Home Random Page


CATEGORIES:

BiologyChemistryConstructionCultureEcologyEconomyElectronicsFinanceGeographyHistoryInformaticsLawMathematicsMechanicsMedicineOtherPedagogyPhilosophyPhysicsPolicyPsychologySociologySportTourism






Study 2: Motives and OCB

This study addressed the major theoretical question in this research effort: Do motives play a unique and significant role in OCB? Municipal employees completed self-report measures of constructs that had been previously found to be correlated with OCB (see the Method section below) and the CMS measure. Because ratings of OCB may differ as a function of the rater ( Morrison, 1994 ), self-, peer, and supervisor ratings of OCB were obtained. We predicted that the zero-order correlations between the motives and the ratings of the dimensions of OCB would be significant. More important, we predicted that the motives would differentially correlate with the dimensions of OCB. Specifically, we expected that (a) PV motives would correlate most strongly with the altruism dimension, or OCBs directed at individuals in the organization, and (b) OC motives would correlate most strongly with the conscientiousness dimension, or OCBs directed at the organization. We made no hypotheses about the correlates of IM motives because there was no clear theoretical basis for predicting differential correlations between IM motives and specific OCB dimensions.

Finally, we predicted that motives would account for unique variance in OCB. That is, when motives are entered into a hierarchical multiple regression after other predictors of OCB, they will add significant accounts of explained variance in OCB ratings. The other predictors were selected on the basis of two criteria. First, the variable was already known to correlate with one of the dimensions of OCB, which enabled us to pit motives against other variables that were "shooting" at the same "targets." Second, there was some presumed conceptual association between each variable and at least one of the motives, which enabled us to use variables other than OCB to examine the construct validity of the CMS.

Method Participants

The participants were 145 individuals (88 male, 57 female) employed by a city government in Florida. They represented 73% of the 200 city employees who had agreed to participate in the study and received questionnaire packets. Ninety-one percent of the sample self-identified as White or European American. Ten percent of the sample were less than 30 years old, 30% were between 30 and 40 years old, 45% were between 40 and 50 years old, and the remainder were more than 50 years old. Thirteen percent had completed high school, 53% had completed 1 or more years of college, 21% had graduated from college, and 12% had completed some postgraduate education. Four percent had worked for the city for 1 year or less, 23% for 1 to 5 years, and 73% for longer than 5 years.

Measures Organizational justice.

There is a substantial prior literature that suggests a positive association between perceived organizational justice and the commission of OCB (see, e.g., Moorman, 1991 , Organ & Ryan, 1995 ). Two aspects of organizational justice were assessed. The first aspect was distributive justice–the degree to which the rewards received by employees were perceived to be related to performance inputs. Distributive justice was measured with a six-item scale developed by Price and Mueller (1986) , in which participants rate the degree to which they feel they are fairly rewarded by the organization. The second aspect was procedural justice–the degree to which fair procedures were used in the organization. Procedural justice was measured with a seven-item scale developed by Moorman (1991) . A 5-point response format was used for both scales. Cronbach's alphas for the two scales were .94 and .89, respectively.



Positive mood.

George (1991) and several other researchers (e.g., Borman et al., 2001 ; Facteau, Allen, Facteau, Bordas, & Tears, 2000 ; Van Scotter & Motowidlo, 1996 ) have found that job-related moods are positively associated with OCB. Positive mood was measured by the Positive Affect subscale of the Job Affect Scale ( Brief, Burke, George, Robinson, & Webster, 1988 ). The subscale consists of six positive mood adjectives: active, strong, excited, enthusiastic, peppy, and elated . Participants rated the extent to which each adjective reflected their current feelings at work by using a 5-point scale. Cronbach's alpha for this scale was .92.

Prosocial Personality Battery.

Several studies have found that Penner, Fritzsche, Craiger, and Freifeld's (1995) measure of prosocial dispositions, the Prosocial Personality Battery (PSB), correlates with self-reports and peer reports of OCB (see Allen, 1998 , 1999 ; Borman et al., 2001 ; Facteau et al., 2000 ; Midili & Penner, 1995 ). The PSB is a 56-item self-report measure. When factor analyzed, it yields two related factors ( Penner et al., 1995 ). The first factor, called Other-Oriented Empathy, concerns prosocial thoughts and feelings. The second factor, called Helpfulness, concerns prosocial behaviors. The factors are positively correlated (ranging from .20 to .40, depending on the sample). The factor structure is virtually invariant across gender, age, and educational level. Participants used a 5-point scale to respond to the PSB items. Cronbach's alphas for the two factors were .82 and .72, respectively.

Motives for OCBs.

The 30-item CMS, described in Study 1, was administered to participants. The respondents were given a description of OCB and asked to rate, using a 6-point scale, how important each of the 30 motives would be in their decision to engage in this behavior. As we discussed earlier, the scale contains three correlated but distinct subscales: PV, OC, and IM. Cronbach's alphas for the subscales were .88, .93, and .89, respectively.

OCB.

OCB was measured using a scale developed by Podsakoff, MacKenzie, Moorman, and Fetter (1990) . We selected this scale because it has been used in several previous studies of OCB. It contains five subscales that purport to measure the five aspects of OCB proposed by Organ (1988) : altruism–helping behaviors directed at specific individuals, conscientiousness–doing things that benefit the organization (e.g., being punctual, obeying rules), courtesy–trying to prevent work-related interpersonal problems from occurring, sportsmanship–tolerating less than ideal circumstances on the job without complaining, and civic virtue–responsibly involving oneself in and being concerned about the life of the company. It is generally agreed that altruism and courtesy concern OCBs directed at individuals (i.e., altruism) and the other subscales concern OCBs directed at the organization (i.e., conscientiousness). The OCB scale contains 24 items and uses a 5-point response format. Subscale scores are the simple sum of the relevant items.

To obtain independent ratings of each employee's OCB, we gave this scale to the target employees, two of their peers, and one supervisor; all of whom rated the target employee. The instructions and the wording of the items were changed to reflect the different perspectives of the three kinds of raters. The time frame for the ratings was the employee's behavior over the past week. In the self-ratings, the Cronbach's alphas for the three OCB subscales were less than .70: .60 for Sportsmanship, .65 for Courtesy, and .69 for Conscientiousness; the other alphas were .71 or higher. The subscales' alphas ranged from .74 to .86 in the peer ratings and from .80 to .87 in the supervisor ratings. The self-ratings of OCB were significantly correlated with peer ratings on four of the five subscales, and the average convergent correlation was .26. They were also significantly correlated with supervisor ratings on four of the five subscales; the average convergent correlation was .22. All five of the peer and supervisor subscale ratings were significantly correlated; the average convergent correlation was .38. 3

Procedure

Participants were recruited through mail solicitations and personal visits. Employees who initially agreed to participate in the study received a packet of questionnaires. Each packet contained the following items: (a) the five self-report questionnaires (described above in the Method section) that were completed by the participant (to control for possible order effects, half of the participants were asked to complete the CMS before the OCB questionnaire, and the other half were asked to complete the CMS after the OCB questionnaire 4 ), (b) two OCB rating forms to be completed by peers of the participant, and (c) one OCB rating form to be completed by a supervisor of the participant. Participants were assured that all of these data would be anonymous and that no one in the organization would have access to them. To achieve this anonymity, each form in the packet was assigned a unique code number known only to Sheila M. Rioux and the participant.

After the participants completed the self-report measures, they asked two coworkers with whom they had regular contact and their immediate supervisor to complete the appropriate OCB rating form. To make it clear to the peers and supervisors whom they were rating, the participants wrote their names on a "post-it" note and placed it on the rating forms they distributed. After completing the questionnaires, all individuals (participants, peers, and supervisors), removed the post-it notes from the questionnaires, placed them in sealed envelopes, and returned the sealed envelopes to a central office in the organization through interoffice mail. Memos were periodically distributed to participants indicating what forms the researcher had received for each code number. The data were collected over a 6-week period.

Results Psychometric Properties of the CMS

The factor structure and the individual factor loadings were virtually identical to those obtained with the student samples in Study 1. The coefficients of congruence between the solutions from the student and the worker samples were .97 for all three factors. The pattern of correlations among the factors was also similar to the pattern found in the student samples. None of the subscale means in the two samples differed by more than 3 points. However, relative to the student sample, the OC and PV means were significantly higher, and the IM mean was significantly lower.

Motives and OCB Self-ratings.

Before considering motives and OCB, we briefly discuss the intercorrelations among the predictor variables (i.e., procedural justice, prosocial personality, mood, and motives). These correlations provide data that speak to the construct validity of the CMS independent of its relationship with OCB. We expected that OC motives would correlate more strongly than the other two motives with variables that concerned affective reactions to the organization, such as organizational justice and mood. In contrast, PV motives would correlate more strongly with the prosocial personality measure. For the most part, these expectations were supported. The correlations between OC motives and procedural justice (.44) and mood (.49) were significant and higher than the correlations between either PV motives and these variables (.24 and .21) or IM motives and the same variables (.07 and .12). The correlations between PV motives and the other-oriented empathy and helpfulness dimensions of the PSB (.46 and .31, respectively) were also significant and higher than the comparable correlations for the OC motives (.27 and .02) and the IM motives (-.02 and .00). The only nonsignificant difference was PV—other-oriented empathy versus OC—other-oriented empathy. 5

Turning to OCB, significant correlations were obtained between procedural justice and two of the OCB measure's subscales (Altruism and Sportsmanship); mood correlated significantly with all of the subscales except Conscientiousness. Other-oriented empathy and helpfulness correlated significantly with the Altruism and Courtesy subscales of OCB. Other-oriented empathy also correlated significantly with the Conscientiousness and Civic Virtue subscales.

The first hypothesis was that the motives would differentially correlate with OCB dimensions. There were significant correlations between PV motives and Altruism (.44) and Courtesy (.43; both altruism-related OCB subscales). As we expected, these correlations were significantly higher than those between PV and Conscientiousness (.13) and Sportsmanship (.03; both conscientiousness-related OCB subscales). However, contrary to the hypotheses, there was a significant correlation (.27) between PV motives and Civic Virtue (another conscientiousness-related subscale). OC motives correlated significantly with Civic Virtue (.55) and Sportsmanship (.25). The OC motives—Civic Virtue correlation was significantly higher than the PV motives—Civic Virtue correlation. However, OC motives also correlated significantly with Altruism (.32) and Courtesy (.25). (The unexpected findings are addressed below.)

The other major hypothesis was that the motives would account for unique variance in the OCB dimensions. To test this hypothesis, we carried out hierarchical linear regressions in which the criterion variables were the five OCB subscales and the predictors were procedural justice, distributive justice, mood, the two factors of the prosocial personality measure, and the three motives, entered last (PV, followed by OC and IM). PV motives accounted for significant amounts of unique variance in the regressions involving Altruism and Courtesy (both D R 2 s = .06). However, they also accounted for significant variance in a conscientiousness-related subscale, Civic Virtue (D R 2 = .03). OC motives accounted for unique variance in Civic Virtue (D R 2 = .13). IM motives accounted for significant variance in Sportsmanship (D R 2 = .04). None of the motives (or any other predictors) had significant regression weights on the Conscientiousness subscale.

These results of correlational and regression analyses were generally consistent with the hypothesis about how the motives would map onto the OCB dimensions. However, an unexpected finding was that PV motives correlated significantly with Civic Virtue and OC motives correlated significantly with Altruism and Courtesy. Also in the regression analysis, PV motives accounted for significant variance in Civic Virtue. To better understand these findings, we carried out additional analyses in which the substantial correlation (.57) between the PV and OC subscales was partialed out from these variables' relationships with OCB. Specifically, two hierarchical multiple regressions, using only motives, were performed on each of the OCB measure's subscales. In one, PV motives were entered before OC motives; in the other, this order was reversed. (IM motives were always entered last.) Whether entered first or second, PV motives accounted for significant amounts of variance in the Altruism subscale (D R 2 s = .19 and .10). OC motives accounted for significant variance in this subscale only when they were entered before the PV motives. Turning to Civic Virtue (a conscientiousness-related subscale), OC motives accounted for significant variance irrespective of the order in which they were entered (D R 2 s = .31 and .24). However, PV motives accounted for significant variance in Civic Virtue only when they were entered before OC motives.

Peer ratings.

Two peers rated 135 of the participants on OCB. Eighty-seven percent of the peers had worked with the target person for at least 1 year, and 86% said that they had observed the target person at least 10 times in the past week. Averaged peer ratings were used in the analyses of this rating source. The correlations involving peer ratings of OCB are presented in Table 2 . Procedural justice correlated significantly with all of the OCB subscales except Courtesy, and distributive justice and mood correlated significantly with Sportsmanship ( r s = .17 and .22, respectively).

The PV motives—Altruism correlation (.24) was significantly higher than the PV motives—Conscientiousness (-.04) and PV motives—Sportsmanship (.11) correlations but did not differ from the PV motives—Civic Virtue correlation (.23). The OC motives correlated significantly with Civic Virtue (.34) and Sportsmanship (.24). The OC motives—Civic Virtue correlation was significantly higher than the OC motives—Altruism (.08) and OC motives—Courtesy (.05) correlations and nonsignificantly larger than the PV motives—Sportsmanship (.11) correlation. No significant correlations were found for the IM motives.

The peer ratings on the five OCB subscales were regressed onto the predictor variables in hierarchical regressions, with the motives entered last (see Table 3 ). Three of the five multiple regressions (Altruism, Civic Virtue, and Sportsmanship) yielded a significant R 2 . Again, the PV motives accounted for significant amounts of variance in the prediction of Altruism and Civic Virtue. The OC motives accounted for significant amounts of unique variance in Civic Virtue, and the IM motives accounted for significant amounts of unique variance in Sportsmanship.

Regression analyses were again used to partial out the covariation between PV and OC motives from their relationships with the OCB subscales. The unexpected significant relationship between PV motives and Civic Virtue disappeared when PV motives were entered after OC motives, but irrespective of the order in which OC motives were entered, they accounted for significant variance in Civic Virtue.

Supervisor ratings.

Supervisors rated 135 participants on OCB. Procedural justice correlated significantly with Sportsmanship ( r = .21), and mood correlated significantly with Conscientiousness ( r = .19). The PV motives correlated significantly (.21) with Altruism; this correlation was nonsignificantly larger than the correlations involving PV motives and Conscientiousness (.16), Civic Virtue (.17), and Sportsmanship (.14). As we predicted, the OC motives correlated significantly with the Conscientiousness, Civic Virtue, and Sportsmanship subscales ( r s = .25, .25, and .20, respectively). The first two correlations were significantly greater than the OC motives—Courtesy correlation (.03). None of the correlations involving IM motives were significant. None of the hierarchical regressions were significant. However, when entered into the regression, OC motives produced a significant increment in the variance accounted for in Civic Virtue (D R 2 = .04).

The regressions conducted only on the motives explicated the relationships between them and the OCB subscales. PV motives accounted for unique variance in Altruism irrespective of the order in which they were entered (both D R 2 s = .04). However, the relationships between PV motives and Conscientiousness, Civic Virtue, and Sportsmanship (i.e., conscientiousness-related subscales) dropped to near zero when PV motives were entered after OC motives . OC motives accounted for significant variance in Conscientiousness and Civic Virtue irrespective of whether they were entered first or second (D R 2 s = .06 and .04 for Conscientiousness; D R 2 s = .06 and .03 for Civic Virtue).


Date: 2015-01-29; view: 1028


<== previous page | next page ==>
Study 1: Scale Development | Discussion
doclecture.net - lectures - 2014-2024 year. Copyright infringement or personal data (0.008 sec.)