Jourrl al of Applied Psychology Copyright 1989 by the American Psychological Association, Inc.

1989, Vol. 74, No. I, 152-156 0021-9010/89/$00.75

Organizational Commitment and Job Performance

It's the Nature of the Commitment That Counts
John P. Meyer, Sampo V. Paunonen, Ian R. Gellatly, Richard D. Goffin, and Douglas N. Jackson.
University of Western Ontario, London, Ontario, Canada

In this study, we examined relations between the performance of first-level managers in a large food
service company and their affective commitment (i.e., emotional attachment to, identification with,
and involvement in the organization), continuance commitment (i.e., perceived costs associated with
leaving the company), and job satisfaction. Commitment and satisfaction scores were correlated
with three indexes of performance obtained from the managers' immediate supervisors. As pre-
dicted, affective commitment correlated positively and continuance commitment correlated nega-
tively with all three measures of performance. Job satisfaction did not correlate significantly with
performance ratings. The findings are interpreted as illustrating the importance of distinguishing
between commitment based on desire and commitment based on need and as supporting organiza-
tional efforts to foster affective commitment in their employees.

Interest in organizational commitment has been stimulated ing) might include the loss of attractive benefits and seniority,
largely by its demonstrated negative relation to turnover: Com- the disruption of personal relations created by moving to an-
mitted employees have been found to be less likely to leave an other location, the effort of seeking a new job, and so on.
organization than those who are uncommitted (e.g., Angle & Meyer and Allen (1984; Allen & Meyer, 1987) used the terms
Perry, 1981; Porter, Steers, Mowday, & Boulian, 1974). Because affective commitment and continuance commitment, respec-
turnover can be costly to organizations, commitment is gener- tively, to characterize Porters' and Becker's discrepant views of
ally assumed to be a desirable quality that should be fostered in the construct. It was noted that, although both affective and
employees. There has been comparatively little research, how- continuance commitment reflect links between the employee
ever, examining the link between organizational commitment and the organization that decrease the likelihood of turnover,
and work-relevant behavior other than turnover. Employees'job the nature of the links are quite different. Employees with a
performance is arguably as important as, or more important strong affective commitment remain with the organization be-
than, whether they ultimately stay or leave. We conducted the cause they want to, whereas those with strong continuance com-
present study, therefore, to examine the relation between em- mitment remain because they need to do so. Consequently, one
ployees' commitment to the organization and supervisors' eval- might expect the on-the-job behavior of those who are affec-
uations of their performance. tively committed to the organization to differ from that of em-
Before we discuss the expected commitment-performance ployees whose primary link to the organization is based on con-
relation, it is important to review differing conceptualizations tinuance commitment. Those who value and want to maintain
ofcommitment. Two quite different definitions of commitment membership should be willing to exert considerable effort on
have been popular in the empirical literature: one provided by behalf of the organization (Mowday et al., 1982). In contrast,
Porter and his associates (Mowday, Porter, & Steers, 1982; Por- those who feel compelled to remain to avoid financial or other
ter et al., 1974), and the other by Becker (1960). According to costs may do little more than the minimum required to retain
Porter et al., commitment is "the strength of an individual's their employment.
identification with and involvement in a particular organiza- The most widely used measure of affective commitment to
tion" (p. 604). Becker, on the other hand, described commit- date has been the Organizational Commitment Questionnaire
ment as the tendency to engage in "consistent lines of activity" (see Mowday et al., 1982). There is some evidence that scores
(p. 33) because of the perceived cost of doing otherwise. In the on this measure correlate positively with individual- and group-
case of commitment to the organization, the activity referred to level indexes of performance (e.g., Mowday, Porter, & Dubin,
by Becker involves staying with the organization, and the per- 1974; Steers, 1977). Thus, employees who are affectively com-
ceived costs associated with discontinuing the activity (i.e., leav-
mitted to the organization tend to perform at a higher level than
those who are not.
Becker's (1960) definition of commitment has typically been
This research was supported by Social Sciences and Humanities Re-
operationalized by the Ritzer and Trice (1969) or Hrebiniak
search Council of Canada research grants 410-87-1235 to John E
Meyer,410-87-0577 to Sampo V. Paunonen, and 411-83-00 ! 4 to Doug- and Alutto (1972) scales. Both measures require respondents to
las N. Jackson. indicate the likelihood that they would leave their organization,
Correspondence concerning this article should be addressed to John given various inducements from another company (e.g., in-
P. Meyer, Department of Psychology, The University of Western On- creases in pay, status, freedom, or promotional opportunities).
tario, London, Ontario, Canada, N6A 5C2. The appropriateness of these measures as an index of cost-in-
152
 ORGANIZATIONAL COMMITMENT AND JOB PERFORMANCE 153

duced c o m m i t m e n t , however, has been questioned (Meyer & ure in the position of unit manager was 4.1 years. The majority were
Allen, 1984; Stebbins, 1970). Meyer and Allen (1984; Allen & married (63%) and had at least a 12th-grade education (65%). Approxi-
Meyer, 1987) developed an alternate questionnaire to measure mately 55% of the unit managers were men.
continuance c o m m i t m e n t . Their scale consists o f eight items
reflecting the perceived costs associated with leaving the organi- Measures
zation, both in terms o f personal loss or sacrifice and lack of Organizational commitment. Commitment was measured with the
comparable alternatives. Using this scale, Allen and Smith eight-item Affective Commitment and Continuance Commitment
(1987) and Meyer and Allen (1986) found negative correlations scales used in previous research by Meyer and Allen (1984; Allen &
between continuance c o m m i t m e n t and self-report measures o f Meyer, i 987) and McGee and Ford (1987). Responses were made on 7-
motivation and performance. Thus, employees who indicated point disagree-agree scales, and scale total scores were computed by
that it would be costly for t h e m to leave their current employers averaging across items. The Affective Commitment and Continuance
described themselves as contributing less to organizational Commitment scale items are reported by Allen and Meyer (1987) and
effectiveness than those who perceived lower costs associated McGee and Ford (1987). Internal consistency estimates (alpha coetti-
dents) obtained in previous research ranged from .84 to .88 for the
with leaving.
Affective Commitment Scale, and from .70 to .84 for the Continuance
To summarize, some evidence exists to suggest that affective
Commitment Scale. Meyer and Allen (1984) reported a correlation of
c o m m i t m e n t is positively related, and continuance c o m m i t - -.01 between affective and continuance commitment, suggesting that
m e n t negatively related, to performance. We tested this hypoth- the two constructs are approximately orthngonal. Furthermore, they
esis further in the present study by measuring the affective and reported a correlation of.86 (p < .001), between the Affective Commit-
continuance c o m m i t m e n t o f first-level managers in a large food ment Scale and the widely used Organizational Commitment Question-
service company and by relating these scores to ratings o f per- naire (Mowday et al., 1982). The Continuance Commitment Scale was
formance obtained from supervisors. The j o b satisfaction of found to correlate -.06 with the Organizational Commitment Ques-
these managers was also measured and related to performance. tionnaire.
Comparison o f the relation between j o b satisfaction and perfor- Job satisfaction. We measured satisfaction in this study by using the
Index of Organizational Reactions (Smith, 1976). The Index of Organi-
mance with the relation between affective c o m m i t m e n t and per-
zational Reactions consists of eight subscales assessing satisfaction with
formance may aid in interpreting the latter.
supervision, company identification, kind of work, amount of work,
coworkers, physical work conditions, financial rewards, and career fu-
Method ture. Responses to the 42 items composing this measure were made on
5-point rating scales. In the following analyses, we examined only the
Subjects and Procedures composite scores reflecting overall satisfaction. We computed these
scores by averaging across items and subscales, excluding company
Participants in the study were unit (first-level) managers in a large identification. The company identification subscale was excluded be-
food service organization and their district managers. The data were cause of its conceptual overlap with affective commitment.
collected as part of a larger study undertaken to assess the concurrent Performance. Each district manager rated his or her unit managers
validity of several instruments being considered for use in the selection on six performance dimensions as well as on overall performance and
of unit managers. promotability. The six dimensions were effectiveness in (a) customer,
We sent materials by mail to 27 district managers throughout Can- client, and public relations, (b) administration and accounting prac-
ada. We included performance appraisal instruments to be completed tices, (c) preparation of written reports and verbal communication, (d)
by the district manager as well as a variety of instruments to be distrib- training and management of unit personnel, (e) following of operational
uted by the district managers to those unit managers under their super- policies and procedures, and (f) conducting ofroutinejob tasks. Ratings
vision. The number of unit managers selected for participation varied were made on 100-point scales, marked in intervals of 10 from 0 to 100,
from 2 to 6 per district, for a total of 114. Where there were more than in which a rating of 50 was labeled average for unit managers in the
six units in a district, six managers were selected at random in order to company. District managers placed a slash through a line to represent
keep the supervisor ratings at a reasonable number. each unit manager's relative effectiveness on each dimension of perfor-
District managers were instructed to complete performance apprais- mance. The number of unit managers rated by each district manager
als of their unit managers prior to visiting the units and to mail the ranged from 2 to 6, with an average of 4.0. Although this performance
evaluation forms directly to the researchers at The University of West- appraisal instrument was developed specifically for the present study,
ern Ontario. District managers then delivered materials to the units in the six dimensions were those that had been identified by the Human
person, including a personal data sheet and measures of verbal and nu- Resource Department as especially important to the company. These
merical aptitude, personality, vocational interest, organizational com- dimensions, furthermore, serve as the basis for their annual perfor-
mitment, and job satisfaction. With the exception of the personal data mance appraisal.
sheet and two timed aptitude measures, all forms were left with the unit Ratings on the six specific performance dimensions correlated sig-
managers to be completed privately and then returned by mail to the nificantly with one another (average r = .54; range = .38 to .75). We
university. computed a composite performance index by averaging scores across
Participation in the study by unit managers was voluntary and mate- the six performance dimensions. Although the overall performance and
rials were identified by code number only. Prior to their participation promotability ratings also correlated highly with ratings on the six per-
in the study, we assured subjects of complete confidentiality of their formance dimensions (Ms = .67 and .47, respectively), they were not
questionnaire responses and informed them that only aggregate data included in this composite measure. The overall performance rating
would be reported to the company. Twenty-three district managers was retained as a separate measure because it reflected supervisors' sub-
(85%) returned the performance appraisals and 65 unit managers (57%) jective weighting of performance dimensions and possibly included di-
returned the work attitude inventory, including the commitment and mensions other than the six that were assessed directly in this study.
satisfaction scales. The average age of these unit managers was 36.7 Promotability ratings, although related to performance, were consid-
years, average tenure with the company was 6.3 years, and average ten- ered to measure a different construct.
154 MEYER, PAUNONEN, GELLATLY, GOFFIN, AND JACKSON

Table 1
Means, Standard Deviations, Reliabilities, and Correlations o f the Measures

Measure 1 2 3 4 5 6 7 8 9 M SD

I. Age -- 36.41 10.15

2. Organizational tenure 55** -- 5.96 5.11
3. Job tenure 51"* 62** -- 3.79 3.70
4. AffectivecommitmenP -04 04 - 14 74 10 58** 15 25* 24* 4.94 1.00
5. Continuanoe commitmenP 20 04 31"* 01 69 27* -26* -27* -42** 3.45 1.14
6. Job satisfaction b 12 03 11 54** 30** 89 -01 09 02 3.81 0.34
7. Composite performance c - 12 06 04 15 -25* -03 87 85** 63** 61.10 13.30
8. Overall performance c -06 15 13 23* -25* -07 85** -- 67** 64.61 13.82
9. Promotability~ -46** - 15 - 18 23* -46** -04 62** 63** -- 47.51 20.27

Note. Listwise deletion for missing data reduced the sample size to 61 for these analyses. Decimal points are omitted in the correlation matrix.
Numbers in the diagonal that appear in boldface are reliability estimates (coefficient alpha), those below the diagonal are zero-order correlations,
and those above the diagonal are third-order partial correlations controlling for age, organizational tenure, and job tenure.
a Measured on a 7-point scale, b Measured on a 5-point scale, c Measured on a 100-point scale.
*p<.05. **p<.01.

Results mance and promotability ratings. In contrast, the correlations

between continuance commitment and the three performance
To determine whether the sample of unit managers who com- measures were all significantly negative. Job satisfaction scores
pleted and returned the attitude inventory was representative did not correlate significantly with any of the performance rat-
of the larger sample whose participation was solicited, respon- ings. As predicted, the direction o f the relations between com-
dents were compared with nonrespondents on age, tenure, and mitment and performance varied as a function of the nature of
performance. The two groups did not differ in age, tenure, or the commitment. Supervisor ratings of performance and pro-
ratings of promotability. There was, however, a significant motability tended to increase with increased affective commit-
difference on the performance composite, t(89) = 2.03, p < .05, ment o f unit managers and to decrease with increased continu-
and on the rating of overall performance, t(89) = 2.13, p < .05. ance commitment.
Those who returned the attitude questionnaires received higher
The partial correlations between the commitment and per-
performance ratings than those who did not. Although this
formance measures, with age, organizational tenure, and job
difference might be expected to produce a restriction in range
tenure controlled, were almost identical to the zero-order corre-
that would affect correlations with performance, comparison of
lations. The relations between commitment and performance,
the standard deviations of the three performance measures for
therefore, cannot be attributed to differences in age and tenure.
the total and reduced samples did not reveal meaningful differ-
Moreover, the fact that j o b satisfaction did not correlate sig-
ences.
nificantly with performance suggests that the correlations in-
volving affective commitment reflected relations between per-
Commitment and Performance formance and employees' affective orientation toward the orga-
Table I contains means, standard deviations, reliability esti- nization rather than toward the j o b or work in general.
mates (coefficient alpha), and intercorrelations for the major In a factor analysis of the Meyer and Allen (1984) affective
variables in the study. Note that, although acceptable, the reli- and continuance commitment items, McGee and Ford (1987)
ability estimates for the Affective Commitment and Continu- found that items from the Continuance Commitment Scale
ance Commitment scales are slightly lower than those obtained loaded on two separate factors. Items reflecting availability of
in previous research. This might reflect a greater homogeneity j o b alternatives loaded on one factor, and items reflecting per-
in commitment scores in the present sample. sonal sacrifice associated with leaving the organization loaded
The correlation matrix in Table 1 contains zero-order corre- on a second factor. For exploratory purposes, therefore, we di-
lations below the diagonal, and third-order partial correlations, vided the Continuance Commitment Scale into the alternatives
controlling for age, organizational tenure, and job tenure, above and personal-sacrifice subscales suggested by McGee and Ford.
the diagonal. To aid in the comparison of correlations, particu- The subscales correlated significantly with the full scale (r = .84
larly the corresponding zero-order and partial correlations, we for both subscales) and with one another (r = .51). The pattern
performed a listwise deletion of missing data prior to analysis. of correlations with the performance and promotability mea-
This reduced the sample size for the analyses reported in Table sures for both subscales was almost identical to that for the full
I to61. scale. Correlations involving the alternatives subscale were
Of greatest interest to this study are the correlations between slightly larger than those involving the personal-sacrifice sub-
the work attitudes expressed by unit managers and performance scale.
ratings made by their supervisors. The correlations between The results of the correlational analyses are generally consis-
affective commitment and the three performance measures tent with our hypotheses. We have, however, assumed a simple
were all positive and were significant for the overall perfor- additive influence of affective and continuance commitment on
 ORGANIZATIONAL COMMITMENT AND JOB PERFORMANCE 155

performance. To test for possible interaction effects, we first ex- Table 2

amined the mean scores for composite performance, overall Moderated Multiple Regression Analyses of Performance and
performance, and promotability for subgroups of unit manag- Promotability Criteria (P) on Affective Commitment (A),
ers. Four subgroups were created by dividing managers at the Continuance Commitment (C), and Their Product (AC)
means on the two commitment dimensions. The highest scores
on the composite performance, overall performance, and pro- Multiple correlation
motability measures were obtained by those managers (N = 17) Criterion Rp.^" RI'.Ac Rr'.AC~^c)
who were high in affective commitment and low in continuance
commitment (Ms = 65.94, 71.78, and 61.89, respectively). The Composite performance .15 .30* .30
t-test comparisons of means in this cell with those in each of Overall performance .23* .34* .39
Promotability .23* .50** .52
the remaining cells revealed significant (p < .05) differences for
each comparison for overall performance and promotability Note. N = 61. Probability levelsrefer to the significanceof increments
ratings but not for composite performance scores. We found no in multiple correlation due to added predictor variable.
significant differences among the means for the low-affective, a One-tailed test.

low-continuance (N = 11; Ms = 60.11, 59.55, and 45.82); low- *p < .05. **p <.01.
affective, high-continuance (N = 18; Ms = 58.21, 61.56, and
38.72); and high-affective, high-continuance (N = 17; Ms =
57.38, 61.59, and 40.65) subgroups. the costs associated with leaving, as Becker (1960) described it,
These results indicate that affective and continuance commit- the benefits of reduced turnover may be obtained at the price
ment interact in their effects on performance. The subgrouping of relatively poor performance.
method of analysis, however, has been criticized as a means of Our data suggest that it is important for organizations to ex-
evaluating moderator effects (e.g., Paunonen & Jackson, 1985; amine the policies they implement to increase commitment.
Zedeck, 1971). We therefore conducted moderated multiple re- Some companies explicitly attempt to bind employees to the
gression analyses as a further test of the interaction effects. It organization through rapid promotion, nonvested pension
has been argued that this multivariate approach is superior to plans, organization-specific skills training, and so on. Although
subgroups analysis for this purpose. Among its several advan- this undoubtedly makes it difficult for employees to decide to
tages are greater statistical power in testing for moderator effects leave, it may not instill in them the desire to contribute to organ-
and immunity to certain spurious effects that can characterize izational effectiveness. Instead, some employees may find them-
subgroups analyses (Paunonen & Jackson, 1985). selves in the position in which they have little desire to remain
The results of the regression analyses are reported in Table with the company but simply cannot afford to do otherwise.
2. Consistent with the previous correlational analyses, affective Such employees may be motivated to do little more than per-
commitment contributed significantly to the prediction of both form at the minimum level required to maintain the jobs on
the ratings of overall performance and promotability. Further- which they have become dependent. It may be more prudent,
more, the linear component of continuance commitment added although more difficult perhaps, for companies to foster affec-
significantly to the prediction of all three performance criteria. tive commitment in their employees than to develop continu-
In no case, however, did inclusion of the interaction term (i.e., ance commitment. Employees who intrinsically value their as-
the product of affective and continuance commitment scores) sociation with the organization are more likely not only to re-
in any regression equation add significantly to prediction of main with the company but to work toward its success.
performance beyond the linear contributions of the commit- There has been a fairly substantial body of research address-
ment scores. Thus, it does not appear from these analyses that ing the development of affective commitment in personnel. Par-
affective and continuance commitment interact in their effects ticular attention has been paid to the nature of the work experi-
on job performance. Although at variance with the findings of ences associated with commitment, for example, confirmed ex-
the subgroups analyses, we would argue, for reasons outlined pectations, job challenge, organizational dependability, and role
here (also see Paunonen & Jackson, 1988), that the results of clarity (Mowday et al., 1982). The results of recent longitudinal
the multiple regression analyses are more defensible. studies provide some evidence for the causal effects of these
work experiences on affective commitment, particularly in the
Discussion first few months of employment (e.g., Meyer & Allen, 1987,
1988). Therefore, organizations wishing to foster long-term
The results of this study generally supported our predictions. affective commitment in their employees should manage these
Affective commitment of employees to a food services organi- early work experiences carefully (see Bray, Campbell, & Grant,
zation was positively related to their measured job perfor- 1974; Buchanan, 1974).
mance, whereas continuance commitment was negatively re- Although the results of this study were based on a relatively
lated. The value of commitment to the organization, therefore, small sample, the consistency in the pattern of correlations and
may depend on the nature of that commitment. When commit- the fact that the commitment and performance data were ob-
ment reflects an identification with and involvement in the tained from independent sources increase our confidence in
company, as conceptualized by Porter and his associates (e.g., their reliability. The findings underscore the need for more re-
Porter et al., 1974), the organization may benefit both in terms search examining relations between commitment and work-re-
of reduced turnover and superior performance. In contrast, lated behavior other than turnover and, furthermore, empha-
when commitment is primarily on the basis of a recognition of size the need to distinguish clearly the nature of the commit-
156 MEYER, PAUNONEN, GELLATLY, GOFFIN, AND JACKSON

ment construct being considered, both in empirical research Meyer, J. P., & Allen, N. J. (1987). A longitudinal analysis of the early
and in practical applications. development and consequences of organizational commitment. Ca-
nadian Journal of Behavioural Science, 19, 199-215.
Meyer, J. P., & Allen, N. J. (1988). Links between work experience and
References organizational commitment during the first year of employment: A
longitudinal analysis. Journal of Occupational Psychology,, 61, 195-
Allen, N. J., & Meyer, J. P. (1987). The measurement and antecedents 209.
of affective, continuance, and normative commitment to the organiza- Mowday, R. T, Porter, L. W., & Dubin, R. (1974). Unit performance,
tion. Manuscript submitted for publication. situational factors, and employee attitudes in spatially separated work
Allen, N. J., & Smith, J. (1987, June). An investigation of"extra-role'" units. Organizational Behavior and Human Performance, 12, 231-
behaviours within organizations. Paper presented at the annual meet- 248.
ing of the Canadian Psychological Association, Vancouver, British Mowday, R. T, Porter, L. W., & Steers, R. M. (1982). Employee-organi-
Columbia, Canada. zation linkages: The psychology of commitment, absenteeism, and
Angle, H., & Perry, J. (I 981). An empirical assessment of organizational turnover. New York: Academic Press.
commitment and organizational effectiveness. Administrative Sci- Paunonen, S. V., & Jackson, D. N. (1985). Idiographic measurement
ence Quarterly, 26, i- 14. strategies for personality and prediction: Some unredeemed promis-
Becker, H. S. (1960). Notes on the concept of commitment. American sory notes. PsychologicalReview, 92, 486-51 I.
Journal of Sociology, 66, 32-42. Paunonen, S. V., & Jackson, D. N. (1988). Type I error rates for moder-
Bray, D. W., Campbell, R. J., & Grant, D. L. (1974). Formative years ated multiple regression analysis. Journal of Applied Psychology, 73,
in business. New York: Wiley-Interscience. 569-573.
Buchanan, B. (1974). Building organizational commitment: The social- Porter, L. W., Steers, R. M., Mowday, R. T., & Boulian, P. V. (1974).
ization of managers in work organizations. Administrative Science Organizational commitment, job satisfaction, and turnover among
Quarterly, 19, 533-546. psychiatric technicians. Journal of Applied Psychology, 59, 603-609.
Hrebiniak, L. G., & Alutto, J. A. (1972). Personal and role-related fac- Ritzer, G., & Trice, H. M. (1969). An empirical study of Howard Beck-
tors in the development of organizational commitment. Administra- er's side-bet theory. Social Forces, 47, 475-479.
tive Science Quarterly, 17, 555-573. Smith, E J. (1976). Index of Organizational Reactions (IOR). JSAS Cat-
McGee, G. W., & Ford, R. C. (1987). Two (or more?) dimensions of alog of Selected Documents in Psychology,,6(54, No. 1265).
organizational commitment: Reexamination ofthe affective and con- Stebbins, R. A. (1970). On misunderstanding the concept of commit-
tinuance commitment scales. Journal ofApplied Psychology, 72, 638- ment: A theoretical clarification. Social Forces, 48, 526-529.
641. Steers, R. M. (1977). Antecedents and outcomes of organizational com-
Meyer, J. P., & Allen, N. J. (1984). Testing the "side-bet theory" of or- mitment. Administrative Science Quarterly, 22, 46-56.
ganizational commitment: Some methodological considerations. Zedeck, S. ( 1971). Problems with use of"moderator" variables. Psycho-
Journal of Applied Psychology, 69, 372-378. logical Bulletin, 76, 295-310.
Meyer, J. P., & Allen, N. J. (1986, June). Development and consequences
of three components of organizational commitment. Paper presented Received January 5, 1988
at the annual meeting of the Administrative Sciences Association of Revision received June 20, 1988
Canada, Whistler, British Columbia, Canada. Accepted June 23, 1988 •