Download as pdf or txt
Download as pdf or txt
You are on page 1of 20

This article was downloaded by: [Princeton University]

On: 27 July 2013, At: 05:55


Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number:
1072954 Registered office: Mortimer House, 37-41 Mortimer Street,
London W1T 3JH, UK

Human Performance
Publication details, including instructions
for authors and subscription information:
http://www.tandfonline.com/loi/hhup20

More Evidence on
Relationships Between the
Work Environment and
Job Performance
Darlene M. Olson & Walter C. Borman
Published online: 12 Nov 2009.

To cite this article: Darlene M. Olson & Walter C. Borman (1989)


More Evidence on Relationships Between the Work Environment and
Job Performance, Human Performance, 2:2, 113-130, DOI: 10.1207/
s15327043hup0202_3

To link to this article: http://dx.doi.org/10.1207/s15327043hup0202_3

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of


all the information (the “Content”) contained in the publications
on our platform. However, Taylor & Francis, our agents, and our
licensors make no representations or warranties whatsoever as to the
accuracy, completeness, or suitability for any purpose of the Content.
Any opinions and views expressed in this publication are the opinions
and views of the authors, and are not the views of or endorsed by
Taylor & Francis. The accuracy of the Content should not be relied
upon and should be independently verified with primary sources of
information. Taylor and Francis shall not be liable for any losses,
actions, claims, proceedings, demands, costs, expenses, damages,
and other liabilities whatsoever or howsoever caused arising directly
or indirectly in connection with, in relation to or arising out of the use
of the Content.
This article may be used for research, teaching, and private study
purposes. Any substantial or systematic reproduction, redistribution,
reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access
and use can be found at http://www.tandfonline.com/page/terms-
and-conditions
Downloaded by [Princeton University] at 05:55 27 July 2013
HUMAN PERFORMANCE, 2(2), 113-130
Copyright @ 1989, Lawrence Erlbaum Associates, Inc.

More Evidence on Relationships


Between the Work Environment
and Job Performance
Downloaded by [Princeton University] at 05:55 27 July 2013

Darlene M. Olson
US. Army Research Institute
Alexandria, VA

Walter C. Borman
Personnel Decisions Research Institute
Minneapolis, MN

The purposes of this research were: (a) to develop a work environment question-
naire appropriate for assessing the situational constraints and facilitators of per-
formance in the US.Army and (b) to evaluate relationships between individual
aspects of the work environment and a wide array of performance measures.
The Army Work Environment Questionnaire (AWEQ) was developed and adrnin-
istered to 5,080 first-term Army enlisted personnel working in nine jobs. Also
available for each member of this sample were peer and supervisor performance
ratings, along with scores on job knowledge and work sample task proficiency
tests. The results of factor analyses indicated that a five-factor solution provides a
parsimonious explanation of the underlying Army work environment. Correla-
tions of about .20 were found: (a) between ratings on some of the performance
dimensions and the environmental factors, Job/'Ttlsk Importance and Supervisor
Support; and (b) between performance on task proficiency measures and the
environmental construct relating to relevance of training for present job assign-
ment. Discussion focuses on the value of maintaining separate environmental
constructs when examining environment-performance relationshipsand addresses
issues related to interpreting the work environment questionnaire responses.

Individuals' work performance is a function primarily of knowledges, skills,


abilities, and other personal characteristics that may contribute to or detract
from effective functioning on the job (e.g., Dunnette, 1976) and environmen-

Requests for reprints should be sent to Darlene M. Olson, U.S. Army Research Institute,
5001 Eisenhower Avenue, Alexandria, VA 22333-5600.
114 OLSON AND BORMAN

tal influences that may constrain or enhance performance (Peters & O'Con-
nor, 1980). To better understand effectiveness on jobs, it is important to learn
more about links between job performance, the person, and situation factors.
Focusing on situation factors, environmental and organizational variables
present in the work setting are known to influence performance in two ways.
First, they can influence performance through constraint (Naylor, Pritchard,
& Ilgen, 1980). The environment can inhibit, interfere with, or set limits on
the range of work behaviors that are displayed which, in turn, potentially
affects task performance and the relationship between ability and perfor-
mance. Second, the environment can impact on performance through affec-
tive responses to the work setting (Naylor et al., 1980). For example, the work
Downloaded by [Princeton University] at 05:55 27 July 2013

environment provides information about the organizational reward system


which subsequently may arouse motives, affective reactions, and expectations
for certain consequences of designated behaviors.
Regarding influences of the work environment on performance, Peters and
O'Connor (1980) initiated an important research program to identify situa-
tional constraints that might: (a) impact directly and negatively on work per-
formance and (b) moderate ability performance relationships. In a series of
laboratory experiments, these researchers demonstrated that constraints on
performance can adversely affect task performance and affective reactions
to the job (e.g., frustration and dissatisfaction; Peters, Chassie, Lindholm,
09Connor,& Kline, 1982; Peters, O'Connor, & Rudolf, 1980). Results of an
initial field study (O'Connor et al., 1984) conducted using managers in civilian
work settings found significant but weak relationships between overall con-
straint scores and supervisory performance ratings (rs = - .12). Similar lower
correlations emerged between constraint scores and performance ratings in
an Air Force sample (Peters, O'Connor, Eulberg, & Watson, 1988; Watson,
O'Connor, Eulberg, & Peters, 1983). Reasons offered for the different results
based on laboratory and field studies include: (a) work environments in-
vestigated in the field had few constraints, restricting the range of environmen-
tal constraint scores; and (b) raters providing performance ratings in the field
studies somehow take into account and "allow for" differences in constraints
for different ratees.
More substantial relationships between work environment and performance
were evident in a field study conducted by Steel and Mento (1986). They found
significant effects of high versus low situational constraint environments on
supe.~isoryappraisals, self-ratings, and a measure of objective performance
for a sample of branch managers in a finance company. Further, Steel and
Mento (1986) demonstrated that a composite constraint score explained more
criterion variance in the ratings (r2= .13) than in the objective performance
scores (r2= .01).
Although this early work has been informative in identifying possible
dimensions of work environments and in beginning to understand relation-
ships between environmental variables and job performance, more research
is needed to develop environmental taxonomies that might operate in other
occupational settings (eg., Army jobs). Also, because previous research has
focused on supervisory ratings of overall performance as the criterion measure
(with the exception of Steel & Mento, 1986), it seems important to evaluate
relationships between environment and a broader array of subjective and ob-
jective performance measures that may tap a wider variety of different criterion
constructs (e.g., technical proficiency, job knowledge, effort on the job, etc.).
Finally, because of the likely differt .!tiation among environmental constructs
measuring distinct aspects of the environment, it is important to examine rela-
tionships between measures of these individual environment constructs and
Downloaded by [Princeton University] at 05:55 27 July 2013

each of the different criterion elements reflected in the array of performance


measures.
Accordingly, the purpose of this research was: (a) to investigate dimen-
sions of the work environment for U.S. Army enlisted jobs and (b) to exam-
ine patterns of relationships between individual work environment factors
and different kinds of performance measures. Our work builds on previous
efforts (Peters & O'Connor, 1980) and extends that research by: (a) focusing
on individual dimensions of the work environment when exp10rhg environment-
performance relationships, with the assumption that some elements of the
environment might relate more strongly to performance than others; and (b)
including in the set of performance measures a wider range of ratings (i.e.,
supervisor and peer evaluations) and objective measures (tapping job
knowledge and task proficiency) than has been previously investigated.
In this research, the AWEQ was developed using a modified critical in-
cidents procedure. The AWEQ was pilot-tested and revised and then ad-
ministered to 5,080 first-tour U.S. Army enlisted personnel working in nine
different jobs. Also administered were performance rating scales to peers and
supervisors of the subject soldiers, and the same soldiers completed job
knowledge and task proficiency work sample tests. The following sections
describe development of the AWEQ, its underlyim structure, and relation-
ships between AWEQ scales and the performance measures.

METHOD

Developing the WEQ


Critical incidenb gcs~lemtionand of th w r k wir0r)ment
taxrw#wny. The critical incidents method (matta%;a.n, 1954)was used as a start-
ing point for development of the taxonomy. Specifically, 67 commissioned
and noncommissioned officers (NCOs) in a wide range of U.S. Army jobs
were asked to write stories about instances where effective or ineffective per-
formance on the part of a first-tour soldier occurred due to situational fac-
tors beyond his or her control. W o examples of incidents follow:
116 OLSON AND BORMAN

TABLE 1
14 Dimension Taxonomy of the Army Work Environment
- --

I . Resources/tools/equipment
2. Workload/time availability
3. Training in MOS skills/opportunity to improve MOS skills
4. Physical working conditions
5. Job relevant authority
6. Job relevant information
7. Perceived job importance
8. Work assignment
9. Changes in job procedures and equipment
10. Reward/recognition/ feedback
Downloaded by [Princeton University] at 05:55 27 July 2013

1 1 . Discipline practices
12. Individual support
13. Job support
14. kaderlpeer role models

A soldier was doing extremely well at his assigned tasks. Another soldier
in the work group was not performing nearly as well. In addition to
his own job responsibilities, the high-performing soldier was required
to help the low-performing soldier with his assignments, so that the
supervisor could be assured both jobs were completed. The performance
of the higher achiever dropped off as the situation continued.
Soldiers performed well during tank gunnery exercises. The brigade com-
mander called soldiers from each company with outstanding individual
performance to the front of the brigade formation and verbally recog-
nized their contributions to success on the exercise. Overall these soldiers'
performance improved on later field exercises.
The officers and NCOs generated a total of 282 such incidents. Six PhD
industrial psychologists familiar with research on work constraints and with
the Army then independently sorted these critical incidents into categories
according to their conceptual similarity. After the critical incidents were con-
tent analyzed, the psychologists, serving as judges, discussed and reconciled
any differences in the dimensions generated. Table 1 presents the taxonomy
of the 14 environmental dimensions resulting from the content analysis of
the critical incidents. The first nine dimensions in the taxonomy are directly
job related, and the remaining five are more climate oriented.
Next, the 14 dimensions were carefully defined by the combined group of
judges according to the critical incidents representative of the specific dimen-
sion. %o example definitions are shown in Table 2. A feature of these defini-
tions that distinguishes this effort somewhat from previous work in this area
(Eulberg, O'Connor, Peters, & Watson, 1984) is that bipolar constructs are
represented, which range from inhibiting or constraining performance to
enhancing or facilitating performance. In a sense, this approach expands the
WORK ENVIRONMENT AND JOB PERFORMANCE 117

TABLE 2
Examples of Definitions for Individual Work Environment Dimensions
- --

Job Dimension

Workload/Time Availability
Workload too heavy, assigned additional details (e.g., training and inspection prepara-
tion), after duty hours; required to work longer shifts due to personnel shortages; good
performers given others' tasks to complete in addition to their own.
Too little time, given unreasonable time limit to complete a specific job, or the assigned
workload consistently too great for time limit; no scheduled time for tasks that are low
priority but essential (cg., maintenance); frequent interruptions (e.g., special duties) conflict
with task completion.
Downloaded by [Princeton University] at 05:55 27 July 2013

Workload too light, too many personnel assigned to a job; unit tasked with too little
work; soldiers must perform "busy work."
Versus
Workload is commensurate with available time limits. It is usually possible to finish all
assigned tasks within the scheduled time limit. Workload is distributed evenly across unit
members.
Assignments are carefully scheduled so that low-priority items can be completed during
slow periods. To the extent possible, training activities and special details are scheduled to
coincide with slack time in the work schedule.

Climate Dimension

Rewards/Recognition/PositiveFeedback
Good performance ignored, inconsistently or inequitably rewarded either due to Army-wide
policies or leadership practices.
Versus
Good performance consistently and fairly rewarded/recognized by chain of command
(e.g., at command level, awards, soldier of month, local recognition; at supervisor level,
praise, favorable assignments, promotion recommendation, passes, etc.).
- -

Note. Dimensions represented here are from the 14 dimension work environment taxonomy;
subsequent factor analysis work yielded five summary environment constructs.

constraint to no-constraint continuum to consider the facilitation half of the


continuum for each dimension.
After the environmental dimensions were defined, the entire set of critical
incidents was resorted into the 14 dimensions by four of the previous judges.
The critical incidents were retranslated correctly into their respective dimen-
sions 76% of the time.

item writing. The next step in questionnaire development was to write


items for each of the 14 dimensions. Items included were descriptive of the
environment rather than evaluative of it and represented the content of the
target environmental dimension. One hundred ten items were generated in
all; the number of items for individual dimensions ranged from a low of 6
to a high of 11. Items on the AWEQ were evaluated on a 5-point frequency
rating scale, ranging from very seldom or never (1) to very often or always
118 OLSON AND BORMAN

TABLE 3
Description of Sample

Army M Months
Job n Men Women in Company

Infantryman
Cannon crewman
Armor crewman
Radio operator
Light wheel vehicle mechanic
Motor transport operator
Administrative specialist
Downloaded by [Princeton University] at 05:55 27 July 2013

Medical specialist
Military police

(5). Respondents were asked to indicate how often each environmental situa-
tion described in a questionnaire item occurred on their present job. For ex-
ample, items consisted of statements such as: "In your job, changes in equip-
ment are introduced with little or no explanation" (changes in job procedures
and equipment); or "If you needed help, you could depend on your co-workers
to help you perform your required job tasks" (job support). For the entire
AWEQ, 53 of the items were worded negatively (as in the first example) and
57 positively (as in the second example).

Field tests of the WEC?. Before administration to the main sample, the
AWEQ was field tested extensively and revised b& on results of those field
tests. Briefly, the 110-item version of the AWEQ was administered to two waves
of first-tour soldiers (1-3 years service): 548; personnel in four different jobs
and 821 soldiers in five other jobs (see %ble 3 for a listing of these jobs).
The soldiers were stationed in seven different locations in the continental U.S.
and four locations in Europe Principal components factor analpis with
varimax rotation was applied to AWEQ data from each wave. Thew solu-
tions were highly similar to the five-factor solution obtained in the main
sample (N = 5,08Q) and discussed in more detail in the results section.
The number of items in the AWEQ was reduced from 110 to 53 between
the second field test and the main sample administration. Criteria for item
deletion were that in one or both field tests: (a) the itsm did not load highly
on an interpretable factor or (b) the item had two or more high lo&dhgs or
low commonalities. A further restriction here was that, because of limits on
the testing time avdabk for the main sample administmation, the total number
of items had to be reduced to approximately 50. Therefore, the items that
remained after the deletion step met the two criteria just mentioned and, as
a total set, reflected as much as possible all the elements of the factors from
field test factor analysis results. (A more complete description of scale develop-
ment work can be found in Olson and Borman, 1986.)

Main Sample
As already mentioned, the main sample consisted of 5,080 soldiers in nine
different Army jobs. The sample size for each job in addition to the gender
composition and experience level in present company is depicted in Table 3.

Administering the AWEQ and the Performance Measures


Downloaded by [Princeton University] at 05:55 27 July 2013

Description of the performance measures. Criterion development work


for Project A was organized around a theory of performance, which postulates
that job performance is multidimensional. Job performance is defined by
the diverse set of behaviors or activities individuals engage in that are judged
as important for accomplishing the goals of the organization. For entry-level
occupations in the Army, two major types of job performance can be
described-Army-wide performance and job-specific performance. Army-wide
performance refers to components in which all soldiers are required to display
proficiency. Examples of Army-wide performance include first aid procedures,
peer leadership, and land navigation skills. Job-specific performance is im-
portant for technical competence in a designated job. For example, prepar-
ing a howitzer for firing would be a performance element for a cannon
crewman, but not for an administrative clerk.
Specifically, performance measures were: (a) rating scales relevant for
evaluating soldiers in any first-tour Army job (Army-wide scales: Borman,
Motowidlo, Rose, & Hanser, 1984; Pulakos & Borman, 1986); (b) job-specific
rating scales poquam et al., 1986); and (c) hands-on task proficiency measures
and job knowledge tests (Campbell, Campbell, Rumsey, & Edwards, 1986).
A more complete description of the theory of performance advocated by this
research and the performance measures used in this research can be found
in the Project A Annual Report for Fiscal Year 1986 (Campbell, 1988).
The Army-wide scales were developed using Behaviorally Anchored Rating
Scale methodology (BARS; Smith & Kendall, 1963) and focus on performance
dimensions appropriate for all Army jobs (e.g., following rules, regulations,
and orders and maintaining equipment). A principal components factor
analysis (with varimax rotation) of pooled peer and supervisor ratings' on
the 10 Army-wide scales resulted in an interpretable three-factor solution: (a)
Effort and Leadership (b) Personal Discipline and (c) Military Bearing

'Peer and supervisor ratings were pooled because there was reasonably good interrater agree-
ment between the two sources and the patterns of relationships with other variables in the research,
which were very similar for the two sets of ratings.
120 OLSON AND BORMAN

(Pulakos & Borman, 1987). Unit weighted composites derived from this solu-
tion were used in subsequent analyses, along with an overall effectivenesscom-
posite formed by unit weighting ratings on each dimension.
Job-specific rating scales were developed using the same BARS
methodology; they focused on performance areas more narrowly relevant to
particular jobs (e.g., loading cargo and transporting personnel for the motor
transport operator job). For the peer/supervisor job-specific ratings, unit
weighted composites, like the one described for the Army-wide effectiveness
index, were formed for each of the nine job rating scales.
Finally, for each of the nine jobs separately, paper-and-pencil, multiple-
Downloaded by [Princeton University] at 05:55 27 July 2013

choice items were developed to assess knowledge about how to perform 30


important tasks for the job, and job-specific hands-on task proficiency
measures were developed to tap skills in completing 15 of these 30 tasks for
each of the jobs. For our purposes, a single overall job knowledge test score
and a single summary task proficiency test score represented these two per-
formance areas for individual soldiers. Accordingiy, a percentage of items
correct index was formed for the job knowledge area, and a percentage of
performance steps passed index served as the overall task proficiency score.

AdminEstmtion pmedures. The rating scales were administered to groups


of about 15 or fewer peers or supervisors of the target ratees after they were
trained using a combination error and accuracy training approach (e.g.,
Pulakos, 1984). On average, 1.90 supervisor raters and 3.26 peer raters per
ratee provided these performance evaluations on the Army-wide and job-
specific behavior rating scales. During the peer rating sessions, raters (who
were ratees and members of the sample as well) also responded to the AWEQ.
Hands-on task proficiency was assessed by administering to each soldier
in the sample 15 individual work samples, rqresenting 15 of the most im-
portant and representative tasks for that job. Experienced job incumbents
or supervisors were trained as hands-on scorers and used a relatively objec-
tive checklist to evaluate each soldier on each work sample task associated
with the job (Campbell et al., 1986). Each of the nine job knowledge tests,
consisting of about 300 items in all, were administered to groups of 15-30
soldiers.

Analyses
Internal consistencies of the resulting AWEQ factors were computed using
Cronbach's coefficient alpha formula. Data analyses also addressed the in-
ternal consistency of the hands-on and job knowledge tests as well as inter-
rater reliabilities of the various rating measures.
Data reduction focused first on factor analyses of the AWEQ. As men-
tioned, field test analyses with samples of 548 and 821 soldiers resulted in
the final 53-item version of the AWEQ for administration to the main sample.
WORK ENVIRONMENT AND JOB PERFORMANCE 121

This 53-item AWEQ was administered to 5,080 soldiers in the nine jobs
represented in Table 3. Item responses were intercorrelated and then factored
using the principal factors method with varimax rotation. Likewise on the
performance side, factor analyses were conducted to evaluate the structure
of the ratings.
Unit homogeneity indices for the AWEQ were computed to determine the
degree of within-Army unit agreement (compared to across-unit agreement)
in perceptions of the work environment for groups that might be expected
to have similar environments. Specifically, we used the form of the intraclass
correlation, ICC (I), suggested by James (1982) and designed to assess level
of agreements between individual unit members in their scores on each en-
Downloaded by [Princeton University] at 05:55 27 July 2013

vironmental dimension. Data from 2,358 individuals who could be definitively


identified as being a member of 1 of 258 company-sized units provided the
unit homogeneity estimates.
Regarding relationships between AWEQ scales and the performance
measures, zero-order correlations were computed between each environmen-
tal scale and each performance construct measure. Also, multiple correlations
were calculated, regressing environmental scales against individual perfor-
mance measures.

RESULTS

Results appear in four parts. First, means, standard deviations, and reliabilities
of all the measures are summarized. Second, AWEQ factor analysis results
are presented. Third, the issue of objective reality versus individual percep-
tion interpretations of AWEQ scores is addressed by presenting intraclass cor-
relation findings. And fourth, relationships between environment scale scores
and performance construct scores are examined.

Psychometric Properties of the Measures


Table 4 summarizes the means, standard deviations, and reliabilities of AWEQ
factors and the performance measures. Reliabilities presented for the AWEQ
are coefficient alpha estimates of internal consistency. Most were reasonably
high, with reliabilities ranging from .65 to .84. The interrater reliabilities were
approximately S O for the Army-wide scales and slightly lower for the job-
specific scales. The median split-half reliabilities for the hands-on and job
knowledge measures were .66 and .85, respectively.

Factor Analysis Results


Table 5 presents results of the AWEQ factor analysis. The varimax rotated
factors were interpretable, and each is named and defined herein. Results from
122 OLSON AND BORMAN

TABLE 4
Means, Standard Deviations, and Reliobilities of the W M Factors
and Performance Measures
Reliabilitiesb
Number of
Research Measures Items M" SD Range Median

Work Environment Factors


General Constraints
Supervisor Support
Training
Job/Task Importance
Downloaded by [Princeton University] at 05:55 27 July 2013

Cohesion/Peer Support
Environment Composite
Performance Criteria
Overall effectivenessc
Task proficiency (hands-on)
Job knowledge
Army-Wide Rating Compositese
Effort and leadership
Personal discipline
Military bearing

"AWEQ ratings were made on a 5-point scale; performance ratings were made on a 7-point
scale. A percentage of items correct index was constructed for the job knowledge tests, and a
percentage of performance steps passed index served as the task proficiency score. bRanges
of reliabilities are reported for jobs. Work environment reliabilities are coefficient alphas (measures
of internal consistency); hands-on and job knowledge reliabiiities are split-half indices; inter-
rater reliabilities are shown for the Army-wide rating factors. Interrater reliabilities are across
supervisor and peer raters. The overall effectiveness rating is the sum across the 10 Army-wide
rating dimensions. dScoreson the job knowledge and hands-on measures were standardized
within job to M = 50 and SD = 10. CEffort and leadership composite was derived from five
Army-wide rating dimensions: technicalknowledge, IGadership, effort, self-develop~mt,and main-
taining equipment. Personal Disciplne composite was derived from three Army-wide rating dimen-
sions: following regulations, self-control, and integrity. The Military &aring composite con-
sists of Army-wide dimensions related to military appearance and physical fitness.

this factor analysis along with internal consistency reliability analyses guided
further revisions to the AWEQ and indicated that a 38-item, five-factor solu-
tion provided the most parsimonious statistical explanation of the underly-
ing Army work environment constructs. Specifically, items on the AWEQ were
reduced from 53 to 38 by retaining those items with the largest factor loadings,
which maintained the conceptual integrity of the original instrument. Com-
posite environmental dimensions based on this 38-item, five-factor solution
and utilizing item unit weighting were employed subsequently when relation-
ships between environmental constructs and performance were examined. The
five AWEQ factors are defined as:

1. General Situational Constraints include a lack of: (a) resources, such


as tools and equipment, to help complete work; fb) information necessary
WORK ENVIRONMENT AND JOB PERFORMANCE 123

to do job or receiving wrong information; and (c) sufficient time to com-


plete tasks (reverse scored to be consistent with the other four factors).

TABLE 5
Summary of AWEQ Factor Analysis Results
-

Rotated Factor Loadings

Items I a 2 3 4 5

1. Your job is made harder because you are not


given enough of the necessary materials,
Downloaded by [Princeton University] at 05:55 27 July 2013

supplies, or parts to complete a job.


2. You cannot get your job done on time because
you are not notified in advance of schedule/
deadline changes.
3. There are unscheduled activities to work on
which keep you from getting your job done.
4. When a squad member is having problems coping
with Army life or the job, your supervisor
tries to help him or her.
5. You get recognition from supervisors for the
work you do.
6. Your immediate supervisor has a real interest
in your personal welfare.
7. You are assigned to work you were not trained
for in Advanced Individual Iltaining (AIT).
8. You are assigned to do the kind of work the
Army trained you to do.
9. You are assigned to a job that is outside of
your Military Occupational Specialty (MOS).
10. The tasks you perform are important to you
and to others.
11. The tasks you perform do not require much
skill-"anyone" could do them.
12. Your skills and abilities are important
for getting the job done.
13. If you need help, you can depend on your
co-workers to help you perform your required
job tasks.
14. The soldiers in your work group help each
other out when they have personal problems.
15. You can rely on your work group to help you
out on the job during difficult times.

Notes. Appearing in the table are marker items, reflecting the three highest loadings for
each factor. The highest factor loadings are underlined in the Table. Median loading on target
factor for the 23 items not appearing in this table is 33. Scores for the negatively worded items
are reflected so that high scores are always interpreted as positive for the factor.
"Factor labels correspond with the following: 1 = General Situational Constraints;
2 = Supervisor Support; 3 = Training/Opportunity to Use Skills; 4 = Job/Task Importance; and
5 = Unit Cohesion/Peer Support.
124 OLSON AND BORMAN

2. Supervisor Support includes (a) receiving proper guidance from super-


visor, as required; and (b) getting recognition and support from supervisor.
3. Training/Opportunity to Use Skills includes: (a) receiving appropriate
training for present assignment, both in Army schools and on-the-job train-
ing; and (b) being assigned tasks relevant to own specialty.
4. Job/Task Importance includes: (a) performing a job important to the
unit and the Army and (b) having a job that requires significant skills and
abilities.
5. Unit Cohesion/Peer Support includes: (a) working in a cohesive unit
and (b) working with a unit wherein members cooperate and support each
Downloaded by [Princeton University] at 05:55 27 July 2013

other.

intraclass Results
In this research, intraclass correlations (ICCs) were computed to index the
degree of within-unit agreement in environment scores. The ICC(1)s obtained
for the 2,358 soldiers in the 258 company-sized units were: 0.07 (General Con-
straints), 0.09 (Supervisor Support), 0.10 (Training), 0.10 (Job/Task Impor-
tance), 0.03 (Cohesion/Peer Support), and 0.13 (Environment Composite) for
the five environmental factors and composite, respectively. These are generally
even lower than the median level of reliabilities (r = .12) found by Jones and
James (1979) in their review of within-unit agreement in organizational climate
perceptions. Accordingly, the individual perception model appears more
plausible here than the objective reality explanation in interpreting work en-
vironment scores, although certain problems with interpreting low intraclasses
are noted in the Discussion.

Relationships Bdween Work Environment Factors


and Performance
Table 6 presents relationships between the work environment constructs
and performance measures. It should be noted that the work environment-
performance correlations vary considerably in magnitude across different
aspects of the environment. Supervisor Support and Job/Task Importance
correlate near .20 with performance as measured by the Army-wide factors
and overall effectiveness, whereas the other three environmental factors cor-
relate lower with these rating constructs. Also, 'Ikaining/Opportunity to Use
Skills correlates .23 with task proficiency test scores. None of the other work
environment constructs correlates above .09 with either of the two objective
performance measures.
Environment-performance relationships also depend on the particular per-
formance measure employed. The Army-wide rating constructs and the overall
soldier effectiveness composite relate more substantially to the environrnen-
TABLE 6
Downloaded by [Princeton University] at 05:55 27 July 2013

Correlations Between Work Environment Constructs and Performance Measures

Work Environment Constructs


- -- - - -

Training/ Overall" Adjusted"


General Supervisor Opportunity Job/Task Unit Cohesion/ Environment Multiple
Performance Measures Constraints Support to Use Skills Importance Peer Support Composite Correlation

Army-wide ratings
Effort and leadership 05 17 07 20 09 18 22
Personal discipline 12 24 06 20 12 22 23
Military bearing 14 18 06 16 10 19 17
Overall soldier effectiveness 10 22 07 22 12 22 26
Job-specific ratings
Overall job performance 00 12 12 12 09 14 21
Objective performance measures
Job knowledge - 07 05 06 07 09 05 17
Task proficiency (work sample) -06 00 23 06 05 09 21

Note. Correlations are based on an N = 4951.


"The Overall Environment Composite is the unit-weighted sum of the five scale scores on the AWEQ. bThe five environmental constructs are in-
cluded in the multiple correlations computation, and Overall Environment Composite is not included; multiple R is adjusted using the Wherry correc-
tion. cCorrelations greater than .05 are significantly different from 0, p < .001.
126 OLSON AND BORMAN

tal variables than do the other three performance measures, according to the
multiple correlation results.

DISCUSSION

A taxonomy of the work environment for first-term soldiers was developed


during this research, using a variant of the critical incidents technique. Items
were written to tap each of 14 environmental variables, resulting in the AWEQ.
Subsequent factor analyses of the AWEQ yielded an interpretable five-factor
Downloaded by [Princeton University] at 05:55 27 July 2013

solution, presumably reflecting underlying work environment constructs.


Regarding relationships between work environment and performance, scores
on two of the five summary environmental factors, Supervisor Support and
Job/Task Importance, correlated about .20 with rated performance on the
Army-wide scales. This result extends previous work by identifying individual
elements of the work environment that are more highly related to performance
than the other elements. These findings suggest that the use of an interpretable
factor solution, as opposed to summed environmental composite scores
utilized in previous constraint research, may provide better conceptual and
operational descriptions of environment-performance relationships.
The magnitude of environment-performance relationships also depends
on the performance measures employed. The Army-wide rating d e s generally
had the highest relationships with environmental variables. These Army-wide
measures presumably tap criterion constructs beyond job proficiency and task
accomplishment to include aspects of organizational commitment, socializa-
tion, and citizenship (Borman et al., 1984). It could be that perceptions of
the environment are more closely tied to these aspects of effectiveness than
to task performance.
Considerable attention should be paid to the meaning of environment per-
formance relationships in interpreting results of our research. For example,
positive correlations between favorability of the environment and performance,
what we might term the traditional view of these relationships, show that the
work environment influences performance. However, the opposite causative
path could also explain such results. Better performers may get placed in more
favorable work environments because of their effective performance.
Either of these broad interpretations treats scores on work environment
factors as somehow substantive and real (the objective reality model), but
there is no denying that work environment questionnaire scores reflect percep-
tions of the environment, quite possibly with a substantial evaluative com-
ponent (the individual perception model).
The distinction is very similar to that made in the organizational climate
literature, in which some theorists and researchers interpret climate scores at
the organizational level, implying a substantive, "actual environment" inter-
pretation of the pooled individual perceptions of climate (e.g., Forehand &
Gilmer, 1964; Payne & Pugh, 1976). Many feel that climate is an intraindividual
phenomenon, with organization members individually reacting to organiza-
tional/situational stimuli and assigning psychological meaning to these fea-
tures of organizations. The relatively individualized perceptions of organiza-
tional features then become important for understanding the behavior of per-
sons in organizations (e.g., James, 1982; Jones & James, 1979; Schneider,
1975). For example, a possible interpretation of this study's results is that good
performers have more favorable views of their work environments, perhaps
as a result of their performance.
Regarding the objective reality versus individual perception models of en-
vironmental questionnaire response, our intraclass results support the latter
Downloaded by [Princeton University] at 05:55 27 July 2013

model. Relatively low within-unit agreement was evident for AWEQ scores.
It should be noted, however, that two additional issues are relevant for ad-
dressing the question of which model more accurately describes the meaning
of work environment questionnaire scores. First, identifying the proper level
of aggregation is a problem (e.g., James, 1982). What is a unit for purposes
of assessing level of homogeneity in work environments? Different sized units
may, in fact, be appropriate for this assessment, depending on the environmen-
tal factor being examined. For example, for Job/Task Importance, small
groups with very similar jobs may be the most reasonable level of aggrega-
tion; for Cohesion/Peer Support, a larger unit that works together consistently
might be appropriate for aggregation purposes.
A second point is that low agreement within the unit on perceptions of
certain work environment factors does not necessarily reflect a lack of ob-
jectivity or accuracy on the part of the respondents. Studies of supervisor-
subordinate dyads (e.g., Graen, 1976) suggest that supervisors often act
substantially differently toward different subordinates (eg., providing different
levels of personal and work-related support). Herein this means that low agree-
ment within a unit on a work environment factor like Supervisor Support
may reflect real differences in these levels of support, not simply mispercep-
tions on the part of some or all respondents.
Nonetheless, low within-group agreement generally suggests a view of work
environment questionnaire responses as perceptions and interpretations of this
environment rather than as reports of objectively verifiable phenomena. The
individual perception view is worthy of investigation in its own right. Specifi-
cally of interest is how these perceptions of different aspects of the work envi-
ronment are formed. How are such perceptions sustained or changed?
The perceptual view further suggests a search for person variable predic-
tors of environmental perceptions. A folk notion here might be that generally
agreeable, self-confident, and nonneurotic persons view their work environ-
ment or some aspects of it as relatively favorable. A study designed to hold
facets of the work environment as constant as possible, so that person variable
measures are predicting reasonably unconfounded differences in perceptions,
would be compelling for investigating this notion.
128 OLSON AND BORMAN

Interpretations of work environment scores, other than what we called the


traditional view, also have important implications for those who would con-
sider correcting or adjusting performance scores for differemes in the
favorability of work environments. For example, one could conceive of ad-
justing criterion scores in selection research for differences in the difficulty
levels of the work environment. Persons in relatively unfavorable environments
would have their performance scores adjusted upward, for e-pk This no-
tion seems fair and compelling, on the one hand, but its appropriateness
depends on our interpretations of work environment-performance relation-
ships, on the other hand.
If the traditional view of environment "causing" performance is correct,
Downloaded by [Princeton University] at 05:55 27 July 2013

an adjustment in performance scores seems warranted. However, if causa-


tion is in the other direction-the more effective performers are placed in
either more favorable work environments or tend to display more positive
perceptions of the work setting as a result of effective performance-then
such an adjustment is clearly inappropriate. Of course, the intraclass results
obtained in this remuch cast doubt on either of these interpretations. Likewise,
if the individual perception model provides an accurate interpretation of en-
vironmental scale scores, adjusting performance scores is inappropriate and
may actually lead to lower predictor-criterion correlations than when unad-
justed performance scores are employed. Consider, for example, the hypo-
thetical case we presented earlier. Agreeable, self-confident, and nonneurotic
persons may tend to see their environment as more favorable than their less
agreeable and less confident counterparts. To the extent that correlates of
work environment perceptions, such as temperament or personality factors,
are also valid predictors of performance, adjustment of performance scores
according to these perceptions of the environment will unwittindy reduce
predictor-criterion relationships. Considerably more work is needed to better
understand and properly interpret work environment questionnaire responses.
As conceptualized in this research, job performance is mdtidimensional
and a product of: (a) individual attributes, abilities, and skills which are
measurable at the time an individual first enters the organization: (b) en-
vironmental and organizational variables that impact on the person after job-
entry; and (c) the person's attitudes, perceptions, and motivation to perform.
Thus, in order to better understand predictor-criterion relationships, future
research should investigate all potential influences on performance, not ex-
clusively job-entry predictors.

ACKNOWLEDGMENTS

We thank John P. Campbell, Lawrence M. Hanser, Jane M. Arabian, and


Michael G. Rumsey for providing constructive cmments on earlier versions
of this article. We also thank Loriann Roberson, Sharon Rose, John Kamp,
Betty Smith, Helen Sperling, and Linda Schultz-Shiner for their help in
developing the AWEQ. A special thanks to Betty Shelly for her skillful typ-
ing of this manuscript.
Support for the research was provided by the U.S. Army Research Institute,
Contract No. MDA903-82-D-0531. This research project (Project A) is a long-
term, large-scale effort concerned with improving the selection and classifica-
tion of enlisted soldiers in the U.S. Army.
The opinions expressed in this article are ours and do not necessarily reflect
the view of the Army Research Institute or the Department of the Army.
Downloaded by [Princeton University] at 05:55 27 July 2013

REFERENCES

Borman, W. C., Motowidlo, S. J., Rose, S. R., & Hanser, L. M. (1984). Development of a model
of soldier effectiveness (Tech. Rep. No. 660). Alexandria, VA: U.S. Army Research Institute.
Campbell, J. P. (Ed.). (1988). Improving the selection, classification, and utilization of Army
enlistedpersonnel: Annual report. I986 fiscal year (ARI Tech. Rep. 813101). Alexandria, VA:
U.S. Army Research Institute.
Campbell, C. H., Campbell, R. C., Rumsey, M. G., &Edwards, D. C. (1986). Development and
field test of Project A task-basedMOS-specific criterion measures (Tech. Rep. No. 717). Alex-
andria, VA: U.S. Army Research Institute.
Dunnette, M. D. (1976). Aptitude, abilities, and skills. In M. D. Dunnette (Ed.), Handbook of
industrial and organizational psychology (pp. 473-520). Chicago, IL: Rand McNally.
Eulberg, J. R., O'Connor, E. J., Peters, L. H., & Watson, T. W. (1984). Situational constraints
upon performance: A selective review of relevant literatures (AFHRL Tech. Paper 83-48).
Brooks Air Force Base, TX: Air Force Human Resources Laboratory.
Flanagan, J. C. (1954). The critical incident technique. Psychological Bulletin, 51, 327-358.
Forehand, G., & Gilmer, B. V. (1%4). Environmental variation in studies of organizational behavior.
Psychological Bulletin, 62, 361-382.
Graen, G. (1976). Role-making process within complex organizations. In M. D. Dunnette (Ed.),
Handbook of industrial and organizational psychology (pp. 1201-1245). Chicago: Rand
McNally.
James, L. R. (1982). Aggregation bias in estimates of perceptual agreement. Journal of Applied
P S Y C ~ O ~67(2),
O ~ Y 219-229.
,
Jones, A. P., & James, L. R. (1979). Psychological climate: Dimensions and relationships of
individual and aggregated work environment perceptions. Organizational Behavior and Human
Performance, 23, 201-250.
Naylor, J. D., Pritchard, R. D., & Ilgen, D. R. (1980). A theory of behavior in organizations.
New York: Academic.
O'Connor, E. J., Peters, L., Pooyan, A., Weekley, J., Frank, B., & Erenkrantz, B. (1984). Situa-
tional constraint effects on performance, affective reactions and turnover: A field replication
and extension. Journal of Applied Psychology, 69(4), 663-672.
Olson, D. M., & Borman, W. C. (1986). Development and field tests of the Army Work En-
vironment Questionnaire (Tech. Rep. 737). Alexandria, VA: U.S. Army Research Institute.
Payne, R. L., & Pugh, D. S. (1976). Organizational structure and climate. In M. D. Dunnette
(Ed.), Handbook of industrial and organizationalpsychology (pp. 1125-1173). Chicago: Rand
McNally.
Peters, L. H., Chassie, M. B., Lindholm, H. R., O'Connor, E. J., & Kline, C. R. (1982). The
joint influence of situational constraints and goal setting on performance and affective out-
comes. Journal of Management, 8, 7-20.
130 OLSON AND BORMAN

Peters, L. H., & O'Connor, E. J. (1980). Situational and work outcomes: The influences of a
frequently overlooked construct. Academy of Management Review. 5 , 391-397.
Peters, L. H., O'Connor, E. J., Eulberg, J. R., &Watson, T. W. (1988). An examination of situa-
tional constraints in Air Force work settings. Human Performance, 1, 133-144.
Peters, L. H., (YConnor, E. J., & Rudolf, C. J. (1980). The behavioral and afFerting consequences
of performance-=levant situation variables. Organizational Behavior and Human Performance,
25, 79-95.
Pulakos, E. D. (1984). A comparison of two rater training programs: Error training versus ac-
curacy training. Journal of Applied Psychology, 69, 581-588.
Pulakos, E. D., & Borman, W. C. (Eds.).(1986). Development andfield test of Army-wide rating
scales and the rater orientation and training program. Alexandria, VA: U.S. Army Research
Institute.
Downloaded by [Princeton University] at 05:55 27 July 2013

Pulakos, E. D., & Borman, W. C. (1987). Developing the basic criterion scores for Army-wide
and MOS-specifc ratings. Alexandria, VA: U.S. Army Research Institute.
Schneider, B. (1975). Organizational climate: An essay. &rsonnel Psychologv, 28, 447-479.
Smith, P. C., & Kendall, L. M. (1%3). Retranslation of expectations: An approach to the con-
struction of unambiguous anchors for rating scales. Journal of Applied Psychology, 47,
149-155.
Steel, R. P., & Mento, A. J. (1986). Impact of situational constraints on subjective and objective
criteria of managerial job performance. OrganizationalBehavior and Human Decision Pro-
cesses, 3 7, 254-265.
Toquam, J. L., McHenry, J. J., Corpe, V. A., Rose, S. R., Larnmlein, S. E., Kemery, E., Bor-
man, W. C., Mendel, R., & Bosshardt, M. J. (1986). Devefopment andfield tmt of behaviorally
anchored mting scales for nine MOS (Tech. Rep. No. 776). !&%andria, VA: U.S Army &%%irch
Institute.
Watson, T. W., O'Connor, E. J., Eulberg, J. R., &Peters, L. H. (1983, October). Measurement
and assessment of situational constraints in Air Force work environments: A brief summary.
Proceedings of the 25th Annual Conferenceof the Mtlifary nsting Amociation. 25,662-627.

You might also like