship physical danger stress-ineffective job performance
reached even minimal statistical significance (ro = •15).
Confirmation of the French model's prediction of a negative relation-
ship between social support and job strain was only found for supervisory
support-job dissatisfaction (rp = -.55) and co-worker support-job dis-
satisfaction (rp = -.29). With length of service partialed out, no signifi-
cant relationships were found between social support and ineffective job
performance. Other important correlations in Table 1 dealt with length of
service, such that length of service was: (a) positively related to physical
danger; (b) negatively related to passenger/intracompany stress; (c) posi-
tively related to superintendent social support; and (d) negatively related
to both job strains (ineffective job performance and job dissatisfaction).
Results listed in Table 2 indicate that job stress factors, superintendent
social support, and co-worker social support account for additive effects
in predicting job dissatisfaction beyond what is predicted by length of
service. These results support the French model's prediction that job
stress and social support from work are important in considering job
strain. However, no such additive effects were found for off-job social
support or concerning the prediction of ineffective job performance.
Length of service accounted for almost all of the variance concerning
ineffective performance.
The moderated regression analysis of the prediction that social support
can buffer relationships between job stress and job strain, yielded one
significant interaction (see Table 2): superintendent social support buff-
ered the relationship between passenger/intracompany stress and in-
effective job performance. Although subgroup analysis of this inter-
action revealed significant correlational differences b e t w e e n bus
operators with perceived low vs. high superintendent social support, such
differences were not in the expected direction. Bus operators with lower
superintendent social support showed a relationship between passenger/
intracompany stress and ineffective job performance of r = -.05, while
290 GARY BLAU

TABLE 2
MULTIPLE CORRELATIONS OF LENGTH OF SERVICE, STRESS FACTORS, SOCIAL SUPPORT
MEASURES, AND INTERACTION TERMS FOR PREDICTING JOB STRAINSa

Ineffective
Job job
Measure dissatisfaction performance

SERVLEN .22 .53


SERVLEN + PD Stress .27* .54
SERVLEN + PD Stress + SUPSOCS .59** .54
SERVLEN + PD Stress + SUPSOCS + INT b .59 .54
SERVLEN + PD Stress + COWSOCS .40** .54
SERVLEN + PD Stress + COWSOCS + 1NT .41 .54
SERVLEN + PD Stress + OFJSOCS .28 .55
SERVLEN + PD Stress + OFJSOCS + INT .28 .55
SERVLEN + P/I Stress .49** .54
SERVLEN + P/I Stress + SUPSOCS .62** .54
SERVLEN + P/1 Stress + SUPSOCS + INT .63 .57*
SERVLEN + P/I Stress + COWSOCS .55* .54
SERVLEN + P/I Stress + COWSOCS + INT .55 .54
SERVLEN + P/I Stress + OFJSOCS .50 .55
SERVLEN + P/I Stress + OFJSCOS + INT .50 .55
SERVLEN + S/A Stress .39** .54
SERVLEN + S/A Stress + SUPSOCS .60** .54
SERVLEN + S/A Stress + SUPSOCS + INT .60 .55
SERVLEN + S/A Stress + COWSOCS .46* .54
SERVLEN + S/A Stress + COWSOCS + INT .46 .54
SERVLEN + S/A Stress + OFJSOCS .39 .55
SERVLEN + S/A Stress + OFJSOCS + INT .39 .55

Note. PD = physical danger; P/I = passenger/intracompany; S/A = scheduling/assistance;


SUPSOCS = superintendent social support; COWSOCS = co-worker social support;
OFJSOCS = off-job social support.
" All multiple correlations are significant in this table. However, those with asterisks indi-
cate a significant contribution of additional linear effects or interaction effects.
b INT = interaction term (stress factor × social support measure).

bus operators with higher superintendent social support revealed a rela-


t i o n s h i p o f r = .30.
Regarding the investigation of social support buffering the relationship
between job stress and job dissatisfaction, an initial investigation revealed
n o i n t e r a c t i o n e f f e c t ( T a b l e 2). H o w e v e r , b e c a u s e it w a s p o s s i b l e t h a t t h e
c o m p l e x i t y o f t h e j o b d i s s a t i s f a c t i o n m e a s u r e (15 d i f f e r e n t f a c e t s c a l e s )
was masking any potential smaller interactions, 3 relevant facet dissatis-
faction scales were pulled out and separately investigated: working
conditions, co-workers, and supervision-human relations. Out of the
moderated regression analyses performed on l0 combinations (see Table
3), 1 s i g n i f i c a n t i n t e r a c t i o n w a s f o u n d : s u p e r i n t e n d e n t s o c i a l s u p p o r t b u f f -
ered the relationship between physical danger stress and dissatisfaction
EMPIRICAL INVESTIGATION 291

TABLE 3
MULTIPLE CORRELATIONS OF LENGTH OF SERVICE~ STRESS FACTORS,
SOCIAL SUPPORT MEASURES AND ~NTERACTION TERMS FOR
PREDICTING THREE FACET SATISFACTION SCALES

Measure Work condition diassatisfaction

SERVLEN ,13
SERVLEN + PD Stress ,18
SERVLEN + PD Stress + SUPSOCS .40
SERVLEN + PD Stress + SUPSOCS + INT ° .44*
SERVLEN + PD Stress + COWSOCS .22
SERVLEN + PD Stress + COWSOCS + 1NT .22
SERVLEN + S/A Stress .31
SERVLEN + S/A Stress + SUPSOCS .42
SERVLEN + S/A Stress + SUPSOCS + INT .43
SERVLEN + S/A Stress + COWSOCS .32
SERVLEN + S/A Stress + COWSOCS + INT .34
Co-worker dissatisfaction
SERVLEN .02 •
SERVLEN + PD Stress .03
SERVLEN + PD Stress + COWSOCS .45
SERVLEN + PD Stress + COWSOCS + INT .46
SERVLEN + P/I Stress .02
SERVLEN + P/I Stress + COWSOCS .46
SERVLEN + P/I Stress + COWSOCS + 1NT .46
SERVLEN + S/A Stress .03
SERVLEN + S/A Stress + COWSOCS .46
SERVLEN + S/A Stress + COWSOCS + I N T .46
Supervision (human relations)
dissatisfaction
SERVLEN .18
SERVLEN + PD Stress .25
SERVLEN + PD Stress + SUPSOCS .63
SERVLEN + PD Stress + SUPSOCS + 1NT .63
SERVLEN + P/I Stress .52
SERVLEN + P/1 Stress + SUPSOCS .67
SERVLEN + P/I Stress + SUPSOCS + INT .67
SERVLEN + S/A Stress .36
SERVLEN + S/A Stress + SUPSOCS .63
SERVLEN + S/A Stress + SUPSOCS + I N T .63

" All multiple correlations greater than . 15 are significant in this table. However, those
with asterisks indicate a significant contribution of interaction effects.
0 INT = interaction term (stress factor x social support measure).
* p < .05.

with working conditions. Subgroup analysis of this interaction revealed


expected directional differences. Bus operators with lower superinten-
dent social support showed a relationship between physical danger stress
and working condition dissatisfaction of r = .33, while bus operators with
higher superintendent social support revealed a relationship of r = - . 11.
292 GARY BLAU

DISCUSSION
In addition to discussing the implications of the study's main results,
there are several other topics of concern. Since the P - E fit approach to
job stress is based upon using discrepancy scores, several methodological
criticisms need to be immediately analyzed in order to establish the credi-
bility of the study's job stress measure. Then, results concerning the
tested relationships of the French model will be discussed. Finally,
broader issues concerned with the construct validity of P - E fit job stress
will be dealt with before conclusions about the French model are drawn.

Methodological Comments
Among the methodological criticisms made against the use of discrep-
ancy or difference scores are included: increased unreliability and
unnecessary complexity in accounting for variance vs components (Cron-
bach & Furby, 1970) and contamination between component (P, E) mea-
sures (Caplan et al., 1975). In an attempt to avoid such criticisms, several
researchers (e.g., Wall & Payne, 1973) have proposed that operation-
alizing the logic behind discrepancy scores while avoid its pitfalls could
take the form of: " H o w much more would you like than you have now?"
Unfortunately, this approach suffers the glaring liability of not having a
reference point for what the individual has.
Cronbach and Furby's (1970) argument concerning unreliability is that
because a difference score can magnify the unreliability of the score's
components, it is possible to obtain a P - E discrepancy score which is less
reliable than the respective P and E components used to compute the
discrepancy. However, in this study a favorable comparison was found
between the reliabilities of the P - E discrepancy score stress factors
(ranging from r = .65 to r = .77) and their P and E components (ranging
from r = .62 to r = .82). In answer to Caplan et al. (1975), the contamina-
tion between P and E measures was found to be minimal (r = .22).
A final concern about discrepancy scores (Cronbach & Furby, 1970) is
that when linear variance in outcomes is predicted, only component
scores need to be used, since the more complex discrepancy scores offer
no empirical advantage. Subsequent internal analysis of the P - E discrep-
ancy score stress factors in terms of the P and E components showed that
both components contributed strongly to two of the stress factors:
passenger/intracompany stress (E = -.64, P -- .69) and scheduling/assist-
ance (E = -.52, P = .77). Hence, the usefulness of utilizing both P and E
components in some combination for at least two of the three job stress
factors was shown. In the case of the physical danger stress factor, the P
component contributed little (r = -.05) because when asked how much
physical danger they needed to have a good job, almost all bus operators
EMPIRICAL INVESTIGATION 293

gave the response of none. Thus, one condition for meaningfully using a
P - E fit approach to job stress involves measuring the type of job prob-
lems where individual differences in preference (need) are likely. Where
very minor individual differences in preference regarding job problems
are predicted, a simple environmental measure should suffice.
The use of component scores vs discrepancy scores in accounting for
linear variance can be examined from another perspective. As Table 4
indicates, P - E fit job stress factors (except for physical danger) ac-
counted for additional variance in job dissatisfaction beyond: (1) separate
P, E components and (2) additively combined P, E components (P and E).
Neither separate component scores nor any combination of component
scores ( P - E or P and E) were meaningfully related to ineffective job
performance (except for physical danger as measured by P - E ) .
Not only is the additive model (P and E) less effective than the discrep-
ancy model ( P - E ) in accounting for job dissatisfaction variance, but a
conceptual problem and a methodological problem become apparent
when working with the additive model. The conceptual problem involves
questioning the theoretical validity of an additive (P and E) model be-
cause the P - E fit definition of job stress is based upon adiscrepancy idea.
According to Murray (1938), individual need frustration results from the
environment not satisfying the individual's needs, while need gratification
is due to proper environmental supplies. The methodological problem
with an additive model is that such an approach masks the contribution of
the components to the product. For example, a P and E product 10 could
be derived in several ways from two 7-point P, E scales: E = 5, P = 5; E =
3, P = 7; E = 7, P = 3; E = 6, P = 4; or E = 4, P = 6. Thus, results from
this study indicate that the methodological criticisms made against
discrepancy scores can be handled while utilizing a P - E fit approach to
job stress. Other researchers interested in P - E fit job stress (e.g., Harri-
son, 1978) have also found favorable results to mitigate such criticisms.

Major Findings
As the results indicated, support for the French model's hypothesized
relationships between job stress and job strains (job dissatisfaction and
ineffective job performance) was limited. Although there was very little
relationship found between job stress factors and ineffective job perfor-
mance, stronger support was given to the job stress-job dissatisfaction
relationship. However, the fact that the job stress-job dissatisfaction
relationship was verified while the job stress-ineffective job performance
relationship was not strongly suggests that the former relationship is par-
tially capitalizing on common method variance. Earlier cited research
evidence (Caplan et al., 1975) which found a strong job stress-job dis-
t~
4~

TABLE 4
CORRELATIONS AMONG JOB STRAINS AND E, P, P AND E, AND P - E STRESS FACTOR INDICES WITH LENGTH OF SERVICE CONTROLLED

Environment Person Misfit


(E) (P) (P and E) (P-E)

Strains PD P/I S/A PD P/I S/A PD P/I S/A PD P/I S/A

Ineffective
job performance .12 -.11 -.13 -.04 .01 -.01 .05 -.08 -.08 .15" .10 .08
Job
dissatisfaction .16 -.38** -.19" -.06 .18" .28** .14 -.12 .11 .14 .45**'" .38**'"

Note. PD = physical danger; P/I = passenger/intracompany; S/A = scheduling/assistance.


* p < .05.
** p < .01.
" = Significantly larger (p < .05) than corresponding E, P, P and E correlations,
EMPIRICAL INVESTIGATION 295

satisfaction relationship but no job stress-physiological-strain relation-


ship further supports the common method variance criticism.
Future research efforts must broaden their methods of measuring job
stress and job strain variables. Objective measures of P - E fit job stress
such as archival records (e.g., Coburn, 1975) or observational ratings can
be gathered. Additional methods for measuring job strains could include:
qualified judges' ratings of job performance (Andrews & Farris, 1972), or
customer observations and evaluations of employee job satisfaction
(Parkington & Schneider, 1979). Multiple independent measures of job
stress and job strain variables will allow the effect of common method
variance to be most clearly assessed.
Superintendent and co-worker social support were significantly nega-
tively related to job dissatisfaction. Also, superintendent social support
was negatively related to two job stress factors (passenger/intracompany,
and scheduling/assistance), while co-worker social support was nega-
tively related to passenger/intracompany stress. Thus, work-related social
support appears to be a useful means for not only offsetting negative
worker feelings about their jobs but also for offsetting some perceived job
stresses. Contrary to the stratified sample finding across 23 occupations
by Caplan et al. (1975), no significant relationship was found between
off-job social support and job dissatisfaction. Thus, either off-job social
support is not very important for the particular occupation of bus opera-
tor, or if Caplan et al. (1975) had partialed out the effect of length of
service (as this study did) from their results, no relationship would remain.
Very little research up to this point has focused upon the relationship of
social support measures to worker job performance. This study investi-
gated this relationship and did not find any significant results. However,
since the bus operator's actual job performance occurs in isolation from
any kind of social support, the lack of results may be due to the occupa-
tion sampled. Since the relationship between social support and job per-
formance is an important part of the French model, future research should
expand the investigation of this relationship to other work settings. In
addition, all job stress factors and both supervisory and co-worker social
support were found to account for significant additive effects upon job
dissatisfaction beyond length of service. However, since the social sup-
port, job stress, and job dissatisfaction measures were all self-report, the
effect of common method variance upon this relationship is unknown and
should be investigated in future research.
All of the above-mentioned relationships involving job stress, social
support, and job dissatisfaction were found while controlling for length of
service. It appears that no other research studies have controlled for
length of service while investigating such relationships. Although length
of service is an important variable to consider as evidenced by its ac-
296 GARY BLAU

counting for almost all variance with ineffective job performance and
changing several relationships when partialed out, length of service by
itself cannot account for the majority of relationships found within the
French model.
However, several interesting correlational results dealing with length of
service were found in this study. The first was that the job stress factor of
physical danger increased as length of service increased. This could be
explained by the fact that the longer a bus operator has been on his/her
job, the more exposure he/she has probably had to dangerous situations
(despite the fact that with increasing seniority, bus operators get more
choice as to which bus run they will drive). That passenger/intra-company
stress decreases as length of service increases could be explained by the
fact that as Argyris (1957) noted, increased job longevity allows an indi-
vidual to better adapt to his/her job environment through adjustment of
expectations/needs. The strong positive relationship between length of
service and perceived superintendent social support could be a result of
veteran bus operators having more of a chance to "get to know" each bus
district superintendent beyond his/her primarily disciplinary role.
The strong negative relationships between length of service and both
job strains (ineffective job performance and job dissatisfaction) suggest
that length of service appears to result in the self-selection of bus
operators who are able to reduce the strains associated with their jobs. In
this study, length of service was by far the strongest predictor of ineffec-
tive job performance. A plausible explanation for length of service leading
to better job performance would be through learning and practice effects.
Some combination of changing needs/expectations (e.g., Hulin & Smith,
1965) or adaptation to environmental conditions would explain why in-
creasing service length leads to less job dissatisfaction. However in this
study, increasing length of service leads to greater subject choice over
certain job aspects, i.e., type of bus run, time of bus run, days off, etc.
Such a fact could also explain a large portion of the reduction in job
strains due to increasing length of service.
Finally, another plausible and simpler (vs. the developmental hypothe-
sis) explanation for the relationships between length of service and job
strains and job stress may be due to selective attrition or turnover. Bus
operators for whom the job stress was too great, or who were too dis-
satisfied or too incompetent are less likely to still be employed. Unfor-
tunately, specific longitudinal turnover data, such as causes for turnover,
were not available. However, anecdotal evidence from transit company
personnel revealed a low turnover rate among bus operators. Instead,
probation periods and temporary j o b transfers (e.g., to information
booths) were utilized as often as possible. Further analysis of the length of
EMPIRICAL INVESTIGATION 297

service variable showed that it did not moderate or buffer any job
stress-job strain relationships.
Investigation of the idea that social support can act as a buffer for job
stress-job strain relationships has generated controversial findings (e.g.,
House & Wells, 1977 vs Pinneau, 1976). This study found two significant
superintendent social support interactions between job stress-job strain
relationships. Superintendent social support buffered the relationships
between passenger/intracompany stress and ineffective job performance
and between physical danger stress and dissatisfaction with working
conditions (a facet satisfaction scale). In view of the facts that: (1) a large
number of regression equations (see Tables 2 and 3) were analyzed to find
these two interactions, and (2) a subgroup analysis of the passenger/intra-
company stress-ineffective-job-performance relationship revealed re-
sults in the direction opposite to that predicted, these interaction findings
are best explained as chance results. Therefore, in agreement with the
findings of Pinneau (Note 3) and LaRocco & Jones (1978) and in contrast
to what the French model predicts, social support does not appear to
buffer or moderate the relationship between job stress and job strain. The
effects of social support are only additive.

The Construct Validity of P - E Fit Job Stress


Although the theoretical foundation of the P-E-fit job stress construct
is well-grounded in the theories of Lewin (1938) and Murray (1938), ques-
tions do arise regarding its validity. At a paradigmatic level, one must be
wary of such a broad conceptualization of job stress. Job stress defined in
terms of n e e d - s u p p l y fit can potentially subsume all types of organiza-
tional phenomena (e.g., supervision, job design, work environment,
interpersonal aspects). Without an identifiable context of application
which allows for comparison, there is the danger that job stress defined in
terms of n e e d - s u p p l y fit could be reduced from a theoretical construct to
an atheoretical tool or method of investigation based on simply calculating
P - E discrepancy scores. It is important for job stress researchers to be
aware of such a problem and take steps to avoid it (e.g., limiting the
context of application).
At the operationalization level of P - E fit job stress several concerns
emerge. For example, a problem arises when comparing the implied
theoretical construct definition of stress (as sustained motive arousal ten-
sion when the environment does not supply enough of what an individual
needs) with the French model's operationalization as a discrepancy in
either direction (oversupply or undersupply). Part of this problem can be
resolved by understanding the types of needs being considered in a situa-
tion.
298 GARY BLAU

The theoretical-construct definition emphasizes the kind of needs


where an oversupply would not be likely to generate motive tension.
House (1972) cites the example of an individual with a certain need level
for personal growth. As the person is presented with more growth oppor-
tunities than he/she could want or use, motive arousal should remain
stable (no stress induced). The majority of items in this study were also of
this type (see Appendix). For other types of needs, however, job stress
can logically result from either an oversupply or undersupply. As an
example, Harrison (1978) cites job complexity. Insufficient job complex-
ity as well as too much job complexity (relative to one's needs) can be
stress producing. However, one may question if the psychological mean-
ings of equally differing discrepancies (oversupply vs undersupply on job
complexity) are equally indicative of job stress. It seems that no research
attention has been given yet to this issue.
A similar concern arises when considering the demand-ability defini-
tion of P - E fit job stress. Also as McGrath (1970) notes, the nature of
stressful imbalance raises temporal questions regarding potential fluctua-
tions is an individual's capabilities (or needs) regardless of environmental
changes. For example, is a demand such that: (a) the person could not
handle it under any circumstances, or (b) it could normally be handled, but
it is beyond momentary capability due to either a temporary resource
depletion or other concurrent demands. P - E fit job stress in terms of
either demand-ability or need-supply dimensions must be able to deal
with extreme situations as well as "slow-fuse" situations (e.g., through
gathering longitudinal data).
Also, further analysis of the relationship between need-supply and
ability-demand dimensions of P - E fit job stress would seem called for.
The French model implies (e.g., Harrison, 1978) that ability-job demand
fit is a prerequisite to achieving need-supply fit. Although this order
would seem to be a logical sequence in most work situations, there are
unaccounted for exceptions (e.g., where a worker is unable to meet his/
her job demands but is not dismissed because of union pressure, and
furthermore achieves need-supply fit in some areas).
Finally, the author is aware of a discrepancy score definition of job
satisfaction (Porter, 1961), and the argument of a tautological connection
with P - E fit job stress. Job satisfaction is most commonly defined and
measured as a general affective response which a person makes to his/her
job (e.g., Brayfield & Rothe, 1951; Smith, Kendall, & Hulin, 1969; Locke,
1976). However, a discrepancy score definition of job satisfaction em-
phasizes why individuals become satisfied (the rewards meet or exceed
the perceived equitable level) rather than the idea of an affective re-
sponse.
EMPIRICAL INVESTIGATION 299

Although P - E fit job stress and the discrepancy score definition of job
satisfaction come from the same theoretical background of p e r s o n -
environment interaction (Lewin, 1938), there are several differences
in the operationalization of the respective constructs. One difference
is that the discrepancy score definition of job satisfaction is concerned
with the amount of discrepancy from perceived equity (Should there
b e - I s now); while the P - E fit definition of job stress emphasizes a dis-
crepancy from fulfillment of desires (Would you need/like-Is now). Also,
the discrepancy score definition of job satisfaction only defines job dis-
satisfaction in terms of undersupply (insufficient supplies for perceived
equity). However as we have noted, the P - E fit definition of job stress
can deal with job components (e.g., job complexity) in which both under-
supply and oversupply are indicative of job stress.
Furthermore, there is a difference in specificity between discrepancy
score job satisfaction measures and P - E fit job stress measures. Porter's
(1961) discrepancy score job satisfaction measure is centered around
Maslow's (1955) need heirarchy and has only been applied to managerial
jobs. P - E fit job stress measures are based upon situationally specific job
components (e.g., job complexity, quantitative workload) and have been
applied to a broad cross section of jobs (Caplan et al., 1975).

CONCLUSION
This study provides only limited support for the French model of job
stress. A P - E fit approach to job stress was found to be methodologically
defensible, the additive role of work-related social support in accounting
for job dissatisfication was verified, and a strong relationship was found
between job stress and job dissatisfaction (a psychological strain). How-
ever, no relationship was found between job stress and ineffective job
performance (a behavioral strain). Such conflicting job stress-job strain
relationship results emphasize the need for multiple methods of measuring
job stress and job strain variables in future research so that the effect of
common method variance can be evaluated. Furthermore, in contrast to
the French model's prediction, no type of social support (supervisor,
co-worker, off-job) acted as a buffer between any job stress-job strain
relationships.
In closing, several specific suggestions for improvement of the French
job stress model should be considered. At present, the model does not
allow for measuring the importance of any particular P - E fit discrepancy
to the subject. Such a measure could help point out the relative contri-
butions of "slow-fuse" situations vs extreme situations to job stress.
Also, the incorporation of the variable length of service into the model
would seem justified.
300 GARY BLAU

APPENDIX: FACTOR LOADINGS FOR JOB STRESS


(PERSON-ENVIRONMENT MISFIT)

Factor I Factor II Factor III


(Passenger/ (Scheduling/ (Physical
Item content intracompany) assistance) danger)

l, Operator cooperation .51 .16 ,01


2. Passenger respect/obeying rules .56 .16 ,01
3. Management concern .69 .13 .01
4. Fair company rules .61 .21 .02
5. Operator participation in rule
setting .55 .18 .13
6. Bus in proper driving condition .63 .24 .10
7. Helpful field supervision .57 .27 .05
8. Knowledgeablepassengers .59 .38 .O9
9. Quick transit police response .61 .32 .05
10. Flexible bus schedule .28 ~ .05
11. Enough layover time .34 .02
12. Proper bus service .35 .02
13. Helpful radio control assistance .37 .10
14. Time to get used to bus run
before moving on .15 .06
15. Properly arranged cross-line
relief points .15 .06
16. Serious accidents .04 .07
17. Serious situation (e.g., robbery) .03 .06
18. Being physically threatened .01 .01
Eigenvalues 6.09 2.19 1.53

REFERENCES
Andrews, F., & Farris, G. Time pressure and performance of scientists and engineers: A
5-year panel study. Organizational Behavior and Human Performance, 1972, g,
185- 200.
Argyris, C. Personality and organization. New York: Harper & Row, 1957.
Beehr, T., & Newman, J. Job stress, employee health, and organizational effectiveness: A
facet analysis, model, and literature review. Personnel Psychology, 1978, 31,665-679.
Brayfield, A., & Rothe, H. An index of job satisfaction. Journal of Applied Psychology,
1951, 35, 307-311.
Caplan, R. D., Cobb, S., French, J. R. P., Jr., Harrison, R. V., & Pinneau, S. R., Jr. Job
demands and work health (U.S. Department of Health, Education, and Welfare,
NIOSH Publication No. 75-160). Washington, D.C.: U.S. Government Printing Office,
1975.
Cobb, S. Social support as a moderator of life stress. Psychosomatic Medicine, 1976, 38,
300-314.
Coburn, D. J o b - w o r k e r incongruence--Consequences for health. Journal of Health and
Social Behavior, 1975, 16, 198-212.
Cooper, C., & Payne, R. (Eds.) Stress at work. New York: Wiley, 1978.
Cronbach, L., & Furby, L. How should we measure " c h a n g e " - - o r should we? Psychologi-
cal Bulletin, 1970, 74, 68-80.
Dawis, R., England, G., & Lofquist, L. Minnesota Studies in Vocational Rehabil#ation
(No. 30): A theory of Work adjustment. Minneapolis, University of Minnesota, 1968.
Finkehnan, J., & Glass, D. Reappraisal of the relationship between noise and human per-
formance by means of a subsidiary task measure. Journal of Applied Psychology, 1970,
54, 211-213.
EMPIRICAL INVESTIGATION 301

Fleishman, E. A. The leader opinion questionnaire. In R. M. Stogdill - - A. E. Coons (Eds.),


Leader behavior: Its description and measurement. Ohio State University: Research
Monograph No. 88, 1957.
French, J. R. P., Jr. Person-role fit. In A. McLean (Ed.) Occupational stress. Springfield,
II1.: Thomas, 1974.
French, J. R. P., Jr., & Caplan, R. D. Organizational stress and individual strain. In A.
Marrow (Ed.) The failure of success. New York: AMACOM, 1972.
French, J. R. P., Jr., & Kahn, R. L. A programmatic approach to studying the industrial
environment and mental health. Journal of Social Issues, 1962, 18, 1-47.
French, J. R. P., Jr., Rogers, W., & Cobb, S. A model of person-environment fit. In G. V.
Coetho, D. A. Hamburgh, & J. E. Adams (Eds.), Coping and adaptation. New York:
Basic Books, 1974.
Gross, E. Work, organization, and stress. In S. Levine and N. Scotch, (Eds 3, Social stress.
Chicago: Aldine, 1970.
Halpin, A., & Winer, B.J. A factorial study of leader behavior description. In. R. M.
Stogdill & A. E. Coons (Eds.), Leader behavior: Its description and measurement.
Ohio State University: Research Monograph No. 88, 1957.
Harrison, R. V. Person-environment fit and job stress. In C. Cooper & R. Payne (Eds.),
Stress at work. New York: Wiley, 1978.
Hill, R. Generic features of families under stress. Social Casework, 1958, 39, 139-150.
House, J. S. The relationship of intrinsic and extrinsic work motivation to occupational
stress and coronary heart disease risk. (Doctoral dissertation, University of Michigan,
1972) Dissertation Abstracts International, 1972, 33, 2514A.
House, R. J., & Rizzo, J. R. Role conflict and ambiguity as critical variables in a model of
organizational behavior. Organizational Behavior and Human Performance, 1972, 7,
464- 505.
Hulin, C., & Smith, P. A linear model of job satisfaction. Journal of Applied Psychology,
1965, 49, 209-216.
Janis, I. Psychological stress. New York: Wiley, 1958.
Johnson, T. W., & Stinson, J. E. Role ambiguity, role conflict and satisfaction: Moderating
effects of individual differences. Journal of Applied Psychology, 1975, 60, 329-333.
Kahn, R. L., Wolfe, D. M., Quinn, R. P., Snoek, J. D., & Rosenthal, P. A. Organizational
stress: Studies in role conflict and ambiguity. New York: Wiley, 1964.
Katz, D., Maccoby, N., & Morse, N. C. Productivity, supervision, and morale in an office
situation. Ann Arbor, Mich.: Institute for Social Research, 1950.
Landy, F. J., & Trumbo, D. A. Psychology of work behavior. Homewood, I11.: Dorsey
Press, 1976.
La Rocco, J. M., & Jones, A. P. Co-worker and leader support as moderators of s t r e s s -
strain relationships in work situations. Journal of Applied Psychology, 1978, 63,
629-634.
Lewin, K. The conceptual representation and measurement of psychological forces.
Durham, N.C.: Duke Univ. Press, 1938.
Locke, E. The nature and causes of job satisfaction. In M. Dunnette (Ed.), Handbook of
industrial and organizational psychology. Chicago: Rand McNally, 1976.
Maslow, A. Deficiency motivation and growth motivation. In M. Jones (Ed.), Nebraska
symposium on motivation. Lincoln, Univ. of Nebraska Press, 1955.
Matteson, M., & Ivancevich, J. Organizational stressors and heart disease: A research
model. Academy of Management Review, 1979, 4, 347-357.
McClelland, D., & Apicella, F. Reminiscence following experimental induced failure. Jour-
nal of Experimental Psychology, 1947, 37, 159-160.
McGrath, J. E. A conceptual formulation for research on stress. In J. E. McGrath (Ed.),
Social and psychological factors in stress. New York: Holt, Rinehart & Winston, 1970.
302 GARY BLAU

McLean, A. Occupational stress--A misnomer. In A. McLean (Ed.), Occupational stress.


Springfield, Ili.: Thomas, 1974.
Murray, H. Explorations in personality. New York: Oxford Univ. Press, 1938.
Parkington, J., & Schneider, B. Some correlates of experienced job stress: A boundary role
study. Academy of Management Journal, 1979, 22, 270-281.
Porter, L. W. A study of perceived need satisfiers in bottom and middle management jobs.
Journal of Applied Psychology, 1961, 45, 1-10.
Schuler, R. Definition and conceptualization of stress in organizations. Working paper,
College of Administrative Science, Ohio State University, 1979.
Selye, H. The stress of life. New York: McGraw-Hill, 1956.
Smith, P., Kendall, L., & Hulin, C. The measurement of satisfaction in work and retire-
ment. Chicago: Rand McNally, 1969.
Wall, T., & Payne, R. Are deficiency scores deficient?Journal of Applied Psychology, 1973,
58, 322-326.
Weiss, D. J., Dawis, R. V., England, G., & Lofquist, L. H. Manual for the Minnesota
Satisfaction Questionnaire. Minnesota Studies in Vocational Rehabilitation (No. 22).
Minneapolis: Univ. of Minnesota, 1967.
Yerkes, R.~ & Dodson, J. D. The relationship of stimulus to rapidity of habit formation.
Journal of Comparative Neurological Psychology, 1908, 18, 459-482.
Zedeck, S. Problems with the use of moderator variables. PsychologicalBulletin, 1971, 76,
295-310.

REFERENCE NOTES
1. French, J. R. P., Jr. Job demands and work health: Introduction. Paper presented at the
84th Annual Convention of the American Psychological Association, September
1976.
2. House, J. S., & Wells, J. A. Occupational stress, social support, and health. Paper
presented at the Conference on Reducing Occupational Stress, White Plains, New
York, May 1977.
3. Pinneau, S. R., Jr. Effects of social support on occupational stress and strain. Paper
presented at the 84th Annual Convention of the American Psychological Associa-
tion, September, 1976.
4. House, J. S., McMichael, A.J., Kaplan, B., & Wells, J. A. Effects of occupational
stress on the health of rubber workers. Paper presented at the American Public
Health Association meetings, Durham, N.C., 1975.

RECEIVED: February 28, 1980

You might also like