Psychological Empowerment and Job Satisfaction
Page content transcription ( If your browser does not render page correctly, please read the page content below )
Group & Organization Management Volume 34 Number 3 June 2009 271-296 Psychological © 2009 SAGE Publications 10.1177/1059601108330089 Empowerment http://gom.sagepub.com hosted at and Job Satisfaction http://online.sagepub.com An Analysis of Interactive Effects Guangping Wang Peggy D. Lee Penn State University This research investigates the interactive effects of the psychological empow- erment dimensions on job satisfaction. Using data collected from employees of multiple organizations, the authors find intriguing three-way interactions among the dimensions. Choice has a weak but negative effect on job satis- faction when both competence and impact are high or low but has a strong positive effect when one of the two dimensions is low and the other is high. Impact has no effect on job satisfaction when choice and competence are both high or both low. The effect of impact is positive only when one of the two dimensions is high and the other is low. In addition, high levels of choice and competence reinforce the positive effect of meaning on job satisfaction. The results offer important insights for future theory development on psy- chological empowerment. Keywords: psychological empowerment; job satisfaction; three-way interaction he concept of psychological empowerment has gained wide acceptance T in both management theory and practice (Conger & Kanungo, 1988; Donovan, 1994; Hall, 2008; Kanter, 1989; Spreitzer, 1995; Thomas & Velthouse, 1990). A substantial body of research has accumulated dur- ing the past two decades refining the conceptual domain of psycholog- ical empowerment and investigating its antecedents and consequences. It is generally recognized that the construct consists of four dimensions— meaning, competence, choice, and impact (Spreitzer, 1995; Thomas & Velthouse, 1990)—and is related to various work behaviors, attitudes, and performance (Fulford & Enz, 1995; Hall, 2008; Koberg, Boss, Senjem, & 271
272 Group & Organization Management Goodman, 1999; Spreitzer, 1995, 1996; Suzik, 1998). Several fundamental questions that have yet to be answered in empowerment research are related to the interactive effects of these dimensions on important job outcomes. For instance, do the dimensions influence job outcomes in an additive fash- ion? Do they affect job outcome independently, or are there synergistic or suppressive effects of these variables? Do the effects of some of the dimen- sions depend on the levels of the other dimensions? Additive effects suggest the influence of one dimension is independent of the other dimensions and each adds linear variance to measured out- comes. Interactive effects occur when the total effect of empowerment is greater or less than the sum of the individual dimensional effects (for a general discussion on interaction, see Aiken & West, 1992). Although interactive effects are often examined in organizational research (Gomez-Mejia & Balkin, 1989; House, Shane, & Herold, 1996; Kravitz, Bludau, & Klineberg, 2008; Valentine, 1999; Wated, Sanchez, & Gomez, 2008), studies on psychological empowerment have addressed only addi- tive effects. It is important to both management research and practice to understand potential dimensional interactions. The additive model sug- gests that more empowerment is always better if the individual dimen- sional effects are positive. Managerially, to enhance focal job outcomes, companies could simply increase empowerment on each and all of the individual dimensions. An interactive model, on the other hand, implies that the dimensions may interfere with each other and one dimension may either reinforce or suppress the effects of the other dimensions. When there is a reinforcing effect, the positive effect of one dimension will be greater in the presence of higher lev- els of the other dimensions. Alternatively, if there is a suppressive effect, the presence of a high level of one dimension might decrease the effect of another dimension (Daniels & Guppy, 1994; Karasek, 1979). Thus, theorists and practitioners will be challenged to identify the best combination of the empowerment dimensions to maximize the job outcomes. If management is to use empowerment effectively, it is important to understand potential inter- actions among the empowerment dimensions. Accordingly, the goal of this research is to examine the interactions among empowerment dimensions. We focus on job satisfaction as an outcome variable because it is an outcome of fundamental importance for organizational performance (Ng & Sorensen, 2008; Spreitzer, Kizilos, & Nason, 1997). Job performance, motivation, turnover, and organizational commitment have all been shown to be related to job satisfaction (Judge, 1993; Martin & Bennett, 1996; Williams & Anderson, 1991). In the following sections, we develop several interaction
Wang, Lee / Psychological Empowerment and Job Satisfaction 273 hypotheses and then report the results of an empirical study. We conclude with a discussion of theoretical and managerial implications. Conceptual Background and Hypotheses Psychological Empowerment Psychological empowerment is defined as a set of motivational cogni- tions shaped by a work environment and reflecting an individual’s active orientation to his or her work role (Spreitzer, 1995). Building on the work of Conger and Kanungo (1988), Thomas and Velthouse (1990) argued that four cognitive assessments represent a comprehensive task-specific evalua- tion and interpretation that determines intrinsic task motivation, hence, psy- chological empowerment. These four assessments are meaning, competence, choice, and impact. Briefly, meaning refers to the value of a task goal or purpose, judged in relation to an individual’s own ideals or standards. It reflects intrinsic inter- est in a task and involves a fit between work role requirements and one’s beliefs and values (Brief & Nord, 1990; Hackman & Oldham, 1980). The congruence between personal value and work role expectations contributes to the belief that work is an end in itself. Competence stems from Bandura’s (1986) work on self-efficacy and is the degree to which an employee feels he or she is able to perform tasks with skill (Gist, 1987; Thomas & Tymon, 1994). Social cognitive theory and empirical evidence from diverse fields suggest that competence has strong direct effects on performance (Bandura, 1986; Gecas, 1989; Gist & Mitchell, 1992; Harackiewicz, Sansone, & Manderlink, 1985; Locke, 1991; Locke, Frederick, Lee, & Bobko, 1984; Ozer & Bandura, 1990; Wang & Netemeyer, 2002). Choice is the sense of autonomy in initiating and regulating work and reflects the degree of self-determination in work behaviors and processes (Bell & Staw, 1989; Deci, Connell, & Ryan, 1989; Spector, 1986). Choice is a key component of intrinsic motivation, leading to learning, interest, and resilience in the face of adversity (Deci et al., 1989; Deci & Ryan, 1985). Impact is the degree to which an individual feels that he or she can influ- ence strategic, administrative, or operating outcomes at work (Ashforth, 1989). Impact is associated with high performance and an absence of with- drawal from difficult situations (Ashforth, 1990). Individuals who believe that they can affect the system in which they are embedded and influence organizational outcomes tend to be more motivated (Spreitzer et al., 1997).
274 Group & Organization Management Job Satisfaction Job satisfaction is “a pleasurable or positive emotional state” that is “a function of the perceived relationship between what one wants from a job and what one perceives it is offering” (Locke, 1976, p. 1300). The job char- acteristics model (Hackman & Oldham, 1980) proposes that critical psycho- logical states such as experienced meaningfulness, feelings of responsibility, and knowledge of work results influence job satisfaction. Although Thomas and Velthouse (1990) and Conger and Kanungo (1988) did not explicitly include job satisfaction in their models of empowerment, Thomas and Tymon (1994) argued that empowerment is more likely to manifest at higher levels of job satisfaction. They further argued that assessments of empowerment generate intrinsic rewards and thus should be positively related to job satisfaction. Empirical support varies regarding the relationships between the indi- vidual empowerment dimensions and job satisfaction. Most available evi- dence is related to additive effects and is in the form of simple correlations. First, there seems to be strong evidence of a positive association between meaning and job satisfaction (Hackman & Oldham, 1980; Spreitzer et al., 1997; Thomas & Tymon, 1994). According to Herzberg (1966), an impor- tant determinant of job satisfaction is personal meaning. Kanter (1983) sug- gests that perceived meaningfulness results in greater commitment and concentration of energy. Job satisfaction results from fulfillment of desired work values (Locke, 1976). Lack of meaning in the workplace has been linked to apathy and job dissatisfaction (Thomas & Velthouse, 1990). Second, researchers have suggested that choice is a psychological need and that meeting this need results in job satisfaction (Conger & Kanungo, 1988; Greenberger, Strausser, Cummings, & Dunham, 1989; Parker, 1993). Studies by Liden, Wayne, Sparrowe, and Bradway (1993) and Thomas and Tymon (1994) show that higher levels of personal control are related to job satisfaction. These results are consistent with Spector’s (1986) seminal review, in which he presented evidence for a positive association between job autonomy and job satisfaction. Third, with regard to the impact–satisfaction relationship, strong and consistent evidence is yet to emerge. Ashforth (1989, 1990) suggested that perceived lack of opportunity to have an impact on the organization might be related to job satisfaction, and Thomas and Tymon (1994) reported a positive relationship between impact and job satisfaction, but Spreitzer et al.’s (1997) study did not support the hypothesized effect of perceived impact on satisfaction.
Wang, Lee / Psychological Empowerment and Job Satisfaction 275 Finally, the literature has yet to establish a consistent link between com- petence and satisfaction. Carless (2004) reported that competence was nega- tively related to job satisfaction, whereas Spreitzer et al. (1997) reported that competence is positively related to job satisfaction among subordinates but not among supervisors. Other research has reported no relationship between these variables (Holdsworth & Cartwright, 2003; Siegall & Gardner, 2000; Thomas & Tymon, 1994). In summary, the published research suggests a positive correlation between job satisfaction and two of the dimensions (i.e., meaning and choice), but the record is less consistent regarding the relationship between satisfaction and the other two dimensions (i.e., impact and competence). There is, however, little consideration for the possibility of interaction among the four dimensions in affecting job satisfaction, which could par- tially account for the inconsistent and sometimes contradictory findings with regard to impact and competence. Interaction Among Dimensions The literature on empowerment, job design, stress, person–environment fit, status inconsistency, and mental health suggests possible interactions among empowerment dimensions. Although individually each of the dimensions of meaning, competence, choice, and impact may induce job satisfaction as the current empowerment research indicates, a combination of different levels of the empowerment dimensions may interact in a complicated fashion that could lead to job satisfaction or dissatisfaction. Spreitzer (1995; Spreitzer et al., 1997) argued that the four dimensions of empowerment combine to form a gestalt. A dictionary definition of a gestalt reads “any of the integrated structure or pat- terns that make up all experience and have specific properties which can nei- ther be derived from the elements of the whole nor considered simply as the sum of these elements” (Neufeldt & Guralnik, 1997, p. 567). Although Spreitzer was not explicit as to whether the four dimensions affect job out- comes additively or interactively, the notion of gestalt implies that overall expe- rience is not simply the sum of all individual elements. A lower or higher level of one dimension might change the overall constellation and the whole experi- ence might be affected disproportionately. In fact, various research streams suggest possible interactive effects among psychological empowerment dimensions. A number of researchers including French, Caplan, and Van Harrison (1982), Csikszentmihalyi (1975), and Warr (1987) have advocated the idea
276 Group & Organization Management of optimal levels of job content. That is, employees have a preferred level of certain job characteristics. In particular, Warr’s vitamin model of job characteristics and mental health suggests that extreme (i.e., overly high or low) levels of job characteristics are harmful. The vitamin model is based on an analogy with the fact that vitamins are required for physical health up to certain level; after attainment of that level, increased vitamin intake can be harmful. Consider job autonomy (i.e., choice), for instance. When autonomy level is too low, employees have little choice or freedom and are constrained to such a degree that they feel suffocated. However, when autonomy is too high, employees can feel a lack of direction or that they have too much responsibility, causing them to experience role stress. The same can be said for impact, another empowerment dimension. When a person’s job has low impact, the employee can feel demotivated because he or she believes that his or her job has no significance to the organization. A job that the employee perceives as having an impact incommensurate with the employee’s role can be overwhelming and intimidating. Researchers on job stress have investigated the buffering effect of job design on the stress process quite extensively. In particular, Karasek’s (1979) job strain model postulates that psychological strain results from the joint effects of work demands and decision-making latitude. The negative effect of job demand is greatest when decision latitude (choice) is minimal, and it decreases as choice increases. Daniels and Guppy (1994) found com- plex three-way interactions among job autonomy, locus of control, and stressors on workers’ psychological well-being. The person–environment fit model builds on the notion that job charac- teristics have their primary influence through the equivalence or discrep- ancy between preferred and perceived environmental levels (Edwards & Cooper, 1990). When the person and the job fit well with each other, both productivity and psychological well-being improve. Discrepancies in either direction (i.e., preferred level being greater or less than perceived level) have negative influences on employees’ well-being. Among the empower- ment dimensions, only meaning directly addresses the correspondence between the person and the job, that is, the fit between the job role require- ments and the employee’s beliefs and values (Brief & Nord, 1990; Hackman & Oldham, 1980). We can thus expect meaning to be positively associated with job satisfaction regardless of the levels of the other dimen- sions, based on the person–environment fit model. Unlike meaning, competence encompasses not only perceived congru- ence between skills required and skills possessed but also perceived positive
Wang, Lee / Psychological Empowerment and Job Satisfaction 277 discrepancy between the two when skills possessed exceed skills required. A competent person may be just right for the job or overqualified. When overqualification occurs, dissatisfaction may result. Impact and choice are perceived job characteristics that are not necessarily related to person–job fit. A person may or may not desire a high level of impact and may or may not be happy with a low choice job that has little responsibility and decision- making latitude. Extremely high levels of impact and choice can present overwhelmingly high role expectations that induce role stress and role ambi- guity to some but may sound challenging and exciting to others. More seriously, unfit situations may rise when the various job dimensions and characteristics present conflicting expectations, which cause stress, uncer- tainty, perceived status inconsistency, and frustration (Bacharach, Bamberger, & Mundell, 1993; Edwards & Cooper, 1990; Harrison, 1978). In the context of psychological empowerment, unfit situations may occur at different con- stellations of the four dimensions. First, from a job design standpoint, choice may be viewed as a job demand because of the magnitude of decisions that have to be made, as a reward to the employee because of the intrinsic need associated with being able to make one’s own decisions, and/or a necessary enabling condition for certain jobs to get the job done (Hackman & Oldham, 1980; Parker, 1993). When competence is low and impact is also low, high choice may constitute a conflicting situation that leads to the perception of an excessive demand (because of low competence) that offers little intrinsic reward (because of low impact), and thus high choice may be just as undesirable as low choice when both competence and impact are low. Alternatively, when competence is high and impact is high as well, high choice may be viewed as a neces- sary enabling condition, thus the effect of choice on satisfaction may be null. When competence is high but impact is low, high choice is likely to be perceived as a compensation or reward for one’s overqualification, a bal- ancing factor for an otherwise unfit situation. Likewise, when competence is low but impact is high, high choice may indicate high job demand, but this demand can be viewed in a positive light as rewarding because of high job impact. Thus, it is possible that the choice–satisfaction relationship is positive when competence is high but impact low or when competence is low but impact high. Therefore, we hypothesize, Hypothesis 1: The choice–satisfaction relationship will depend on the levels of competence and impact such that the relationship is positive only when one of them is high and the other is low but is null when both competence and impact are high or low.
278 Group & Organization Management Following the choice–satisfaction relationship, we examined impact, which may be viewed as an intrinsic reward, a responsibility, a demand, and/or a status indicator. On one hand, when there is little choice on the job and the employee feels incompetent for the job, one can hardly imagine the impact may have any positive effect on satisfaction, as any possible positive motivational effect of impact could be overshadowed by the lack of choice and competence. On the other hand, when choice and competence are both high and the employee has the ability and skill to do the job as he or she sees fit, a corresponding level of impact on the larger scope should be expected as a package that comes with the job to avoid perceptions of status incon- sistency (Bacharach et al., 1993), operating much as a hygiene factor in the motivation–hygiene theory (Herzberg, Mausner, & Snyderman, 1959). As such, the marginal effect of perceived impact may be minimal. We therefore expect impact not to be related to satisfaction when choice and competence are both low or both high. However, when choice is low but competence is high, there is an imbal- ance or status inconsistency in the person–job situation. Impact can be a pivotal factor that influences job satisfaction. Low impact will be detri- mental as it exacerbates the imbalanced situation, the employee’s feeling of overqualification will be stronger, and the employee will have a hard time justifying the situation. High impact, on the other hand, may operate as a reward that compensates the unfit situation, brings a balancing factor, and signals a justifiable reason for the job. Another imbalanced or nonfit person–job situation arises when choice is high but competence is low. In this case, the employee is stretching his or her ability to handle the decision-making demands associated with high choice. The job may be perceived as too demanding for the employee. If this is accompanied by low impact, the experience can be characterized as too much effort with too little reward. However, if the perceive level of impact is high, the employee may feel that it is worthwhile to manage the high job demand. Thus, we hypothesize, Hypothesis 2: The impact–satisfaction relationship will depend on the levels of competence and choice such that the relationship is positive only when one of them is high and the other is low but is null when both competence and choice are high or low. Finally, as discussed earlier, because meaning reflects the fit between the employee and the job, we expect the overall meaning–satisfaction relation- ship to be positive regardless of the levels of the other dimensions.
Wang, Lee / Psychological Empowerment and Job Satisfaction 279 However, the relationship strength may vary depending on other empower- ment dimensions. When the other three dimensions are high, the effect of meaning may be reinforced by a more balanced situation. When some dimensions are low but others are high, the imbalanced situation may sup- press the positive influence of meaning. For instance, when competence is high but choice and impact are low, there is a strong feeling of overqualifi- cation, which may offset the positive effect of meaning to certain degree. When choice is high but competence is low, the job demand may be per- ceived as overwhelming and thus may decrease the positive influence of meaning (Karasek, 1979). As such, we hypothesize, Hypothesis 3: The meaning–satisfaction relationship will be positive overall but will be stronger when other empowerment dimensions are all high than when some of the other dimensions are low. Method Sample and Data Collection Survey questionnaires were distributed to the employees in local for- profit organizations (service, retail and distribution, research, and manufac- turing) through part-time MBA students of a state university in the northeastern United States. A total of 510 surveys from employees were returned in postage-paid envelopes directly to the researchers’ university address. We randomly selected 30 (6%) of the respondents and telephoned them to validate that they actually responded, and all of them were true respondents. A total of 21 responses were discarded because of excessive missing data, resulting in 485 useable responses. The one third of responses that were returned first and one third returned last were compared on key study variables, and no significant differences were detected via t tests. About half of the respondents (52%) were male, and 76% of them were between 20 and 49. Approximately 70% of the sample received at least some college education, whereas 14% received graduate education. Slightly fewer than half (44%) were supervisors. The sample represented more than 300 manufacturing and service firms from diverse industries that included software (22%), automobile (20%), banking (13%), construction (13%), communications (10%), restaurants (8%), accounting (6%), retailing (3%), and others (5%). The sizes of the companies also varied greatly, with 50% of the companies employing 300 or fewer employees.
280 Group & Organization Management Table 1 Construct Measures and Corrected Item to Total Correlations Completely Standardized Factor Measurement Items Loadings Meaning 1. The work I do is very important to me. .89 2. My job activities are personally meaningful to me. .84 3. The work I do is not very meaningful to me personally. (R) .43 Competence 1. I do not have enough confidence in my ability to do my job. (R) .47 2. I am self-assured about my capabilities to perform my work activities. .75 3. I have mastered the skills necessary for my job. .74 Choice 1. I do not have enough autonomy in determining how I do my job. (R) Dropped 2. I can decide on my own how to go about doing my work. .79 3. I have considerable opportunity for independence and freedom .86 in how I do my job. Impact 1. My impact on what happens in my department is large. .59 2. I have a great deal of control over what happens in my organization. .93 3. I have significant influence over what happens in my organization. .91 Job satisfaction 1. I feel fairly happy with my present line of work. .79 2. All things considered (i.e., pay, promotion, supervisors, coworkers, etc.), .82 I am very satisfied with my line of work. 3. I feel a great sense of personal fulfillment from my line of work. .89 Note: R = reverse-coded item. Measures Psychological empowerment was measured with the 12 items adapted from Spreitzer (1995) on 5-point Likert scales (1 = strongly disagree to 5 = strongly agree). To minimize response set bias, three of the items (Meaning No. 3, Competence No. 1, and Choice No. 1) were modified to be reverse worded. One item (Choice No. 1) was deleted in the final analy- sis because of the fact that its factor loading was too low (.31). Job satis- faction was assessed with three items from Netemeyer, Boles, McKee, and McMurrian (1997). The items asked the respondents the degree to which they were happy, satisfied, and felt a sense of personal fulfillment with their present line of work. The items were anchored on 5-point Likert-type
Wang, Lee / Psychological Empowerment and Job Satisfaction 281 scales (1 = strongly disagree to 5 = strongly agree). The measurement items for key constructs are presented in Table 1. We also included a number of covariates in the survey to control for extra- neous variances in such a cross-sectional study. Perceived organizational sup- port (Eisenberger, Cummings, Armeli, & Lynch, 1997), organizational climate for innovation (Scott & Bruce, 1994), and innovative personality (Hurt, Joseph, & Cook, 1977) have been shown to be relevant to empower- ment as well as job satisfaction. The scales all achieved satisfactory reliabil- ity coefficients (Cronbach’s alpha ranged from .74 to .86). In addition, demographic information such as gender, age, education, supervisory status, and employment status (full-time or part-time) was also collected. Analysis and Results Confirmatory factor analysis was first conducted to ensure satisfactory psychometric property of the key measures. We then computed a summated scale for each construct and used hierarchical multiple regression (Aiken & West, 1992) to test our hypotheses. Measurement and Dimensionality We first estimated a confirmatory factor model that included 15 manifest indicators for 5 key latent constructs (job satisfaction, meaning, compe- tence, choice, and impact). The model was evaluated with model fit, dis- criminant validity of the constructs, and internal consistency among the construct measures. Model fit was evaluated using the chi-square statistic, the comparative fit index (CFI), the Tucker–Lewis index (TLI), and the root mean square error of approximation (RMSEA). Values of .90 and above have been recommended for CFI and TLI, and values of .08 and less have been suggested for RMSEA (Browne & Cudeck, 1993; Hu & Bentler, 1995). The fit indices for the measurement model were χ2 = 256.65 (df = 80), CFI = .96, TLI = .94, and RMSEA = .072. Although the model fit appeared satisfactory, an examination of the factor loading structure revealed that the reverse-worded empowerment items suffered from relatively low factor loadings. The first item in the choice dimension was especially low at .31. The item was subsequently dropped, and a new measurement model with 14 manifest indicators and 5 latent constructs was estimated. The fit indices were slightly better than the previous model: χ2 = 200.16 (df = 67), CFI = .97, TLI = .95, and
282 Group & Organization Management Table 2 Means, Standard Deviations, Internal Consistency, and Correlations M SD Alpha AVE 1 2 3 4 5 6 7 8 1. Satisfaction 3.61 1.06 .88 .70 — 2. Meaning 3.79 0.92 .70 .56 .55*** — 3. Competence 4.33 0.70 .65 .44 .07 .13*** — 4. Choice 3.91 0.95 .81 .67 .34*** .28*** .25*** — 5. Impact 3.22 1.07 .83 .68 .41*** .37*** .06 0.43*** — 6. Age 2.57 0.87 — — .25*** .17*** .10** .17*** .14*** — 7. POS 3.58 1.00 .86 .70 .48*** .23*** –.05 .24*** .39*** .02 — 8. OCI 3.03 0.99 .74 .61 .31*** .17*** .05 .25*** .20*** –.01 .37*** — Note: AVE = average variance extracted; POS = perceived organizational support; OCI = organizational cli- mate for innovation. **p < .05. ***p < .01. RMSEA = .068. All but three items had completely standardized factor loadings greater than .70 (see Table 1). Further testing was conducted to ensure that psychological empower- ment was best represented by four dimensions as separate constructs. We tested a series of models that specified one-factor, two-factor, three-factor, four-factor, and second-order four-factor structures. The four-factor model fit significantly better than any other specifications. The four-factor model had a chi-square of 117.35 with df = 38. The second-order four-factor model had the second best fit (χ2 = 143.03, df = 40), but the chi-square dif- ference from that of the four-factor model was significant (∆χ2 = 25.68, df = 2, p < .01). The chi-square values for the 6 three-factor models ranged from 341.71 to 489.35 with df = 41; the one-factor model had the worst fit (χ2 = 1006.86, df = 44). Thus, we conclude that the four-factor model fits the data best. Table 2 contains the correlations among the model constructs, construct means, standard deviations, and internal consistency estimates. As mea- sures of internal consistency, Cronbach’s alpha ranged from .65 to .88, and the average variance extracted estimates (AVE) ranged from .44 to .70. Discriminant validity was assessed by comparing the square of the correla- tion (phi-square) between two constructs and their average AVE. Evidence supporting the constructs’ discriminant validity is present when phi-square is less than the average AVE (Fornell & Larcker, 1981). The criterion has been considered the most stringent test of discriminant validity and was met for all possible pairs of constructs. These results indicate an acceptable fit for the measurement model.
Wang, Lee / Psychological Empowerment and Job Satisfaction 283 Table 3 Regression Results (Standardized Coefficients) Independent Variable Model 1 Model 2 Model 3 Model 4 Age 24*** .15*** .15*** .15*** POS .43*** .31*** .31*** .32*** OCI .16*** .10*** .10*** .10*** Meaning .38*** .35*** .35*** Competency –.01 –.01 .04 Choice .08** .13*** .07* Impact .07* .08* .11*** Meaning × choice –.06* –.07* Meaning × impact –.11*** –.10*** Competency × choice × impact –.13*** Meaning × competency × choice .07* R2 .32 .48 .50 .52 Note: N = 489. POS = perceived organizational support; OCI = organizational climate for innovation. *p < .10. **p < .05. ***p < .01. Hypothesis Testing Given the satisfactory results for the measurement items, we computed a summated scale for each construct and tested the hypotheses via estimating a series of hierarchical regression models with job satisfaction as the dependent variable. Model 1 included the control variables only, that is, demographics (age, education, income, employment status [full-time or part-time], supervi- sory status, and firm size), perceived organizational support, innovative per- sonality scale, and organizational climate for innovation. Only age, perceived organizational support, and organizational climate for innovation were statisti- cally significant. Therefore, we included only these three variables in the subsequent models. Model 2 included the four empowerment dimensions (meaning, competence, choice, and impact) in addition to the three control vari- ables. We then added two-way and three-way interaction terms to the model to form Models 3 and 4, respectively. The cross-product interaction terms were created after the empowerment dimension variables were mean centered. In Table 3, we report the results that include the significant interaction terms. Model 1 shows that age, perceived organizational support, and organiza- tional climate for innovation all have a positive association with job satis- faction. The model explains 32% of the variance in job satisfaction. The next
284 Group & Organization Management Figure 1 Plot of Competence and Impact Versus Choice Low Competence 0.3 0.2 Job Satisfaction 0.1 0 Low Impact High Impact –0.1 –0.2 –0.3 Choice Panel A High Competence 0.3 0.2 Job Satisfaction 0.1 Low Impact 0 High Impact –0.1 –0.2 –0.3 Choice Panel B model (Model 2) shows that three of the four empowerment dimensions are significantly and positively related to job satisfaction, but competence is not a significant predictor of job satisfaction. Overall, the results are consistent with the extant empirical findings and lend a general support to the notion that psychological empowerment contributes to positive work outcomes such as job satisfaction. Model 3 contains two significant two-way interaction terms: meaning by choice and meaning by impact. The final model, Model 4, consists of the 2
Wang, Lee / Psychological Empowerment and Job Satisfaction 285 Figure 2 Plot of Choice and Competence Versus Impact Low Choice 0.3 0.2 Job Satisfaction 0.1 0 –0.1 –0.2 –0.3 Impact Low Competence Panel A High Competence High Choice 0.3 0.2 Job Satisfaction 0.1 0 –0.1 –0.2 –0.3 Impact Low Competence Panel B High Competence two-way and 2 three-way interaction terms in addition to the main effect terms of the four dimensions and the control variables. The changes in the F statistic were all significant at the .01 level as additional terms were added from Model 1 to Model 4. Compared to the main effect model (Model 2), Model 4 explains 4% more variance in job satisfaction. Based on Cohen (1988), we calculated the effect size for interaction to be .08.
286 Group & Organization Management Figure 3 Plot of Choice and Competence Versus Meaning Low Choice 0.6 0.4 Job Satisfaction 0.2 0 –0.2 –0.4 –0.6 –0.8 Meaning Low Competence High Competence Panel A High Choice 0.3 0.2 Job Satisfaction 0.1 0 −0.1 −0.2 −0.3 Meaning Low Competence Panel B High Competence Because Cohen suggests that interaction effect sizes are considered small at .02, medium at .15, and large at .35, the interaction effect size in this study is between small and medium. To facilitate interpretation, we plotted the regression slopes in Figures 1 to 3 based on coefficient estimates from Model 4. The high and low levels of the explanatory variables were set at one standard deviation above and below the mean (Cohen & Cohen, 1983).
Wang, Lee / Psychological Empowerment and Job Satisfaction 287 Hypothesis 1 proposes that the choice–satisfaction relationship is null when competence and impact are both high or both low but is positive when one is high and the other is low. Panel A of Figure 1 shows that when com- petence level is low, choice is positively related to job satisfaction only when impact is high. When impact is low, the choice–satisfaction relation- ship is slightly negative. That is, when employees have low perceived com- petence, a high level of choice (perceived as a job demand that leads to role stress) should be accompanied by a high level of impact (as a reward to compensate for the job demand) to generate a high level of job satisfaction. When impact is low, high autonomy will generate greater role stress and lead to lower job satisfaction. Panel B of Figure 1 shows that at a high competence level, the choice–satisfaction relationship is positive only when impact is low and is slightly negative when the impact is high. That is, when employees perceive themselves as highly competent, choice is more important when impact on larger scope is low. When perceived impact is high, a high level of choice becomes an enabling condition that is expected and has little effect on sat- isfaction. Although Panels A and B show slightly negative relationship under conditions of high competence and high impact and under low com- petence and low impact, the slopes are rather flat, and therefore support for Hypothesis 1 is evident. Hypothesis 2 posits that the impact–satisfaction relationship depends on the levels of competence and choice, such that the relationship is null when competence and choice are both high or low and is positive when one is low and the other is high. Panel A of Figure 2 suggests that at low choice level, the impact–satisfaction relationship is positive only when competence is high. When competence is low, impact has almost no relationship with sat- isfaction. Panel B of Figure 2 shows that at a high choice level, the impact–satisfaction relationship is positive only when competence is low. When competence is high, impact has almost no relationship with satisfac- tion. Thus, the results support Hypothesis 2. Hypothesis 3 suggests that the positive meaning–satisfaction relation- ship is stronger when other dimensions are all high than when some of the dimensions are low. The two panels in Figure 3 show that the meaning– satisfaction relationship is always positive regardless the levels of choice and competence. However, this positive relationship is stronger when choice and competence are both low (Panel A) or both high (Panel B) than when one is high and the other is low. This suggests some support for Hypothesis 3 as far as choice and competence are concerned. It is interesting that the meaning–satisfaction relationship is also stronger when both choice and
288 Group & Organization Management competence are low, possibly because meaning becomes the only bonding element between the job and the person when other dimensions are low. Discussion Summary and Theoretical Implications This study applies the vitamin model, the person–environment fit model, and the literature on job stress, job demand, and status inconsistency to the research on psychological empowerment and finds support for significant interactive effects. Although research on job design has long noticed the potential interaction between job characteristics and work context (Ferris & Gilmore, 1984) and between the job and the person (Edwards & Cooper, 1990), research on empowerment appears to have only examined linear and additive effects of psychological empowerment on job outcomes (e.g., Spreitzer et al., 1997; Thomas & Velthouse, 1990). Our theoretical analysis and empirical findings contribute to the literature by demonstrating the complicated relationship among empowerment dimensions and their inter- active effects on a critical job outcome. Overall, the data offer substantial support for our three interaction hypotheses, thus improving our under- standing of the dynamics of psychological empowerment. In this sense, the research not only offers empirical evidence for Spreitzer’s (1995) argument that the four dimensions combine to form a gestalt to influence employee well-being but also provides a sense what the gestalt looks like. The find- ings provide a fresh, more sophisticated perspective on empowerment and open new avenues for future research. In support of Hypothesis 1, we find that choice has a weak but negative effect on job satisfaction when both competence and impact are high or low but has a strong positive effect when one of the two dimensions of compe- tence and impact is low and the other is high. Consistent with Hypothesis 2, we find impact has no effect on job satisfaction when choice and com- petence are both high or low. The effect of impact is positive only when one of the two dimensions is high and the other is low. Also, supporting Hypothesis 3 in general, our data suggest that high levels of choice and competence reinforce the positive effect of meaning on job satisfaction. It is evident that the dimensions may reinforce each other in affecting job outcomes. For instance, in a low competence situation, high perceived impact coupled with high choice is able to generate greater job satisfaction (Figure 1, Panel A). The combination of intrinsic reward (impact) and job demand
Wang, Lee / Psychological Empowerment and Job Satisfaction 289 (choice) seems to work nicely. Similarly, in a low choice setting, high compe- tence coupled with high impact has a strong positive effect on job satisfaction (Figure 2, Panel A). In a high choice setting, high competence makes the pos- itive effect of meaning on satisfaction even stronger (Figure 3, Panel B). Such findings may help explain why the literature is inconsistent on the effect of impact and competence on job outcomes (e.g., Carless, 2004; Holdsworth & Cartwright, 2003; Siegall & Gardner, 2000; Spreitzer et al., 1997; Thomas & Tymon, 1994), because the effects of these dimensions are most likely contin- gent on other empowerment dimensions and/or other job contexts. We also demonstrate that the empowerment dimensions may suppress each other. In a low competence and low impact situation, greater choice leads to less, rather than more, satisfaction (Figure 1, Panel A). Similarly, a combination of high competence and high impact suppresses the effect of choice on satisfaction; greater choice actually leads to slightly lower satis- faction (Figure 1, Panel B). When low choice is combined with low com- petence, impact has little effect on satisfaction (Figure 2, Panel A), as when high choice is accompanied by high competence (Figure 2, Panel B). The finding that a high level on one dimension can reduce the positive effect of another is alarming. Contrary to the accepted wisdom that psy- chological empowerment improves employees’ psychological well-being and other job outcomes (Chen & Chen, 2008; Hall, 2008; Spreitzer, 1995), this study echoes research on job stress (Karasek, 1979) and job content (Warr, 1987) and demonstrates that empowerment may sometimes be a source of stress that can even cause lower satisfaction through a compli- cated interplay among the four empowerment dimensions. Because choice, impact, and competence can be interpreted differently under varying situa- tions, consideration of balance and counterbalance of the stress generators (e.g., choice as a job demand), reward (e.g., impact), and one’s perceived ability to handle the stress (e.g., competence as qualification or overquali- fication) seems more important than achieving a high level on any or all empowerment dimensions. The varying constellations of the levels of choice, competence, meaning, and impact shed the dimensions in different lights so that they lead to different perceptions of the job in terms of stress, demand, and intrinsic rewards, which will then affect job satisfaction. In sum, although the empowerment dimensions, when looked at sepa- rately, may in general have positive main effects on job satisfaction, individ- ual dimensions interact with each other to enhance or reduce the influences of other dimensions on job outcomes. It appears the received view of the pos- itive effect of psychological empowerment is overly simplistic and incom- plete. Our research suggests that the distinct dimensions may be viewed as
290 Group & Organization Management motivators, stressors, enablers, a part of the work context, and/or an evalua- tion of the job–person fit. The way one dimension affects job outcomes is not constant and additive but may be enhanced or suppressed by the levels of other dimensions or the combination of other dimensions. Thus, empower- ment theorists must focus on identifying optimal combinations that represent balanced or fit situations in different work contexts, which are essential to employee well-being and other job outcomes. Managerial Implications We observe that choice has a positive association with job satisfaction when the competence is low but impact is high (Figure 1, Panel A). When the competence and impact are both low, the choice–satisfaction relationship becomes slightly negative. Furthermore, when impact is low but competence is high, the choice–satisfaction relationship is also positive, but when both competence and impact are high the relationship is again slightly negative (Figure 1, Panel B). Managerially, when empowerment intervention efforts are geared toward increasing employees’ perceived choice level, such efforts should also focus on improving the felt competence if the impact of the job is perceived to be low or improving the perception of job impact if the per- ceived competence is low. When impact and competence are both low or high, it is probably not a good idea to focus on increasing the perceived level of choice, as choice has little effect, or even a negative effect, on satisfaction. Likewise, from Figure 2, when choice is low, empowerment programs should aim for the improvement of both competence and impact to gen- erate greater job satisfaction, and when competence is low improvement in both choice and impact should be the dual means of increasing job sat- isfaction. When choice and competence are both low or high, intervention programs should not focus on impact, as perceived impact has little effect on job satisfaction. Instead, in these situations, improving the perceived job meaning would be the most helpful in boosting job satisfaction. As shown in Figure 3, the positive association between job meaning and job satisfaction is stronger when competence and choice are both low or both high. Nevertheless, job meaningfulness is the most important dimension among the four empowerment dimensions for job satisfaction, which has a positive effect on satisfaction regardless of other dimensional levels. Job training leading to greater felt competence should accompany increases in job choice or impact. If employees’ felt competence is not improved in low impact situations or if felt impact is not improved in low competence situations, increased perceived choice may create problems
Wang, Lee / Psychological Empowerment and Job Satisfaction 291 such as role ambiguity and role stress that they do not feel to have the capa- bility or motivation to handle. This could contribute to negative job out- comes such as lower job satisfaction. The moderating role of competence and impact in the choice–satisfaction relationship should be carefully mon- itored in practice, as should the role of competence and choice in the impact–satisfaction relationship. In sum, our findings suggest that empowerment programs should focus on reaching optimal levels on the individual dimensions to create a balanced combination that will result in the most positive outcomes. Although all four dimensions of empowerment are important, as Spreitzer et al. (1997) and many others have suggested, it is essential to achieve balanced combinations to create maximal job outcomes. The meaning dimension appears to have a consistently strong positive effect on satisfaction, which is also consistent with much of the extant research. As such, a program that emphasizes a high level of meaning should generally do well in terms of generating employee job satisfaction. However, at an average level of job meaning, a combination of too high levels of choice, competence, and impact may actually work against employee well-being (e.g., job satisfaction). Limitations and Future Research Directions Our analysis on interactive effects is a significant addition to the exist- ing empowerment literature, where only linear relationships and additive effects are documented. Like all research, however, this study has limita- tions. First, the use of cross-sectional data precludes any inference of causality. Although it makes theoretical sense to argue that psychological empowerment affects job satisfaction, the reverse could very well be true. Job satisfaction could make the individual feel that his or her job is mean- ingful and motivate him or her to learn and develop higher levels of skills and thus feel more empowered. As such, longitudinal studies such as that of Laschinger, Finegen, Shamian, and Wilk (2004) are highly encouraged. A second limitation concerns the possible common method variance (CMV) because all constructs were measured with a paper and pencil sur- vey instrument on the same respondent at the same point of time. As it is dif- ficult to identify the specific sources of CMV such as social desirability, positive or negative affectivity, and/or acquiescence, we relied on Podsakoff, MacKenzie, Lee, and Podsakoff’s (2003) recommendation to use a single common method factor approach to assess the extent of this problem. We used the latent method factor technique advanced by Williams and Anderson (1994) and compared two structural equation models that included a
292 Group & Organization Management same-source factor in addition to the four predictor constructs (empower- ment dimensions) and job satisfaction. The two comparison models were the constrained model with factor loadings of all indicator items on the same-source factor constrained to zero and the unconstrained model with loadings of all items on the same-source factor freely estimated. Although the unconstrained model fit better than the constrained model as indicated by a chi-square difference test, the path coefficient estimates from empow- erment dimensions to job satisfaction were almost identical and mirrored those obtained from the regression model. As such, we conclude that CMV might not be a serious problem for our study. Nevertheless, future studies should strive to minimize the CMV by procedural as well as statistical remedies (Podsakoff et al., 2003). Third, because we collected data from multiple organizations, there could be intraorganizational random effects that have not been accounted for in our models. To assess the extent of this problem, we ran random intercept models via Hierarchical Linear Model 2 (HLM2) at Level 2 (orga- nizational level) and found that although the intercept and some of the slopes were significantly different across organizations (χ2 ranged from 13.7 [p > .10, not significant] to 25.7 [p < .01], df = 10), based on the 11 cases that had sufficient data, the intraclass correlations were calculated to be less than .10, which is deemed rather small. The fixed effect coefficients were fairly identical to those obtained from the Ordinary Least Square (OLS) models. We conclude that although there may be some company- level variances in the data, the majority of the variances are at the individ- ual level. Last, this research focuses on the interaction effects on one job out- come variable, that is, job satisfaction. Future research should examine other workplace outcome variables such as performance, learning, team work, and creativity in relation to these dimensions. Antecedents of psy- chological empowerment should also be investigated. Simultaneous esti- mation of models that incorporate a comprehensive set of antecedents and outcomes will allow researchers to gain a fuller understanding of psycho- logical empowerment in the workplace. In addition, situational variables that may potentially moderate the dimensional effects on outcome vari- ables should be identified and investigated. Such research will not only enrich our theoretical understanding of psychological empowerment but also shed important light on managing workplace attitudes and behavior through meaningful empowerment programs and interventions.
Wang, Lee / Psychological Empowerment and Job Satisfaction 293 References Aiken, L. S., & West, S. G. (1992). Multiple regression: Testing and interpreting interactions. Newbury Park, CA: Sage. Ashforth, B. E. (1989). The experience of powerlessness in organizations. Organizational Behavior and Human Decision Processes, 43, 207-242. Ashforth, B. E. (1990). The organizationally induced helplessness syndrome: A preliminary model. Canadian Journal of Administrative Sciences, 7, 30-36. Bacharach, S. B., Bamberger, P., & Mundell, B. (1993). Status inconsistency in organizations: From social hierarchy to stress. Journal of Organizational Behavior, 14, 21-36. Bandura, A. (1986). Social foundations of thought and action: A social-cognitive view. Englewood Cliffs, NJ: Prentice Hall. Bell, N. E., & Staw, B. M. (1989). People as sculptors versus sculpture. New York: Cambridge University Press. Brief, A. P., & Nord, W. R. (1990). Meanings of occupational work. Lexington, MA: Lexington Books. Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model fit. Newbury Park, CA: Sage. Carless, S. A. (2004). Does psychological empowerment mediate the relationship between psy- chological climate and job satisfaction? Journal of Business and Psychology, 18(4), 405-406. Chen, H., & Chen, Y. (2008). The impact of work redesign and psychological empowerment on organizational commitment in a changing environment: An example from Taiwan’s state-owned enterprises. Public Personnel Management, 37(3), 279-302. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum. Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlation analysis for the behav- ioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum. Conger, J., & Kanungo, R. (1988). The empowerment process: integrating theory and practice. Academy of Management Review, 13(3), 471-482. Csikszentmihalyi, M. (1975). Beyond boredom and anxiety. San Francisco: Jossey-Bass. Daniels, K., & Guppy, A. (1994). Occupational stress, social support, job control, and psy- chological well-being. Human Relations, 47(12), 1523-1544. Deci, E. L., Connell, J. P., & Ryan, R. M. (1989). Self-determination in a work organization. Journal of Applied Psychology, 74, 580-590. Deci, E. L., & Ryan, R. M. (1985). The support of autonomy and control of behavior. Journal of Personality and Social Psychology, 53, 1024-1037. Donovan, M. (1994). The empowerment plan. Journal for Quality and Participation, 17(4), 12-14. Edwards, J., & Cooper, C. (1990). The person-environment fit approach to stress: Recurring problems and some suggested solutions. Journal of Organizational Behavior, 11, 293-307. Eisenberger, R., Cummings, J., Armeli, S., & Lynch, P. (1997). Perceived organizational support, discretionary treatment, and job satisfaction. Journal of Applied Psychology, 82(5), 812-820. Ferris, G. R., & Gilmore, D. C. (1984). The moderating role of work context in job design research: A test of competing models. Academy of Management Journal, 27(4), 885-892. Fornell, C., & Larcker, D. F. (1981, February). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18, 39-50. French, J. R. P., Caplan, R. D., & Van Harrison, R. (1982). The mechanisms of job stress and strain. New York: John Wiley.
294 Group & Organization Management Fulford, M. D., & Enz, C. A. (1995). The impact of empowerment on service employees. Journal of Managerial Issues, 7(2), 161-175. Gecas, V. (1989). The social psychology of self-efficacy. Annual Review of Psychology, 15, 291-316. Gist, M. E. (1987, April). Self-efficacy: Implications for organizational behavior and human resource management. Academy of Management Review, 17, 183-211. Gist, M. E., & Mitchell, T. R. (1992). Self-efficacy: A theoretical analysis of its determinants and malleability. Academy of Management Review, 17(2), 183-211. Gomez-Mejia, L. R., & Balkin, D. B. (1989). Effectiveness of individual and aggregate com- pensation strategies. Industrial Relations, 28, 431-445. Greenberger, D. B., Strausser, S., Cummings, L. L., & Dunham, R. (1989). The impact of per- sonal control on performance and satisfaction. Organizational Behavior and Human Decision Processes, 43, 29-51. Hackman, J. R., & Oldham, G. R. (1980). Work redesign. Reading, MA: Addison-Wesley. Hall, M. (2008). The effect of comprehensive performance measurement systems on role clar- ity, psychological empowerment and managerial performance. Accounting, Organizations and Society, 33(2-3), 141-163. Harackiewicz, J. M., Sansone, C., & Manderlink, G. (1985). Competence, achievement moti- vation and intrinsic motivation. Journal of Personality and Social Psychology, 48, 493-508. Harrison, R. (1978). Person-environment fit and job stress in blue collar stress (C. Cooper & M. Smith, Eds.). Chichester, UK: Wiley. Herzberg, F. (1966). Work and the nature of man. Cleveland, OH: World. Herzberg, F., Mausner, B., & Snyderman, B. (1959). The motivation to work (2nd ed.). New York: John Wiley. Holdsworth, L., & Cartwright, S. (2003). Empowerment, stress and satisfaction: An exploratory study of a call centre. Leadership & Organization Development Journal, 24(3), 131-140. House, R. J., Shane, S. A., & Herold, D. M. (1996). Rumors of the death of dispositional research are vastly exaggerated. Academy of Management Review, 21(1), 203-224. Hu, L., & Bentler, P. M. (1995). Evaluating model fit. Thousand Oaks, CA: Sage. Hurt, H. T., Joseph, K., & Cook, C. D. (1977). Scales for the measurement of innovativeness. Human Communication Research, 4(1), 58-65. Judge, T. A. (1993). Does affective disposition moderate the relationship between job satis- faction and voluntary turnover? Journal of Applied Psychology, 78(3), 395-401. Kanter, R. M. (1983). The change masters. New York: Simon & Schuster. Kanter, R. M. (1989). The new managerial work. Harvard Business Review, 66, 85-92. Karasek, R. A., Jr. (1979). Job demands, job decision latitude, and mental strain: Implications for job redesign. Administrative Science Quarterly, 24, 285-309. Koberg, C. S., Boss, R. W., Senjem, J. C., & Goodman, G. E. (1999). Antecedents and out- comes of empowerment. Group & Organization Management, 24, 71-91. Kravitz, D. A., Bludau, T. M., & Klineberg, S. L. (2008). The impact of anticipated conse- quences, respondent group, and strength of affirmative action plan on affirmative action attitudes. Group & Organization Management, 33, 361-391. Laschinger, H. K. S., Finegen, J. E., Shamian, J., & Wilk, P. (2004). A longitudinal analysis of the impact of workplace empowerment on work satisfaction. Journal of Organizational Behavior, 25, 527-545. Liden, R. C., Wayne, S. J., Sparrowe, R. T., & Bradway, L. (1993). Empowerment and effective- ness study: Feedback report (Technical report 1). Urbana-Champaign: University of Illinois. Locke, E. A. (1976). The nature and causes of job satisfaction. Chicago: Rand McNally.
Pages you may also like
You can also read
Next part ... Cancel