JOURNAL OF APPLIED PSYCHOLOGY

MONOGRAPH

Vol. 55, No. 3 June 1971

EMPLOYEE REACTIONS TO JOB CHARACTERISTICS

J. RICHARD HACKMAN2 AND EDWARD E. LAWLER III
Yale University

A conceptual framework specifying the conditions under which jobs will facilitate
the development of internal motivation for effective performance was developed and
tested. The 5s were 208 employees of a telephone company who worked on 13
different jobs. Primary independent variables were: (a) a measure of strength of
desire for the satisfaction of "higher order" needs (e.g., obtaining feelings of accom-
plishment, personal growth); and (6) descriptions of jobs on four core dimensions
(variety, autonomy, task identity, feedback). It was predicted and found that when
jobs are high on the four core dimensions, employees who are desirous of higher
order need satisfaction tend to have high motivation, have high job satisfaction, be
absent from work infrequently, and be rated by supervisors as doing high quality
work. A number of supplementary analyses were reported, and the implications of
the results for future research on job effects and for the design of jobs were discussed.

Researchers and managers alike are increas- establish fair rates of pay, for training pur-
ingly attending to the way jobs are designed poses, and in personnel selection (see, e.g.,
as an important factor in determining the Ghiselli & Brown, 1955; Lytle, 1946; Stigers
motivation, satisfaction, and performance of & Reed, 1944). The general expectation of the
employees at work. This is not to say that scientific management approach was that by
jobs previously have been seen as irrelevant simplifying jobs, work could be carried out
to organizational administration. On the con- more efficiently; less-skilled employees would
trary, earlier in this century when scientific be required; the control of management over
management was in its prime, considerable production would be increased; and,ultimately,
research effort was expended to find ways that organizational profits would be enhanced.
jobs could be simplified, specialized, standard- In recent years, numerous scholars have
ized, and routinized. At the same time, in- documented a number of unintended and un-
dustrial psychologists were developing rather fortunate consequences of the trend toward
complex and sophisticated procedures for de- work simplification (e.g., Argyris, 1964; Blau-
scribing and analyzing jobs in terms of their ner, 1964; Davis, 1957; Friedmann, 1961;
simplest components, as a means of evaluating Guest, 1955; Herzberg, Mausner, & Snyder-
the skill levels required for different jobs. The man, 1959; Walker, 1950; Walker & Guest,
results of job analyses have been used to 1952). In brief, it has been shown that simple,
routine nonchallenging jobs often lead to high
iThis research was supported in part by United
States Air Force Office of Scientific Research Grant employee dissatisfaction, to increased absentee-
AF-AFOSR-68-1600, Effects of Task Characteristics ism and turnover, and to substantial difficulties
on Performance (J. Richard Hackman, principal in- in effectively managing employees who work on
vestigator). The cooperation and assistance of em-
ployees and managers in the company where the re- simplified jobs.3 The expected increases in
search was done is gratefully acknowledged, as is the profitability from work simplification have not
assistance of Ann Garvin and Lloyd Suttle in collecting materialized as had been hoped, and the
and analyzing the data. Clayton Alderfer and Chris
Argyris provided helpful comments on an earlier draft reasons apparently have very much to do
of 2this article. 3
Requests for reprints should be sent to J. Richard These observations have not, however, gone un-
Hackman, 2 Hillhouse Ave., Yale University, New challenged. See, for example, Kilbridge (1961) and
Haven, Connecticut 06520. MacKinney, Wernimont, and Galitz (1962).
259
260 J. RICHARD HACKMAN AND EDWARD E. LAWLER III

with the human problems encountered when stances, and empirical research which is de-
jobs are standardized and simplified. signed explicitly to test these propositions.
Partially in response to the above findings, This article proposes one way of conceptualiz-
a number of researchers began experimentally ing the impact of job characteristics on indi-
enlarging various jobs to determine whether vidual work behavior and attitudes. It then
or not worker productivity and satisfaction reports data which were collected to provide a
would increase if jobs were designed so as to preliminary test of that conceptualization.
be more generally meaningful and challenging
to employees. By and large, those job enlarge- Previous Theory and Research
ment experiments which have been reported Some progress toward the development of
in the literature have been considered success- theory relevant to job design has been made
ful (see, e.g., Biganne & Stewart, 1963; Conant in recent years. The well-known two-factor
& Kilbridge, 1965; Davis & Valfer, 1965; theory of Herzberg (Herzberg, Mausner, &
Ford, 1969; Kilbridge, 1960; Pelissier, 1965). Snyderman, 1959; Herzberg, 1966), for ex-
With few exceptions, however, job enlargement ample, can be used to derive general proposi-
experiments have been case studies and often tions regarding conditions on the job which
have lacked appropriate experimental con- will be motivating and satisfying to employees.
trols. Hulin and Blood (1968) review the re- In particular, the theory suggests that a job
search literature on job enlargement in some should enhance employee motivation to the
detail and are especially attentive to possible extent that it provides opportunities for (a)
difficulties in procedure and methodology which achievement, (b) recognition, (c) responsibil-
may cast doubt on the generality or the va- ity, (d) advancement, and (e) growth in com-
lidity of the findings reported. petence. These principles have given rise to a
Perhaps equally as disturbing as the uneven series of generally successful job enlargement
level of methodological rigor which has char- experiments in the American Telephone and
acterized job enlargement studies is the almost Telegraph Company (summarized by Ford,
total absence of any systematic conceptual or 1969).
theoretical basis for the studies which have Unfortunately, a number of researchers have
been done. As a result, after dozens of experi- been unable to provide empirical support for
ments, little cumulative knowledge has been some of the major tenets of the theory from
gained regarding the effects and effectiveness which the principles used in the AT&T studies
of job redesign. Job enlargement experiments, were derived (e.g., Dunnette, Campbell, &
for example, typically have involved a number Hakel, 1967; Hinton, 1968; King, 1970), and
of simultaneous changes—such as in the the general conceptual status of the theory
amount of variety in the work, the amount must presently be considered uncertain. Fur-
of responsibility required, the degree to which ther, the theory has not yet been elaborated
working with others is an important part of to specify how characteristics of workers
the enlarged job, etc. Very little is known interact with the presence or absence of the
about which of these (or of other) aspects of five motivating conditions in determining
the redesigned job are in fact responsible for worker performance and satisfaction. Finally,
observed behavioral and attitudinal changes. the theory in its present form does not specify
Further, the generality of job enlargement how the presence or absence of the motivating
effects is largely unknown (e.g., whether they conditions can be measured for existing jobs.
are effective only for certain types of workers This makes it very difficult to test the theory
or whether they are relevant only to certain and to generate unambiguous predictions from
kinds of jobs). More case studies are not it about the effects of specific changes which
likely to contribute very much to the develop- may be contemplated for existing jobs.
ment of answers to crucial questions such as The problem of measuring job character-
these. Instead, what appears to be needed are istics has been explicitly and carefully dealt
conceptual frameworks which generate test- with by Turner and Lawrence (1965). These
able propositions about how job character- authors developed operational measures of
istics affect employees under various circum- six "requisite task attributes" which, on the
 EMPLOYEE REACTIONS TO JOB CHARACTERISTICS 261

basis of a review of existing literature and an factories in the Turner and Lawrence study),
a priori conceptual framework, were predicted more complex jobs should be responded to pos-
to be positively related to worker satisfaction itively. When employees are alienated from
and attendance. The six attributes are: (a) these norms (as might be expected of urban
variety, (b) autonomy, (c) required interaction, workers), more complex jobs should be re-
(d) optional interaction, (e) knowledge and sponded to negatively. Blood and Hulin
skill required, and (/) responsibility. Scores (1967) provide data supporting this general
on each attribute for each of 47 different jobs proposition and propose a three-dimensional
were obtained from field observations and response surface (Hulin & Blood, 1968) which
interviews by the researchers, and precise specifies the expected interrelationships among
specification of how scores on each of the worker alienation, job level, and satisfaction
attributes is determined is provided. Exami- with work.
nation of the relationships among the six The data of Turner and Lawrence (1965)
requisite task attributes for the 47 jobs re- and of Blood and Hulin (1967) are sufficiently
vealed that the attributes were very closely compelling that the generality of the strong
related to one another. Therefore, Turner arid proposition that enlarged jobs (in the sense of
Lawrence developed a summary measure called being high, e.g., on the dimensions which make
the Requisite Task Attribute Index (RTA up the RTA Index) lead to improved satis-
Index) by formulating a linear combination faction, attendance, and/or performance on
of the six separately measured attributes. This the job must be called into question. Instead,
summary index was then used in ascertaining it appears that certain characteristics of the
the relationships between the attributes of employees themselves must be taken into
the jobs and worker job satisfaction and account simultaneously with the character-
attendance. istics of their jobs in order to generate valid
The authors' expectation that employees predictions about the behavioral and affective
working on jobs which were high on the RTA responses of employees at work.
Index would have higher job satisfaction aiid If the above conclusion is accepted, thorny
lower absenteeism was not fully supported. questions then arise about how the relevant
Instead, it appeared that the predicted relja- differences among employees are to be con-
tionship between the RTA Index and employee ceptualized and measured. Both Turner and
reactions held only for workers from factories Lawrence (1965) and Hulin and Blood (1968)
located in small towns. Workers in urban choose to deal with individual differences on
settings reported less satisfaction with their a subcultural or sociological level (i.e., in
jobs when the jobs were high on the RTA terms of differences between town and city
Index, and the RTA Index was unrelated to workers or in terms of the alienation of city
absenteeism for urban workers. Turner and workers from middle-class work norms).
Lawrence (1965) argued that the obtained An alternative strategy would be to attempt
differences in reactions to good (i.e., high RTJA to conceptualize and measure the relevant
Index) jobs were substantially moderated by individual differences directly at the individual
differences in the cultural backgrounds of level of analysis. The town-city conceptualiza-
employees. tion assumes a substantial homogeneity of
Blood and Hulin (1967) and Hulin and worker characteristics and response tendencies
Blood (1968) provide additional data on the for employees within the two cultural settings.
importance of subcultural factors in deter- To the extent that there are substantial in-
mining worker responses to the makeup .of dividual differences among town workers and
their jobs. These authors hypothesize that an among city workers, an attempt to measure
important moderating factor is alienation from relevant individual differences directly at the
the traditional work norms which character- individual level would seem to have consider-
ize the middle class. When employees hcjld able merit. The difficulty in implementing this
traditional values regarding the value of work alternative approach, of course, is that it
and achievement in work settings (as would requires prior specification on a conceptual
be expected of the employees in small town level of what specific differences among people
262 J. RICHARD HACKMAN AND EDWARD E. LAWLER III

are responsible for the results reported by 3. Thus, to the extent that conditions at
Turner and Lawrence (1965) and Blood and work can be arranged so that employees can
Hulin (1967) (i.e., what it is about people that satisfy their own needs best by working effec-
moderates the way they react to their jobs). tively toward organizational goals, employees
In the following paragraphs, we will propose will in fact tend to work hard toward the
such a conceptualization, and derive from it a achievement of these goals (McGregor, 1960).
number of predictions about the effects of 4. Most lower level needs (e.g., physical
job characteristics on employee satisfaction well-being, security) can be (and often are)
and motivation. reasonably well satisfied for individuals in
contemporary society on a continuing basis
Jobs and Individuals: A Conceptual Framework and, therefore, will not serve as motivational
incentives except under unusual circumstances.
The present conceptualization of the inter- This is not the case, however, for certain
action between job characteristics and indi- higher order needs (e.g., needs for personal
vidual differences is based primarily on the growth and development or for feelings of
expectancy theory of motivation, as formu- worthwhile accomplishment). A person may
lated by Lewin (1938) and Tolman (1959), experience higher order need satisfaction on a
and as applied to work settings by Vroom continuing basis without the strength of desire
(1964), Porter and Lawler (1968), and others. for additional satisfaction of these needs di-
In particular, five propositions based on ex- minishing. Indeed, it may be that additional
pectancy theory are suggested below, which satisfaction of higher order needs actually
address the specific problem of how employee increases their strength (Alderfer, 1969). This
motivation can be enhanced through the design. is an important possibility since it suggests
of jobs. that the opportunity for the development of
1. To the extent that an individual believes continuing (and possibly even increasing)
that he can obtain an outcome he values by motivation is much more a reality when higher
engaging in some particular behavior or class order needs are engaged than is the case for
of behaviors, the likelihood that he will more easily satisfied lower order needs. There
actually engage in that behavior is enhanced. is, of course, a major cost associated with
Relevant valued outcomes can be both in- any motivational approach in which higher
trinsic (e.g., feelings of accomplishment or of order needs are central: Not all employees can
personal worth) and extrinsic (e.g., material or will respond to opportunities for the satis-
goods); the only requirement is that the out- faction of higher order needs, and thus moti-
comes be valued by the individual. When an vational approaches based on these needs
individual anticipates obtaining some valued cannot be applied indiscriminantly. Maslow
outcome as a result of a contemplated action or (1943, 1954) and Alderfer (1969, 1971) discuss
course of action, that outcome may be termed in much more complete detail the nature of
an incentive for engaging in the action. higher order needs and their motivational
2. Outcomes are valued by individuals to implications.
the extent that they satisfy the physiological 5. Individuals who are capable of higher
or psychological needs of the individual, or to order need satisfaction will in fact experience
the extent that they lead to other outcomes such satisfaction when they learn that they
which satisfy such needs or are expected by have, as a result of their own efforts, accom-
the individual to do so. Such need satisfaction plished something that they personally believe
need not, of course, be in the objective best is worthwhile or meaningful (see Argyris, 1964;
interest of the individual. People frequently Lewin, Dembo, Festinger, & Sears, 1944).
strive for satisfying states of affairs which are Specifically, individuals who desire higher
quite inconsistent with their long-term well- order need satisfactions should be most likely
being (Locke, 1969). Nevertheless, if an out- to obtain them when they work effectively on
come is not somehow linked to satisfaction, meaningful jobs which provide feedback on
it will not continue to be valued and therefore the adequacy of their personal work activities.
cannot continue to serve as an incentive. To establish conditions for internal work moti-
 EMPLOYEE REACTIONS TO JOB CHARACTERISTICS 263

vation, then, it appears that a job must: (a) of his work. What is accomplished must be
allow workers to feel personally responsible through the individual's own efforts. He must
for an identifiable and meaningful portion of realize that the work he does is his own. And
the work, (i) provide work outcomes which he must believe that he personally is responsi-
are intrinsically meaningful or otherwise ex- ble for whatever successes and failures occur
perienced as worthwhile, and (c) provide feed- as a result of his work. Only if what is accom-
back about performance effectiveness. The plished is seen as one's own can an individual
harder and better an individual works on such experience a feeling of personal success and a
a job, the more opportunities he will have to gain in self-esteem. This does not mean, of
experience higher order need satisfactions and course, that feelings of personal responsibility
the more incentive there can be for continued for work outcomes cannot occur in team
effective performance. Higher order need satis- projects; all that is required is for team mem-
factions, therefore, are seen both as (a) a bers to feel that their own efforts are important
result of (rather than a determinant of) effec- in accomplishing the task at hand.
tive performance (Lawler & Porter, 1967), The autonomy dimension, as specified by
and (6) an incentive for continued efforts to Turner and Lawrence (1965), would seem to
perform effectively.4 tap the degree to which workers feel personal
The five propositions outlined above lead responsibility for their work. In jobs high on
to the conclusion that it may be possible under measured autonomy, workers will tend to feel
specifiable conditions simultaneously to achieve that they own the outcomes of their work;
high employee satisfaction and high employee in jobs low on autonomy, a worker may more
effort toward organizational goals. Specifically, often feel that successes and failures on the
the long term congruence of high satisfaction job are more often due to the good work (or
and high effort is seen as depending upon (a) to the incompetence) of other workers or of
the existence of employee desires for higher his supervisor.6
order need satisfaction and (b) conditions on 2. The job must provide outcomes which are
the job such that working hard and effectively intrinsically meaningful or otherwise experi-
toward organizational goals will bring about enced as worthwhile to the individual. If a
satisfaction of these needs. worker feels that the results of his efforts are
not very important, it is unlikely that he will
Characteristics of Motivating Jobs feel especially good if he works effectively.
Instead, he must achieve something that he
The three general job characteristics identi- personally feels is worthwhile and important
fied above as central in developing a congruence if he is to be able to experience positive
between individual need satisfaction and orga- feelings about himself as a result of his efforts.
nizational goal achievement must be describ- It clearly is not possible to indicate for people
able in more measurable terms if the validity in general what kinds of job characteristics
of the conceptualization proposed here is to be will be likely to provide outcomes seen as
tested. In the following paragraphs, therefore, meaningful and worthwhile; people differ too
each of the three general characteristics much in the kinds of things they value for
are examined in somewhat more detail. In any statement of such generality to be made.
addition, it will be proposed that four of It is possible, however, to provide some such
the requisite task attributes proposed by 6
Turner and Lawrence (1965) are likely to Having high autonomy on the job does not, of
course, necessarily imply that one will have major
be useful as measures of the three general control over the work outcomes achieved. There may
job characteristics. be a number of factors in the work environment which
1. The job must allow a worker to feel affect the nature of work outcomes, over which the
personally responsible for a meaningful portion worker has little meaningful control. For example, a
football quarterback has high autonomy in selecting
4
It should be noted that only higher order satisfac- plays but only a moderate level of personal control
tions are predicted to be increased by effective work on over the outcomes obtained from execution of the
a job with the characteristics outlined above; other plays. Thus, work autonomy is probably best viewed
satisfactions (e.g., pay satisfaction, satisfaction with as a necessary but not sufficient condition for feeling
supervision) may not be affected. personal responsibility for work outcomes.
264 J. RICHARD HACKMAN AND EDWARD E. LAWLER III

specifications~for*lndividuals""who have high worker will be expected to be experienced as

desires, for higher order need satisfaction and, meaningful to workers with desires for higher
of course, these are the individuals to whom order need satisfaction; screwing many differ-
the present conceptualization is intended to ent sizes of nuts on many different colors of
apply. bolts, if this could be considered variety,
There are at least two ways that work can would not be expected to be experienced as
come to be experienced as meaningful for meaningful.6
employees with relatively high desires for To summarize, it may be that jobs can come
higher order need satisfaction. The first is for to be experienced as meaningful to employees
the job to be a sufficiently whole piece of work to the extent that they involve doing a whole
that the worker can perceive that he has piece of work of some significance (i.e., have
produced or accomplished something of con- high task identity) and, at the same time, to
sequence. In terms of a Turner and Lawrence the extent that they give employees the chance
task attribute, this would be expected to be to use their valued skills and abilities (i.e., to
the case when a job is high on task identity. be challenged) in doing the work. In many
According to Turner and Lawrence (1965, p. cases the latter condition may be met on jobs
157), jobs high on task identity are character- which have high variety.
ized by (a) a very clear cycle of perceived 3. The job must provide feedback about
closure—the job provides a distinct sense of what is accomplished. Even if the two general
the beginning and ending of a transformation conditions discussed above are met, an em-
process, (b) high visibility of the transforma- ployee cannot experience higher order need
tion to the worker, (c) high visibility of the satisfaction when he performs effectively un-
transformation in the finished product, and less he obtains some kind of feedback about
(d) a transformation of considerable magni- how he is doing. Such feedback may come
tude. For a worker who has high needs for from doing the task itself (e.g., when a tele-
developing and using his competence, a job phone operator successfully completes a long
with such characteristics generally would be distance person-to-person call), but perfor-
expected to be experienced as highly meaning- mance feedback also may come from some
ful and worthwhile. other person—an esteemed co-worker, a super-
In addition, the experienced meaningfulness visor, etc. The crucial condition is that feed-
of work may be enhanced when a job provides back be present in a form that is believable to
a worker with the opportunity to accomplish the worker, so that a realistic basis exists for
something by using skills and abilities which the satisfaction (or frustration) of higher order
he personally values. For example, a strongly needs.
motivated duffer feels good when he hits a It should be emphasized that, for all of the
solid tee shot, even though the broader sig- job characteristics discussed above, it is not
nificance of this event is doubtful. His golfing their objective state which affects employee
skills are on the line when he steps to the tee; attitudes and behavior, but rather how they
those skills are important to him; he performs are experienced by the employees. Regardless
well—and that, in itself, is enough. of the amount of feedback (or variety, or
Jobs high on the Turner and Lawrence autonomy, or task identity) a worker really
(1965) dimension of variety would be expected has in his work, it is how much he perceives
to provide opportunities for workers to experi- that he has which will affect his reactions to the
ence this kind of meaningfulness on the job, 6
since high variety jobs typically tap a number It is also possible, of course, for a job to have too
much variety. Activation theory (e.g., Scott, 1966)
of different skills which may be important to suggests that when variety is too high, employees may
the employee. Thus, working on high variety experience a general state of muscular and mental
jobs may become personally meaningful to hypertension which can greatly handicap performance
some employees through a process very analog- effectiveness. In addition, Hall and Lawler (1970)
found that among research scientists, high job variety
ous to that which makes golf meaningful to can be associated with low job satisfaction, apparently
the duffer. It should be noted, however, that because jobs with high variety also tended to be low
only variety which does in fact challenge the in task identity and feedback.
 EMPLOYEE REACTIONS TO JOB CHARACTERISTICS 265

job. Objective job characteristics are impor- forming on the job; («) dealing with others,
tant because they do affect the perceptions the degree to which a job requires employees
and experiences of employees. But there are to deal with other people (either customers,
often substantial differences between objective other company employees, or both) to com-
job characteristics and how they are perceived plete the work; (/) friendship opportunities,
by employees, and it is dangerous to assume the degree to which a job allows employees
that simply because the objective character- to talk with one another on the job and to
istics of a job have been measured (or changed) establish informal relationships with other em-
that the way that job is experienced by em- ployees at work.
ployees has been dealt with as well. The latter two dimensions were included to
In summary, then, it has been argued that permit exploration of the impact of the inter-
the characteristics of jobs can establish con- personal characteristics of job design. These
ditions which will enhance the intrinsic motiva- dimensions were adapted with very minor
tion of workers who desire higher order need revision from the task attributes "required
satisfaction. In particular, it has been sug- interaction" and "optional interaction" pro-
gested, in terms of a subset of the Turner and posed by Turner and Lawrence (1965). They
Lawrence (1965) dimensions, that such indi- are not, however, directly relevant to the con-
viduals will be able to obtain meaningful ceptualization about job-based work motiva-
personal satisfaction when they perform well tion proposed above, and no specific predic-
on jobs which they experience as high on tions regarding them were made. Thirteen
variety, autonomy, task identity, and feedback. different jobs were described on these six di-
Further, the harder and better an individual mensions by the researchers, by employees
performs on a job which is perceived as high who worked on the jobs, and by members of
on these dimensions (hereafter will be called the the management of the telephone company in
four core dimensions), the more satisfaction he which the research was carried out.
is likely to feel. 2. A measure was developed which was
expected, on an a priori basis, to reflect the
Strategy of the Present Research level of employee desire for the satisfaction of
The conceptualization presented above pro- higher order needs.
vides the basis for the present empirical 3. Based on the mean scores of the employees
research on the relationships among job char- on the measure of need strength, predictions
acteristics; individual differences in need were made regarding the expected relation-
strength; and employee motivation, satisfac- ships between the job characteristics as mea-
tion, performance, and absenteeism on the sured by the four core dimensions and the
job. In particular, the research to be reported dependent variables: satisfaction, performance,
here follows the strategy steps listed below: and absenteeism. Relevant data were collected
1. Measures of the following six job dimen- from 208 employees, and correlations between
sions were developed: (a) variety, the degree each of the four core dimensions and each of
to which a job requires employees to perform the dependent measures were computed.
a wide range of operations in their work 4. The theory outlined above indicates that
and/or the degree to which employees must how a job is experienced or perceived by an
use a variety of equipment and procedures in individual employee should determine his re-
their work; (b) autonomy, the extent to which actions to it, rather than the objective char-
employees have a major say in scheduling acteristics of the job. This possibility can be
their work, selecting the equipment they will examined by analyzing the relationship be-
use, and deciding on procedures to be fol- tween the characteristics of a single job (as
lowed; (c) task identity, the extent to which they are perceived by employees) and the
employees do an entire or whole piece of work behavioral and attitudinal reactions of indi-
and can clearly identify the result of their viduals who hold that job. By restricting the
efforts; (d) feedback, the degree to which scope of the analysis to individuals working
employees receive information as they are on the same job, it is possible to rule out ob-
working which reveals how well they are per- jective differences in jobs as an explanation for
266 J. RICHARD HACKMAN AND EDWARD E. LAWLER III

obtained empirical relationships—and, thereby, METHOD

to address the possibility that the perceived
characteristics of jobs affect employees in the Research Setting and Subjects
same general fashion as do objectively mea- The research was carried out in an eastern telephone
sured characteristics. Therefore, relationships company, and focused on employees who worked on 13
different jobs in the plant and traffic departments of
between perceived job characteristics and the the company. The jobs were selected so as to include
dependent variables of interest were computed (a) several varieties of operators, installers, central
separately for employees who worked on each office repairmen, and cable splicers; (b) to range widely
of the 13 objectively different jobs included in in complexity and in the level of employee skill re-
the study, and these relationships were com- quired; and (c) to be located in both rural and urban
settings.
pared to those obtained in the overall analyses Data were collected from 208 employees and 62
involving all 5s and all 13 jobs. supervisors. Employees in the traffic department
5. The theory implies that satisfaction, per- (about one-third of the sample) were female; all plant
formance, and attendance should be highest department employees were male.
when all four of the core dimensions are
present. That is, being high on, say, variety Procedure
and task identity but low on autonomy and All data were collected on site at each of the 13 job
feedback should not provide the conditions locations. One to three days were spent at each loca-
that are necessary for positive behavioral and tion, and all data were collected from each location
before moving on to another location. Data were col-
affective responses. Under such circumstances, lected over a period of about 9 mo. At each location,
a worker would be expected to have neither the following five procedural steps were followed
a sense of personal ownership of his work (although the order of the steps sometimes varied
activities and outcomes (low autonomy) nor because of local circumstances).
1. Local second- or third-level management was
a trustworthy gauge of the adequacy of his visited to obtain permission to collect data from
performance (low feedback). The conditions employees working on a particular job. When permis-
for the development of internal work motiva- sion was secured (it was never denied), the managers
tion specified earlier would therefore not be were interviewed about the general nature of the job
met. The importance of having all four core as they perceived it.
2. First-level supervisors on the local job were inter-
dimensions present was tested by comparing viewed about the nature of the job and employees'
the dependent variable scores of employees reactions to it and were given a questionnaire which
who saw their jobs as moderately high on all tapped the supervisors' perceptions of the employee
four core dimensions both with the scores of job in a format similar to that used for obtaining the
employees' own job perceptions.
employees who saw their jobs as high on some 3. Employees working on the job were observed and
dimensions and low on others, and with the interviewed informally. These observations and inter-
scores of employees who saw their jobs as views were conducted by two researchers and con-
moderately low on all four dimensions. tinued until the researchers felt that they were no
longer obtaining substantial new information about the
6. The theory states that individual differ- job. Observation typically lasted approximately one
ences in desire for higher order need satisfac- working day. Ratings of the job characteristics by the
tions should moderate the relationships be- researchers were made on the basis of these observa-
tween job characteristics and the dependent tions and interviews.
4. A questionnaire was administered to a sample of
variables. In order to test this possibility, 15-20 employees on the job. The questionnaire took
relationships were computed separately and between |-1 hr. to complete and usually was admin-
compared for the third of the employees high- istered to employees in groups of three or four. The
general nature of the research was explained to each
est on desire for higher order need satisfaction group of employees before they began work on the
and for those employees lowest on desire for questionnaire, although the hypotheses of the study
higher order need satisfaction. and the dependent variables to be analyzed were not
mentioned. It was emphasized to each individual that
7. Finally, exploratory analyses were made participation was voluntary, and a few individuals did
of the relationships between the two inter- decline to participate. In addition, employees were
personal job dimensions (dealing with others, told that putting their names on the questionnaires,
while desirable for research purposes, also was volun-
and friendship opportunities) and the depen- tary. About 10% of the employees who participated
dent variables. declined to give their names.
 EMPLOYEE REACTIONS TO JOB CHARACTERISTICS 267

5. Ratings of the performance of those employees Very little; I often work for long stretches without
who had taken the questionnaire and provided their finding out how I am doing.
names were made by first-line supervisors. Absence Moderately; sometimes I know how I am doing
data were collected later from company records. and other times I do not.
Very much; I get almost constant "feedback" on
Instruments and Measures my performance as I work.
e. To what extent do you have the opportunity to
Job descriptions. Each of the 13 jobs examined in the talk informally with other employees while at
research was described in terms of the six general work. (That is, is your job arranged so that you
dimensions described earlier (variety, autonomy, task can chat with other workers while on the job—
identity, feedback, dealing with others, and friendship even though the job does not require you to talk
opportunities). The jobs were described on all six to these people?)
dimensions using four different methods: (a) a sample Very little; there is almost no chance to talk to
of employees on the job completed several question- other employees except about "business."
naire scales relevant to the six dimensions; (b) first- Moderately; there is some chance to talk, but
and second-level management on the job completed you may have to arrange it ahead of time.
identical questionnaire scales; (e) the researchers, after Very much; there is almost always an opportunity
observing the job and interviewing employees and to talk with other employees about non-business
supervisors, completed similar scales; and (d) the re- topics.
searchers utilized a set of objective coding procedures f. To what extent is dealing with other people a part
adapted from Turner and Lawrence (1965) to obtain of your job?
a set of nonsubjective assessments of each job. Very little; working with other people is not a
The job descriptions completed by employees and very important part of my job.
supervisors involved three separate questionnaire items Moderately; I have to deal with some other
for five of the six dimensions; two items were used for people, but this is not a major part of my job.
the dimension "dealing with others." One section of Very much; probably the single most important
the questionnaire involved making direct ratings of the part of my job is working with other people.
job on six 7-point scales. The scales used in this section
of the employee questionnaire, together with the labels A second section of the questionnaire asked em-
attached to each of the scale end points and the mid- ployees on 7-point scales how much of the various job
point, are listed below. attributes was actually present on their jobs. These
a. How much variety is there in your job? scales ranged from "none or a minimum amount"
Very little; I do pretty much the same things over (scored 1) through "a moderate amount" (scored 4)
and over and use the same pieces of equipment to "a maximum amount" (scored 7). Scales which were
and procedures almost all the time. (Scored 1) used in this section of the questionnaire include:
Moderate variety (Scored 4) (a) For variety: The amount of variety in my job.
Very much; I do many different things and use a The opportunity to do a number of different
wide variety of equipment and procedures. things.
(Scored 7) (b) For autonomy: The opportunity for indepen-
b. How much autonomy do you have on your job; dent thought and action.
how much are you left on your own to do your The freedom to do pretty much what I want on
own work? my job.
Very little; I have almost no "say" about schedul- (c) For-task identity: The opportunity to do a job
ing my work; the work and the procedures are all from the beginning to end (i.e., the chance to do
laid out for me in detail. a whole job).
Moderate autonomy; I make some of the decisions The opportunity to complete work I start.
about my work, but many of them are made for (d) For feedback: The opportunity to find out how
me. well I am doing in my job.
Very much; I have almost all of the "say" about The feeling that I know whether I am perform-
the scheduling of my own work; I alone decide ing my job well or poorly.
what procedures will be used. (e) For dealing with others: The opportunity, in
c. To what extent do you do a "whole" piece of work my job, to give help to other people.
(as opposed to doing part of a job which is (/) For friendship opportunities: The opportunity
finished by some other employee) ? in my job to get to know other people.
I do one small part of a job; there are many The opportunity to develop close friendships in
others who do other parts of the job; I may not my job.
see the final result.
I do a moderate size "chunk" of the work; there The items described above were randomly spread
are others involved too, but my contribution is throughout a 23-item section of the questionnaire
clear. which asked about the presence or absence of a number
I do an entire piece of work; I do the job from of heterogeneous aspects of the work setting.
start to finish, and what is done is clearly "mine." Scores on all items describing each of the six job
d. To what extent do you find out how well you are dimensions were averaged to arrive at a set of six
doing on the job as you are working? summary scores from the employee ratings described
268 J. RICHARD HACKMAN AND EDWARD E. LAWLER III

TABLE l regarding this dimension is that some feedback is

ESTIMATED RELIABILITIES OF THE Six provided directly by the job as an employee works
DESCRIPTIVE DIMENSIONS (e.g., an installer successfully installing and testing a
telephone), while additional feedback may be provided
by supervision (e.g., a foreman telling an installer
Source of ratings that he has taken too long to do a particular job). It
may be that employees, supervisors, and researchers
Dimension Turner- were not attending to the same aspect of the feedback
Em- Super- process in making their judgments, thereby lowering
visors Lawr-
ployees the level of agreement.
ence
In addition to the problem with the feedback dimen-
Variety .90 .91 .86 sion, there is a tendency for correlations involving
Autonomy .77 .68 .89 supervisors to be lower than others. This is especially
Task identity .77 .86 .95 true for the two dimensions involving interpersonal
Feedback .75 .75 .97 relationships: dealing with others, and friendship op-
Dealing with others .47 .17 .88 portunities. Examination of the mean scores of em-
Friendship opportunities .43 .42 .92 ployees and supervisors on these two jobs reveals no
consistent differences. Supervisors did not see the jobs
Note.—Reported reliabilities for employee and supervisor as having either consistently more or consistently less
ratings are estimated internal consistencies. Each dimension of an interpersonal component than employees—they
score for these ratings consists of the average of three question-
naire items, except for "dealing with others" which is the aver- simply did not agree well with the employees (or with
age of two items. The reliability reported for the Turner-Lawr- the researchers) about the extent of interpersonal
ence procedure is the estimated reliability of the average of two activities relevant to the job. Part of the reason for
judges, adjusted by Spearman-Brown procedures. Ratings by
the researchers (the fourth procedure used) were made col- this lack of agreement may be the low internal con-
laboratively and reliability data are not available. sistency reliability of supervisors on the dimension
"dealing with others." In addition, on-site observations
above. The questionnaire items given to supervisors indicated that, by and large, supervisors were con-
were identical in every respect, except that supervisors siderably more attentive to and concerned with the
were asked to describe "the job you supervise" rather technical and production aspects of the jobs they
than "your job" as was the case for the employees. supervised than with interpersonal issues. This ten-
Internal consistency reliabilities for the items making dency may partially account for both the unreliability
up each of the six job dimension scores are presented of the supervisors on the "dealing with others" dimen-
separately for employees and supervisors in Table 1. sion and the low level of agreement between supervisors
Also included are interjudge reliabilities for the re- and the other raters on the two interpersonal dimensions.
searchers' assessments of the jobs. With the exception Table 2 shows that employees themselves provide
of supervisors' data on the dimension "dealing with judgments of the characteristics of their jobs which, in
others," the reliabilities are adequate. general, agree quite well with those made by outsiders
Comparison of the scores obtained for each job by and (with the exceptions noted above) by their super-
the four rating procedures reveals high agreement on visors. It is not, of course, possible to demonstrate
all but one of the dimensions. These correlations are conclusively that the employee judgments are objec-
presented in Table 2. There is no agreement about tively accurate, because no unambiguous standard of
which jobs are high and which jobs are low on the accuracy is available. Nevertheless, the strong con-
feedback dimension. One possible source of confusion vergence of the employee judgments with the assess-

TABLE 2
CORRELATIONS AMONG THE FOUR RATING PROCEDURES

Job dimensions

Correlation between Dealing Friend-

Task ship op-
Variety Autonomy Feedback with
identity portuni-
others
ties

Employees & supervisors .87* .85* .65* .09 .31 .49*

Employees & researchers .94* .94* .79* -.22 .91* .73*
Employees & Turner-Lawrence .89* .72* .63* -.66 .95* .52*
Supervisors & researchers .94* .79* .63* -.15 .37 .38
Supervisors & Turner-Lawrence .84* .49* .62* .23 .26 .35
Researchers & Turner-Lawrence .90* .80* .69* .16 .93* .67*

Note.—N = 13 jobs.
* p < .05 (one-tailed test).
 EMPLOYEE REACTIONS TO JOB CHARACTERISTICS 269
TABLE 3
CORRELATIONS AMONG THE Six DESCRIPTIVE DIMENSIONS : EMPLOYEE RATINGS

Task Dealing Friendship

Dimensions Variety Autonomy Feedback with opportuni-
identity
others ties

Variety
Autonomy .67* —
Task identity .17 .24* —
Feedback .11 .06 .12 —
Dealing with others .05 -.05 ,19* .26* —
Friendship opportunities .41* .21* .07 .32* .13 —

Note.—N — 208 employees.

* t < .01 (two-tailed test).

ments obtained from the researchers, from supervisors, Individual need strength. To obtain a measure of the
and from the Turner-Lawrence procedures does suggest degree to which 5s were desirous of obtaining higher
that the employees were able to provide generally order need satisfactions from their work, a question-
nondistorted descriptions of the characteristics of their naire was administered which asked how much of
jobs. In most of the analyses reported in this article, various opportunities and attributes the employees
employee judgments are used as the primary measures "would like" to have on their job. In a space beside
of job characteristics. This is appropriate and necessary, each item, employees entered a number ranging from
given that it was argued earlier that employees' per- 1 (would like to have none or a minimum amount) to
ceptions of their jobs (rather than objective job char- 7 (would like to have a maximum amount). Content
acteristics) are causal in affecting the reactions of of the questionnaire ranged widely, and included items
employees to their work. Fortunately, the data in relevant to pay, promotion, security, working condi-
Table 2 establish that employee perceptions have tions, peer relationships, and supervisory relationships.
convergent validity as descriptors of jobs—and, there- Twelve of the items were judged on an a priori basis
fore, that they are probably reasonably well-grounded to measure desire for higher order need satisfactions.
in reality. These items are:
The relationships among the six job dimensions
themselves are presented in Table 3. Although there (a) The opportunity for personal growth and de-
is some tendency for the six dimensions to be positively velopment on my job.
related to one another, only two of the correlations are (6) The opportunity for independent thought and
of substantial magnitude: jobs seen as having high action on my job.
variety also are seen as being high in autonomy and (c) The opportunity to find out how I am doing.
in friendship opportunities. The level of interrelation- (d) The opportunity to complete work I start.
ship among the six dimensions as measured in the (e) The opportunity to do challenging work.
present research is lower than that reported by Turner (/) The feeling that I know whether I am perform-
and Lawrence (1965), and does not mitigate against ing my job well or poorly.
use of the six dimensions separately as descriptors of (g) The opportunity to do a number of different
job characteristics. things.
Table 4 presents a listing of the 13 jobs studied, the (h) The opportunity to do a job from the beginning
number of employees on each job who participated in to the end (that is, the chance to do a whole job).
the research, and the mean employee ratings of the (i) The freedom to do pretty much what I want on
characteristics of each job.7 my job.
7
(j) The amount of variety in my job.
Descriptions of the 13 jobs on Turner-Lawrence (h) The feeling of worthwhile accomplishment in my
dimensions are not included in Table 4 for two reasons. job.
First, the Turner-Lawrence procedures were modified (/) The opportunity, in my job, for participation in
somewhat for use in the present study, and therefore the determination of methods, procedures, and
the scores obtained are not directly comparable to goals.
those reported by Turner and Lawrence (1965) for
their sample of 47 jobs. Second, the primary use of the Scores of each employee on the 12 items were
Turner-Lawrence scores in the present research was to summed to obtain an overall measure of the level of
demonstrate the convergence of employee assessments
of their jobs with more objectively based job descrip- in the telephone company and with employee sex
tions (see Table 2). Substantive analyses using the (see Table 4). It was concluded on these grounds that
modified Turner-Lawrence scores are not reported, the sample of 13 jobs was not adequate for analysis
since the scores of specific jobs on the six dimensions of relations between objective job descriptions and mean
are confounded both with traffic versus plant functions dependent variable scores using jobs as observations.
270 J. RICHARD HACKMAN AND EDWARD E. LAWLER III

TABLE 4
SUMMARY or THE CHARACTERISTICS or THE JOBS STUDIED

X employee ratings of job characteristics

Sex Deal- Friend-

Job description N Task ing ship
Vari- Auton- Feed-
iden- with oppor-
ety omy back
tity tuni-
others
ties

Director assistance (DA) operators F 16 3.08 3.64 5.66 4.88 6.06 4.13
Toll operators F 18 2.87 3.06 5.15 4.80 6.03 3.69
Combined DA & toll operators F 17 3.97 3.29 5.87 5.59 6.32 5.02
Traffic service position (TSP) operators F 18 2.78 2.91 5.06 5.09 6.14 4.23
TSP operators with "enriched" jobs F 19 2.94 2.96 5.26 5.08 5.90 4.05
Tip & ring (residence) installers M 18 4.78 5.23 6.38 4.41 5.23 4.14
Key (commercial) installers M 19 5.72 5.06 6.26 4.87 5.60 4.85
Private branch exchange installers M IS 6.31 5.61 5.25 4.11 5.07 4.87
Cable splicers M 16 5.14 4.67 5.04 5.12 4.59 4.67
Combined tip & ring & key installers M 17 4.94 5.09 6.11 4.23 4.85 4.45
Central office framemen M 8 4.50 3.89 4.42 5.43 5.12 5.44
Central office repairmen : Step equipment M 11 5.30 4.84 5.32 4.30 4.77 4.70
Central office repairmen : Crossbar equipment M 16 5.91 4.92 5.03 5.15 5.47 5.19

higher order need strength. Internal consistency re- pressure (e.g., "It is not especially important for a
liability of the 12-item scale was .89. worker on this job to do high quality work").
Employee reactions to their jobs. Four types of em- 2. Job involvement: three questionnaire items, in
ployee reactions to their jobs were obtained from the the same format as the intrinsic motivation items,
questionnaires employees completed, for use as depen- were averaged to yield a score indicative of the degree
dent variable measures. The four types of reactions are employees felt personally involved in their work. The
described separately below. items were selected from those used by Lodahl and
1. Experienced work motivation: measures were Kejner (1965). Internal consistency reliability of the
derived from the questionnaire to reflect both the scale was .81. The three items are: (a) The most im-
amount of intrinsic motivation employees experienced portant things that happen to me involve my work, (b)
on the job and the focus of their motivation. The I live, eat, and breathe my job. (c) I am very much
amount of intrinsic motivation was assessed by three personally involved in my work.
questionnaire items: (a) I feel a great sense of personal 3. General job satisfaction: employees completed
satisfaction when I do my job well, (b) Doing my job three questionnaire items designed to tap overall job
well increases my feeling of self-esteem, (c) I feel bad satisfaction. The three items also were in the same
when I do my job poorly. The items were placed on format as the intrinsic motivation items and were
7-point scales ranging from "strongly disagree" (scored averaged to provide an index of general satisfaction.
1) to "strongly agree" (scored 7). Scores on the three Internal consistency reliability of the 3-item scale was
items were averaged to yield a score indicative of the .76. The items are: (a) Generally speaking, I am very
degree to which an employee experiences positive satisfied with my job. (b) I frequently think of quitting
affective outcomes when he performs well, and nega- my job (reversed scoring), (c) Generally speaking, I
tive affective outcomes when he performs poorly am very satisfied with the kind of work I have to do
(Lawler & Hall, 1970). Internal consistency reliability on my job.
of this scale was .72. The nine items used in the general satisfaction, job
The focus of employee motivation was tapped by involvement, and intrinsic motivation scales were
three questionnaire items which addressed the kinds randomly spread throughout a 23-item section of the
of internal pressures for performance which were employee questionnaire which addressed various types
experienced on the job. Three items were used, which of reactions to work. Lawler and Hall (1970) have
dealt with three different (but not mutually exclusive) argued that it is useful, on a conceptual level, to treat
foci of work motivation: (a) Being personally responsi- satisfaction, involvement, and intrinsic motivation
ble for what you do; checking your own work, (b) separately. In addition, they demonstrated that the
Producing a large quantity of work, (c) Doing high three variables do have discriminant validity in pre-
quality work. Each item was on a separate 7-point dicting other employee attitudes and behaviors—even
scale, which ranged from high experienced pressure though the variables are moderately positively inter-
(e.g., "It is extremely important for a worker on this correlated. The correlations obtained in the present
job to do high quality work") to low experienced research among intrinsic motivation, general satisfac-
 EMPLOYEE REACTIONS TO JOB CHARACTERISTICS 271
tion, and the three items reflecting the specific focus absent were used to discount the effects of single long
of employee motivation are included in Table 5. periods of absence. For example, if an employee were
4. Specific satisfaction items: Employees indicated ill for a 2-wk. period, the data would show the event
on 12 questionnaire items the degree to which they as one occasion of absence, rather than as 10 working
were satisfied with particular aspects of their job. The days of absence.
items ranged from "extremely dissatisfied" (scored 1) Summary of measures. The various measures used in
to "extremely satisfied" (scored 7). The 12 items were the research are summarized below. The means,
not summed but were retained as separate indicators standard deviations, and intercorrelations of all de-
of specific satisfactions of the job. Median intercorrela- pendent variable measures are presented in Table 5.
tion among the 12 items was .40. The items queried The measures include: (a) Job descriptions: variety,
how satisfied employees were with: autonomy, task identity, feedback, dealing with others,
(a) The feeling of self-esteem or self-respect a person and friendship opportunities. Descriptions were made
gets from being in my job. by employees, by supervisors, by the researchers using
(b) The opportunity for personal growth and de- the Turner-Lawrence procedures, and by the re-
velopment in my job. searchers subjectively after job observations and inter-
(c) The prestige of my job inside the company views. Because there was generally high convergent
(that is, the regard received from others in the validity among the four sets of job descriptions, and
company). because the conceptual basis of the study suggests
(d) The amount of close supervision I receive. that jobs as experienced by employees should be most
(e) The opportunity for independent thought and directly causal of employee reactions to their jobs,
action in my job. the employee job descriptions are used in the analyses
(/) The feeling of security in my job. to be reported in subsequent sections. (6) Level of
(g) The pay for my job. higher order need strength of employees: a summary
(h) The feeling of worthwhile accomplishment in my score of employee reactions to 12 need strength items.
job. (c) Employee reactions to their jobs and work. Ques-
(i) The opportunity, in my job, for participation in tionnaire-derived measures of: the amount of intrinsic
the determination of methods, procedures, and motivation experienced by employees, and the focus
goals. of their motivation; general job satisfaction; personal
(J) The opportunity to develop close friendships in job involvement; and 12 specific satisfactions or dis-
my job. satisfactions with the job or the work situation, (d)
(k) The opportunity for promotion. Rated performance effectiveness of employees, in terms
(/) The amount of respect and fair treatment I of quantity of work produced, work quality, and over-
receive from my boss. all performance effectiveness, (e) Absenteeism, mea-
sured by the number of occasions an employee was
Performance measures. Supervisors rated the per- absent during a 12-mo. period.
formance of each employee on several dimensions,
three of which are used in this report: (a) quantity of
work produced, (6) quality of work produced, and
RESULTS
(c) overall performance effectiveness. General Relationships between Job Character-
The quantity and quality scores for each employee istics and Employee Reactions
were obtained by averaging the ratings he received on
7-point rating scales from his supervisors (when more According to the conceptual position out-
than one supervisory rating was available). The overall lined earlier, the nature of the relationships
performance measure was obtained by summing aver-
age supervisory ratings across seven separate per- between job characteristics and employee re-
formance scales, including quantity and quality. Inter- actions to their work (including satisfaction,
judge reliabilities of the quantity and quality scales performance, and absenteeism) will depend
were .71 and .79, respectively, and the internal con- upon the need states of the employees. In
sistency reliability of the overall performance effective- particular, it was predicted that, if employees
ness scale was .92. The correlation between the quantity
and the quality scales was .63. are desirous of higher order need satisfactions,
Anchors for the performance ratings were phrased in there should be a positive relationship between
absolute terms, and supervisors were encouraged to the four core dimensions (variety, autonomy,
assess each employee against the highest conceivable task identity, and feedback) and motivation,
level of performance possible for the job in making his
assessments. Nevertheless, the degree to which ratings
satisfaction, performance, and attendance.
made by different supervisors on different jobs are The mean score of the 208 telephone com-
directly comparable is open to question, and results pany employees on the 12-item scale indicative
involving the performance ratings must be interpreted of higher order need strength was 6.01. Given
with caution. that the maximum possible score on the scale
Absenteeism measure. The number of occasions an
employee was absent during the 12-mo. period during is 7.0, it would appear that the employees who
which the study took place was derived from company participated in the present research had strong
payroll records. Occasions absent rather than days higher order needs, or at least that they felt
 to
TABLE 5 ~i
to
INTERCORRELATIONS AMONG DEPENDENT VARIABLES

Correlations
Variable X SD
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

1. Level of intrin-
sic motivation 5.96 .87 —
Focus of motivation.
2. Taking personal
responsibility 6.39 .99 .30* —
3. Doing large
quantities of
work 5.21 1.53 .01 .17* —
4. Doing high qual-
ity work 5.22 1.09 .25* .44* .08 —
Rated performance
5. Quantity 5.24 1.45 .04 -.09 .09 -.05 —
6. Quality 5.24 1.27 .13 .05 .08 -.04 .63* — o
7. Overall effec-
tiveness 5.53 1.11 .18* .02 .10 .06 .79* .85* —
8. General job >
satisfaction 4.97 1.26 .39* .18* .01 .23* .07 .08 .16* —
9. Job involvement 3.00 1.37 .39* .20* .04 .16* .07 .07 .11 .44* —
10. Absenteeism (no.
of times absent) 2.69 2.16 -.23* -.14* -.10 -.06 -.12 -.26* -.33* -.10 -.15* —
Specific satisfaction D
items
11. Self-esteem ob- W
tained from job 4.34 1.44 .23* .04 -.01 .06 -.07 -.02 -.04 .49* .34* -.06 — O
12. Personal growth
and development 4.33 1.48 .26* .18* .04 .11 .07 .03 .07 .58* .38* -.02 .66* —
13. Prestige of job
inside company 3.95 1.44 .16* -.01 .10 .04 -.09 -.03 -.03 .44* .29* -.09 .63* .50* —
14. Amount of close
supervision re-
ceived 4.46 1.45 .25* .14* -.01 .13 .04 .06 .10 .28* .25* -.07 .31* .45* .30* —
15. Independent
thought and
action 4.62 1.58 .21* .13 -.11 .10 .11 .12 .17* .36* .34* -.09 .39* .50* .32* .62* —
16. Security 5.32 1.47 .15* .13 .02 .16* -.01 .06 .09 .47* .23* —.08 .43* .46* .39* .40* .40* —
17. Pay 3.68 1.74 .07 .04 -.02 .10 -.06 -.13 -.11 .44* .16* .02 .41* .43* .39* .21* .15* .45* —
18. Feeling of worth-
while accom-
plishment 4.82 1.61 .45* .22* .01 .19* .09 .14* .11 .58* .39* -.02 .58* .58* .41* .32* .49* .53* .42* —
19. Participation in
job-related de-
cisions 3.86 1.50 .23* .08 .05 .13 .04 .04 .07 .31* .28* —.04 .38* .52* .39* .37* .46* .30* .31* .50* —
20. Development of
close friendships 4.78 1.28 .06 .06 -.11 -.02 -.09 -.04 -.14* .19* .05 .15* .24* .34* .21* .25* .31* .28* .22* .33* .33* —
21. Promotion 3.98 1.64 .16* .16* .08 .15* .07 -.04 .03 .39* .34* .09 .43* .60* .42* .38* .29* .44* .48* .43* .51* .22* —
22. Respect and fair
treatment from
boss 5.34 1.47 .23* .08 -.06 .14* .10 .06 .15* .33* .10 —.07 .42* .40* .35* .43* .41* .57* .34* .45* .35* .25* .42* —

Note.—N = 208.
* p < .05 (two-tailed test).
 EMPLOYEE REACTIONS TO JOB CHARACTERISTICS 273

TABLE 6
GENERAL RELATIONSHIPS BETWEEN JOB CHARACTERISTICS AND EMPLOYEE REACTIONS

Interpersonal
Core dimension dimension
Dependent variable
Dealing Friendship
Variety Autonomy Task Feedback with opportuni-
identity others ties

Level of intrinsic motivation .32* .30* .16* .18* .07 .09

Focus of motivation
Taking personal responsibility .14* .12* .19* .06 .08 .05
Doing large quantities of work -.10 -.12 .01 .02 .06 -.17
Doing high quality work .16* .12* .13* .10 .04 .12*
Rated performance
Quantity -.03 .13* .05 .00 -.02 -.18
Quality .17* .16* .07 .02 -.11 -.02
Overall effectiveness .20* .26* .11* -.03 -.07 -.09
General job satisfaction .38* .39* .20* .28* .17* .21*
Job involvement .24* .22* .12* .24* .03 .16*
Absenteeism (no. of times absent) .02 -.14* -.22* -.10 .01 -.05
Specific satisfaction items
Self-esteem obtained from job .32* .32* .15* .35* .15* .27*
Personal growth and development .36* .34* .14* .31* .11* .29*
Prestige of job inside company .30* .25* .15* .35* .14* .28*
Amount of close supervision received .31* .35* .13* .30* .07 .16*
Independent thought and action .53* .62* .25* .15* .00 .25*
Security .22* .27* .19* .39* .15* .28*
Pay .04 .05 .04 .34* .25* .24*
Feeling of worthwhile accomplishment .29* .32* .28* .42* .23* .31*
Participation in job-related decisions .28* .27* .20* .34* .12* .25*
Development of close friendships .25* .12* .09 .29* .09 .47*
Promotion .17* .20* .15* .34* .21* .19*
Respect and fair treatment from boss .19* .26* .22* .35* .14* .24*

Note—2V = 208.
* p < .05 (one-tailed test).

it was appropriate to express high desires for sions, the more employees tend to report feel-
these need satisfactions. Therefore, generally ing internal pressures to take personal respon-
positive relationships would be expected be- sibility for their work and to do high quality
tween job descriptions on the four core dimen- work. And, in fact, when jobs are described
sions and the dependent variables. as being higher on variety, autonomy, and
Correlations between the six job description task identity, employees are rated as doing
measures obtained from employees and the higher quality work and as being generally
scores of employees on the dependent variable more effective performers on the job.
measures are presented in Table 6. Results are Further, the data provide support for the
discussed below separately for the four core di- interpretation of these relationships which
mensions and the two interpersonal dimensions. was proposed earlier in this paper: when jobs
are high on the core dimensions, employees
Core Dimensions: Toward Internal Motivation report having higher intrinsic motivation to
In general, positive relationships were ob- perform well. That is, employees indicate that
tained between the four core dimensions and when they perform well on the job, they ex-
dependent measures indicative of motivation, perience positive internal feelings, and when
satisfaction, performance, and attendance. they perform poorly, they feel badly. Appar-
Performance motivation and actual perfor- ently jobs seen by employees as being high
mance. The higher jobs are on the core dimen- on the core dimensions do establish conditions
274 J. RICHARD HACKMAN AND EDWARD E. LAWLER

on the job so that it is possible for workers to ing the kinds of effects that the core dimensions
obtain personally rewarding experiences by have on employees' reactions to their jobs.
doing well on the job. The data suggest, The four specific satisfaction items most
moreover, that "doing well" is interpreted in strongly related to the core dimensions are
the job context as having much more to do (in descending order): (a) The opportunity
with high quality performance than producing for independent thought and action in my job.
large quantities of work. The core dimensions (b) The feeling of worthwhile accomplishment
do not relate either to felt pressure for high in my job. (c) The opportunity for personal
quantity production, or to the actual quantity growth and development in my job. (d) The
of work which is produced. This is consistent self-esteem and self-respect a person gets from
with the notion that employees with strong being in my job.
higher order needs feel positively when they The four items least strongly related to the
have "accomplished something that they feel core dimensions are (in ascending order): (a)
is meaningful"; it is not unreasonable that The pay for my job. (b) The opportunity to
such workers would see doing high quality develop close friendships in my job. (c) The
work as much more of a meaningful "accom- opportunity for promotion, (d) The amount
plishment" than simply turning out large of respect and fair treatment I receive from
quantities of work. my boss.
General satisfaction, job involvement, and It appears, as expected, that the four core
absenteeism. The core dimensions are, as ex- dimensions seem to be most strongly related
pected, strongly and positively related to to the satisfaction of higher order needs. The
overall job satisfaction and to the degree that satisfaction items which are most weakly re-
employees feel personally involved in their lated to the four core dimensions seem to be
work. It is not, therefore, surprising to find most relevant to the satisfaction of needs
that employees whose jobs are seen as high classified as lower-order in the hierarchies of
on the core dimensions tend to have better Maslow (1943) and Alderfer (1969).
absence records as well. Work apparently is
a satisfying place to be for employees with Two Interpersonal Dimensions
jobs high on the four core dimensions, and one Also included in Table 6 are the relation-
way they behaviorally demonstrate this is by ships between two primarily interpersonal job
coming to work regularly. dimensions (dealing with others, and friend-
Specific satisfactions. Nearly all of the spe- ship opportunities) and the dependent variable
cific satisfaction items were significantly posi- measures. The nature of the relationships ob-
tively correlated with the descriptions of jobs tained are strikingly different from those ob-
on the core dimensions. This is to be expected tained for the four core dimensions. The two
from the fact that the core dimensions relate interpersonal dimensions do not relate very
substantially to overall job satisfaction—since consistently or strongly either to employee
overall satisfaction is likely to be strongly in- affective responses to the job, or to their
fluenced by satisfaction with the particular actual work performance. There is no sig-
aspects of the work situation addressed by the nificant relationship between the dimension
12 specific satisfaction items. "dealing with others" and any measure of
It is instructive, however, to note which of motivation, performance, or absenteeism.
the specific satisfaction items are especially Relationships involving the dimension
strongly related to the core dimensions, and "friendship opportunities" also are generally
which have negligible relationships. Items with negligible. The dimension does not relate to
consistently strong relationships to the core the level of intrinsic motivation employees
dimensions should be considered especially report, or to any measure of performance or
sensitive to differences on these dimensions, absenteeism.
whereas those items with weak relationships There are positive relationships (moderately
should be considered not as responsive to large for friendship opportunities, small for
differences in job design. These data should dealing with others) between the two inter-
be helpful, therefore, in more fully understand- personal dimensions and job satisfaction. The
 EMPLOYEE REACTIONS TO JOB CHARACTERISTICS 275

single largest relationship between any of the istics and the employee reactions—and not
satisfaction items and either of the inter- because they caused the perceived job char-
personal job dimensions is not a very sur- acteristics which, in turn, caused the reactions.
prising one: The more friendship opportuni- By examining the relationships between em-
ties are present on a job, the more employees ployee perceptions of their jobs and the
are satisfied with the chances they have to dependent variables separately for workers on
develop close friendships on the job. This each of the 13 jobs, it is possible to determine
relationship suggests that the kinds of con- if perceptual factors, largely uncontaminated
sequences to be expected from having jobs by objective between-job differences, bear a
with high interpersonal components (as mea- significant relationship to the dependent vari-
sured in this research) are primarily social in ables. Such an analysis, therefore, would
nature—rather than being relevant to the per- provide a relatively unconfounded test of the
formance and motivation of employees as is second link in the above diagram. Within-job
the case for jobs high on the core dimensions. relationships should, of course, be lower in
magnitude than those obtained using all em-
Within-Job Relationships between Job Dimen- ployees working on all jobs, since there is
sions and Employee Reactions little meaningful variation in actual job char-
acteristics when data are obtained only from
It was proposed earlier that job design
a single job, and as a result the variance
factors affect employee attitudes and behavior in employee perceptions should be severely
because of their impact on the perceptions
attenuated.8
employees have of their jobs. This may be
Thirteen matrices of relationships were ob-
represented schematically as follows:
tained by correlating job descriptions and
Objective job 1 Perceived job 2 Work attitudes dependent variables separately for employees
characteristics characteristics and behavior working on each of the 13 jobs. The median
of the 13 correlations between each job dimen-
Data presented in Table 2 show that em-
sion and each dependent variable was selected,
ployee perceptions of their jobs have substan-
and the matrix of median correlations is pre-
tial convergence with the assessments of ob-
sented in Table 7.
jective job characteristics made by the re-
searchers and by company supervisors. If it As expected, the median within-job correla-
is assumed that the assessments by the tions generally are consistent with the correla-
researchers and supervisors do reflect the tions computed across all employees and jobs
"objective" character of the jobs, these data (see Table 6). The order of magnitude of the
strongly suggest that employee job perceptions median within-job correlations, also as ex-
are based (at least in major part) on objective pected, is lower than that of the correlations
job characteristics—the first link in the chain. based on data from all employees. These re-
Results reported in Table 6 and discussed im- sults suggest, therefore, that employees' per-
mediately above show that employee job per- ceptions of their jobs are of central importance
ceptions also are related to work attitudes and in affecting job attitudes and behaviors, but
behavior—the second link in the chain. How- that the major determinant of such percep-
ever, these latter analyses are based on job de- tions is the objective make-up of the job
scriptions provided by employees who worked itself.
on 13 different jobs—jobs which have been 8
To the extent that there are differences in the actual
shown (see Table 4) to vary substantially on jobs assigned to employees who work on the same
the six job dimensions. These large objective objective job, the usefulness of within-job analyses
differences among the 13 jobs may, therefore, for testing the impact of strictly perceptual factors on
the dependent variables is limited. Although some
have been primarily responsible for the rela- differences in actual work assignments were noted by
tionships which were obtained. In other words, the researchers among employees working on the same
the results in Table 6 could have come about objective job, these differences invariably were of little
consequence. Therefore, the within-job analyses prob-
because the objective job characteristics di- ably do reflect primarily differences in employee job
rectly caused both the perceived job character- perceptions rather than actual differences in job design.
276 J. RICHARD HACKMAN AND EDWARD E. LAWLER III

TABLE 7
MEDIAN WITHIN-JOB CORRELATIONS BETWEEN JOB CHARACTERISTICS AND EMPLOYEE REACTIONS

Core dimension Interpersonal

dimension
Dependent variable
rp 1
Dealing Friendship
Variety Autonomy iask Feedback with opportuni-
identity others ties

Level of intrinsic motivation .31 .23 .11 .35 .03 .06

Focus of motivation
Taking personal responsibility .00 .00 .18 .09 .15 .00
Doing large quantities of work -.05 -.06 .10 .01 .07 -.07
Doing high quality work .05 .22 .19 .05 .10 .06
Rated performance
Quantity .06 .21 -.10 -.05 -.02 -.18
Quality .06 .18 .06 .03 -.10 -.05
Overall effectiveness .12 .21 .06 .04 .00 -.11
General job satisfaction .38 .48 .21 .39 .22 .21
Job involvement .34 .28 .20 .19 .22 .04
Absenteeism (no. of times absent) -.10 -.03 .03 .00 .17 .14
Specific satisfaction items
Self-esteem obtained from job .26 .31 .15 .44 .18 .27
Personal growth and development .57 .40 .18 .34 .25 .21
Prestige of job inside company .44 .44 .17 .42 .24 .26
Amount of close supervision received .19 .39 .08 .31 .17 .07
Independent thought and action .40 .55 .31 .30 .13 .20
Security .21 .28 .28 .35 .11 .25
Pay .24 .33 .17 .33 .19 .18
Feeling of worthwhile accomplishment .40 .37 .41 .48 .25 .26
Participation in job-related decisions .34 .34 .39 .31 .31 .25
Development of close friendships .27 .07 .25 .35 .26 .51
Promotion .19 .33 .13 .32 .26 .06
Respect and fair treatment from boss .20 .29 .24 .35 .09 .30

Note.—Because the correlations reported in this table are medians, no indication of statistical significance is given.

Exploration of Condition for Enhancing Motiva- core dimensions, employees have the oppor-
tion and Satisfaction tunity to find out (feedback) that they
personally (autonomy) have accomplished
Results presented in the preceding discussion something meaningful (task identity and va-
indicate that when jobs are high on the core riety) when they perform well. The implica-
dimensions, employees who have strong desires tion of this assertation is that, for maximum
for higher order need satisfaction will be highly motivation, jobs should be high simultane-
motivated and well satisfied on the job. In ously on all four of the core dimensions
the following discussion, two analyses will be
reported which extend these findings and pro- To test this possibility, ,Ss were partitioned
vide additional documentation of the condi- into three groups: (a) those who described
tions which seem necessary to enhance em- their jobs as being above the 60th percentile
ployee motivation and satisfaction through job on all four core dimensions; (b) those who
design. described their jobs as being below the 40th
percentile on all four core dimensions; and (c)
the majority of 5s, who typically described
Must Jobs Be High on All Four Core their jobs as being high on some of the core
Dimensions'? dimensions and low on others. A one-way,
The theory on which the present research is missing-data, unequal N analysis of variance
based specifies that when jobs are high on the was used to compare the mean scores of em-
 EMPLOYEE REACTIONS TO JOB CHARACTERISTICS 277
TABLE 8
COMPARISON or JOBS DESCRIBED AS HIGH ON ALL CORE DIMENSIONS, HIGH ON SOME CORE
DIMENSIONS, AND Low ON ALL CORE DIMENSIONS

Dependent variable X

Dependent variable Jobs low on Jobs high on Jobs high on

F ratio all core some core all core
dimensions0 dimensions1" dimensions"

Level of intrinsic motivation 13.72** 5.28 S.99 6.46

Focus of motivation
Taking personal responsibility 6.08** 5.90 6.40 6.88
Doing large quantities of work .24 5.48 5.17 5.29
Doing high quality work 3.04* 6.00 6.20 6.71
Rated performance
Quantity .15 5.19 5.23 5.39
Quality 3.36* 4.97 5.57 5.89
Overall effectiveness 3.40* 5.14 5.54 5.94
General job satisfaction 28.97** 3.78 4.99 6.21
Job involvement 18.40** 2.30 2.94 4.43
Absenteeism (no. of times absent) 1.62 3.11 2.70 2.00
Specific satisfaction items
Self-esteem obtained from job 27.06** 3.33 4.30 6.00
Personal growth and development 24.45** 3.43 4.27 6.00
Prestige of job inside company 21.65** 3.24 3.88 5.59
Amount of close supervision received 15.97** 3.57 4.44 5.65
Independent thought and action 31.66** 3.24 4.62 6.35
Security 9.75** 4.52 5.32 6.24
Pay 2.95 3.38 3.63 4.47
Feeling of worthwhile accomplishment 35.59** 3.05 4.88 6.41
Participation in job-related decisions 21.09** 2.81 3.85 5.29
Development of close friendships 9.56** 4.14 4.77 5.65
Promotion 13.98** 3.24 3.92 5.47
Respect and fair treatment from boss 14.64** 4.14 5.40 6.24

"N = 21.
1>2V = 170.
° JV = 17.
*p < .05 (one-tailed test).
** p < .01 (one-tailed test).

ployees in the three groups on each of the and involvement would be high, and (e)
dependent variables. Means and F ratios are absenteeism would be low. All of these ex-
presented in Table 8. pectations were borne out substantially and
Results are consistent with those reported statistically significantly, except for the one
in Table 6 and, in addition, tend to be more involving absenteeism. While absenteeism was
substantial in magnitude and more statisti- lowest when jobs were seen as high on all
cally reliable.9 It was expected that when jobs four core dimensions, differences among the
were described as high on all four core dimen- three groups were not statistically significant.
sions: (a) experienced pressures to take per- Although these results support the conten-
sonal responsibility for one's work and to do tion that motivation, satisfaction, and per-
high quality work would be high, (b) intrinsic formance are enhanced when jobs are high
motivation would be high, (c) rated perfor- simultaneously on all four core dimensions,
mance quality and overall performance effec- they do not imply that a disjunctive model
tiveness ' would be high, (d) job satisfaction for combining the core dimensions is neces-
'The analytic procedures used, unfortunately, do
sarily the optimal one. Therefore, the relative
not permit computation of the amount of variation usefulness of other ways of combining the
controlled, for comparison with the correlational data dimensions was explored. In particular, (a) an
reported earlier. unweighted sum of the four dimensions was
278 J. RICHARD HACKMAN AND EDWARD E. LAWLER III

computed for each S and correlated with each for those employees whose higher order need
of the dependent variables, (5) the four dimen- strength scores were in the top one-third of the
sions were used as independent variables in a distribution of scores for all 5s, and for those
multiple-regression analysis predicting each of employees whose scores were in the bottom
the dependent variables, and (c) the product one-third of the same distribution. The mean
of the four dimensions (Variety X Autonomy higher order need strength score for the em-
X Task Identity X Feedback) was correlated ployees in the top third was 6.78, and the
with each of the dependent variables. The mean score for employees in the bottom third
latter score is a continuous measure which is was 5.09. Since 5.09 is still one full point above
congruent with the disjunctive combinatory the midpoint of the scale, it was expected that
model specified by the theory. That is, for a even for these "low" need strength employees,
product score to be high, all of the four core relationships between the core dimensions and
dimensions must be moderately high, and, the dependent variables would be in the posi-
therefore, the measure reflects the degree to tive direction, but lower in magnitude than
which a job is high simultaneously on all four the relationships for 5s in the top third of the
core dimensions. Correlations involving the distribution. It also was expected that the
product score were compared to those ob- relationship found for the top third should be
tained using the other two combinatory models. stronger than those reported in Table 6 for all
Correlations between the unweighted sum 5s. Results are presented in Table 9.
score and the dependent variables ranged from For variety, autonomy, and, to a lesser
— .08 to .61, with a median of .30. Multiple extent, feedback, the expectation of a differ-
correlations ranged from .13 to .65, with a ential relationship was confirmed. Relation-
median of .37. Correlations using the product ships between these job dimensions and the
score ranged from — .05 to .60, and also had a dependent variables are larger, and often
median of .30. Since some shrinkage would be strikingly so, for the 5s with high higher order
expected in the magnitude of the multiple need strength than for 5s with moderate
correlations obtained upon cross-validation, it higher order need strength. Further, the cor-
seems appropriate to conclude that the results relations for higher need strength employees
did not meaningfully differentiate among the are typically larger than those reported in the
three means of combining the core dimensions. general analysis involving all 5s (see Table 6),
Thus, while the theory-specified disjunctive and the correlations based on the lower need
model is not disconfirmed by the data, neither strength 5s are typically lower than the gen-
has it been shown to represent a more adequate eral relationships.
means of combining the dimensions than the The moderating effect of higher order need
other models tested. strength was not obtained for task identity.
Part of the reason for this may have to do with
Does High Order Need Strength Moderate the nature of the particular jobs included in
Obtained Relationships'? this study. Among the jobs in this study
It has been maintained throughout this which were high on task identity were those
article that jobs which are high on the core of telephone operators. By and large, when
dimensions should be motivating only to indi- an operator takes a toll call she does a "whole"
viduals who are desirous of the intrinsic re- piece of work. She is involved in the call from
wards that the jobs provide, namely, higher the time the customer comes on the line until
order need satisfactions. the called party answers and confirms his
The analysis reported in this section tests identity. The operator will do and redo this
the possibility that 5s differing in higher order task and similar tasks many times during a
need strength do in fact show differential typical working day. Observations by the re-
responsiveness to jobs high on the core dimen- searchers strongly suggest that an operator
sions, despite the generally high mean of all with strong higher order need strength quickly
,5s on the need strength measure. Correlations becomes bored and impatient with so many
between each of the core dimensions and the repetitions of this small but "whole" piece of
dependent variables were computed separately work. If this is so, then it should not be sur-
 EMPLOYEE REACTIONS TO JOB CHARACTERISTICS 279
TABLE 9
MODERATING EFFECTS OF HIGHER ORDER NEED STRENGTH

Variety Autonomy Task identity Feedback

Dependent variable
High Low High Low High Low High Low
strength strength strength strength strength strength strength strength
Level of intrinsic motivation .56* .15 .49* .07 .19 .22* .41* .25*
Focus of motivation
Taking personal responsibility .41* .00 .39* .07 .30* .23* .21* -.02
Doing large quantities of work -.07 -.20 -.19 -.06 -.18 .03 .10 .09
Doing high quality work .17 .19 .18 .03 .07 .01 .21* .09
Rated performance
Quantity -.02 -.24 .09 -.09 .01 .15 .04 .23*
Quality .15 .05 .17 -.05 -.02 .31* .12 .03
Overall effectiveness .09 .05 .16 -.01 .01 .26* -.04 .06
General job satisfaction .41* .28* .43* .29* .27* .18 .17 .33*
Job involvement .30* .17 .35* .19 .16 .26* .30* .15
Absenteeism (no. of times absent) -.29* -.18 -.24* .01 .21 -.26* .04 .06
Specific satisfaction items
Self-esteem obtained from job .48* .06 .39* .07 .38* .12 .34* .29*
Personal growth and development .55* .03 .47* -.01 .34* .04 .39* .33*
Prestige of job inside company .41* .17 .32* .07 .32* .18 .41* .24*
Amount of close supervision received .33* .15 .43* .10 .06 .07 .45* .28*
Independent thought and action .62* .42* .75* .45* .28* .29* .37* -.01
Security .29* .04 .38* .08 .20* .24* .41* .33*
Pay .24* -.10 .15 .09 .07 .03 .26* .36*
Feeling of worthwhile accomplishment .53* .06 .49* .15 .38* .33* .42* .40*
Participation in job-related decisions .41* .20* .32* .18 .15 .22* .41* .25*
Development of close friendships .56* -.04 .40* -.35* .38* -.03 .36* .13
Promotion .29* -.09 .35* -.10 .14 .10 .40* .34*
Respect and fair treatment from boss .39* .06 .45* .03 .21* .22* .38* .38*

Note.—IV = 67 in each group.

* p < .05 (one-tailed test).

prising to find that there is little difference in strength than for 5s low in need strength for
the relationship between task identity and the nearly all of the measures of motivation, satis-
dependent variables for employees with high faction, and performance. The difference be-
versus low higher order need strength.
Indeed, it would be expected from the theory TABLE 10
that differences between high and low need MODERATING EFFECT OF HIGHER ORDER NEED
strength employees should be maximized when STRENGTH FOR JOBS HIGH ON ALL
jobs are moderately high on all four core FOUR CORE DIMENSIONS
dimensions. The job of the toll operator
High Low
clearly does not meet this criterion: while it Dependent variable need need
strength strength
is reasonably high on task identity and feed-
back, it is very low on both variety and Level of intrinsic motivation .54* .23*
Focus of motivation
autonomy (see Table 4). Taking personal responsibility
Doing large quantities of work
.37*
-.09
.12
-.01
To examine the validity of this argu- Doing high quality work .21* .12
Rated performance
ment, therefore, the product score (Variety X Quantity .07 -.04
Quality .23* .02
Autonomy X Task Identity X Feedback) de- Overall effectiveness .15 .05
General job satisfaction .48* .40*
scribed earlier was correlated with the depen- Job involvement .45* .28*
dent variables separately for the third of ,5s Absenteeism (no. of times absent) -.26* -.08
Specific satisfaction items
highest in higher order need strength and for Self-esteem obtained from job .54*
.57*
.13"
.16"
Personal growth and development
the third of 5s lowest in higher order need Prestige of job inside company .50* .26*
Amount of close supervision received .48* .27*
strength. Since the product score reflects the Independent thought and action ,70* .45*°"
Security .49* .25*»
degree to which a job is seen as being simul- Pay .27* .09
taneously high on all four core dimensions, Feeling of worthwhile accomplishment .59* .32*"
Participation in job-related decisions .44* .37*
substantial differences would be expected in Development of close friendships
Promotion
.54*
.45*
-.19"
,08>
the magnitude of correlations obtained for Respect and fair treatment from boss .48* .20*"
high versus low need strength employees. Re-
Note.—N = 67 in each group.
sults are presented in Table 10. • The difference between the correlations for high need
strength 5s and for low need strength 5s is significant at p
Again, obtained relationships were much <.05.
higher for 5s who were high in relevant need *p< .05 (one-tailed test).
280 J. RICHARD HACKMAN AND EDWARD E. LAWLER III

tween the correlations for high and low need In both of these studies a sociological-level
strength 5s was tested for statistical signifi- variable (urban vs. rural background) was
cance, and these results are included in the shown to moderate the relationship between
table. Even though most of the correlations job characteristics and employee satisfaction.
were positive (because of the relatively high The relationship between job level (i.e., jobs
mean need strength for even the "low" need which would be high on the four core dimen-
strength employees), a number of pairs of cor- sions) and satisfaction was found to be high
relations were significantly different from one for rural workers and low for urban workers.
another. All in all, the data make a strong case The present study demonstrates that individ-
for the moderating effect of individual higher ual higher order need strength also moderates
order need strength in determining the effects the relationship between job level and satis-
of job characteristics on employee behavior faction (as well as the relationship between
and attitudes at work. job level and other dependent variables). In
particular, individuals with strong desires for
DISCUSSION AND IMPLICATIONS higher order need satisfaction respond much
more positively to high level jobs than do
Job Design and Individual Differences individuals who have weaker higher order
The results of this study suggest that there needs.
are important interdependencies among the It may be that urban workers desire higher
characteristics of individuals and the char- order need satisfaction less than do rural
acteristics of jobs which must be taken ac- workers—in which case the findings of the
count of in the development of any full under- earlier studies and the present study would be
standing of the impact of various kinds of job highly congruent. This possibility was tested
designs. Both the advocates of a "scientific with the present data and a small difference
management" approach to job design (make in the expected direction was in fact found:
the work routine, simple, and standardized) workers with rural backgrounds were higher
and the more recent supporters of "job en- on higher order need strength than were
largement" (make the work complex, challeng- workers with urban backgrounds (t — 1.47,
ing, and demanding of individual responsibility p < .10). Thus, the present research has taken
and decision making) appear to have attached the previous finding regarding urban-rural dif-
insufficient importance to individual-job inter- ferences from the sociological level to the
actions in determining affective and behavioral psychological level, refined it, and successfully
reactions to jobs. Those of the scientific man- replicated it. The present results also sub-
agement persuasion, for example, have tended stantially extend the previous results in that
to assume that the typical employee will be the moderating effect of individual differences
content, if paid judiciously for his cooperation, has been shown to apply to a number of
to work on jobs which provide little or no dependent variables in addition to job satis-
opportunity for personal feelings of accom- faction (i.e., job involvement, work motiva-
plishment or achievement. Those of the job tion, quality of performance). Presumably it
enlargement school, on the other hand, have now should be possible to make considerably
tended to assume that most employees are more specific predictions about how different
desirous of such opportunities and will work individuals will respond to jobs with various
hard and effectively when they have a job characteristics.
which provides them. The present research It has been assumed throughout the above
suggests that, depending on the character- discussion that job characteristics actually
istics of the workers involved, both points of cause the differences in employee satisfaction,
view would lead to job design practices which motivation, performance, and absenteeism
are appropriate some of the time and inappro- which were observed. Although the predic-
priate other times. tions which were made (and confirmed by
The present findings and conclusions fit the data) were based on a conceptual frame-
well with the previous research of Turner and work which includes causal propositions, the
Lawrence (1965) and Hulin and Blood (1968). study design was correlational and at no point
 EMPLOYEE REACTIONS TO JOB CHARACTERISTICS 281

were the causal links in the theory directly the important conceptual and methodological
tested. There is, however, reason to believe advantages associated with the use of inde-
that the causal directions specified in the pendently and objectively described indepen-
theory may be correct. While the simple cor- dent variables are lost.10
relations between job characteristics and the There are at least two strategies for dealing
dependent variables (see Table 6) are open to with this problem in research on task and job
alternative causal interpretations, this appears effects on behavior. One, which was used in
to be much less the case for those data which the present research, is to employ subjective
support other, more specific, predictions of the assessments of the tasks or jobs by the per-
theory. In particular, it is difficult even after formers themselves, but simultaneously to
the fact to imagine an alternative causal model develop means of determining the relationship
which would predict simultaneously that (a) between these assessments and others, includ-
all four of the core dimensions need to be ing objective measures when possible. Thus,
present to realize the most positive affective in addition to employee descriptions of the
and behavioral reactions on the part of em- characteristics of their jobs in the present
ployees, and that (b) the strength of desire for study, jobs were rated by three other methods:
higher order need satisfaction will substan- by company managers, by the researchers, and
tially moderate the relationships between the by an adaptation of the Turner and Lawrence
core dimensions and the dependent variables. (1965) procedures for describing jobs in opera-
It appears, therefore, that a cautious interpre- tional terms. As was shown in Table 2, there
tation of the results in terms of the causal was high convergence among the four methods
impact of job characteristics may be reason- for all job dimensions except feedback. Thus,
able. This does not deny the fact that any while the job descriptions used in the analyses
theory can best be tested by experimental are those actually experienced by the em-
alteration of the independent variables, and ployees (and thus presumably are more di-
manipulation of the core dimensions in an ex- rectly causal of their behavior than would be
perimental setting would seem to be a clear the case for objectively described character-
next step in research on job effects. istics), they also have been shown to have
substantial convergent validity and to be
Use of Perceived Job Characteristics based upon the objective job characteristics.
One of the major conceptual and methodo- A second strategy for dealing with the re-
logical problems which pervades studies of definition problem has been proposed by
task and job effects on behavior has to do Hackman (1970) and, while methodologically
with the differences between task materials as more difficult to implement, probably has more
they exist in objective reality and as they are potential for the development of a general
perceived by individual performers. Tasks and understanding of the effects of tasks and jobs.
jobs are invariably redefined by the individuals In essence, this strategy suggests that the
who perform them, sometimes deliberately and redefinition process should be viewed as the
sometimes without full awareness by the per- first stage of the performance process itself,
formers of the changes or re-emphases that and the redefined task should be conceived
are being made. Further, it is the redefined of as a potentially measurable intervening
task rather than the objective task which the variable in the causal chain between the ob-
individual tries to perform, and thus only jective task input and the dependent variables
those aspects of tasks or jobs which are of interest (e.g., performance, satisfaction).
actually perceived or experienced by a per- Just as individual differences in need strength
former can have an impact on his performance have been shown in this research to moderate
and attitudes. This would seem to argue for the effects of job characteristics on employee
the use of task and job characteristics as behavior and attitudes, so would individual
described by the performers themselves in re- needs, values, and goals be expected to inter-
search aimed at ascertaining the effects of 10
The conceptual and methodological implications of
these characteristics on performance. Yet when the task redefinition phenomenon are discussed in more
such subjective assessments are used, many of detail elsewhere (Hackman, 1969b; Weick, 1965).
282 J. RICHARD HACKMAN AND EDWARD E. LAWLER III

act with the objective task or job in influencing task identity present. Finally, it may be that
task redefinition. It might be, for example, increases in variety can serve two different
that individuals with a particular pattern of functions, depending on the amount initially
needs tend to redefine the tasks they perform present on a given job. For low initial levels
to be more consistent with those needs than of variety an increase may serve mainly to
is actually the case. These individuals might decrease the monotony of the work and make
then develop hypotheses about appropriate it possible for perceived meaningfulness to
on-the-job behavior which they expect to lead vary as a function of the amount of task
to need-satisfying outcomes, but which, be- identity present, as suggested above. At some
cause of the lack of congruence between the point, however, variety may assume a different
objective and the redefined task, might in fact function, namely, introducing challenge into
serve only to set the stage for disillusionment the work which, when successfully dealt with,
and frustration. In any case, it appears that can be satisfying to many individuals. Clearly,
research on the ways different individuals additional research on the psychological impact
redefine tasks and jobs may have considerable of these dimensions is called for.
potential for furthering general understanding Additional work also needs to be done on
of the effects of tasks and jobs on the behavior the operationalization of the feedback dimen-
and attitudes of performers. sion, since no convergence was obtained among
the several methods for measuring this dimen-
Nature of the Four Core Dimensions sion. The fact that the dimension operated as
The results of the present research show predicted in the data analyses indicates that
that, in general, employees with moderately the employee judgments of the dimension
high desires for higher order need satisfaction have some validity, but because the various
tend to work harder and be more satisfied groups who rated the jobs did not agree on
when they perceive their jobs as being rela- the amount of feedback present, the exact
tively high on the four core dimensions. In meaning of the dimension remains highly
addition, it was shown that for the most uncertain.
favorable outcomes, jobs need to be at least Finally, the generality of the moderately
moderately high on all four of the dimensions. high correlation (r = .67) between variety and
Further refinement of the core dimensions autonomy should be tested. The descriptions
and exploration of their impact on individuals of the scales used to measure these two dimen-
clearly is called for by the research. For ex- sions were quite dissimilar, and it may be that
ample, the relationship between variety and the high correlation was obtained simply be-
task identity needs further examination. In cause jobs which had high variety in the par-
one sense, since both of these dimensions are ticular organization where the study took
viewed as enhancing the meaningfulness of place also tended to be jobs which provided
the work (although by different means), it considerable autonomy. The fact that most
might be argued that it is not necessary for telephone traffic jobs (i.e., various kinds of
both to be present (i.e., that one can effec- operators) were low on both variety and au-
tively substitute for the other). The finding of tonomy would certainly facilitate obtaining a
Hall and Lawler (1970), that jobs which were positive correlation between the dimensions,
very high on variety and low in task identity and might suggest that the high correlation is
are not associated with high quality perfor- to some extent organization specific. There is
mance among researchers, would tend to argue reason, however, to believe that the positive
against this possibility, however. An alterna- relationship between variety and autonomy
tive possibility which would not be contrary is at least moderately general. Turner and
to the Hall and Lawler findings is that some Lawrence (1965) found an analogously high
moderate level of variety is essential simply relationship between these two dimensions
to keep the employee from being bored with across 47 different jobs, and Alderfer (1967)
his work (Scott, 1966), and that once this is obtained a very high correlation between
achieved, experienced meaningfulness of the variety and "decision time" (which has some
work will vary directly with the amount of conceptual similarity to autonomy) across a
 EMPLOYEE REACTIONS TO JOB CHARACTERISTICS 283

sample of 30 jobs which included management They do affect satisfaction and it is also pos-
jobs. It appears well worth examining the sible that they affect other variables which
possibility that the psychological meaning of were not considered in this study (e.g., prob-
the two dimensions is similar in ways that lem-solving effectiveness, turnover). It should
have not been attended to in the present also be noted that all of the jobs in the present
concep tualization. study were at least moderately high on the
interpersonal dimensions; it is very likely that
Interpersonal Components of Jobs important negative reactions might be ob-
The two job dimensions reflective of the tained from employees on jobs which were
interpersonal components of jobs (dealing with very low or totally lacking in interpersonal
aspects. Further, the present data address
others, and friendship opportunities) did not
relate to employee work motivation or perfor- only the degree to which the existing structure
mance. The dimensions did relate positively of jobs requires or provides opportunities for
to certain kinds of satisfaction, but the rela- interpersonal activities. Alderfer (1967) has
tionships were not as substantial as those in- shown that when jobs are changed, interper-
volving the four core dimensions. sonal relationships (especially between the
According to the conceptualization on which worker and his supervisor) are markedly af-
the present study is based, the degree to which fected, and that "relationship problems" clearly
have the potential of negating or reversing
jobs require interpersonal activities should re-
late to work motivation only when (a) workers increases in motivation and satisfaction which
have high desires for the satisfaction of social are anticipated as a result of job enlargement.
needs, and (b) working hard on the job can Nature of the Impact of Jobs on Behavior
lead to the satisfaction of these needs. Even
those jobs which scored relatively high on the Until relatively recently, research attention
two interpersonal dimensions (e.g., some op- given to jobs has focused mainly on the ways
erators jobs) fail to meet the latter criterion. jobs can be designed for most efficient pro-
Operators reported in fact that they could duction, or on the analysis of jobs into com-
obtain social satisfactions best when they were ponent parts to facilitate employee selection,
not "working hard" by company standards placement, and compensation. Since the mid-
(i.e., completing a large number of calls), 1950s, however, research attention gradually
since when the load of calls was heavy they has shifted toward examination of the effects
had little or no time for meaningful inter- jobs can have on the people who do them,
personal activities with either customers or and how jobs can be designed so that these
fellow employees. effects are desirable for both employees and
Even if jobs were designed so that relating organizations.
meaningfully to others was an integral part Many studies, including this one, have shown
of doing the job well, there is reason to doubt that jobs can and often do have a substantial
whether such jobs would have long-term mo- impact on employee behavior and attitudes;
tivational payoffs. The reason for skepticism this finding is by now rather well documented.
is the arguments of Maslow (1953) and of It appears that additional understanding of
Alderfer (1971) (presented earlier in this the nature and implications of job effects will
article) that when individuals have had ample require more direct investigation of the process
opportunities to satisfy their social needs, the by which these effects take place.
level of desire for additional social satisfaction The present study represents some progress
will decrease and the level of desire for higher in this direction, by providing data which
order need satisfaction will increase. When show that certain job characteristics and
this occurs, of course, then jobs should be certain individual differences interact in deter-
high on the four core dimensions (rather than mining behavior and satisfaction. Apparently,
high on an interpersonal dimension) to pro- individuals who are desirous of higher order
vide conditions for internal work motivation. need satisfactions will, when working on a job
All of the above is not to suggest that the high on the core dimensions, gradually develop
interpersonal aspects of work are unimportant. and/or verify the hypothesis that personally
284 J. RICHARD HACKMAN AND EDWARD E. LAWLER III

valued rewards can be obtained by working tribute most effectively to organizational goals
hard and effectively on the job. When this (and simultaneously to satisfy their own needs)
occurs, the job can be said to have influenced if they are placed on jobs which are high on
the behavior of these employees by affecting the four core dimensions. Other employees,
their personal hypotheses about what kinds of course, who may be neither desirous of
of behaviors will lead to favorable outcomes. higher order need satisfactions nor capable of
Thus, one general way that jobs can influence dealing with complex jobs requiring consider-
behavior is by creating conditions whereby able autonomy, would be ineffective on such
employees are likely to develop and validate jobs and dissatisfied with them.
specific behavioral hypotheses about what they It appears from research cited earlier in this
can do on the job to obtain work outcomes article that many organizations err rather con-
favorable to themselves (or to avoid unfavor- sistently by designing jobs which are too low
able outcomes). on the core dimensions. The present study
The general process outlined briefly above supports this conclusion. It suggests that there
represents only one of several possible ways are many workers who want to obtain more
that jobs can influence behavior and attitudes higher order need satisfactions from their work,
on the job. For example, jobs also can affect but few who are overwhelmed by the psy-
behavior by influencing the level of activation chological demands of their jobs. The implica-
of employees at work (Scott, 1966), or by tion of this argument, of course, is that orga-
arousing (or depressing) particular employee nizations might be well advised to consider
need states themselves. The general question redesigning many of their jobs.
of how jobs and tasks influence behavior is When job enlargement is carried out, the
discussed in some detail by Hackman (1969a; question often arises whether the changes
1970). The point of emphasis here is that job should be toward horizontal enlargement (i.e.,
effects can and do come about via a number increasing the number of different things an
of different (albeit interrelated) psychological employee does) or toward vertical enlarge-
processes, only one of which was given focal ment (i.e., increasing the degree to which an
attention in the present research. It is likely employee is responsible for making most major
that sometimes these processes will operate decisions about his work) or both. Lawler
simultaneously and will be mutually reinforc- (1969) has reviewed the literature regarding
ing; other times they may work at cross the effects of vertical and horizontal job en-
purposes. Further work toward a more gen- largement and concludes that simultaneous
eral understanding of the several processes by enlargement in both directions may be op-
which jobs affect behavior and how they timal in most cases. The results of the present
interact with individual differences would seem study provide some support for this contention.
to be warranted. Only if a job is enlarged vertically is an em-
ployee likely to feel personally responsible for
Implications for Organizational Practice his work outcomes; and only if a job has some
amount of horizontal enlargement is he likely
Standard organizational selection and place-
to experience his work as meaningful—although
ment procedures attempt to match the skills
it should be kept in mind that too much hori-
and abilities of a prospective employee with
zontal enlargement apparently can cause prob-
the skill requirements of the job for which he
lems. Simultaneous vertical and horizontal en-
is being considered. The results of the present
largement should increase the likelihood that
research suggest that it may be equally critical
a redesigned job will be high on all four of the
for long-term organizational effectiveness to core dimensions.
achieve a match between the psychological It should be re-emphasized in conclusion,
makeup of the prospective employee and the however, that while jobs appear to be highly
psychological demands and opportunities of potent in determining employee motivation
the job. In particular, the present results and satisfaction, there is no single best way to
suggest that individuals who desire higher design a job. Instead, the results of the present
order need satisfaction will be likely to con- research suggest that the substantial motiva-
 EMPLOYEE REACTIONS TO JOB CHARACTERISTICS 285

tional potential of jobs can be realized only psychological factors in stress. New York: Holt,
when the psychological demands and oppor- Rinehart & Winston, 1970.
HALL, D. T., & LAWLER, E. E., Ill Job design and job
tunities of jobs mesh well with the personal pressures as facilitators of professional-organization
needs and goals of employees who work on integration. Administrative Science Quarterly, 1970,
them. This kind of matching can be developed IS, 271-281.
through selection and placement of employees, HERZBERG, F. Work and the nature of man. Cleveland:
through job redesign, or (perhaps optimally) World, 1966.
HERZBERG, F., MAUSNER, B., & SNYDERMAN, B. The
by attempting to fit people to jobs and jobs motivation to work. New York: Wiley, 1959.
to people simultaneously and continuously as HINTON, B. L. An empirical investigation of the
both the organization and the characteristics Herzberg methodology and two-factor theory. Orga-
of its employees change over time. nizational Behavior and Human Performance, 1968,
3, 286-309.
HULIN, C. L., & BLOOD, M. R. Job enlargement,
REFERENCES individual differences, and worker responses. Psycho-
logical Bulletin, 1968, 69, 41-55.
ALDERFER, C. P. An organizational syndrome. Admini- KILBRIDGE, M. D. Reduced costs through job enlarge-
strative Science Quarterly, 1967, 12, 440-460. ment : A case. Journal of Business of the University
ALDERFER, C. P. An empirical test of a new theory of of Chicago, 1960, 33, 357-362.
human needs. Organizational Behavior and Human KILBRIDGE, M. D. Turnover, absence, and transfer
Performance, 1969, 4, 142-175. rates as indicators of employee dissatisfaction with
ALDERFER, C. P. Human needs in organizational settings. repetitive work. Industrial and Labor Relations Re-
\ New York: The Free Press of Glencoe, 1971. view, 1961, 15, 21-32.
ARGYRIS, C. Integrating the individual and the organiza- KING, N. A clarification and evaluation of the two-
tion. New York: Wiley, 1964. factor theory of job satisfaction. Psychological Bul-
BIGANNE, J. F., & STEWART, P. A. Job enlargement: A letin, 1970, 74, 18-31.
case study. (Research Series No. 25) State University LAWLER, E. E., III. Job design and employee motiva-
of Iowa, Bureau of Labor and Management, 1963. tion. Personnel Psychology, 1969, 22, 426-435.
BLAUNER, R. Alienation and freedom. Chicago: Uni- LAWLER, E. E., Ill, & HALL, D. T. The relationship
versity of Chicago Press, 1964. of job characteristics to job involvement, satisfaction
BLOOD, M. R., & HULIN, C. L. Alienation, environ- and intrinsic motivation. Journal of Applied Psy-
mental characteristics, and worker responses. Journal chology, 1970, 54, 305-312.
of Applied Psychology, 1967, 51, 284-290. LAWLER, E. E., Ill, & PORTER, L. W. Antecedent
CONANT, E. H., & KILBRIDGE, M. D. An interdis- attitudes of effective managerial performance. Orga-
ciplinary analysis of job enlargement: Technology, nizational Behavior and Human Performance, 1967,
costs, and behavioral implications. Industrial and 2, 122-142.
Labor Relations Review, 1965, 3, 377-395. LEWIN, K. The conceptual representation of the measure-
DAVIS, L. E. Job design and productivity: A new ment of psychological forces. Durham: Duke Uni-
approach. Personnel, 1957, 33, 418-429. versity Press, 1938.
DAVIS, L. E., & VALFER, E. S. Intervening responses to
LEWIN, K., DEMBO, T., FESTINGER, L., & SEARS, P.
changes in supervisor job designs. Occupational
Level of aspiration. In J. McV. Hunt (Ed.), Person-
Psychology, 1965, 39, 171-189.
DUNNETTE, M. D., CAMPBELL, J. P., & HAKEL, M. D. ality and the behavior disorders. New York: Ronald
Press, 1944.
Factors contributing to job satisfaction and job
dissatisfaction in six occupational groups. Organiza- LOCKE, E. A. What is job satisfaction? Organizational
tional Behavior and Human Performance, 1967, 2, Behavior and Human Performance, 1969, 4, 309-336.
143-174. LODAHL, T. M., & KEJNER, M. The definition and
FORD, R. N. Motivation through the work itself. New measurement of job involvement. Journal of Applied
York: American Management Association, 1969. Psychology, 1965, 49, 24-33.
FRIEDMANN, G. The anatomy of work. New York: The LYTLE, C. W. Job evaluation methods. New York:
Free Press of Glencoe, 1961. Ronald Press, 1946.
GHISELLI, E. E., & BROWN, C. W. Personnel and MACKINNEY, A. C., WERNIMONT, P. F., & GALITZ, W.
industrial psychology. New York: McGraw-Hill, 1955. O. Has specialization reduced job satisfaction?
GUEST, R. H. Men and machines: An assembly-line Personnel, 1962, 39, 8-17.
worker looks at his job. Personnel, 1955, 31, 3-10. MASLOW, A. H. A theory of human motivation. Psy-
HACKMAN, J. R. Nature of the task as a determiner of chological Review, 1943, 50, 370-396.
job behavior. Personnel Psychology, 1969,22,435-444. MASLOW, A. H. Motivation and personality. New York:
(a) Harper, 1954.
HACKMAN, J. R. Toward understanding the role of MCGREGOR, D. The human side of enterprise. New
tasks in behavioral research. A eta Psychologica, 1969, York: McGraw-Hill, 1960.
31, 97-128. (b) PELISSIER, R. F. Successful experience with job design.
HACKMAN, J. R. Tasks and task performance in re- Personnel Administration, 1965, 28, 12-16.
search on stress. In J. E. McGrath (Ed.), Social and PORTER, L. W., & LAWLER, E. E., III. Managerial
286 J. RICHARD HACKMAN AND EDWARD E. LAWLER III

attitudes and performance. Homewood, 111.: Irwin, VROOM, V. H. Work and motivation. New York :Wiley,
1968. 1964.
SCOTT, W. E. Activation theory and task design. WALKER, C. R. The problem of the repetitive job.
Organizational Behavior and Human Performance, Harvard Business Review, 1950, 28, 54-58.
1966, 1, 3-30.
WALKER, C. R., & GUEST, R. H. The man on the as-
STIGERS, M. F., & REED, E. G. The theory and practice
of job rating. New York: McGraw-Hill, 1944. sembly line. Cambridge: Harvard University Press,
TOLMAN, E. C. Principles of purposive behavior. In S. 1952.
Koch (Ed.), Psychology: A study of a science. Vol. 2. WEICK, K. E. Laboratory experimentation with orga-
New York: McGraw-Hill, 1959. nizations. In J. G. March (Ed.), Handbook of orga-
TURNER, A. N., & LAWRENCE, P. R. Industrial jobs and nizations. Chicago: Rand-McNally, 1965.
the worker. Boston: Harvard University Graduate
School of Business Administration, 1965. (Received March 30, 1970)