Behavioral Intentions as Predictors of Job Attitudes - APA PsycNET

8 downloads 0 Views 670KB Size Report
A. B. Freeman School of Business. Tulane University. Jennifer M. George. Department of Management. Texas A&M University. Hypotheses regarding behavioral ...
Copyright 1991 by the American Psychological Association, Inc. 0021-9010/91/S3.00

Journal of Applied Psychology 1991. Vol. 76, No. 1,40-45

Behavioral Intentions as Predictors of Job Attitudes: The Role of Economic Choice Veronica K. Stone

Lucinda I. Doran

New \brk University

College of Business Administration Northeastern University

Jennifer M. George

Arthur P. Brief

Department of Management Texas A&M University

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

A. B. Freeman School of Business Tulane University

Hypotheses regarding behavioral intentions as antecedents of job attitudes were tested longitudinally with a sample of 126 retail salespeople. Hierarchical regression analysis revealed that workers' intentions to leave at organizational entry predicted subsequent job satisfaction on 19 of 21 satisfaction scales. In general, these relationships were moderated by perceived choice, construed here as the absence of externally imposed financial requirements or economic pressures to stay on the job. Consistent with cognitive dissonance theory, the intent-to-leave-job-satisfaction relationships were stronger when economic choice was higher (i.e., financial requirements were lower). The practical and theoretical implications of these findings for existing models of organizational behavior are discussed.

dissonance. The core of this theory is as follows: (a) Dissonance exists when an individual holds a cognition that is inconsistent with his or her other cognitions in the same domain; (b) dissonance will give rise to measures to reduce, as well as to avoid increases in, the dissonance; and (c) one way in which the individual can reduce dissonance is by altering the discrepant cognition to bring it in line with his or her other cognitions. In other words, individuals strive for consistency in their cognitions. Note that the term cognition as used by Festinger pertains to "any knowledge, opinion, or belief about the environment, about oneself, or about one's behavior" (1957, p. 3). Thus, it may be said that individuals strive for consistency in cognitions about their attitudes, behavioral intentions, and behavior. In applying Festinger's (1957) theory, we proposed that workers' initial intentions to quit are systematically related to their subsequent job attitudes. Specifically, we predicted that workers who entered an organization with the intent to leave would subsequently report lower levels of job satisfaction than would individuals who entered with the intent to stay. In addition, we proposed that this intention-satisfaction relationship is a moderated one. The moderating construct of interest, the perception of free choice, is derived from Brehm and Cohen's (1962) extension of cognitive dissonance theory (cf. Cooper & Fazio, 1984). Brehm and Cohen asserted that, for dissonance to be aroused, a "person must feel he has some control about whether or not to put himself into a situation in which inconsistent cognitive elements occur" and, furthermore, that the magnitude of dissonance is a direct function of the degree of this choice (Brehm & Cohen, 1962, p. 203). (Also see, for example, Davis & Jones [1960], Under, Cooper, & Jones [1967], and Sherman [1970].) Thus, pressures toward consistency are greater in situations in

Breast-feeding, marijuana smoking, and church attendance represent a few of the behaviors that social psychologists have found can be predicted from behavioral intentions (Ajzen, Timko, & White, 1982; King, 1975; Manstead, Proffitt, & Smart, 1983). The conceptual importance of such behavioral intentions is captured, for example, in the influential theory of reasoned action (Ajzen & Fishbein, 1980; Fishbein & Ajzen, 1975). That is, "according to the theory of reasoned action, attitudes follow reasonably from the beliefs people hold about the object of the attitudes, just as intentions and actions follow reasonably from attitudes" (Ajzen, 1988, p. 32). The efficacy of the notion that behavioral intentions mediate attitude-behavior relationships is evidenced by the fact that this idea has been incorporated into several models of organizational behavior. For example, Mobley, GrifFeth, Hand, and Meglino (1979) postulated that intent to quit is the immediate precursor of turnover and that a primary determinant of this intention is job satisfaction. The focus of attention in this study also was on behavioral intentions; however, we construed intent to be, not a consequence of attitudes, but an antecedent. Our conception of the intentions-attitudes relationship—or more precisely the intent-to-quit-job-satisfaction relationship —is influenced heavily by Festinger's (1957) theory of cognitive

We would like to thank Ray Aldag, Joan Brett, Rob Folger, and two anonymous reviewers for their comments on earlier versions of this article. Correspondence concerning this article should be addressed to Arthur P. Brief, A. B. Freeman School of Business, Tulane University, New Orleans, Louisiana 70118-5669, or to Lucinda I. Doran, College of Business Administration, Northeastern University, Boston, Massachusetts 02115.

40

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

ECONOMIC CHOICE

which perceived choice is high than when it is low. Recent evidence suggests that such a choice effect also may hold when aroused dissonance is due to the felt responsibility for aversive consequences, regardless of cognitive inconsistency (Scher & Cooper, 1989). That is, perceptions of free choice and felt responsibility for aversive consequences can arouse dissonance even in the absence of a discrepancy. In line with Steers and Mowday (1981), we construed perceived choice in the present context as the absence of externally imposed constraints or pressures to stay on the job. Examples of such constraints include "economic constraints (e.g., investments in the pension system), family constraints (e.g., dual career families), [and] no available alternatives" (Steers & Mowday, 1981, p. 258). In the present context, the constraint of interest was the worker's financial requirements (i.e., money needed to support oneself and, if any, one's dependents). We chose to examine financial requirements as an external constraint because many, if not most, workers engage in paid work to survive economically or to establish a sense of financial security. In other words, work serves important economic functions in people's lives (e.g., Brief & Aldag, 1989; Brief & Atieh, 1987; Fein, 1976; George & Brief, in press; Neff, 1968; Nord, 1977; Strauss, 1963; Wool, 1973). However, workers' economic dependency on their jobs varies. For example, one worker may be dependent on earned income for basic subsistence level necessities, such as food and shelter, whereas another worker's job-related income may be used primarily for luxury or discretionary items. Essentially, workers with higher financial requirements may be more economically dependent on their job-related income. Hence, workers with low financial requirements ought to perceive more choice in employment-related decisions—because their economic circumstances are less of a constraint— than will workers with high financial requirements. Put simply, perceptions of free choice in employment contexts are likely to increase as financial requirements decrease. Thus, following Brehm and Cohen (1962), one would expect there to be greater pressures for consistency between initial intentions and subsequent attitudes for workers with relatively low financial requirements. Conversely, workers with high financial requirements may not be driven to form attitudes consonant with initial intentions because their perceptions of free choice in the employment context are likely to be lower. Therefore, we predicted that the relationship between turnover intentions at organizational entry and subsequent job satisfaction would be stronger for workers with low financial requirements than for workers with high financial requirements. In summary, two hypotheses have been advanced: (a) Workers' intentions to leave at organizational entry are negatively related to subsequent job satisfaction; and (b) this relationship is stronger for workers whose financial requirements are lower. Support for the first hypothesis would indicate that a reformulation of some current theories of organizational behavior, most especially those pertaining to employee turnover, may be necessary (such a reformulation would have to account for a reciprocal relationship between job attitudes and behavioral intentions). This is one of several possible implications of our hypotheses.

41

Method Subjects and Setting The data for this research came from a larger study conducted in two locations of a major department store in the northeastern United States. The design of the larger study called for all salespeople hired during the fall of 1987 and spring of 1988 to be administered several questionnaires; the first of these (Time 1 questionnaire) was administered the morning of each new employee's first day of employment. The Time 2 questionnaire was administered after each employee had gained experience on the selling floor, approximately 4 weeks posthire. The initial sample comprised 332 salespeople (80% were women). However, because of turnover (17% of the original subjects left the organization before the administration of the second questionnaire) and a relatively low response rate to the second questionnaire (45%, largely due to administrative problems), the sample for the present study fell to 126 (81% were women). Subjects' mean age was 28.2 years (SD = 10.2), and 25% of the subjects were married. There were no substantive demographic differences between subjects at the two store locations nor between respondents and still-employed nonrespondents.

Measures Intent to leave. The three-item Intention to Turn Over subscale of the Michigan Organizational Assessment Questionnaire (Cammann, Fichman, Jenkins, & Klesh, 1979; Seashore, Lawler, Mirvis, & Cammann, 1982) was used to measure intent to leave at Time 1. Responses to the items were made on a 7-point scale. The subscale mean was 2.31 (SD = 1.46); coefficient alpha (Cronbach, 1951) was .76. Financial requirements. An additive index, adapted from George and Brief (in press), was used to measure financial requirements at Time 1. The index takes into account each respondent's marital status, spouse's employment status, number of children aged 22 and under, and housing arrangements. These variables were all measured with single-item, objectively worded questions. Following the logic of George and Brief (in press), we viewed single respondents and married respondents whose spouses worked full-time as having fewer financial requirements than married respondents whose spouses worked part-time or not at all. Thus, the former received an initial financial requirements score of 0, and the latter, a score of 1. Added to this initial score was the number of children aged 22 and under because, conceptually, the greater this number, the greater the financial requirements. Among a sample of managerial and professional employees of an insurance company, George and Brief (in press) found, as hypothesized, that financial requirements (indexed in the above manner) moderated the relationship between pay satisfaction and life satisfaction. Mean financial requirements in George and Brief's sample was 1.60. George and Brief acknowledged that other factors may also affect perceived financial requirements. Because perceived requirements ought to have a greater effect on affective reactions than objective requirements, George and Brief concluded that the use of this objective indicator probably resulted in an understatement of the observed moderating effect. The fact that financial requirements, so indexed, had the hypothesized moderating effect lends indirect support for the construct validity of the index. Because a substantial portion of net income is taken up by housing costs (30%; Bureau of Labor Statistics, 1988), we deemed it desirable to augment George and Brief's financial requirements index by adding scores on a housing arrangements variable. This variable was scored as follows: reside with parents or other relatives (0), rent room (1), and own or rent house or apartment (2); all other housing arrangements received the median score of 1. Thus, for example, respondents who lived with

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

42

DORAN, STONE, BRIE!; AND GEORGE

their parents were viewed as having fewer financial requirements than respondents who owned their own home. Total scores on the financial requirements index ranged from 0 to 6, with a mean of 1.47 (SD = 1.40). Job satisfaction. The 100-item Minnesota Satisfaction Questionnaire (MSQ; Weiss, Dawis, England, & Lofquist, 1967) was used to measure job satisfaction at Time 2. The MSQ taps respondents' satisfaction with 20 job facets; 5 items comprise each facet scale. Responses to each item are made on a 5-point scale ranging from very satisfied (5) to very dissatisfied (I). The facets include the following: ability utilization, achievement, activity, advancement, authority, company policies and practices, compensation, co-workers, creativity, independence, moral values, recognition, responsibility, security, social service, social status, supervision (human relations), supervision (technical), variety, and working conditions. The MSQ also taps respondents' general job satisfaction with a scale comprising 20 items, one from each facet satisfaction scale. Coefficient alpha (Cronbach, 1951) was .94 for the general job satisfaction scale and ranged from .86 to .95 for the facet satisfaction scales.

Results

Data Analysis We used hierarchical regression analysis (e.g., Cohen & Cohen, 1983; Stone & Hollenbeck, 1984) to test both the main effect (Hypothesis 1) and the two-way interaction effect (Hypothesis 2) on each of the 21 MSQ satisfaction scales. Intent to leave scores were entered in the first hierarchical step, financial requirements were entered in the second, and the two-way interaction term was entered in the final step. For each scale, the variance explained by intent to leave at Step 1, the incremental variance explained by financial requirements at Step 2, and the incremental variance explained by the two-way interaction term at Step 3 were each tested for statistical significance. The results of these analyses, including the beta weights for the variables entered at the final step and the cumulative explained variance of each scale, as well as means and standard deviations for all variables, are presented in Table 1.

Hypothesis 1 We hypothesized that workers' intentions to leave at organizational entry would predict subsequent job satisfaction. The results summarized in column 3 of Table 1 indicate that this hypothesis generally was well supported.1 Intent to leave accounted for a statistically significant portion of the variance in 19 of the 21 satisfaction scales. In addition, the direction of the main effect on all 21 dimensions of job satisfaction was in the hypothesized negative direction, indicating that the greater workers' intentions to leave were at the onset of employment, the lower their subsequent job satisfaction was. The significant portion of variance accounted for by intent to leave ranged from 3% (variety) to 15% (advancement).

Hypothesis 2 We hypothesized that the intent-to-leave-job-satisfaction relationship would be moderated by workers' external financial requirements. Specifically, we stated that the relationship would be stronger when external financial requirements were

lower. The results summarized in column 7 of Table 1 indicate that this hypothesis also was well supported. The two-way interaction term (Intent to Leave X Financial Requirements) accounted for a statistically significant portion of the incremental variance in 15 of the 21 satisfaction scales; the effect on all 21 scales was in the same direction. The significant portion of incremental variance accounted for by the interaction term ranged from 4% (activity, advancement, compensation, and independence) to 10% (achievement and authority). The correlation between the components of the interaction term was minimal (r = -.16, p < .05), indicating the relative independence of the two variables. The nature of these interactions is illustrated in Figure 1, in which intent-to-leave-general-satisfaction relationship is plotted separately for respondents with high (« = 56) and low (n = 44) financial requirements. Group membership was assigned by median split on the financial requirements index (Mdn = 2.0). In support of Hypothesis 2, the relationship between initial intentions to leave and subsequent general job satisfaction was stronger for workers with low financial requirements than for workers with high financial requirements. This effect also was evidenced by a strong, negative correlation between intent to leave and general satisfaction in the low-financial-requirements subgroup (r = — .61, p < .001); the corresponding correlation in the high-financial-requirements subgroup was not significant (r = .005). These subgroup correlations differ significantly (z=3.43, p