Review and conceptual analysis of the employee turnover process

Psychological Bulletin 1979, Vol. 86, No. 3, 493-522 Review and Conceptual Analysis of the Employee Turnover Process W. H. Mobley, R. W. Griffeth, H. H. Hand, and B. M. Meglino Center for Management and Organizational Research University of South Carolina Research on employee turnover since the Porter and Steers analysis of the literature reveals that age, tenure, overall satisfaction, job content, intentions to remain on the job, and commitment are consistently and negatively related to turnover. Generally, however, less than 20% of the variance in turnover is explained. Lack of a clear conceptual model, failure to consider available job alternatives, insufficient multivariate research, and infrequent longitudinal studies are identified as factors precluding a better understanding of the psychology of the employee turnover process. A conceptual model is presented that suggests a need to distinguish between satisfaction (present oriented) and attraction/expected utility (future oriented) for both the present role and alternative roles, a need to consider nonwork values and nonwork consequences of turnover behavior as well as contractual constraints, and a potential mechanism for integrating aggregate-level research findings into an individual-level model of the turnover process. Employee withdrawal, in the form of turnover, has sustained the interest of personnel researchers, behavioral scientists, and management practitioners. At the macro level, economists and personnel researchers have demonstrated the relationship between turnover rates and the aggregate level of economic activity, employment levels, and vacancy levels (see, e.g., Armknecht & Early, 1972; Forrest, Cummings, & Johnson, 1977; Price, 1977; Woodward, 1975-1976). At the micro level, behavioral research has established a consistent, although generally weak, correlation between job dissatisfaction and turnover (Brayfield & Crockett, 19SS; Locke, 1976; Porter & Steers, 1973; Vroom, 1964; Herzberg, Mausner, Peterson, & Capwell, The preparation of this article was supported by the Office of Naval Research under Contract N0001476-C-0938: NR 170-819. Appreciation is expressed to John Logan and John Cathcart for their assistance in the literature search and to the editor and reviewers lor their constructive comments. Requests for reprints should be sent to William H. Mobley, Center for Management and Organizational Research, University of South Carolina, Columbia, South Carolina 29208. Note 1). While the economic and job dissatisfaction contributions to turnover are well established, they are conceptually simplistic and empirically deficient bases for understanding the employee turnover process. Recently, a number of authors (Forrest et al., 1977; Locke, 1976; Mobley, 1977; Porter & Steers, 1973; Price, 1977) have advocated abandoning further replication of bivariate correlates of turnover, particularly job dissatisfaction, in favor of well-developed conceptual models of the turnover process. Such a model is one objective of this article. Employee turnover is a behavior of interest to many disciplines and is subject to analysis and discussion at many levels of discourse. The approach taken in this article is basically psychological and rests on the belief that turnover is an individual choice behavior. Thus, the individual is the primary unit of analysis. Selecting the individual as the unit of analysis does not mean that turnover research at the unit, organizational, or other aggregate level is not of value and interest. However, to conclude that such studies clarify the individual turnover decision process may be tantamount to what Robinson Copyright 1979 by the American Psychological Association, Inc. 0033-2909/79/8603-0493$00.75 493 Table 1 Studies of Relations Between Personal Factors and Turnover Factor Population N Relation to turnover Age Federico, Federico, & Lundquist (1976) Hellriegel & White (1973) Mangione (Note 2) Credit union females Certified public accountants ISR diverse occupational sample Marsh & Mannari (1977) Japanese electrical company employees Mobley, Horner, & Hollingsworth (1978) Porter, Steers, Mowday, & Boulian (1974) Waters, Roach, & Waters (1976) Tenure Mangione (Note 2) Mobley et al. (1978) Waters et al. (1976) 96 349 294 1,033 Hospital employees Psychiatric technicians Insurance company clerical 203 60 105 ISR diverse occupational sample 295 Hospital employees Insurance company clerical 203 105 Institute for Social Research diverse occupational sample Japanese electrical company employees 293 Younger age at application associated with lower tenure No differences (significance test not reported) Chi-square, p < .001; younger age associated with higher turnover r — —.22**; younger age associated with higher turnover = -.22** Stayers significantly older than leavers** r = -.25* Chi-square, p < .001; lower tenure associated with higher turnover r .25** r = -.30** I w w < a w H K Sex Mangione (Note 2) Marsh & Mannari (1977) Family responsibilities Federico et al. (1976) Mangione (Note 2) Marsh & Mannari (1977) Waters et al. (1976) Education Federico et al. (1976) Hellriegel & White (1973) Mangione (Note 2) 1,033 2! d r = —.31**; women had higher turnover O Credit union females ISR diverse occupational sample Japanese electrical company employees Insurance company clerical Credit union females Certified public accountants ISR diverse occupational sample 96 295 1,033 105 96 349 294 Higher responsibility associated with higher tenure; factors included marital status, number of children, age of youngest child, and age Chi-square, p < .001; single people had higher turnover r = —.22; no or few dependents had higher turnover ns (marital status) Higher education associated with lower tenure No differences (significance test not reported) ns f I EMPLOYEE TURNOVER t_ T3 3 -o O « £ .C u 1-|! & D .2 u 3 > ti g •a a.l> fe -C 3 J3 . ^ o *•* *5 £ (U 4-* B a 3 J3H u <3 S | ** 4J <u *-H U 4> *•* 5 ?* 4j O JJ H c c oc O 4-• 1 i-4> ^S flji- Ga e .2 js 73 £ o .Si» u Ml C D "^ "frt -^ ^ "S t'S'fc SJ 'S ^ §M dj U< ™ 3 C 3 3 5 £ I'S.'S) g _* -<• *c u o -^ oo n oo r~. >o o* fe; I bi o _a cu cj >> > o 6 IE e3 u *•• . » 1-1 * .£ CN bfi ''"S || -G »^ * to o 1 o "a. S "3. C c e U aE c _o 4-• > a & *O en =. 3 </> ti bA rt ,/, MB c g e Cs Ert N8.» iseI all 8l •§ c os gu w < S ae 8 ft 8 ""! tO •c w "3 <v o "53 % ii 0 I rt 1 a —, i—> (2 s o y ta 1 1 vHi 'C .8 •a S 8 O >o rt H Personality Hines (1973) *•• » Distance migrate Marsh & Man C" « mg O ^ | *i • §1 3° II JX is* I J "4-* |1 §s z i i Oi ui t/3 q 1; V J ** 495 (1950) has termed the ecological fallacy. For example, the relationship between aggregate unemployment levels and turnover rates, although well established (see, e.g., Armknecht & Early, 1972; Price, 1977; Woodward, 1975-1976) adds little to understanding individual turnover decisions. A linking mechanism is needed that considers the individuals's perception and evaluation of available alternatives relative to the present position. At the individual level, satisfaction is the most frequently studied psychological variable thought to be related to turnover. However, the satisfaction-turnover relationship, although consistent, usually accounts for less than 16% of the variance in turnover (Locke, 1976; Porter & Steers, 1973). It is apparent that models of the employee turnover process must move beyond satisfaction as the sole explanatory variable. Recently, the constructs of organizational commitment (Porter, Crampon, & Smith, 1976; Porter, Steers, Mowday, & Boulian, 1974; Steers, 1977), organizational attachment (Koch & Steers, 1978), role attachment (Graen, 1976; Graen & Ginsburgh, 1977), and behavioral intentions (Kraut, 1975; Mobley, 1977; Newman, 1974) have been offered as explanatory concepts in the turnover process. However, the conceptual and empirical identity of these concepts and their interrelationships have not always been clear. An additional objective of the present article is to attempt to clarify and integrate these concepts into a general model of the individual employee turnover process. A third objective of this article is to update earlier reviews of the literature. The last major review of turnover from the individual perspective was that of Porter and Steers (1973). For somewhat more limited reviews of certain aspects of turnover, see Goodman, Salipante, and Paransky (1973) on the hardcore unemployed and retention and Pettman's (1973) partial review of the March and Simon (1958) model. More recently, Price (1977), a sociologist, has published a significant book that seeks to codify the turnover literature from a variety of disciplines and cultures. The Price work contains a number of references generally not 496 MOBLEY, GRIFFETH, HAND, AND MEGLINO included in the psychological and management turnover literature cited in the United States; however, it does not deal with post1974 research and is incomplete in its coverage of the psychological and management literature on employee turnover. Forrest etal. (1977) also recently presented a partial review of the turnover literature. However, the latter review, which deals with a broader spectrum of organizational participation behaviors, contains no post-1973 research and has a conceptual rather than an empirical emphasis. The Forrest et al. model is discussed in a later section of this article. In summary, the major objectives of this article are (a) to update the last major reviews and analyses of the turnover literature, (b) to attempt to clarify the distinctions among various constructs that have recently been suggested as explanatory variables in the turnover process, (c) to develop a conceptual model of the individual-level employee turnover process that is consistent with the research literature, and (d) to suggest areas of further research. This article focuses on voluntary, that is, self-initiated, turnover rather than on organization-initiated terminations. This distinction, discussed in a subsequent section, is not always made clear in specific research studies. Additionally, this article does not deal with absenteeism. Whether absenteeism is best thought of as having no consistent relationship to turnover (March & Simon, 1958), as a precursor of turnover (Herzberg et al., Note 1), or as an alternative form of withdrawal behavior (Hill & Trist, 19SS; Rice & Trist, 19S2) is an important research question (Burke & Wilcox, 1972), but is beyond the scope of the present article. Update of Turnover Analyses and Reviews The last major reviews of the turnover literature were by Porter and Steers (1973) and Price (1977). This section summarizes the recent research not included in these reviews and offers the conclusions of the authors of this article. A subsequent section summarizes and integrates the results of this and the two previous reviews. Although no taxonomic schema is entirely satisfactory, the research summary is divided into the following sections: (a) individual demographic and personal variables, (b) overall satisfaction, (c) organizational and work environment factors, (d) job content factors, (e) external environment factors, (f) occupational groupings, (g) recently developed constructs, and (h) multivariate studies. Most studies reviewed take a bivariate approach to turnover, but this emphasis is reflective of the current literature rather than of the present authors' belief in the relative merit of this approach. Individual Demographic and Personal Factors Included in this category are age, tenure, sex, family responsibilities, education, personality, other personal considerations, and weighted application blanks. Table 1 summarizes recent research on these variables, which were not included in the Porter and Steers or Price reviews. Age. Recent research, with the exception of Hellriegel and White (1973) who reported no differences, indicates a negative relationship between age and turnover. However, the amount of variance explained is less than 7%. One should note that since age is correlated with many other variables, it alone contributes little to the understanding of turnover behavior. As is noted later, a conceptual model and multivariate studies are required to adequately comprehend the psychology of the turnover process. This observation also applies to each variable discussed below. Tenure. Three recent studies, cited in Table 1, showed a negative relationship between tenure and turnover. Mangione (Note 2) concluded on the basis of a multivariate study (see Table 12 for a summary of multivariate studies) that length of service is one of the best single predictors of turnover. Sex. Of the two studies relating an individual's sex to turnover, Marsh and Mannari (1977) observed that female Japanese manufacturing employees had higher turnover than males. Mangione (Note 2) found no relationship. EMPLOYEE TURNOVER Family responsibilities. Three of the four studies summarized in Table 1 indicate that family responsibility, including marital status, is associated with decreased turnover. Education. Of the recent studies dealing with education, one found that female credit union employees with higher education had lower tenure (Federico, Federico, & Lundquist, 1976), whereas Mangione (Note 2) and Hellriegel and White (1973) discovered no differences. Lack of variance in education in studies such as Hellriegel and White's, which used certified public accountants, precludes adequate evaluation of the role of education. Weighted application blanks. Weighted application blanks use a procedure for weighting information on an employment application form so as to predict some aspect of job performance, including turnover. Schwab and Oliver (1974) have raised serious questions regarding the utility of weighted application blanks for predicting turnover. In four samples, they found that validities shrunk below statistical significance upon cross-validation. However, in two recent studies Cascio (1976) and Lee and Booth (1974) reported significant relationships that were cross-validated. The utility of weighted application blanks for employee selection continues to require situation-specific validation (and regular cross-validation), and alone they offer little contribution to understanding the psychology of the turnover process. Other personal variables. Table 1 cites other studies that dealt with personality, distance migrated, and number of previous jobs. Because the number of studies is small, no generalizations are possible. 497 CO CO C 0 I* <u E •2 ^ J• S 0 .1i2 Ui J? O +^ 0 § £• ; c CO *i• - 'S ^S •S 2" gj t> "co QJ ^ S v. a 1—1 r H fe; ' 3 II w *l -- 6S s I^ H O M3 ^O •« If 7^ B *• " O # » •3 -2 B * G CN O ^ TJ ' O 1 JS ». -B B! 8 ," S S S . Or--Q f ^ _ _ o » —10 ^ ws ""31^.^ | '^ l' u l' l' f 01 II X II ° II II II t^ ~ t^ ^ f l' II II T3 01 *^ (N OCN~H ?^^»-j | _o 'K Populatio 2 ** 1 o "o. g l i e -T S a. c 2 CO 1 jf , 3 a "i- S § O 35 ,£ -S — S 'C ^J ^, 33 1 iiiiiil! g a<« g o "a | 2 Ui 0) -o *c3 (2 B £ % S. ^ ^ b«'^ e || S |-|-| | ! * < = *&&& Overall Job Satisfaction and Turnover t =3 Studies involving overall job satisfaction are summarized in Table 2. With one exception (Koch & Steers, 1978), these studies indicate a negative relationship between overall satisfaction and turnover. It is important to note, however, that the amount of variance accounted for is consistently less than 14%. As noted in subsequent sections, when satisfaction is included in multiple regressions with variables such as intentions and com- » CN ^ 2 i 3 fi o g 1 Q <« - P.| £r T)' B CO bo S\ ^ oc? S; gr ^- iu "^ C^.^ c> - c «8 ~ rt ,*• J3 1 I Jill S V CA • , * ^ B oy . B °^ q *U B IH 2 to ^ V t2 S SSz^ 5 cs a o S co * VO OO Table 3 Studies of Relations Between Pay-Promotion and Turnover Factor Salary, actual and expected Federico, Federico, & Lundquist (1976) Salary increases Hellriegel & White (1973) Satisfaction with pay Hellriegel & White (1973) Koch & Steers (1978) Kraut (1975) Mangione (Note 2) Mobley, Homer, & Hollingsworth (1978) Newman (1974) Waters, Roach, & Waters (1976) Satisfaction with promotion, advancement Hellriegel & White (1973) Koch & Steers (1978) Kraut (1975) Mobley et al. (1978) Newman (1974) Waters et al. (1976) Perceived chances of promotion Marsh & Mannari (1977) *p < .01. Population Credit union females N Relation to turnover 96 Higher salary associated with longer tenure (salary strictly performance based); the greater the difference between expected and actual, the shorter the tenure; the higher the expectations, the lower the tenure O Cd Certified public accountants 349 Turnovers reported 20% increases in pay on new job W Certified public accountants 349 Turnovers more negative than nonturnovers on attitudes toward pay policies and comparability of salary (significance levels not reported) 77 ns O Nonmanagement entry-level public agency employees Salesmen Institute for Social Research diverse occupational sample Hospital employees Nursing home employees Female insurance company clerical Certified public accountants Nonmanagement entry-level public agency employees Salesmen Hospital employees Nursing home employees Female insurance company clerical Japanese electrical company employees M 911 ns 295 r = -.16* 203 ns 108 ns 105 ns 349 Turnovers more negative about opportunities than nonturnovers (significance levels not reported) 77 ns 911 203 108 105 ns ns ns ns 1,033 r = — .22*; poorer perceived chances of promotion had higher turnover a >z 0 > 2, a g a o t-1 I Table 4 Studies of Relations Between Supervision and Turnover Population Factor Satisfaction with supervision Hellriegel & White (1973) Ilgen & Dugoni (Note 3) Koch & Steers (1978) Mobley, Homer, & Hollingsworth (1978) Newman (1974) Waters, Roach, & Waters (1976) Leader acceptance Dansereau, Cashman, & Graen (1974) Graen & Ginsburgh (1977) * p < .05. N Relation to turnover Certified public accountants 349 Retail clerks Nonmanagement entry-level public agency employees Hospital employees Nursing home employees Insurance company clerical 117 77 203 108 105 ns ns ns Managers 354 r = —.11* (and moderated by expectancy of finding comparable alternative, r = —.21* for high-expectancy group, « = 98) Significant main effect (p < .024) with resignation at 24 months; leader acceptance defined in terms of leader's flexibility in changing employee job and chances of leader using his or her power to help employees solve work problems Turnovers had significantly less favorable attitudes (few significance tests reported) r = -.23* ns o H 90 University service employees 89 I M 500 MOBLEY, GRIFFETH, HAND, AND MEGLINO mitment, its effect on turnover may become nonsignificant (Marsh & Mannari, 1977; Mobley, Homer, & Hollingsworth, 1978). O B 1 '1 K Organizational and Work Environment Factors e ^ >o c u. 2 3 hH O 1-1 3 o -*-* c _o V £ K rt bo 01 en w. _S V O _, U 4) v^. V, oO, ^u. * "s t. K. 1) 0 C -H >H ^ CN B Pi' FT 8 >. : K. 1 Q, ' fe; ^ 2,1? ^'e'a's' fel f•f i o o o II O 6 u 6o ^8 8 6 1^ V 6 T*H QJ )•> fll k^ ^) tj 8 8 8 S S 8 r^ t^ f^ P*5 OO *O IO 8.8 S 2 SS ir> — 0) ^ 8 i>X u g 3 3 f 1 g $ £3 Q, O OH •O 53 S> f-, •g <e '•H <O ^ V §J U «u Pa «• * Social Research al sample trical company loyees e employees npany, clerical "O 8 O C u e 1 >, u, •M C <1) +J 01 § 0 S" OJ O (V] ^ V 5 -M 8 g'S.1 § bO 2 -fl c3 ? o ^ o 3 2 0? i i J3 <u f• » to •—» ffi Z ^5 5J3 [A t (U c <u E.& ^^W £ ^ •1 S- $ 3 M c oo" t^» 3 £ o) a 6 o •3 c 'II c. CJ i >to 01 Dloyees lie accountants •-• 4 "o. g '"3 a "3 1 U4 M & 4-• > i- -a ^ O rT & Cti tn i »i o- Workers, T Effectiven Hellriegel & Pay and promotion. Research dealing with pay and promotion is summarized in Table 3. Federico et al. (1976) found that higher salary was associated with longer tenure, whereas higher salary and the difference between expected and actual salary were associated with shorter tenure. Mangione (Note 2) found a significant negative correlation between pay satisfaction and turnover. Hellriegel and White (1973) discovered that "leavers" had more negative attitudes toward pay than "stayers" and also reported significant increases in pay on their new jobs. Evidence from five other recent studies suggests a lack of relationship between pay satisfaction and turnover. Also evident in Table 3 is the general lack of relationship between satisfaction with promotion and turnover, although Hellriegel and White did find that leavers had more negative attitudes toward promotion than stayers. Marsh and Mannari (1977) reported a correlation of —.22 between perceived chances of promotion and turnover. Supervision. Table 4 summarizes recent studies relating satisfaction with supervision to turnover. In four of the studies a nonsignificant relationship between satisfaction with supervision and turnover was found. Hellriegel and White (1973) and Ilgen and Dugoni (Note 3) found a significant negative relationship. The study by Graen and Ginsburgh (1977) is of particular interest. Here, leadership was significantly associated with turnover. The leadership variable, however, was not satisfaction with supervision but specific aspects of the leader-member exchange. Contrasted with the conclusions of recent studies using satisfaction with supervision as the independent variable, the Graen and Ginsburgh results suggest the need for more detailed study of the leader-member exchange rather than reliance on generalized supervision affect measures. OJ l_ O m 0 oo §7 oa=s g-g £! *^ p-s e S^ " ^ ^ ^2 o" 1S O cs •g l-l 2^ ws ^^ '*•• ^ ^i I §"s-5 oa . c ^ j: & g E £ S |S Illl Table 6 Studies of Relations Between Other Organizational and Work Environment Factors and Turnover Factor Role status Hellriegel & White (1973) Population Certified public accountants N Relation to turnover Perceptions of prestige and social status in the community lower for turnovers than nonturnovers (significance tests not reported) Japanese electrical company employees 1,033 r = —.25**; lower status had higher turnover 349 Marsh & Mannari (1977) Knowledge of organization's procedures Japanese electrical company employees 1,033 r = —.14**; lower knowledge had higher turnover Marsh & Mannari (1977) Control processes Certified public accountants 349 Turnovers "generally" more negative than stays, especially re Hellriegel & White (1973) performance evaluation (significance tests not reported) Role pressures Hellriegel & White (1973) Certified public accountants 349 Not significant on items related to scarcity of time Climate Ilgen & Dugoni (Note 3) Retail clerks 117 Satisfaction with company Salesmen 911 ns Kraut (1975) Satisfaction with hours Ilgen & Dugoni (Note 3) Retail clerks 117 r = -.33** Resource adequacy Mangione (Note 2) Institute for Social Research diverse 295 r .13* occupational sample Satisfaction with comfort Institute for Social Research diverse 295 r = -.20** Mangione (Note 2) occupational sample *p < .05. **p < .01. W S T) 1 W Cn Table 7 O to Studies of Relations Between Job Content Factors and Turnover Factor Satisfaction with work itself Koch & Steers (1978) Population N 77 Kraut (1975) Nonmanagement entry-level public agency employees Salesmen 911 Mobley, Horner, & Hollingsworth (1978) Newman (1974) Porter, Steers, Mowday, & Boulian (1974) Hospital employees Nursing home employees Psychiatric technicians 203 108 60 Waters & Roach (1973) Female insurance company clerical 80 in follow-up Relation to turnover r .31** r = —.14**; short term, 18 months r = —.12**; long term, 5J years r = -.20** ns Only moderate contribution to discriminate functions with organization commitment the strongest variable r = .26* O 60 W O 117 Female insurance company clerical Waters, Roach, & Waters (1976) Amount of work Kraut (1975) Salesmen Job autonomy Marsh & Mannari (1977) Japanese electrical company employees Intrinsic satisfaction Mangione (Note 2) Mirvis & Lawler (1977) in another office r = -.40* 105 r .37** X > 911 z D Institute for Social Research diverse occupational sample Bank tellers 1,033 r — — .09*; low perceived autonomy had higher turnover 295 r = - .22** with challenge 160 r O .20* H-I § Intrinsic motivation Mirvis & Lawler (1977) Bank tellers Herzberg motivator deprivation Karp & Nickson (1973) Black working poor 50 r = .37** with number of jobs held in last 5 years Role orientation Graen, Orris, & Johnson (1973) University clerical probationary period 62 25% attrition for high role orientation; 59% attrition for low role orientation at the end of 10 months; role orientation defined as perceived relevance of the job for the worker's career 160 r = -.16* EMPLOYEE TURNOVER e .1 u| II '£ g_ Ui 3 0 C Jsa) 1 ?1i \v/ -S> o s6-* t* e '5. ° •£ S3 2 "° 5 Si |-8 3 4j *• * c .2 3 u(/) O .3 -M Lc o j= •w 0) 1 fc j: .Sf T3 o! _C a! 1-1 0 •Su, 0 nl (/> > *j Cl C ?- 1 8 ^' » is *o \ a 'M 3'! ic •£ 'i'<* (75 „ « s 7' H ="* >- a^ ' Ui i 3 K, PO fe; 00 —< s tS 8 aJ o .0a 1 oa ft* "a. Eaj io e aE "5. ,! y 8 to E %^ jj & s 4{ *t4)8s '35 "o « '3u s 2 ~ b > c S en "£ <u "^ —, a 'S<u 2 0& a! <3 .si j•^j 4) (/) g a rt S 1 ~ '^ £ |fo | v *—x $ |> <S o jiI '^$^ c ."§ g, a C 3 *O o "?§ 0S « W] +-• * 0 "O .- ^ iS .—i C o o\ ^ Q -B ? 1 .2 1 nl 3 s -By ^c 8«s ?2 ^s 5i^l TT bo g S Perceived performance Marsh & Mannari ( Table 7 (continued) io ia ^ °v 503 Overall, recent studies offer moderate support for the negative relationship between satisfaction with supervision and turnover. However, the number of studies finding no significant relationship between these variables indicates the need to more closely examine the nature of leadership measures, to conduct more microanalyses of the leadermember exchange, and to assess the contribution of supervision in multivariate designs that consider other salient variables. Peer group relations. Research on peer relations and turnover is shown in Table 5. In seven of the nine studies, no significant results were reported. Koch and Steers (1978) found a significant correlation between satisfaction with co-workers and turnover, but only 4% of the variance was explained. The studies summarized in Table S do not support the generalization of a strong relationship between group relations and turnover. Individual differences in variables such as need for affiliation, the role of other variables, for example, required task interaction and external job alternatives and the method of measuring group relations contribute to the difficulty in explicating these findings. Other variables. Table 6 describes recent studies dealing with a variety of other organizational and work environment factors. Both Hellriegel and White (1973) and Marsh and Mannari (1977) found a negative relationship between preceived status and turnover. This status may have come, in part, from the work role or from organizational affiliation. Knowledge of organizational procedures and perceptions of control processes were each shown to be negatively related to turnover. In three separate studies, role pressures, climate, and satisfaction with the company were not significantly related to turnover. Ilgen and Dugoni (Note 3) found a significant negative correlation between satisfaction with hours of work and turnover among retail clerks. Mangione (Note 2) discovered a significant negative correlation between resource adequacy and turnover. * ^ Job Content Factors v Recently job content has become one of the more active areas of industrial-organiza- ^ MOBLEY, GRIFFETH, HAND, AND MEGLINO ca 13 "• C 3 nj v•• ceived expectancy ated the correlatio (office workers, r • ion to turnover ^5 *^ x ^ IM '—' <-< t a-o > ^ ws ° |7f E 'M 12 ja> ' .ti .> II & o c ll V 3 II C .- w uOS *C jp 3 |« c 5 I• *£ >; S5 s *• ! «..c c !H * i acceptable altern; intention to quit ( related significant! tional research. A number of studies in this area have employed turnover as a criterion. Inspection of the recent studies shown in Table 7 indicates that job content factors are significantly related to turnover. Satisfaction with work itself exhibits a uniform negative correlation with turnover, although the amount of variance explained is consistently less than 16%. Additional studies indicate that the perceived intrinsic value of work, intrinsic motivation, and intrinsic satisfaction are all significantly and negatively related to turnover. Graen, Orris, and Johnson (1973) and Graen and Ginsburgh (1977) demonstrated that role orientation, defined as the perceived relevance of the job for the worker's career, was significantly related to turnover. In a particularly interesting study, because it took a somewhat different approach, EkpoUfot (1976) found that self-perception of task-relevant abilities was significantly and negatively associated with auto assembler turnover. mployment, R2= nemployment, vac = .83 504 1 •3 T) u ^-s . 1s a : "•e w O 3 -^ II '-C ^ 35 ^ "3. .2 "c *• o"S 3^ g xi c« S >> 0 e II ? c8 § Ow b/J g >^*4J 3 a• o. 3 S E-H *G O -^ — a 9 — •O B 1— 1 J ifli 1 w "8 t; a^ fe; S$ (S f5 CO 0 o c External Environment c ae I s 8 c 1 ^o 3 1 a (£ S 1 .c 1 °% o .a *«» C3 S * tuo D rt (j g '•M 'C OS 5 >^%> *^.—, Ov w S v^a > ta « 1 ^ u -g __^ ll =1 | | T3 C I-E °v » *j5 '^ 11 H 'o u 0) ^ il u "rt c" 2 O C ^ 8 ^ S £ o "3. I! •^ ^C c4 PQ « Relation The probable role of the availability of alternative jobs in employee turnover has long been recognized; see, for example, March and Simon (1958). Economists and sociologists have documented the aggregatelevel relationship between economic indicators such as employment levels or job vacancy rates and turnover rates (Price, 1977). However, research on individual-level turnover has infrequently assessed perceived alternatives (Forrest et al., 1977; Locke, 1976), Conceptually, the perception and evaluation of alternatives seems to be a crucial variable in the individual turnover process. Empirically, assessment of the relationship between turnover and personal, organizational, job content, or other variables is inexorably bound to consideration of the perception and evaluation of alternatives. Table 8 summarizes the limited amount of recent research dealing with alternatives. Woodward (1975-1976) reaffirmed the aggregate-level negative relationship between unemployment and turnover and the positive ll o S 5s ^^•Q^ W 33 0? o\ ~~" 1 § U) ^q 3^3 08 o 33 0) 2 o S EMPLOYEE TURNOVER relationship between unfilled vacancies and turnover rates. At the individual level, Dansereau, Cashman, and Graen (1974) found that the expectancy of finding an alternative job moderated the correlations between attitude and turnover. Mobley et al., (1978) found that expectancy of finding an acceptable alternative position was significantly and positively related to intention to quit but not to actual quitting, although intention to quit was significantly and positively related to turnover. It is evident that much additional research is required to explicate the role of perception and evaluation of alternatives in the individual turnover process. Occupational Groupings Price (1977) reviewed research on occupational characteristics and found moderate support for the proposition that unskilled blue-collar workers have higher turnover than white-collar workers. He found only weak support for the propositions that nonmanagers have higher turnover than managers, that nongovernment employees have higher turnover than government employees, and that higher professionalism is associated with higher turnover. Since most individual-level turnover studies are carried out within occupational groupings, the present review adds little to Price's analysis. It is apparent, however, that any complete model of individual turnover behavior should be able to account for differences in turnover among occupational groupings. The perception and evaluation of alternatives is one obvious link between the two levels of analysis. A second is exemplified in the research of Herman and Hulin (1972) and Herman, Dunham, and Hulin (197S), which demonstrated that organizational variables such as position level may be better predictors of behavior than demographic or personality variables. The frame of reference provided by position level may influence values, perceptions, and expectations, thus linking organizational variables with individual behavior. 505 Recently Explored Variables and Processes Since the Porter and Steers (1973) review, interest has developed in a variety of additional variables, constructs, and processes, including behavioral intentions, organizational commitment, realistic expectations, and the centrality of work values. The Fishbein Behavioral intentions. (1967) and Fishbein and Ajzen (1975) model of the relationships among beliefs, attitudes, intentions, and behaviors emphasizes the role of intentions in understanding the link between attitudes and behavior. The Locke (1968) model of task motivation also conceives of intention as an immediate precursor of behavior. Drawing on these and other related theoretical models, a number of recent studies have assessed the role of intentions in predicting and understanding turnover. Table 9 summarizes these studies. It is evident from these studies that behavioral intentions to stay or leave are consistently related to turnover behavior. It is also evident that this relationship generally accounts for more variance in turnover than does the satisfaction-turnover relationship. Conceptually this appears appropriate, since satisfaction is an affective or emotional response, whereas intentions are statements regarding the specific behavior of interest, in this case, turnover. It is possible, as Mobley et al. (1978) suggested, that intentions also capture the individual's perception and evaluation of alternatives. Although the relationship between intentions and turnover appears to be consistent and generally stronger than the satisfactionturnover relationship, it accounts for less than 24% of the variance in turnover. Among the possible reasons for this are that intentions do not account for impulsive behavior, that they do not adequately capture the perception and evaluation of alternatives, and that along with personal, organization, and external conditions, they may change between original measurement and the observation of actual behavior. The more specific the behavioral intention statement and the less time its measurement and the behavior, the stronger the relationship should be. However, as Graen and Ginsburgh (1977) noted, Table 9 Studies of Relations Between Intentions and Turnover Population Factor Behavioral intentions Kraut (1975) Institute for Social Research diverse occupational sample Marsh & Mannari (1977) Japanese electrical company employees Mobley, Horner, & Hollingsworth (1978) Hospital employees Attitude toward act of quitting Newman (1974) Normative beliefs re quitting Newman (1974) *p < .01. Relation to turnover £ o 5 Salesmen Mangione (Note 2) Newman (1974) Waters, Roach, & Waters (1976) N Nursing home employees Insurance company clerical 911 Lower intent to remain significantly associated with turnover after 18 months, p < .01; between 18 and 60 months, p < .05 (n = 791; r = —.17* for voluntary turnover in both short- and long-term cases 242 Intentions significantly related to turnover, p < .001 (over 2 years) 1,033 r-b .13* (4 years) 203 r = .49* for intention to quit (47 weeks); r = .29* for intention to search; r = .19* for thinking of quitting 108 r = .39* (over 2 months) 105 r = —.42* (over 2 years with intent to remain) r = .36* (with biographical data partialed out) Nursing home employees 108 r = .30* Nursing home employees 108 r = .32* o » I-H 11 w 2! O § M O Table 10 Studies of Relations Among Organizational Commitment, Involvement, Job Attachment, and Turnover Factor Organizational commitment Marsh & Mannari (1977) Population N Relation to turnover Japanese electrical company employees 1,033 r = —.09*, but no significant contribution in multiple regression with 15 other variables Porter, Steers, Mowday, & Boulian (1974) Psychiatric technicians 60 Commitment had largest standardized weight in discriminant function, which included Job Descriptive Index Porter, Crampon, & Smith (1976) 32 r — .41* for Day 1 measure; r = .43* for measure in last 2 months with turnover data over 15 months; decline in commitment prior to actually leaving Steers (1977) Organizational involvement Mirvis & Lawler (1977) Job attachment Koch & Steers (1978) Management trainees Hospital employees 382 r = —.17** (1 year); commitment and intention to remain, r = .31**; desire to remain, r = .44** Bank tellers 160 r = -.29** (3 months) Nonmanagerial entry-level public agency employees M S W M c » 1 77 r = —.38**; r = —.40** with satisfaction partialed out * p < .05. ** p < .01. Cn O 508 MOBLEY, GRIFFETH, HAND, AND MEGLINO the more specific the intention measure and the closer the person is to actually quitting, the more trivial the prediction. Additionally, without analyses of the precursors of intentions, little knowledge of the psychology of turnover behavior is generated. Also included in Table 9 are two variables related to intentions in the Fishbein (1967) model: attitude toward the act of quitting and normative beliefs regarding quitting. The Newman (1974) study is one of the few that tests the Fishbein model with a turnover criterion. Both variables, as well as intentions, were significantly related to turnover. It is evident that intentions are a significant variable in the turnover process. However, additional research is required on the antecedent and covariates of intentions, the manner in which intentions change over time, and the reasons for the lack of a stronger relationship between intentions and turnover. Organizational commitment, involvement, and job attachment. A number of researchers have recently focused on the antecedents and consequences of organizational commitment (see, e.g., Porter et al. 1974; Steers, 1977). Porter et al. (1974, p. 604) defined organizational commitment as a more global evaluative linkage between the employee and the organization, which includes job satisfaction among its components. More specifically, organizational commitment was defined as the strength of an individual's identification with and involvement in a particular organization and is characterized by (a) a strong belief in and acceptance of an organization's goals and values, (b) a willingness to exert considerable effort on behalf of the organization, and (c) a definite desire to maintain organizational membership. Porter et al. stated that intention to remain is a component of commitment. More recently, Koch and Steers (1978) suggested that job attachment may be a primary precursor of turnover. They defined job attachment as an attitudinal response to one's job characterized by (a) a congruence between one's real and ideal jobs, (b) an identification with one's chosen occupation, and (c) a reluctance to seek alternative employment. Koch and Steers further noted that job attachment is clearly related to organizational commitment, although it focuses more specifically on one's occupation or job than on the organization as a whole, that job attachment should be more closely related to turnover than is satisfaction because of its conative or intentional component, and that it should be influenced relatively more by individual than by job characteristics. Table 10 summarizes recent studies of the relation between these variables and turnover. Porter et al. (1974), Porter et al. (1976), and Steers (1977) all found that commitment was more significantly and negatively related to turnover than was satisfaction. Marsh and Mannari (1977) discovered a significant but weak negative correlation between commitment and turnover among Japanese employees, whereas Mirvis and Lawler (1977) observed that organizational involvement, one component of commitment, was significantly and negatively related to turnover. Koch and Steers (1978) found job attachment to be significantly and negatively related to turnover. It should be noted that role orientation in the previously reviewed Graen et al. (1973) and Graen and Ginsburgh (1977) studies (perceived relevance of the job for the worker's career) is related to at least one aspect of Koch and Steers's definition of job attachment. The Graen studies found role orientation to be significantly and negatively related to turnover. The developing body of research on commitment and attachment suggests that these concepts are significantly and negatively related to turnover and more strongly related to turnover than to satisfaction. However, both commitment and attachment, as defined in the research cited above, are such complex constructs as to make generalizations rather tenuous. For example, is it the inclusion of intentions in the operational definitions of commitment and attachment that accounts for their improved prediction of turnover? Is it not possible for congruence between individual and organizational goals and values to vary independently of the other two components of commitment? Perhaps a more microanalytic treatment of these constructs Table 11 Studies of Relations Between Met Expectations and Turnover Study Population Dunnette, Arvey, & Banas (1973) College graduates Farr, O'Leary, & Bartlett (1973) Female sewing machine operators N Relation to turnover 1,020 Leavers had larger differences than stayers on a Vroom-type (1964) motivation index between expectations and job experiences on last job (significance levels not reported) White applicants administered a work sample test had lower voluntary turnover than group given traditional tests, p < .05; no differences for minority applicants (no direct measure of expectations) Those receiving booklet of realistic information had lower turnover than control group (no direct measure of expectations and treatment after acceptance decision) Met expectations was inconsistently related to turnover, although summary met expectations variable was significantly and negatively related to turnover (r = — .22**) Realistic compared with traditional job preview had lower expectations on Job Descriptive Index Work* and Supervision,** lower thoughts of quitting,** and lower measures on relevant Minnesota Satisfaction Questionnaire items and job survival (P < -20) 160 Ilgen & Seely (1974) New West Point cadets 468 Ilgen & Dugoni (Note 3) Retail clerks 117 Wanous (1973) New telephone company operators *p< .05. **p < .01. 80 M O M 510 MOBLEY, GRIFFETH, HAND, AND MEGLINO would prove useful. A model that incorporates some components of commitment and attachment is discussed in a subsequent section. Met expectations. Porter and Steers (1973) suggested that met expectations provide a conceptual framework for the diverse turnover literature. They viewed this concept as the discrepancy between what a person encounters on the job in the way of positive and negative experiences and what was expected. They predicted that "when an individual's expectations—whatever they are— are not substantially met, his propensity to withdraw would increace" (Porter & Steers, 1973,p. 152). Since the Porter and Steers review, there have been several studies relevant to the met expectations hypothesis. A subset of these studies dealt with expectations at the time of original organizational entry. Table 11 summarizes studies since the Porter and Steers review that are most relevant to met expectations. The Dunnette, Arvey, and Banas (1973) study found that leavers exhibited a greater discrepancy between original expectations and actual experiences than did stayers. However, significance levels were not reported and leavers' perceptions of their last job were retrospective, suggesting the possibility of postdecision distortion. Farr, O'Leary, and Bartlett (1973) and Ilgen and Seely (1974) found some evidence that individuals given realistic information about the job (via a work sample and a booklet) exhibited lower turnover. However, in neither study were expectations or subsequent experiences-directly assessed. Ilgen and Dugoni (Note 3) sought to assess directly the met expectation hypothesis, but found that met expectations were inconsistently related to satisfaction or turnover. Wanous (1973) discovered that a realistic job preview, compared with a more traditional orientation, lowered both expectations and thoughts of quitting, but did not significantly influence turnover. Direct support of the met expectations hypothesis is rather weak. In an insightful discussion of the conceptual and empirical support for met expectations (particularly as related to realistic job previews), Ilgen and Dugoni concluded that to expect realistic job previews to influence satisfaction, and subsequently turnover, through the mechanism of met expectations is naive. Specifically, the hypothesis is theorized to inadequately reflect individual differences in values. However, it should be noted that Porter and Steers (1973) appeared to account for individual differences through "desired expectations." Ilgen and Dugoni also noted that accurate expectations cannot compensate for deficiencies in the immediate job environment. Additionally, the met expectations hypothesis appears to give insufficient attention to the socialization and assimilation processes. Although expectations may play an important role in attachment, satisfaction, and turnover, a more complex conceptualization than the met expectations hypothesis appears necessary. One such conceptual model is proposed in a subsequent section of this article. Multivariate Studies The preceding sections of this review have repeatedly suggested that multivariate studies are necessary in turnover research. Such studies are necessary in order to interpret the relative efficacy of numerous variables and constructs thought to be related to turnover, to resolve apparently contradictory bivariate studies, to attempt to account for a greater proportion of the variance in turnover, and to move toward a more complete understanding of the turnover process. Table 12 summarizes recent multivariate studies that have used turnover as the criterion. Graen and Ginsburgh (1977) found that role orientation, leader acceptance, and their interaction accounted for 23% of the variance in university service employee turnover. That the interaction accounted for 6% of the variance suggests that noncompensatory models of turnover may be required. Mobley et al. (1978) tested a simplified version of a model of possible intermediate linkages between job satisfaction and turnover (Mobley, 1977). Although a number of Table 12 Summary of Recent Multivariate Studies of Turnover Population Study Graen & Ginsburgh (1977) Mangione (Note 2) Marsh & Mannari (1977) Newman (1974) Porter, Steers, Mowday, & Boulian (1974) Note. JDI = Job Descriptive Index. Independent variables and relation to turnover University service employees Mobley, Horner, & Hollingsworth (1978) Waters, Roach, & Waters (1976) N 89 Significant main effects for role orientation (10% variance explained) and leader acceptance (7% variance); significant interaction (6% variance), with resignations at 24 months Hospital employees 203 Intention to quit was the only significant regression coefficient (r2 = 24%) in equation that included intent to search, thinking of quitting, probability of finding acceptable alternative, age, tenure, and satisfaction (47 weeks) Institute for Social Research diverse 295 Using 15 demographic, satisfaction, and occupational varioccupational sample ables, R — .63 (shrunken r2 = 22%); rank ordered by betas, satisfaction with comfort, satisfaction with coworkers, industry, age, tenure, occupation, satisfaction with financial rewards, occupational prestige, satisfaction with challenge, education, marital status, resource adequacy, race, collar, and sex Japanese electrical company employees 784 R = .34 (16 variables); sex, —.27**; organizational status, — .08*; performance, .07*; number previous jobs, .06* (betas); nonsignificant variables: promotion chances, cohesiveness, participation, distance migrated, knowledge of procedures, values, size of community of origin, lifetime commitment, residence, autonomy, and satisfaction Nursing home employees 108 R = .48** for JDI scales, faces scale, attitude toward quitting, normative beliefs regarding quitting, and intentions to quit (not cross-validated; individual betas not reported) Psychiatric technicians 60 Organizational commitment and JDI with age partialed out; significant discriminant function at 2 times closest to actual termination (20.7%) and 21% variance related to termination decision Insurance company clerical * p < .05. ** p < .01. 105 R = 50, with intentions, JDI work, and tenure the only significant variables; age, other JDI scales, and marital status not significant; intentions accounted for 18% of variance; JDI work and tenure added 7% H fO § 512 MOBLEY, GRIFFETH, HAND, AND MEGLINO demographic, satisfaction, and perceived alternative measures exhibited significant bivariate relations with turnover, multiple regression analysis revealed intention to quit as the only significant coefficient (r2 = 24%). Mangione (Note 2) used IS demographic, satisfaction, and occupational variables to predict turnover. He found the strongest regression coefficients to be satisfaction with comfort, satisfaction with co-workers, rewards, industry type, age, tenure, occupation, and satisfaction with financial rewards (r- - 40%; adjusted r2 = 22%). Of particular interest was the fact that three different classes of variables (satisfaction, demographic, and occupational) were represented among the strongest regression coefficients. Satisfaction did not subsume the unique variance in the demographic and occupational variables. Unfortunately, this study did not include perception and evaluation of alternatives. The Marsh and Mannari (1977) study is of particular interest because it dealt with a Japanese sample. These authors found only four variables with significant coefficients (sex, organizational stature, performance, and number of previous jobs). Commitment and satisfaction were among those that did not exhibit significant regression coefficients. This study serves to emphasize the necessity of evaluating models that can generalize beyond the United States and Western industrialized nations. Newman (1974) conducted one of the few direct tests of the Fishbein (1967) model. Although individual regression coefficients were not reported, he discovered that 23% of the variance in turnover was accounted for by satisfaction, attitude toward quitting, normative beliefs regarding quitting, and actual intentions to quit. The multivariate study by Waters, Roach, and Waters (1976) found a coefficient of 25%. Intentions accounted for 18%, whereas the Job Descriptive Index (JDI) Work scale (Smith, Kendall, & Hulin, 1969) and tenure added the additional 7%. Porter et al. (1974) observed that organizational commitment and the JDI accounted for 21% of the variance in turnover at two different points in time with age partialed out. Commitment made the stronger contribution. Several generalizations are possible from these studies. First, each of the studies accounted for more variance in turnover than did satisfaction or any other of the single variables. Thus, satisfaction does not appear to be an adequate composite of other precursors and correlates of turnover. Also, it is evident that intentions, whether measured directly or included in commitment, enhance the prediction of turnover. With the exception of the Mobley et al. (1978) study, other variables when combined with intentions enhanced the prediction. It should be noted that few of the multivariate studies included either perception and evaluation of alternatives or cross-validation. These omissions continue to be major shortcomings of the research. Summary Employee turnover remains a frequently researched phenomenon. This is evident from the number of studies since the Porter and Steers (1973) review. Many of these studies have dealt with only a small subset of the variables potentially relevant to turnover, and many are not based on a clear conceptual model. This precludes making strong summary generalizations of the research studies. Table 13 presents a summary of the Porter and Steers (1973) review, the Price (1977) review, and the present authors' conclusions based on the recent research. In the placement of categories in Table 13 an attempt has been made at maintaining the integrity of the various authors' classification schema and at calling attention to possible overlap in classification groupings. In interpreting the table, negative refers to a negative relationship, that is, the higher the variable the lower the turnover, while positive refers to a positive relationship. In the case of normal variables, the nature of the relationship is specified. The qualifiers consistent, moderate, weak, or inconclusive are used in Table 13. These qualifiers refer to the consistency with which EMPLOYEE TURNOVER a significant relationship was found and to the relative number of studies reporting such a relationship. These qualifiers do not refer to the strength of a relationship in terms of the size of a correlation or variance explained. The present review, in agreement with the earlier reviews of Porter and Steers and Price, found age, tenure, overall job satisfaction, and reaction to job content to be consistently and negatively associated with turnover. Among the more recently studied variables, intentions and commitment-attachment were found to consistently relate to turnover. Because of the relatively few multivariate studies, an ordering of these variables in terms of relative contribution to turnover is tenuous. However, it appears that intentions and commitment-attachment (which includes intentions) made a stronger contribution to turnover behavior than did satisfaction and demographic variables. Further research is needed for an adequate mapping of the antecedents of intentions. Moderately consistent support for the negative relationship between supervisory style and turnover was evident, a somewhat more qualified conclusion than that reached by Porter and Steers (1973). Recent research reveals an inconclusive pattern of results with respect to pay, promotion, and peer group relations. These results stand in contrast with the consistent negative generalization of Porter and Steers's review. Differences in the availability of alternatives, the lack of multivariate studies, the lack of multiple measures of perception or affect, and the lack of a clear conceptual model make interpretation of these differences difficult. The compelling conceptual argument that alternatives are an important variable in the turnover process continues to be supported in aggregate-level studies, but has weak support at the individual level because it has been infrequently studied. Direct support for the met expectations hypothesis is weak. Although realistic job previews have been shown to be a possible aid in reducing turnover, the psychology of this effect is not well understood. 513 Finally, the limited number of multivariate studies indicates that greater variance in turnover can be explained by using multiple variables, that a great deal of variance is still unexplained, that inclusion of intentions significantly enhances the prediction of turnover, and that satisfaction is an inadequate summary variable for capturing the effects of other demographic, organizational, occupational, or external variables. Methodological and Conceptual Comments Predictive Designs It is encouraging to note that an overwhelming majority of recent research designs have been predictive rather than retrospective. However, few studies have used repeated measures of the perceptual or affective variables, the Porter et al. (1976) and Graen and Ginsburgh (1977) studies being substantive exceptions. To the extent that turnover is a dynamic process, longitudinal designs with repeated measures should be of high utility. Linear Models Most research has been based on the assumptions of a linear and compensatory model. Fleishman and Harris (1962) earlier demonstrated the possibility of a nonlinear relationship between supervisory style and turnover. Mangione (Note 2) was one of the few authors to examine possible nonlinear relationships. Additionally, the Graen and Ginsburgh (1977) finding of a significant interaction between role orientation and leader acceptance calls attention to the need for further exploration of interaction terms. Criterion A. troublesome issue in turnover research concerns the definition of turnover rates and types of turnover at both the aggregate and individual levels (see Price, 1975-1976,1977, for an evaluation of a number of aggregate measures of turnover). At the individual level, one of the more troublesome issues is the distinction between voluntary and involuntary turnover. The bulk of the individual- Table 13 Summary of Three Reviews of the Turnover Literature Variable Personal characteristic Age Tenure Similarity of job with vocational interests Personality Family size and responsibilities Sex Education Weighted application blank Overall job satisfaction Organizational and job characteristic Pay Promotion Size of organization Size of work unit Peer group interaction Integration Supervision style Instrumental communication Formal communication Role clarity Job autonomy and responsibility Centralization Task repetitiveness Overall reaction to job content Occupational grouping Blue collar: skilled vs. unskilled Blue collar vs. white collar Nonmanagers vs. managers Nongovernment vs. government Professional ism Porter & Steers (1973) Consistent negative Consistent negative Weak negative Weak negative for extreme traits Generally positive for women; inconclusive for males Price (1977) Consistent negative Consistent negative Inconclusive Weak positive Consistent negative Consistent negative Consistent negative Consistent negative Inconclusive Consistent positive for blue collar; inconclusive for white collar Moderate negative Consistent negative Weak negative Inconclusive Inconclusive Present review Consistent negative Consistent negative Inconclusive Inconclusive Moderate positive Consistent negative g O 60 f M O M H Inconclusive Inconclusive 2! a Inconclusive Consistent negative Consistent negative Consistent negative § W 0 r § Consistent negative Consistent negative Moderate positive Consistent negative f 2. O Moderate negative Consistent negative a 0 Consistent negative Weak positive Consistent negative Moderate (unskilled higher) Moderate (blue collar higher) Weak (nonmanagers higher) Weak (nongovernment higher) Weak positive (professionalism higher) EMPLOYEE TURNOVER j! > '£ SoD 4>> 'S a; a.-*^ §° I-•c S? c S <u «J S'> C <u CL, i- & c $• t 0 C C Jf t (n "« C e cti s u s t-^ <u OJ .° .^4 £ a c u in C O U 1o> U to OH 515 level turnover research focuses on voluntary turnover. Precise definitions of voluntary turnover are infrequently given, and what is included as voluntary may differ across studies. For example, Marsh and Mannari (1977) incorporated pregnancy under voluntary turnover, whereas Mirvis and Lawler (1977) and Waters et al. (1976) excluded pregnancy. The definition of voluntary may well have contributed to Marsh and Mannari's finding of a significant difference in turnover as a function of sex. Much more subtle effects may be associated with results as a function of the definition of voluntary. The categorization of reasons for turnover is frequently taken from company records. Many personnel practitioners readily admit that a variety of factors influence the administratively recorded reason for attrition. Lefkowitz and Katz (1969) reported significant differences in administrative and selfreported reasons for termination. In agreement with Forrest et al. (1977), further efforts to clarify the implications of different operational measurements of turnovers are appropriate. A multiple measure approach to identifying reasons for turnover would be useful. Finally, little research has addressed the relationship between voluntary and involuntary turnovers. To assume that these are completely independent phenomena, especially in the case of discipline-related terminations, appears simplistic. Although turnover is frequently thought of as a "clean" objective criterion, the issues raised above suggest the need for greater attention to the criterion problem in turnover research. 'c 3 Measures of Satisfaction V 3 rt H 4-1 C studied variabl ons to quit itment/attachm pectations "V T-H XI .2 ao.o environment >f employment/ ed r alternatives ! (continued) 4-) 111 4) 4) <y f'l | ; w P4 In recent years, the JDI (Smith et al., 1969) has become the predominant measure of satisfaction with various facets of the job setting. The majority of satisfaction-turnover studies reviewed in this article used the JDI. The careful development of the JDI is well documented, and there is a clear advantage to using a common satisfaction measure across a variety of studies. However, 516 MOBLEY, GRIFFETH, HAND, AND MEGLINO overreliance on any single measure raises the possibility that method variance has contaminated supposedly generalizable relationships. As Gillet and Schwab (197'5) suggested, it seems prudent to use multiple measures of the same construct wherever possible. Time The role of time in turnover research is evident in a number of ways. As noted earlier, there is a consistent negative relationship between tenure (length of time on the job) and turnover. Some studies have ignored tenure, some have partialed out its effect, and others have included it in a multivariate design. Understanding the psychology of the tenure effect is probably best facilitated by the latter treatment. The time variable is also part of the criterion problem to the extent that different studies measure turnover over different lengths of time. Marsh and Mannari (1977) collected their turnover data over a 4-year period, whereas other studies have looked at turnover over a matter of weeks, for example, Newman (1974). The effect of differing lengths of time between measurement of independent variables and the turnover behavior is infrequently studied. Porter et al. (1976) and Waters et al. (1976) are exceptions. This appears to be a topic in need of additional research. Finally, the temporal dimension may be relevant to the extent that different variables or combinations of variables exert a differential influence on turnover as a function of stages in the organizational socialization process. Graen and Ginsburgh (1977) discussed this possibility. Primacy of Work Turnover is generally conceptualized in terms of demographic, organizational, and individual affective factors and on infrequent occasions in terms of perceived alternatives. While such conceptualizations may reflect individual values relative to the work setting, they do not reflect the importance of work- related values relative to other life values and interests. The work of Dubin and his associates (see, e.g., Dubin, Champoux, & Porter, 197S) has demonstrated that differences in central life interests are related to differences in evaluations of the work environment and in organizational commitment. Marsh and Mannari (1977) found a significant negative relationship between primacy of work values and turnover. It appears that future turnover research should deal not only with the work environment and external alternatives but also with the centrality of work relative to other life values and interests. Conceptual Model of Employee Turnover Drawing in part on the review and analysis presented earlier, this section develops a conceptual model of the employee turnover process. A simplified schematic representation of this model is presented in Figure 1. Among the characteristics of this model are the following: 1. It is a model of individual-level turnover behavior. Individual differences in perceptions, expectations, and values are explicitly recognized. Further, individual differences in personal and occupational variables are included. 2. Perception and evaluation of alternative jobs is given explicit treatment. 3. The probable roles of centrality of work values and interests relative to other values and interests, beliefs regarding nonwork consequences of quitting or staying, and contractual constraints are specifically recognized. 4. The possible joint contribution to turnover of job satisfaction (present affect), job attraction (expected future affect), and attraction of attainable alternatives is proposed. 5. Intention to quit is considered to be the immediate precursor of turnover, with impulsive behavior and ithe time between measurement of intentions and behavior attenuating this relationship. The rationale for the model is described starting with turnover behavior and working back through its antecedents. EMPLOYEE TURNOVER 517 Individual Hierarchical level Skill Level Status Professionalism Perionol Age Tenure Education Interest! Personality Soclo-Economlc Family Responsibility Aptitude Economic-Labor Mortcet Vacancy Rates Advertising Levels Recruiting Levels word ol Mouth Communication EXPECTATIONS RE: ALTERNATIVE JOBS: 1. Expectancies re: future job outcomes; 2. Expectancy re: attaining alternative. Centrallty of non-work values; Beliefs re: non-work consequences of quitting: Conttoetuol eonitrolnlt Alternative forms ol withdrawal behavior Immediate vs. delayed gratification Impulsive behavior; Specificity ft time between measures Figure 1. A schematic representation of the primary variables and process of employee turnover. Intentions The immediate precursor of behavior is thought to be intentions (Dulaney, 1961, 1968; Fishbein & Ajzen, 1975; Locke, 1968; Mobley, 1977; Ryan, 1970). Therefore, the best predictor of turnover should be intention to quit (see, e.g., Kraut, 1975; Mobley et al., 1978; Newman, 1974; Waters et al., 1976). The relationship between turnover and intention should be stronger the more specific the intention statement and the closer in time the measurement of the intention and the behavior. Impulsive behavior attenuates the intention-behavior relationship. Graen and Ginsburgh (1977) observed that the more specific and closer to the act the intention measure comes, the more trivial the prediction. Although probably valid, this observation should not be interpreted as indicating that understanding the turnover process is not facilitated by including intentions and evaluating their precursors. Neither should the personnel-planning utility of assessing even more distant intentions (see, e.g., Kraut, 1975) be overlooked. There are at least two intentions of interest, intention to search and intention to quit. Mobley (1977) suggested that intention to search and search behavior should generally precede intention to quit and turnover. Exceptions include impulsive behavior and nonsolicited attractive alternatives. Lack of perceived attractive alternatives or an un- 518 MOBLEY, GRIFFETH, HAND, AND MEGLINO successful search may lead to forms of withdrawal other than turnover intentions and behavior. Relations among alternative forms of withdrawal and the effects of no alternative or an unsuccessful search continue to need additional research. The primary determinants of intentions are thought to be (a) satisfaction, (b) attraction expected utility of present job, and (c) attraction expected utility of alternative jobs or roles. Satisfaction The present review and previous reviews (e.g., Locke, 1976; Porter & Steers, 1973; Herzberg et al., Note 1) have documented the consistent and negative, although moderate, relationship between job satisfaction and turnover. Satisfaction is seen as the affective response to evaluation of the job. This evaluation is considered to be a function of perceptions o.f various aspects of the job relative to individual values (Locke, 1969, 1976). It is important to note that satisfaction is present rather than future oriented. The behavioral implication of satisfactiondissatisfaction is a tendency toward approach-avoidance. However, whether this approach-avoidance tendency is expressed in the form of turnover is thought to be related to at least three other classes of variables: attraction expected utility of the present role, attraction expected utility of attainable alternative roles, and centrality of work values, beliefs regarding nonwork consequences of quitting-staying, and contractual constraints. Failure to consider these classes of variables may explain the absence of a stronger relationship between job satisfaction and turnover. Attraction and Expected Utility of Present Job Whereas satisfaction is present oriented, attraction is considered to be future oriented. Attraction is seen as being based on the expectancies that the job will lead to future attainment of various positively and negatively valued outcomes. When combined with the expectancy of being able to retain the present job, an index can be generated that is analogous to Vroom's (1964) force for a single alternative and the expected utility index of Dachler and Mobley (1973) or Graen (1976). The concept of expected utility is applied in a variety of models. For example, economists and decision theorists frequently use expected utility and expected value, concepts analogous to the expected utility previously described (see March & Simon, 1958, pp. 137-138). In one interdisciplinary analysis, Blau, Gustad, Jessor, Parnes, and Wilcox (1956) conceptualized the evaluation of occupational alternatives as "the individual's valuation of the rewards offered by different alternatives and his appraisal of his chances of being able to realize each of the alternatives" (p. 533). It is thought that attraction expected utility of the present role, like satisfaction, contributes to the tendency toward approachavoidance. Although many studies have analyzed the satis faction-turnover relationship, the dual contribution of satisfaction and attraction expected utility to turnover has not been researched. While there may be some correlation between satisfaction and attraction expected utility, these variables are conceptually distinct and should have separate effects on intentions (to search or to quit) and turnover. For example, individuals may be satisfied (or dissatisfied) with their present job, but may expect the present job to be relevant (or irrelevant) to their subsequent career. Graen and Ginsburgh (1977) found the latter belief to be significantly related to resignation. On the other hand, one may be dissatisfied with one's work group but be attracted to it because of expectations that it will facilitate the future attainment of valued outcomes or goals. The above relationships can easily be extended to include a variety of other job factors, for example, supervision, benefits, and so forth. Just as there are multiple dimensions of satisfaction, there are multiple dimensions of attraction. The salience of these different dimensions is a function of individual differences in values. The values may be related to occupation, position level, age, tenure, and other personal variables. EMPLOYEE TURNOVER 519 One dimension of satisfaction and attraction concerns organizational goals and values. The congruence between individual and organizational goals and values has been defined by Porter et al. (1974) and Steers (1977) as one component of organizational commitment. As noted earlier, these authors also included willingness to exert effort, desire and intention to remain in the organization, and job satisfaction in the definition of commitment. The model suggested here seeks to subdivide the complex variable of commitment. Congruence between individual and organizational goals and values may be an important variable, but can be seen as distinct from, and only one of several focuses of, both satisfaction (present) and attraction (future), which are in turn related to turnover intentions and behavior. If both satisfaction and attraction expected utility contribute to intentions to search and quit, and to turnover, then it is necessary to analyze the conditions under which one or the other makes the most contribution to variance explained. It may be that individual-level variables such as the need for immediate versus delayed gratification (see, e.g., Mischel, 1976) will aid in predicting whether satisfaction (present) or attraction (future) is most strongly related to turnover intentions and behavior for given individuals. The present model suggests that it is not merely the visibility of alternatives but the attraction of alternatives and the expectancy of attaining the alternatives that are most salient. Forrest et al. (1977), Mobley (1977), and Schneider (1976) are among recent authors who make a strong argument for inclusion of the variable attraction of alternatives in turnover research. Attraction of alternatives is defined in terms of expectations that the alternative will lead to the future attainment of various positively and negatively valued outcomes. When combined with the expectancy of being able to attain the alternative, an index can be generated that is analogous to Vroom's (1964) force for a single alternative and to the expected utility index of Dachler and Mobley (1973) or Graen (1976). (See Mobley, 1977, for a microanalytic treatment of the possible role of alternatives in search and turnover intentions and behavior.) As noted in Figure 1, there may well be some covariation among satisfaction, attraction expected utility of the present job, and attraction expected utility of alternatives. This is to be expected, since values are common to all three and the presence or absence of attractive alternatives may result in the revaluation of one's satisfaction with or the attraction of the present role. Attraction and Expected Utility of Alternatives Moderating Variables Considering both satisfaction and attraction expected utility should increase our understanding and prediction of turnover intentions and behavior. However, the attraction and attainability of alternative jobs or roles must also be considered. March and Simon (1958) presented an organizational participation model that gave a prominent role to visibility of alternatives. The March and Simon components of "perceived desirability of leaving the organization" and "perceived ease of movement from the organization" (p. 93) roughly correspond to "expectancies re: future job outcomes" and "expectancy re: attaining alternative" in Figure 1. As noted earlier, these variables have received little attention in turnover research. Although satisfaction, attraction expected utility of the present job, and attraction expected utility of alternatives are considered to be the primary determinants of turnover intentions and behavior, several other variables can be expected to moderate these relationships. To the extent that nonwork values and interests are not central to an individual's life values and interests (Dubin et al., 1975) and to the extent that an individual associates significant nonwork consequences with quitting (see, e.g., Newman, 1974), the relationships among satisfaction, attraction, and turnover intentions and behavior will be attenuated. Additionally, to the extent that an individual is bound by a contract, as for example in professional sports, the military, and certain professions, 520 MOBLEY, GRIFFETH, HAND, AND MEGLINO the relationships will be attenuated during the term of the contract. Under such circumstances, it can be hypothesized that the individual who is dissatisfied, perceives little attraction in the present job, or perceives an attractive alternative may engage in other forms of avoidance and withdrawal behavior. These suggested moderating influences, especially nonwork values and interests and nonwork consequences of turnover behavior, call attention to the need to look beyond the work setting for a complete understanding of the psychology of the turnover process. Antecedents The antecedents of satisfaction and attraction are considered to be organizational variables as perceived by the individual, economic variables related to availability of alternatives as perceived by the individual, and individual-level occupational and personal variables as they influence individual values, perceptions, and expectations. Although a detailed analysis of these antecedents is beyond the scope of this article, it is important to emphasize that the influence of various organizational, economic or labor market, occupational, and personal variables is through individual perceptions, expectations, and values. Research Implications The model described here indicates the need for multivariate research on the turnover process. As noted in the review section of this article, although the negative relationships between both age and tenure and turnover are well established, the amount of variance explained is low and the psychology of the relationships is not well understood. The model proposed here suggests that multivariate research that concurrently assesses values, job-related perceptions, external perceptions, and the previously mentioned moderating variables should facilitate an understanding of the relationship of age and tenure to turnover. Similarly, multivariate research that concurrently assesses individual-level occupa- tional and personal variables, job-related perceptions, external perceptions, individual values, and potential moderating variables provides a framework for integrating and understanding, at the individual level, the aggregate-level effects of various organizational and economic or labor market variables summarized by Price (1977). Graen (1976), among others, noted that neither individuals nor organizations are fixed or static; neither is the economy or labor market. The clear implication is that understanding the turnover process will require longitudinal as well as multivariate research. Longitudinal research, not simply in terms of the collection of criterion data over time but also in terms of repeated measures of the independent variables, as recently exemplified by Graen and Ginsburgh (1977) and Porter et al. (1976), is needed. Conclusions In 1973 Porter and Steers observed that the then existing body of research left much to be understood about the psychology of the employee withdrawal process. Review of the subsequent research leads to a similar observation. The conceptual model suggested here calls attention to the possible main effects of satisfaction (present oriented), attraction expected utility of the current role (future oriented), and attraction expected utility of alternative roles. A number of moderating variables and constraints were suggested. The need for integrative, multivariate longitudinal research is evident if significant progress is to be made in understanding the psychology of the employee turnover process. Reference Notes 1. Herzberg, F., Mausner, B., Peterson, R. O., & Capwell, R. F. Job attitudes: Review of research and opinions. Pittsburgh, Pa.: Pittsburgh Psychological Services, 1957. 2. Mangione, T. W. Turnover—Some psychological and demographic correlates. In R. P. Quinn & T. W. Mangione (Eds.), The 1969-1970 survey of working conditions. Ann Arbor: University of Michigan, Survey Research Center, 1973. 3. Ilgen, D. R., & Dugoni, B. L. Initial orientation to the organization: Its impact on psychological EMPLOYEE TURNOVER processes associated with the adjustment of new employees. Kissimee, Fla.: Academy of Management, August 1977. References Armknecht, P. A., & Early, J. F. Quits in manufacturing: A study of their causes. Monthly Labor Review, 1972, 95, 31-37. Blau, P. M., Gustad, J. W., Jessor, R., Parnes, H. S., & Wilcox, R. C. Occupational choice: A conceptual framework. Industrial and Labor Relations Review, 1956, 8, 531-543. Brayfield, A. H., & Crockett, W. H. Employee attitudes and employee performance. Psychological Bulletin, 19SS, 52, 396-424. Burke, R. J., and Wilcox, D. S. Absenteeism and turnover among female telephone operators. Personnel Psychology, 1972, 25, 639-648. Cascio, W. F. Turnover, biographical data, and fair employment practice. Journal of Applied Psychology, 1976, 61, S76-S80. Dachler, H. P., & Mobley, W. H. Construct validation of an instrumentality-expectancy-taskgoal model of work motivation. Some theoretical boundary conditions. Journal of Applied Psychology, 1973, 58, 397-418. Dansereau, F., jr., Cashman, J., & Graen, G. Expectancy as a moderator of the relationship between job attitudes and turnover. Journal oj Applied Psychology, 1974, 59, 228-229. Dubin, R., Champoux, J., & Porter, L. Central life interests and organizational commitment of blue collar and clerical workers. Administrative Science Quarterly, 197S, 20, 411-421. Dulaney, D. E., Jr. Hypotheses and habits in verbal "operant conditioning." Journal of Abnormal and Social Psychology, 1961, 63, 251-263. Dulaney, D. E. Awareness, rules, and prepositional control: A confrontation with S-R behavior theory. In T. R. Dixon & D. L. Horton (Eds.), Verbal behavior and general behavior theory. Englewood Cliffs, N.J.: Prentice-Hall, 1968. Dunnette, M. D., Arvey, R. D., & Banas, P. A. Why do they leave? Personnel, 1973, 50, 25-38. Ekpo-Ufot, A. Self-perceived abilities relevant in the task (SPART) : A potential predictor of labor turnove • r in an industrial work setting. Personnel Psychology, 1976, 29, 405-416. Farr, J. L., O'Leary, B. S., & Bartlett, C. J. Effect of work sample test upon self-selection and turnover of job applicants. Journal of Applied Psychology, 1973, 58, 283-285. Federico, J. M., Federico, P., & Lundquist, G. W. Predicting women's turnover as a function of extent of met salary expectations and biodemographic data. Personnel Psychology, 1976, 29, 559-566. Fishbein, M. Attitude and the prediction of behavior. In M. Fishbein (Ed.), Readings in attitude theory and measurement. New York: Wiley, 1967. 521 Fishbein, M., & Ajzen, I. Belief, attitudes, intention and behavior. Reading, Mass.: Addison-Wesley, 1975. Fleishman, E. A., & Harris, E. F. Patterns of leadership behavior related to employee grievances and turnover. Personnel Psychology, 1962, IS, 43-56. Forrest, C. R., Cummings, L. L., & Johnson, A. C. Organizational participation: A critique and model. Academy of Management Review, 1977, 2, 586601. Gillet, B., & Schwab, D. P. Convergent and discriminant validities of corresponding Job Descriptive Index and Minnesota Satisfaction Questionnaire scales. Journal oj Applied Psychology, 1975, 60, 313-317. Goodman, P. S., Salipante, P., & Paransky, H. Hiring, training, and retraining the hard-core unemployed: A selected review. Journal of Applied Psychology, 1973, 58, 23-33. Graen, G. Role-making processes within complex organizations. In M. D. Dunnette (Ed.), Handbook of industrial and organizational psychology, Chicago: Rand McNally, 1976. Graen, G., & Ginsburgh, S. Job resignation as a function of role orientation and leader acceptance: A longitudinal investigation of organization assimilation. Organizational Behavior and Human Performance, 1977, 19, 1^17. Graen, G. B., Orris, J. B., & Johnson, T. W. Role assimilation processes in a complex organization. Journal of Vocational Behavior, 1973, 3, 395-420. Hellriegel, D., & White, G. E. Turnover of professionals in public accounting: A comparative analysis. Personnel Psychology, 1973, 26, 239-249. Herman, J. B., Dunham, R. B., & Hulin, C. L. Organizational structure, demographic characteristics and employee responses. Organizational Behavior and Human Performance, 1975, 13, 206232. Herman, J. B., & Hulin, C. L. Studying organizational attitudes from individual and organizational frames of reference. Organizational Behavior and Human Performance, 1972, 8, 84-108. Hill, J. M., & Trist, E. L. Changes in accidents and other absences with length of service: A further study of their incidence and relation to each other in an iron and steel works. Human Relations, 1955, 8, 121-152. Hines, G. H. Achievement motivation, occupations, and labor turnover in New Zealand. Journal of Applied Psychology, 1973, 58, 313-317. Ilgen, D. R., & Seely, W. Realistic expectations as an aid in reducing voluntary resignations. Journal of Applied Psychology, 1974, 59, 452-455. Karp, H. B., & Nickson, J. W., Jr. Motivator-hygiene deprivation as a predictor of job turnover. Personnel Psychology, 1973, 26, 377-384. Koch, J. L., & Steers, R. M. Job attachment, satisfaction, and turnover among public sector employees. Journal of Vocational Behavior, 1978, 12, 119-128. 522 MOBLEY, GRIFFETH, HAND, AND MEGLINO Kraut, A. I. Predicting turnover of employees from measured job attitudes. Organizational Behavior and Human Performance, 1975, 13, 233-243. Lee, R., & Booth, J. M. A utility analysis of a weighted application blank designed to predict turnover -for clerical employees. Journal of Applied Psychology, 1974, SO, 516-518. Lefkowitz, J., & Katz, M. L. Validity of exit interviews. Personnel Psychology, 1969, 22, 445-455. Locke, E. A. Toward a theory of task motivation and incentives. Organizational Behavior and Human Performance, 1968, 3, 157-189. Locke, E. A. What is job satisfaction? Organizational Behavior and Human Performance, 1969, 4, 309-336. Locke, E. A. The nature and consequences of job satisfaction. In M. D. Dunnette (Ed.), Handbook of industrial and organizational psychology. Chicago: Rand-McNally, 1976. March, J. G., & Simon, H. A. Organizations. New York: Wiley, 1958. Marsh, R., & Mannari, H. Organizational commitment and turnover: A predictive study. Administrative Science Quarterly, 1977, 22, 57-75. Mirvis, P. H., & Lawler, E. E., III. Measuring the financial impact of employee attitudes. Journal of Applied Psychology, 1977, 62, 1-8. Mischel, W, Introduction to personality. New York: Holt, Rinehart & Winston, 1976. Mobley, W. H. Intermediate linkages in the relationship between job satisfaction and employee turnover. Journal of Applied Psychology, 1977, 62, 237-240. Mobley, W. H., Homer, S. O., & Hollingsworth, A. T. An evaluation of precursors of hospital employee turnover. Journal of Applied Psychology, 1978, 63, 408-414. Newman, J. E. Predicting absenteeism and turnover: A field comparison of Eishbein's model and traditional job attitude measures. Journal of Applied Psychology, 1974, 59, 610-615. Pettman, B. D. Some factors influencing labour turnover: A review of the literature. Industrial Relations Journal, 1973, 4, 43-61. Porter, L. W., Crampon, W. J., & Smith, F. J. Organizational commitment and managerial turnover: A longitudinal study. Organizational Behavior and Human Performance, 1976, 15, 87-98. Porter, L. W., & Steers, R. M. Organizational, work, and personal factors in employee turnover and absenteeism. Psychological Bulletin, 1973, 80, 151-176. Porter, L. W., Steers, R. M., Mowday, R. T., & Boulian, P. V. Organizational commitment, job satisfaction, and turnover among psychiatric technicians. Journal of Applied Psychology, 1974, 59, 603-609. Price, J. L. The measurement of turnover. Industrial Relations Journal, 1975-1976, 6, 33-46. Price, J. L. The study of turnover. Ames: Iowa State University Press, 1977. Rice, A. K., & Trist, E. L. Institutional and subinstitutional determinants of change in labor turnover. Human Relations, 1952, 5, 347-372. Robinson, W. S. Ecological correlations and the behavior of individuals. American Sociological Review, 1950, IS, 351-357. Ryan, T. A. Intentional behavior: An approach to human motivation. New York: Ronald Press, 1970. Schneider, J. The "greener grass" phenomenon: Differential effects of a work context alternative on organizational participation and withdrawal intentions. Organizational Behavior and Human Performance, 1976,16, 308-333. Schwab, D. P., & Oliver, R. L. Predicting tenure with biographical data: Exhuming buried evidence. Personnel Psychology, 1974, 27, 125-128. Smith, P. C., Kendall, L. M., & Hulin, C. L. The measurement of satisfaction in work and retirement. Chicago: Rand McNally, 1969. Steers, R. M. Antecedents and outcomes of organizational commitment. Administrative Science Quarterly, 1977, 22, 46-56. Vroom, V. H. Work and motivation. New York: Wiley, 1964. Wanous, J. P. Effects of a realistic job preview on job acceptance, job attitudes, and job survival. Journal of Applied Psychology, 1973, 58, 327332. Waters, L. K., & Roach, D. Job attitudes as predictors of termination and absenteeism: Consistency over time and across organizational units. Journal of Applied Psychology, 1973, 57, 341-342. Waters, L. K., Roach, D., & Waters, C. W. Estimate of future tenure, satisfaction, and biographical variables as predictors of termination. Personnel Psychology, 1976, 29, 57-60. Woodward, N. The economic causes of labour turnover: A case study. Industrial Relations Journal, 1975-1976, 6, 19-32. Received January 12, 1978 •