Journal of Applied Psychology

197S, Vol. 60, No. 2, 159-170

Development of the Job Diagnostic Survey

J. Richard Hackman Greg R, Oldham
Yafe University Department of Business Administration,
University of Illinois at Urbana-Champaign
The properties and uses of the Job Diagnostic Survey (JDS) are described.
The JDS is intended (a) to diagnose existing jobs to determine if (and how)
they might be redesigned to improve employee motivation and productivity,
and (b) to evaluate the effects of job changes on employees. The instrument
is based on a specific theory of how job design affects work motivation, and
provides measures of (a) objective job dimensions, (b) individual psycho-
logical states resulting from these dimensions, (c) affective reactions of em-
ployees to the job and work setting, and (d) individual growth need strength
(interpreted as the readiness of individuals to respond to "enriched" jobs).
Reliability and validity data are summarized for 6S& employees on 62 different
jobs in 7 organizations who have responded to a revised version of the
instrument.

As both organizational productivity and There are a number of reasons for this un-
employee alienation from work become in- fortunate state of affairs. Some of them have
creasingly problematic in American society, to do with the adequacy of existing theories
more and more organizations are turning to about how jobs affect people; others derive
the redesign of work as a strategy for orga- from methodological difficulties in carrying
nizational change directed toward solving out job redesign experiments in on-going
these problems (cf. Davis & Taylor, 1972; organizations. Yet perhaps one of the most
Ford, 1969; Maher, 1971). Indeed, one par- compelling explanations for the paucity of
ticular application of work redesign, job en- knowledge about work redesign is also one
richment, seems about to become something of the most basic: namely, that our capability
of a fad among.managers and organizational to measure (and thereby understand) what
consultants. happens when jobs are changed has been very
As yet, however, a solid body of knowl- limited. '
edge about the consequences of job enrich- The present article reports the development
ment has not emerged from behavioral science of a measurement tool, the Job Diagnostic
research. Neither are there abundant data Survey (JDS), which may be helpful in fill-
available about the relative effectiveness of ing this void in research and action projects
various strategies for implementing work re- involving the redesign of work. Specifically,
design projects (Hulin & Blood, 1968; Porter, (a) the conceptual basis of the instrument is
Lawler, & Hackman, 197S, chap. 10). presented; (b) the instrument itself is briefly
described; (c) the empirical properties of the
This report was prepared in connection with re- instrument are presented and discussed; and
search supported by the Office of Naval Research
(Organizational Effectiveness Research Program, (d) the uses and limitations of the instrument
Contract N00014-67A-0097-0026, NR 170-744) and are explored.
by the U.S. Department of Labor (Manpower Ad- It should be kept in mind throughout that
ministration, Grant 21-09-74-14).
Tlie authors express great appreciation to mem-
the instrument is designed to be of use both
bers of the Roy W. Walters & Associates consulting in the diagnosis of jobs prior to their re-
firm for assistance in gaining access to the organiza- design, and in research and evaluation activi-
tions where this research was conducted and to ties aimed at assessing the effects of re-
Kenneth Brousseau, Daniel Feldman, and Linda designed jobs on the people who do them.
Frank for assistance in data collection and analysis.
Requests for reprints should be addressed to J. We believe that use of such an instrument to
Richard Hackman, S6 Hillhouse Avenue, Yale Uni- diagnose the motivational properties of jobs
versity, New Haven, Connecticut 06520. prior to redesign should aid change agents in
159
 J. RICHARD HACKMAN AND GREG R. OLDHAM

wisely planning and implementing work re- outcomes (high internal motivation, high
design projects. Moreover, the availability of work satisfaction, high quality performance,
a standardized instrument for use in evalu- and low absenteeism and turnover) are ob-
ating the effects of such projects should tained when three "critical psychological
facilitate efforts by behavioral scientists to states" are present for a given employee
understand how and why job enrichment (experienced meaningfulness of the work, ex-
works when it does work—and what has gone perienced responsibility for the outcomes of
wrong when it does not. the work, and knowledge of the results of the
work activities). All three of the psychologi-
CONCEPTUAL BASIS OF THE INSTRUMENT cal states must be present for the positive out-
Any measuring device is based on some comes to be realized.
underlying theory of "what's important" re- The theory proposes that these critical
garding the phenomena under consideration psychological states are created by the pres-
(even if such a theory is implicit), and this ence of five "core" job dimensions. Experi-
instrument is no exception. The theory which enced meaningfulness of the work is enhanced
gave rise to the present instrument is based primarily by three of the core dimensions:
on earlier work by Turner and Lawrence skill variety, task identity, and task signifi-
(1965) and by Hackman and Lawler (1971). cance. Experienced responsibility for work
It is sketched briefly below to provide a outcomes is increased when a job has high
context for understanding and interpreting autonomy. Knowledge of results is increased
the measures generated by the instrument. when a job is high on feedback. Following the
For a more detailed description and discus- theory diagrammed in Figure 1, it is possible
sion of the theory itself, see Hackman and to generate a summary score reflecting the
Oldham (Note 1)" overall "motivating potential" of a job in
The basic theory is presented in Figure 1. terms of the core job dimensions. The score
It proposes that positive personal and work is computed as follows:

Motivating Potential Score (MPS)

Skill Task Task 1
Variety Identity Significance
3" " X (Autonomy) X (Feedback).

As can be seen from the formula, an increase "stretched" by it. Therefore, individual
in any of the core dimensions will increase growth need strength is shown in Figure 1 as
the MPS; but (because of the multiplicative a moderator of the other theory-specified
relationship among its components) if any of relationships.
the three major components of MPS is low,
the resulting MPS also must be low. The DESCRIPTION OF THE JOB DIAGNOSTIC
rationale for the MPS measure is discussed in SURVEY
detail by Hackman and Oldham (Note 1). The Job Diagnostic Survey (JDS), taken
A job high in motivating potential will not by employees who work on any given job,
affect all individuals in the same way. In provides for that job measures of each of the
particular, people who strongly value and concepts specified in the theoretical frame-
desire personal feelings of accomplishment work. In addition, the instrument provides
and growth should respond very positively to several supplementary measures of respon-
a job which is high on the core dimensions; dents' reactions to their work.
individuals who do not value personal growth The JDS has its origins in previous meth-
and accomplishment may find such a job odologies developed by Turner and Lawrence
anxiety arousing and may be uncomfortably (1965) and by Hackman and Lawler (1971).
 DEVELOPMENT OP THE JOB DIAGNOSTIC SURVEY 161

CRITICAL
i CORE JOB P E R S O N A L AND
j DIMENSIONS W O R K OUTCOMES
j STATES

S k i l l Voriely ^1
1 Experienced ^ High Internal
Task I d e n t i f y S- > Mean ing fulness Work M o t i v a t i o n
[ of the Work
Task S i g n i f i c a n c e \
High Q u a l i t y
Work Performance
E x p e r i e need
Responsibility >
Aulonom-
u onom,
for Outcomes High S a t i s f a c t i o n
of the Work W i t h the Work

K n o w l e d g e of the

A1- the Work A c t i v i t i e s A and T u r n o v e r

EMPLOYEE GROWTH
NEED STRENGTH

FIGURE 1. A theoretical model relating the core job dimensions, the critical
psychological states, and on-the-job outcomes (as moderated by employee
growth need strength).

Many of the scales and items used by these and by items written in two different for-
researchers are retained, in revised form, in mats, thereby decreasing the degree to which
the JDS. The JDS itself has undergone three substantive content and measurement tech-
major revisions over a 2-year developmental nique are confounded within the instrument.
period, In its various forms, it has been taken Seven-point response scales are used through-
by over 1,500 individuals working on more out (1 = low, 7 = high). The full instrument
than 100 different jobs in about IS different —and a detailed discussion of item content
organizations. and format—are provided in a separate re-
Revisions of the instrument were based on port (Hackman & Oklham, Note 2).
both psychometric and substantive considera-
tions. On the one hand, items were added, Job Dimensions
deleted, and altered to maximize scale relia- The JDS provides measures of the five core
bilities and the empirical discrimination dimensions shown in Figure 1, which are de-
among scales. At the same time, however, fined as follows:
efforts were made to keep the content of the Skill variety. The degree to which a job
; items tapping a given construct as hetero- requires a variety of different activities in
geneous as possible, to maximize the sub- carrying out the work, which involve the use
stantive "richness" of the measures. Through- of a number of different skills and talents of
out the development of the JDS, analyses the employee.
were conducted to assess the validity of the Task identity. The degree to which the job
theory on which the instrument is based— requires completion of a "whole" and identifi-
.and the findings were used to revise and able piece of work—that is, doing a job from
refine the theory simultaneously with the im- beginning to end with a visible outcome.
provement of the instrument itself. Task significance. The degree to which the
The specific measures obtained from the job has a substantial impact on the lives or
JDS are described below. Each class of varia- work of other people—whether in the imme-
bles (except the "specific satisfactions") is diate organization or in the external environ-
measured in two different sections of the JDS ment.
162 J. RICHARD HACKMAN AND GREG R. OLDHAM

Autonomy, The degree to which the job continuous basis, how effectively he or she is
provides substantial freedom, independence, performing the job.
and discretion to the employee in scheduling Scores for the critical psychological states
the work and in determining the procedures are obtained from both self-descriptive and
to be used in carrying it out. projective type items. In the self-descriptive
Feedback from the job itself. The degree to section, respondents indicate their level of
which carrying out the work activities re- agreement with a number of statements about
quired by the job results in the employee their work experiences. In the projective sec-
obtaining direct and clear information about tion, respondents are asked to "think of other
the effectiveness of his or her performance. people in your organization who hold the
In addition, measures are obtained for two same job as you do" and report how accurate
supplementary dimensions which have been they believe a number of statements are in
found to be helpful in understanding jobs and describing the feelings of those people.
employee reactions to them. These are:
Feedback from agents. The degree to which Affective Reactions to the Job
the employee receives clear information about The JDS provides measures of a number
his or her performance from supervisors or of personal, affective reactions or feelings a
from co-workers. (This dimension is not, person obtains from performing the job. These
strictly speaking, a characteristic of the job are viewed, in the context of the theory in
itself. It is included to provide information Figure 1, as the "personal outcomes" ob-
to supplement that provided by the "feed- tained from doing the work. (The instrument
back from the job itself" dimension.) does not measure actual work outcomes:
Dealing with others. The degree to which productivity, employee perceptions of their
the job requires the employee to work closely productivity, turnover, or absenteeism.) The
with other people in carrying out the work personal outcomes are:
activities (including dealings with other General satisfaction. An overall measure of
organization members and with external or- the degree to which the employee is satisfied
ganizational "clients.") and happy with the job.
As noted earlier, scores on the job dimen- Internal work motivation. The degree to
sions are obtained from items in two sections which the employee is self-motivated to per-
of the instrument. In the first section, re- form effectively on the job—that is, the em-
spondents indicate directly on the seven-point ployee experiences positive internal feelings
response scale the amount of each job char- when working effectively on the job, and neg-
acteristic they perceive to be present in their ative internal feelings when doing poorly.
job; in the second section, respondents indi- Specific satisfactions. A number of short
cate the accuracy of a number of statements scales provide separate measures of satisfac-
about the characteristics of their job. tion with: (a) job security, (b) pay and other
compensation, (c) peers and co-workers ("so-
Critical Psychological States cial" satisfaction), (d) supervision, and (e)
The JDS provides measures of each of the opportunity for personal growth and develop-
three psychological states shown in Figure 1 ment on the job ("growth" satisfaction).
as mediating between the core job dimensions Items measuring general satisfaction and
and the outcomes of the work. These are: internal work motivation are intermixed with
Experienced meaning fulness of the work. items tapping the three critical psychological
The degree to which the employee experiences states, in both the self-descriptive and projec-
the job as one which is generally meaningful, tive sections of the instrument. For the five
valuable, and worthwhile. specific satisfactions, respondents report di-
Experienced responsibility for work out- rectly how satisfied (or dissatisfied) they are
comes. The degree to which the employee with various aspects of their jobs.
feels personally accountable and responsible
for the results of the work he or she does. Individual Growth Need Strength
Knowledge of results. The degree to which Finally, the JDS taps the strength of the
the employee knows and understands, on a respondent's desire to obtain "growth" satis-
 DEVELOPMENT OF THE JOB DIAGNOSTIC SURVEY 163

factions from his or her work. This measure zation. The JDS was administered to em-
is viewed as a malleable individual difference ployees in groups ranging in size from 3 to
characteristic which (as shown in Figure 1) 25. Participation was optional for all re-
is predicted to influence how positively an spondents. Assurances of confidentiality were
employee will respond to a job with objec- provided, and few employees declined to par-
tively high motivating potential. ticipate or to provide their names on the
Two separate measures of growth need instrument.
strength are obtained, one from items in a Additional assessments of the character-
"would like" format, and one from items in a istics of each job were obtained from super-
"job choice" format. In the former section of visors of the focal job and from the
the instrument, respondents are asked to indi- researchers—providing three independent
cate directly how much they would like to sources of data about each job. The data
have a number of specified conditions present from supervisors and researchers were ob-
in their jobs, some of which (e.g., "Chances tained using the Job Rating Form. This in-
to exercise independent thought and action strument, designed for use by people who are
in my job") focus on growth-relevant aspects not job incumbents, consists of job descrip-
of the work. In the "job choice" section of tive items nearly identical in form and con-
the instrument, respondents indicate their tent to those on the JDS itself. 2
relative preference for pairs of hypothetical Absence data were obtained from company
jobs (e.g., "A job where you are often re- records. Ratings of the work effectiveness of
quired to make important decisions" vs. "A each employee were obtained from super-
job with many pleasant people to work visors on forms provided by the researchers,
with"). In each item a job with character- using 7-point rating scales for effort, work
istics relevant to growth need satisfaction is quantity, and work quality. Members of the
paired with a job which has the potential for research team spent 1-4 days at each organi-
satisfying one of a variety of other needs. zation collecting data. Since it was not possi-
ble to obtain complete data for all jobs in the
EMPIRICAL PROPERTIES OP THE JOB sample, some of the results reported below
DIAGNOSTIC SURVEY 1 are based on a subset of the total sample.
Methodology Full details of the data collection method-
Results reported here are based on data ology, including a summary of the demo-
obtained from 658 employees working on 62 graphic characteristics of the respondents, are
different jobs in 7 organizations. The jobs provided by Hackman and Oldham (Note
were highly heterogeneous, including blue- 2).
collar, white-collar, and professional work. JDS Scale Reliabilities
Both industrial and service organizations
Table 1 presents the internal consistency
were included in the sample, but all were
reliabilities of each of the scales measured by
businesses. The organizations were located in
the Job Diagnostic Survey. 3 ' 4 Also included in
the east, southeast, and midwest, in both
the table for each scale is the median of the
urban and rural settings. Fifty-nine percent
of the respondents were male; their median - The properties of the Job Rating Form are
age was 29, and their education ranged from described and discussed in a separate report
grade school only to a graduate degree. (Hackman & Oldham, Note 2 ) .
3
Data were collected on site at each organi- The term "scale" is used loosely throughout the
remainder of this report to refer to the summary
score obtained for each variable measured by the
1
A final, "fine-tuning" revision of the JDS was JDS. These scores were obtained by averaging the
made after the data reported here were collected. items written to measure each of the JDS variables,
Therefore, some of the results reported may be and they are not formal "scales" in the technical
slightly discrepant from those which would be ob- sense of the term.
4
tained using the instrument in its final form. When Reliabilities were computed by obtaining the
there is any reason to believe that empirical results median inter-item correlation for all items which are
might be substantially affected by a change which scored on each scale, and then adjusting the median
has been made, notation of that possibility is made by Spearman-Brown procedures to obtain an esti-
in the data table. mate of the reliability of the summary scale score.
164 J. RICHAED HACK.MAN AND GREG R. OLDHAM

TABLE 1 the other job dimension scales). These me-

RELIABILITIES OF THE JOE DIAGNOSTIC SURVEY dian correlations (referred to in the table as
(JDS) SCALES "off-diagonal" correlations) provide one indi-
cation of the discriminant validity of the
Median items.
Internal of/-
JDS scale «" consist- diagonal Internal consistency reliabilities range from*
ency corre- a high of .88 (growth need strength, in
reliability lation1' the "would like" format) to a low of .56
(social satisfaction). The median off-diagonal
Job dimensions correlations range from .12 (task identity) to
.28 (growth satisfaction). In general, the
Skill variety 3 .71 .19 results suggest that both the internal con-
Task identity 3 .59 .12
Task significance 3 .66 .14 sistency reliability of the scales and the dis-
Autonomy 3 .66 .19 criminant validity of the items are satisfac-^
Feedback from the job tory.
itself 3 .71 .19
Feedback from agents 3 .78 .15 Objectivity of the Job Dimensions
Dealing with others 3 .59 .15
As indicated earlier, assessments of the
Psychological states focal jobs on the job dimensions were made
not only by employees who worked on those
Experienced meaningful- jobs, but by supervisors and by the research-
ness of the work 4 .74 .26 ers as well. This was done to provide an in-
Experienced responsibility direct test of the "objectivity" of employee
for the work 6 .72 .23
Knowledge of results 4 .76 .17
ratings of the characteristics of their own
jobs.
Affective responses to the job The relationships among the judgments
made by employees, supervisors, and observ-
General satisfaction 5 .76 .25
Internal work motivation 6 .76 .25
Specific satisfactions TABLE 2
Job security" 2 RELATIONSHIPS AMONG EMPLOYEES', SUPERVISORS',
Pay" 2 AND OBSERVERS' JOB RATINGS
Social .56 .23
Supervisory 3 .79 .25 Correlations between ratings
Growth .84 .28
Job dimension Emploj'ees Employees Supervisors
Growth need strength
and and and
supervisors observers observers
"Would like" format' 1 6 .88
Job choice format' 1 12 .71 — Skill variety .64 .66 .89
a
Task identity .31 .32 .44
b
Number of items composing each scale. Task
The median off-diagonal correlation is the median correla-
lion of the items scored on a given scale with all of the items significance .48 .65 -.14
scored on different scales of the same type of variable. Thus, Autonomy .58 .76 .72
the median off-diagonal correlation for skill variety (.19) is the
median correlation of all items measuring skill variety with all Feedback from
thec items measuring the other six job dimensions. the job itself .33 .58 .47
These scales were added to the JDS after the present data
were
<l
collected, and no reliability data arc yet available. Feedback from
Off-diagonal correlations are not reported for these two agents .07 -.13 .14
scales, since all items were designed to tap tlie same construct.
The scale scores obtained using the "would like" format corre- Dealing with
late .50 with the scale scores obtained using the job choice others .55 .61 .37
format.
Motivating po-
tential score .56 .70 .71
correlations between (a) the items composing Mdn .51 .63 .46
a given scale (e.g., skill variety) and (b) all
of the other items which are scored on dif- Xote. Oata are included only for those jobs for which more
than one set of supervisory ratings were available. Ar ranged
ferent scales of the same general type (i.e., from 12 to 21 jobs.
 DEVELOPMENT OF THE JOB DIAGNOSTIC SURVEY 165

ers are shown in Table 2. The ratings of each some job dimensions (e.g., feedback from
group (i.e., employees, supervisors; observers) agents) for. which the correlations between
were averaged for each job, and then corre- two of the 1 groups are quite low. Moreover,
lations were computed using jobs as observa- the general level of the correlations is lower
tions. The median of the correlations between than those reported for similar job dimen-
employees and supervisors is. .51; between sions by Hackman and Lawler (1971).
employees and observers is .63; and between
supervisors and observers is .46. Means and Variances oj the JDS Scales
Although in general the ratings of the three Means and standard deviations of the JDS
groups converge moderately well, there are scale scores across all 658 respondents are

TABLE 3
MEANS, VARIANCES, AND ANALYSIS op VARIANCE TOR JOB DIAGNOSTIC SURVEY (JDS) SCORES

Analysis of variance across jobs

Total sample (N = 658)
JDS scales X SD Variance Variance
X within between F(49, 563)«
(n = 62) jobs jobs

Job dimensions

Skill variety 4.49 1.67 4.47 1.54 17.70 11.49*

Task identity 4.87 1.43 4.87 1.71 5.90 3.45*
Task significance 5.49 1.29 5.54 1.55 3.22 2.08*
Autonomy 4.80 1.43 4.75 1.5S 7.92 5.11*
Feedback from the job itself 4.98 1.41 4,96 1.76 4.4:! 2.51*
Feedback from agents 3.98 1.65 3.87 2.28 6.82 2.99*
Dealing with others 5.29 1.34 5.27 1.35 6.70 4.96*
Motivating potential score (MPS) 128.31 72.73 120,68 4,112. 19,959. 4.85*

Psychological states

Experienced meaningfulness of the work 5.12 1.10 5.06 1.05 3.19 3.04*
Experienced responsibility for the work 5.48 0.91 5.44 0.70 2.37 2.24*
Knowledge of results S.18 1.09 5.19 1.06 2.57 2.42*

Affective responses to the job

General satisfaction 4.62 1.18 4.57 1.13 4.19 3.71*

Internal work motivation 5.39 0.96 5.34 0.82 2.19 2.67*
Specific satisfactions
Job security8 — — — — — —
Pay" — — — — — —
Social 5.42 0.92 5.42 0.77 1.72 2.23*
Supervisory 5.28 1.27 5.32 1.42 3.81 2.68*
Growth 4.82 1.32 4.77 1.48 4.64 3.14*

Growth need strength

i
"Would like" format 5.62 1.28 5,51 1.30 5.11 3.93*
Job choice format' 1 — — — — — —
ft
b
These scales were added to the JDS after the present data were collected, and normative data arc not yet available.
The response scale for the job choice format was revised from 7 to 5 points after these data were collected. Preliminary indica-
tions
c
are that the mean of the five-point scale will be close to the midpoint (.3,0).
The analysis of variance was conducted on 50 jobs which had five or more respondents.
* P < .01,
166 J. RICHARD HACKMAN AND GREG R. OLDHAM

presented in Table 3. The table also shows scores for the EEOC sample are higher than
the mean JDS scores across the 62 jobs in the the means for the sample of jobs in business
sample (i.e., the scores of respondents who organizations shown in Table 3.
worked on each job were averaged, and the
mean of these averages was computed across Relationships Among the JDS Scales
the 62 jobs for each scale). The scale means Correlations among the JDS scales are
obtained across all respondents are very simi- presented in Table 4. The job dimensions*^
lar to those obtained when averages were themselves are moderately positively inter- ft
computed across jobs, suggesting that the correlated, as has been found previously /
different numbers of respondents who held (Hackman & Lawler, 1971). This is to be*
the various jobs did not substantially affect expected, if it is assumed that "good" jobs
the mean scale scores. often are good in a number of ways, and
Also reported in Table 3 are the results of "bad" jobs often are generally bad. There is
one-way analyses of variance which were no a priori reason to expect that the job
computed for each scale across SO jobs which dimensions would or should be completely
had five or more respondents. As expected, independent, and the moderate level of inter-
between-job differences were statistically sig- correlation among them does not detract from
nificant for all of the JDS scale scores. The their usefulness as separate job dimensions —
data in the table show that the JDS scales so long as the fact of their nonindependence
vary considerably both in the amount of is recognized and accounted for in interpret-
between-job variance present and in the ing the scores of jobs on a given job dimen-
amount of variance present among respon- sion.
\
dents within jobs. The F ratios can be taken As expected, the job dimensions are posi-N
as rough indicators of the sensitivity of the lively related to measures of work satisfac-l
scales to between-job differences (at least for tion and motivation, and are generally inde-J
the set of jobs in the present sample). It pendent of the two measures of growth need
should be kept in mind, however, that the strength. The measures of the critical psycho-\
within-job variance (the denominator of the logical states are strongly related to those J
F ratio) is multiply determined — and in core job dimensions predicted by the theory I
part determined by real differences in actual to affect them. They also are not substan-/
jobs within organizational job categories. tially related to the need strength measures/
That is, some (unknown) amount of the The correlations in Table 4 were computed
within-job variance must be attributed to across all 658 respondents; in addition, inter-
scale unreliability and to individual differences correlations were computed across the 62 jobs
among respondents. At the same time, some (using the average of respondent scores for
(also unknown) amount of the same variance each job as observations). These correla-
is explained by the fact that jobs often are tions (which are presented by Hackman &
individually designed to take account of par- Oldham, Note 2) show a pattern quite simi-
ticular characteristics of the people who do lar to that obtained in the across-respondent
them, or because of the need for certain analysis reported in Table 4. The level of
specialized activities to be performed by some interrelationship among the scales, however,
people within a given job category. There- is substantially higher in the across-job analy-
fore, the ratio of the between- to the within- sis, which may be attributed at least in part
job variance should be interpreted with cau- to the fact that group averages are certain to
tion. be more reliable than the scores of individual
Means for a subset of the JDS scales for respondents.
an entirely different sample of respondents
are provided by VanMaanen and Katz (Note Substantive Validity of the JDS
3). Scores are presented for a group of over The substantive validity of the instrument
3,000 public employees, broken into eight is addressed in detail in a separate report
Equal Employment Opportunity Commission (Hackman & Oldham, Note 1). In general,
(EEOC) job categories. In general, the mean that report shows that the variables measured
 TABLE 4
INTERCORRELATIONS AMONG JOB DIAGNOSTIC SURVEY (JDS) SCALE SCORES o
H

JDS scale 1 2 3 4 5 6 7 8 9 10 11 12 "13 14 15 16 17 18 I

§
£
1. Skill variety B
2. Task identity .16 — 2
rt
3. Task significance .21 .20 —
4. Autonomy .51 .38 .22 — §
5. Feedback from the job itself .32 .26 .26 .34 —
6. Feedback from agents .25 .16 .22 .23 .37 —
7. Dealing with others .46 .02 .24 .29 .24 .26 —
8. Motivating potential score (MPS) .62 .51 .41 .80 .72 .36 .34 —
9. Experienced meaningfulness of the work .51 .26 .43 .46 .41 .31 .33 .57 —
10. Experienced responsibility for the work .40 .34 .34 .41 .37 .23 .24 .53 .64 —
11. Knowledge of results .12 .21 .21 .26 .54 .39 .06 .43 .33 .32 — o
12. General satisfaction 42 22 74 43 37 33 24 49 66 48 34 —
42 o
13. Internal work motivation 77 32 33 36 25 30 46 63 66 25 51 — VI
rt
14. Social satisfaction .31 .17 24 .38 27 31 .36 .40 .41 .38 .32 .40 4.0 — M
O
15 Supervisory satisfaction 15 16 16 32 31 41 13 35 39 32 37 46 31 37 —
16. Growth satisfaction .52 .31 .33 .58 .44 .39 .28 .63 .68 .54 .36 .67 .56 .52 .47 — in
d
17. Growth need strength (would like format) 27 .08 03 10 11 13 .16 .19 10 .21 .07 .04 .19 .08 .07 02 — ta
18 Growth need strength (job choice format) 31 06 — 01 19 13 15 20 '5 15 21 05 13 17 10 10 08 50 —

Note. N = 658. Correlations >.10 are significant at the .01 level (two-tailed).
 168 J. RICHARD HACKMAN AND GHKG R. OLDHAM

by the JDS relate to one another (and to fied relationships among JDS scales (and
external criterion variables) generally as pre- between these scales and behaviorally based
dicted by the theory on which the instrument dependent measures) are in the predicted di-
is based. In particular, the job dimensions rection.
(and the motivating potential score) relate
positively and often substantially to: Diagnostic Use of the JDS
1. The other variables measured by the One of the major intended uses of the Job
JDS which are predicted to be affected by Diagnostic Survey is in diagnosing existing
the job characteristics, including the three jobs as an input to planned job redesign. The
critical psychological states, general satisfac- instrument provides data on the following
tion, growth satisfaction, and internal work issues, each of which is likely to be relevant
motivation (cf. Table 4). to the diagnosis of a job prior to change —
2. Behavioral measures of absenteeism and as well as to evaluative assessment of the
supervisory ratings of work performance effects of job redesign after the change has
effectiveness. (For example, the motivating been implemented:
potential score has a median correlation of 1. The overall level of motivation and sat-
— .25 with absenteeism, and of .24 with a isfaction of employees on -the focal job. The
summary measure of performance effective- internal work motivation scale and the sev-
ness. Both relationships are statistically reli- eral measures of job satisfaction provide
able at p < .OS.) indication of whether or not observed organi-
I
ln addition, and also as predicted by the zational or behavioral "problems" are in fact
theory, the relationships between the job di- rooted in the relationships of employees to
mensions and the dependent measures are their work. In addition, examination of the
stronger for individuals high in growth need level of satisfaction with aspects of organiza-
strength than they are for individuals who do tional life not directly related to the work
not strongly desire growth satisfactions. All itself can signal special opportunities (or po-
of these relationships are explored in more tential difficulties) in the process of imple-
detail in Hackman and Oldham (Note f). menting work redesign. (If, for example, pay
and supervisory satisfaction are very low,
DISCUSSION difficulties in initiating and carrying out a
Empirical Characteristics oj the JDS successful job redesign project might be
Data discussed in the previous section show very significant.)
that the Job Diagnostic Survey has satisfac- 2. The overall motivating potential of ex-
tory psychometric characteristics, and that isting jobs, and how specific aspects of the
the variables it taps relate generally as pre- job contribute to the obtained motivating
dicted to appropriate external criteria. In- potential score. Comparison of the MPS of a
ternal consistency reliabilities are generally focal job with that obtained for other jobs
satisfactory, and the items which compose the (and with normative data such as that pre-
scales show adequate discriminant validity. sented in Table 3) can indicate the degree to
Ratings of job characteristics by employees, which the job is realistically open to improve-
ment through work redesign. Examination of
supervisors, and outside observers show a
moderate level of convergence for most of the each of the core dimensions shows which
job dimensions. Variances of the scales are specific aspects of the job are most in need
generally satisfactory, although some JDS of improvement. These data can guide those
who are planning the job redesign toward
scales show greater sensitivity to between-
those aspects of the work that most effectively
job differences than do others. Relationships can be changed and improved (cf. Hackman,
among the JDS scales are generally positive, Oldham, Janson, & Purely, in press).
indicating that either the concepts tapped by Employees, supervisors, and outside ob-
the instrument or the methodologies used to servers who participated in the present re-
gauge these concepts (or both) are not com- search showed only moderate agreement
pletely independent. In general, theory-speci- about the characteristics of the objective jobs
 DEVELOPMENT OF THE JOB DIAGNOSTIC SURVEY 169

under study (see Table 2 ) , For this reason, grade education or less, or for individuals
it is recommended that job descriptions ob- who do not read English well.
tained using the JDS be supplemented by 2. The instrument is readily fakable, and
independent assessments made by individuals probably should not be used for selection or
who are not incumbants of the focal job. (As placement purposes unless an extraordinarily
mentioned earlier, .the Job Rating Form, high level of trust exists between the re-
Hackman & Olclham, Note 2, is designed spondents and the individuals who will be
explicitly for this purpose.) using the results. Indeed, even when the JDS
It can be argued, of course, that when the is used to diagnose a work system prior to
intent is to predict or understand employee change (or to assess the effects of changes
attitudes or behavior at work, employee rat- which have been made) care should be taken
ings of the job dimensions should be used, to ensure that employees believe that their
since it is the employee's own perception of own interests will be best served if the data
the objective job that is causal of his reac- they provide accurately reflect the objective
tions to it. Yet if work redesign activities are characteristics of the jobs and their personal
to be planned on the basis of the job dimen- reactions to them.
sion scores obtained, it is important to know 3. Related to the above, it probably is pref-
that those scores are reasonably congruent rable for employees to take the JDS under
with objective reality. It might be, for exam- conditions of anonymity. While the research
ple, that for any of a number of reasons em- reported in this paper required the listing of
ployees are systematically misunderstanding names (and names were voluntarily supplied
or distorting the actual characteristics of by nearly all of the respondents), the instru-
their jobs. In such cases, an educational pro- ment was administered by a university-affili-
gram might be much more appropriate as an ated person and it was explicitly explained to
intervention than redesign of the work. By the respondents that the primary use of their
using job descriptions from more than one answers was for research purposes. When the
source and "triangulating" on the objective instrument is administered by members of
characteristics of the job, the diagnostician organizational management for use by man-
can gain reasonable assurance that problem agement, anonymity surely will be important
areas identified in the job actually are rooted for at least some of the respondents.
in the job itself — rather than in faulty per- 4. The instrument is not recommended for
ceptions of the job. use in diagnosing the jobs of single individu-
3. The "readiness" of employees for als. Anonymity, of course, is impossible if the
change. The measures of individual growth individual knows that it is his or her own
need strength provided by the JDS can be individual job that is being diagnosed. But
an important factor in planning job changes, the issue extends beyond that. In developing
since people high in growth needs tend to the JDS, the intent was to develop scales
respond more readily to "enriched" jobs than composed of items with rather heterogeneous
do people with little need for growth. The content — to maximize the substantive "rich-
process by which job changes are introduced ness" of each measure. This was accomplished
and implemented should probably be a more at some cost to internal consistency reliabil-
careful and deliberate one when employees ity. The reliabilities are more than satisfac-
are relatively low in growth need strength. tory when the instrument is used to obtain
average scores of a group of five or more
Cautions in the Use of the JDS individuals who work on a given job. In such
Listed below are a number of issues which, circumstances, the estimated internal con-
if not recognized, could impair the validity sistency of each JDS scale would exceed .85
and the usefulness of die Job Diagnostic Sur- for the average of the group of individuals
vey in some applications. who hold the job. For data collected from a
1. Respondents to the JDS must be mod- single individual, the reliabilities would be as
erately literate. Use of the JDS is not rec- shown in Table 1, which may not be high
ommended for individuals with an eighth enough to warrant job changes (or other ac-
170 J, RICHARD HACKMAN AND GREG R. OLDHAM

tion steps) on the basis of individual scale University, Department of Administrative Sci-
scores. (An exception to this state of affairs ences, 1974.
is the measure of individual growth need 2. Hackman, J. R., & Oldham, G. R. The job diag-
nostic survey: An instrument for the diagnosis
strength. This scale is designed to be a mea- of jobs and the evaluation of job redesign projects
sure of an individual characteristic, and was (Tech. Rep. No. 4). New Haven, Conn.: Yale
constructed so as to be a highly reliable in- University, Department of Administrative Sci-
dicator of individual needs.) ences, 1974.
5. Normative data are still being accumu- 3. VanMaanen, J., & Katz, R. Work satisfaction in
lated on the JDS scales. At this writing, sev- the public sector (Tech. Rep.). Washington,
D.C.: National Training and Development Ser-
eral thousand respondents have taken one or vice, 1974.
another of the preliminary versions of the
JDS. Yet because the instrument itself has REFERENCES
been modified on the basis of those responses, Davis, L. E., & Taj'lor, J. C. Design of jobs. Mid-
a stable normative base has not yet been dlesex, England: Penguin, 1972.
Ford, R. N. Motivation through the work itself.
established. The mean scale scores reported New York: American Management Association,
in Table 3 can be used legitimately to make 1969.
comparisons with scores obtained in other Hackman, J. R., & Lawler, E. E., III. Employee
reactions to job characteristics. Journal of Applied
uses of the instrument. But the populations Psychology Monograph, 1971, 55, 259-286.
from which these data were obtained were Hackman, J. R,, Oldham, G. R., Janson, R., &
not selected systematically enough for the Purdy, K. A new strategy for job enrichment.
California Management Review, in press.
data to be used to generate formal norms Hulin, C. L., & Blood, M. R. Job enlargement, indi-
(i.e., in computing standard scores and a scale vidual differences, and worker responses. Psycho-
of percentil.es for the JDS measures). As ad- logical Bulletin, 1968, 69, 41-55.
Maher, J. R. New perspectives in job enrichment.
ditional data are accumulated from uses of New York: Van Nostrand, 1971.
the final version of the JDS, more complete Porter, L. W., Lawler, E. E., Ill, & Hackman, J.
normative information will be made available. R. Behavior in organizations. New York: Mc-
Graw-Hill, 1975.
REFERENCE NOTES Turner, A. N., & Lawrence, P. R. Industrial jobs
and the worker. Boston: Harvard University
1. Hackman, J. R., & Oldham, G. R. Motivation Graduate School of Business Administration, 1965,
through the design of work: Test of a theory
(Tech. Rep. No. 6). New Haven, Conn.: Yale (Received June 6, 1974)