Perceptions of Organizational Politics as a Moderator

0 downloads 0 Views 773KB Size Report
inquiry by assessing the moderating effect of perceptions of orga- nizational politics on .... Porter (1976) argued that organizational politics is an important area of ...
Journal of Applied Psychology 2000, Vol. 85, No. 3. 472-478

Copyright 2000 by the American Psychological Association, Inc. 0021-9010/00/S5.00 DOI: 10.1037//0021-9010.85.3.472

RESEARCH REPORTS

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.

Perceptions of Organizational Politics as a Moderator of the Relationship Between Conscientiousness and Job Performance Wayne A. Hochwarter

L. A. Witt

University of Alabama

University of New Orleans

K. Michele Kacmar Florida State University

Meta-analytic studies of the relationships between the five-factor model of personality constructs and job performance indicate that conscientiousness has been the most consistent predictor. Recent research has sought to identify situational factors that may explain additional variance beyond what has been reported by simple bivariate relationships. The authors hypothesized that perceptions of organizational politics would moderate the relationship between conscientiousness and job performance. Data collected from 234 male and 579 female workers in 4 organizations indicated that conscientiousness was related to job performance among workers perceiving average to high levels of organizational politics but unrelated to performance among workers perceiving low levels of organizational politics. Moreover, perceptions of organizational politics were negatively related to job performance only among workers of average to low levels of conscientiousness.

Recent literature reviews (e.g., House, Shane, & Herold, 1996) illustrate the advances in research made by examining dispositional predictors of job attitudes and outcomes across a variety of organizational domains. Much of this progress can be attributed to the emergence of the five-factor model (FFM) of personality (Costa & McCrae, 1988). Researchers (e.g., Digman, 1990) have confirmed that the five factors represented in the model—extraversion, agreeableness, conscientiousness, emotional stability, and openness to experience—embody a concise taxonomy for examining personality traits. Meta-analyses (Barrick & Mount, 1991; Hough, Dunnette, Eaton, Kamp, & McCloy, 1990; Salgado, 1997; Tett, Jackson, & Rothstein, 1991; Vinchur, Schippmann, Switzer, & Roth, 1998) indicate that the FFM constructs are valid predictors of performance for some, but not all, occupations. Of the five,

conscientiousness has been the most generalizable predictor (Barrick, Mount, & Strauss, 1993). Given the empirical support for the predictive ability of the conscientiousness dimension for job performance, the next plausible avenue of research is to examine moderating variables that explain variance beyond that attributed to main effects. Barrick and Mount (1993) suggested that "an issue that has yet to be adequately addressed is whether the validities of personality constructs are influenced by the presence of moderator variables" (p. 111). This is an issue of practical as well as theoretical significance. For example, managers want to hire the right employees and put in place an environment that permits them to perform their best. The purpose of the present study is to pursue this line of inquiry by assessing the moderating effect of perceptions of organizational politics on the relationship between conscientiousness and job performance.

Conscientiousness and Job Performance

Wayne A. Hochwarter, Department of Management, University of Alabama; L. A. Witt, Department of Management, University of New Orleans; K. Michele Kacmar, Department of Management, Florida State University. We presented an earlier version of this article in April 1997 during the annual meeting of the Southern Management Association, Atlanta, Georgia. We thank Gerald R. Ferris, Diane E. Johnson, Michael K. Mount, and Greg Stewart for their comments on earlier versions of this article. Correspondence concerning this article should be addressed to L. A. Witt, Department of Management, University of New Orleans, New Orleans, Louisiana 70148-1560. Electronic mail may be sent to [email protected].

Conscientiousness refers to behavior that is "responsible, dependable, persistent, and achievement-oriented" (Barrick & Mount, 1993, p. 111). Gellatly (1996) noted that the traits that differentiate high-conscientious individuals (i.e., ambitious, exacting, methodical, and disciplined) from low-conscientious individuals (i.e., lazy, imprecise, impetuous, and disorganized) lead those in the former group to expect and strive for greater success on the job. It is not surprising, then, that individuals high in conscientiousness are described as having "ideal employee" traits (Mount & Barrick, 1995a) and tend to be higher performers. Indeed, 472

RESEARCH REPORTS

Barrick and Mount (1991) demonstrated that conscientiousness predicts the core task proficiencies of virtually all jobs. To replicate previous findings, we proposed the following:

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.

Hypothesis 1: Conscientiousness scores are positively related to supervisor ratings of performance.

Although corrected validity coefficients average around .30, meta-analytical studies suggest that the observed relationships between conscientiousness and job performance are often in the low teens (Barrick & Mount, 1991; Hough et al., 1990; Salgado, 1997; Tett et al., 1991; Vinchur et al., 1998). In single studies, validity coefficients have ranged from near zero (Grant, 1995) to the high .20s or low .30s (Barrick & Mount, 1993). Barrick and Mount (1991) concluded that the presence of moderator variables may partially account for the modest size of observed personalityjob performance validity coefficients. This is important for at least two reasons. First, identifying moderators of the conscientiousness-job performance relationship would help managers target appropriate supervisory practices for subordinates at different points along the conscientiousness continuum. Second, it would facilitate the use of conscientiousness as a selection criterion in situations in which it matters most. Gellatly (1996) noted that it is "shortsighted to make a selection on the basis of conscientiousness and then place new hires in work environments that potentially inhibit the natural expression" (p. 480) of conscientious behavior. He advocated the use of managerial practices that bring out the best in conscientious workers. Barrick and Mount (1993) demonstrated that providing autonomy is one such managerial practice. They found that increased levels of autonomy were associated with greater levels of performance among highly conscientious managers. However, autonomy was unrelated to the performance of managers who were of average or below-average levels of conscientiousness. Another strategy to increase worker performance may be to address aspects of the work environment that bring out the worst in workers with average to low levels of conscientiousness. An environment that potentially has deleterious effects on employees (thus bringing out their worst) is one fraught with perceived organizational politics. Following Mintzberg (1983), we define organizational politics as behaviors intended to promote self-interest, often without regard to or even at the expense of organizational goals. Examples of organizational politics include providing resources to coworkers who support one's agenda and withholding them from those who do not, assigning higher raises to subordinates who are yes men rather than to those who sometimes question decisions, and sabotaging the work efforts of coworkers who are not advocates of one's positions. Although some theoretical work has considered organizational politics in neutral terms (e.g., Ferris & Judge, 1991), most definitions of organizational politics imply involvement in self-serving actions not sanctioned by the organization (e.g., Parker, Dipboye, & Jackson, 1995). Theorists (Ferris, Russ, & Fandt, 1989; Gandz & Murray, 1980) have suggested that politics is appropriately viewed as a subjective phenomenon. Porter (1976) argued that organizational politics is an important area of inquiry even if it may sometimes represent misconceptions of actual events. Consistent with the theme that situations should be assessed in terms of how they are perceived (e.g., Lewin, 1936), most contemporary studies have treated per-

473

ceptions of politics, not objective indices, as the most critical predictor of attitudes and behaviors (e.g., Cropanzano, Howes, Grandey, & Toth, 1997; Maslyn & Fedor, 1998). Although some work has indicated that political behaviors lead to beneficial outcomes (e.g., Madison, Allen, Porter, Renwick, & Mayes, 1980), the preponderance of literature consists of reported links between perceptions of organizational politics and detrimental individuallevel outcomes, including higher levels of stress, anxiety, and job dissatisfaction and lower levels of organizational commitment and job performance (for a recent literature review, see Kacmar & Baron, 1999). Political behaviors are likely to occur when individuals do not see the environment providing normative guidelines for behaviors (Ferris et al., 1989). For example, when organizational priorities or values are unclear and there is uncertainty about organizational criteria for raises and promotions, some workers may avoid risk and take measures to ensure that they receive their fair share of the available rewards. One such step might be to create the image of a team player who does not rock the boat by questioning management prerogatives. To achieve such an image, workers may go along with plans that they know to be flawed but that are favored by key managers. Workers less motivated by self-interest may put forth effort toward goals that they genuinely believe to be important to the organization but that in fact are viewed as unimportant by most others (Mayes & Allen, 1977). In either scenario, others in the organization may view these workers as acting either in self-interest or against the interests of the organization and thus consider the behaviors to be political (Zhou & Ferris, 1995). In other words, in the absence of widely accepted norms that define priorities and protocol for decisions and criteria for personnel actions, workers are left to their own devices to get the work done and further their career aspirations (Pfeffer, 1981). Similarly, observers are left to their own devices to determine whether others are acting in self-interest or on behalf of the organization. When others believe an individual is acting in self-interest, they may feel the need to engage in their own defensive tactics (Mayes & Allen, 1977). When they see obstacles, workers who are high in conscientiousness seek out and attain higher levels of performance than low-conscientious workers because they are by predisposition more persevering and disciplined (Colquitt & Simmering, 1998), more confident in their abilities (Barrick and Mount, 1991), and more proactive and effective in goal setting (Barrick et al., 1993; Gellatly, 1996). Therefore, they are likely to be effective in identifying key priorities and finding ways of getting things done even when perceiving high levels of politics in the workplace. In contrast, lazy and disorganized low-conscientious workers may be handicapped when seeing themselves in such situations. Lowconscientious workers lack the patience, tenacity, diligence, and sense of urgency (Digman, 1990) to seek out actual organizational priorities and to find ways to get things done in an environment they see as political. Thus, perceptions of high levels of organizational politics may be particularly detrimental to individuals with low levels of conscientiousness. The notion that perceived high levels of politics may promote an opportunity for individual differences in conscientiousness to affect job performance is consistent with previous work by personality theorists. For example, Mischel (1977) suggested that indi-

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.

474

RESEARCH REPORTS

vidual differences in personality determine behavior when individuals "have no clear expectations about the behaviors most likely to be appropriate (normative, reinforced) in that situation" (p. 347). As noted earlier, the perception of high levels of organizational politics is likely to reflect such a condition (Ferris et al. 1989). In contrast, when individuals understand what is expected (a condition likely reflecting the perception of low levels of politics), they tend to conform to the norm, and personality is less important (Mischel & Peake, 1982). Therefore, we suggest that when individuals perceive low levels of politics (i.e., when they see others openly sharing critical work-relevant information, established objectives, and normative guidelines for decision making), conscientiousness may not be a distinguishing factor in performance. Hypothesis 2: Perceptions of organizational politics moderate the conscientiousness-performance relationship. The relationship is stronger among workers perceiving high levels of organizational politics than among those perceiving low levels of politics. Moreover, perceptions of organizational politics are negatively related to job performance primarily among workers low in conscientiousness.

Method Participants and Procedure We collected complete data on 813 individuals from four organizations as part of a larger program of research conducted for both research and applied purposes. We asked that employees report to a training room in small groups at their respective sites. We informed them about the study, provided them with a chance to ask questions, and gave them opportunities to excuse themselves from involvement in the project. At approximately the same time, we asked supervisors to complete performance ratings on their subordinates. A total of 234 (29%) of the participants were men, 579 (71%) were women, and 331 (41%) were ethnic minorities. Sample 1. We asked that all 310 nonsupervisory workers in a distribution services organization participate in the study; 254 (82%) volunteered. We collected complete data on 154 (49%) employees, of whom 32 (21%) were men, 122 (79%) were women, and 93 (60%) were ethnic minorities. To determine the representativeness of the sample, we collected population statistics. The population of 310 consisted of 24% men, 76% women, and 64% ethnic minorities. Sample 2. We asked that all 487 employees of a production organization participate in the study; 390 (81%) volunteered. We collected complete data on 355 (73%) employees, of whom 73 (21%) were men, 282 (79%) were women, and 115 (32%) were ethnic minorities. The population consisted of 16% men, 84% women, and 35% ethnic minorities. Sample 3. We asked that all 410 nonsupervisory sales agents in three different call centers of a telemarketing organization participate in the study; nearly 90% (378) volunteered. In this organization, we were able to collect data from each participant at two different times, roughly 3-4 months apart. This helped reduce the likelihood of problems associated with common method variance. We collected complete data on 207 (50%) participants, of whom 75 (36%) were men, 132 (64%) were women, and 107 (52%) were ethnic minorities. The population consisted of 27% men, 73% women, and 51% ethnic minorities. Sample 4. We sent memoranda to managers of 356 nonsupervisory software engineers in a systems development organization, requesting that they ask their subordinates to participate in our study. We were unable to ascertain how many employees received notification of the opportunity to participate. A total of 126 workers attended one of six small-group meetings in which they completed questionnaires. We collected complete data

on 97 (77%) workers, of whom 67 (69%) were men, 30 (31%) were women, and 16 (17%) were ethnic minorities. The population consisted of 42% men, 58% women, and 12% ethnic minorities.

Measures Perceptions of organizational politics. We measured perceptions of organizational politics by the Going Along to Get Ahead subscale of the Perceptions of Organizational Politics Scale (Kacmar & Ferris, 1991). Items such as "There is no place for yes-men around here; good ideas are desired even if it means disagreeing with superiors" (reverse-scored), "Employees are encouraged to speak out frankly even when they are critical of well-established ideas" (reverse-scored), "It is safer to agree with management than to say what you think is right," and "Rewards come only to those who work hard in this organization" measured the degree to which respondents perceived workers in the organization to be strategically withholding dissenting ideas or information to receive valued rewards. Previous studies (e.g., Cropanzano et al., 1997) have reported average reliabilities for the full scale in the high 80s. Kacmar and Carlson (1997) reported additional validity evidence for the full scale and for each subscale. Higher scores indicate perceptions of higher organizational politics. Conscientiousness. We assessed conscientiousness using the 30-item conscientiousness scale of the Mount and Barrick (1995b) Personal Characteristics Inventory (PCI). For replication, we also administered the 10item detail-conscious scale (a = .70) from Saville, Holdsworth, Nyfield, Cramp, and Mabey's (1984) Occupational Personality Questionnaire (OPQ) to Sample 3. Higher scores indicate greater levels of conscientiousness. Job performance. We developed criterion items based on job analysis results from each sample. In Sample 1, the supervisors rated employee performance using eight items (e.g., "[employee name] does not knowingly repeat mistakes"). The supervisors in Sample 2 rated five items (e.g., "[employee name] takes responsibility for errors"). In Sample 3, we measured sales performance using a three-item scale (e.g., "[employee name] uses cross-selling as an opportunity to benefit customers") completed by the supervisors. As part of their everyday functions, supervisors in this organization frequently monitored calls of their direct reports and thus were capable of making these assessments. Supervisors in Sample 4 rated employee performance using a four-item scale (e.g., "[employee name] complies with standards for documenting projects"). In each sample, supervisors provided responses to the items on a 5-point Likert-type scale ranging from 1 (weak or bottom 10%) to 5 (best or top 10%). Control variables. Because demographic variables, education, and general mental ability (GMA) have been shown to explain variance in performance (Barrick et al., 1993; Hofmann, Jacobs, & Gerras, 1992; Howard, 1986), we included them as control variables in the analyses. We gathered Equal Employment Opportunity Commission-defined ethnic status (converted into ethnic nonminority = 1 and ethnic minority = 2 categories), sex (men = 1, women = 2), age, and organizational tenure from the archives of the respective organizations. We gathered selfreported education levels on the identification page of the PCI. Because of differences in the nature of the jobs, we used two different, timed GMA measures. We used the Wonderlic Personnel Test (Wonderlic Personnel Test, Inc., 1992) in Samples 1, 2, and 4 and the Watson-Glaser Critical Thinking Appraisal (Watson & Glaser, 1994) in Sample 3. Higher scores reflect higher GMA.

Results We analyzed the data from each sample separately and found very similar results. To simplify reporting, we standardized the GMA and performance scores, combined the data across samples, and present results of the analyses conducted on the combined data

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.

RESEARCH REPORTS set. The descriptive statistics, reliability estimates, and intercorrelation matrix are presented in Table 1. Hypothesis 1 was confirmed; workers scoring high in PCI-measured conscientiousness were rated higher in job performance than those scoring low in conscientiousness (r = .16, p < .01). This is slightly higher than the mean observed validity coefficients reported in meta-analytic studies (Barrick & Mount, 1991; Salgado, 1997; Tett et ah, 1991; Vinchur et al., 1998). Hierarchical moderated multiple regression analysis (Cohen & Cohen, 1983) was used to test Hypothesis 2. We first entered the control variables into the equation as one block. We then entered conscientiousness and perceptions-of-politics scores at the second and third steps, respectively. We entered the cross-product term (Conscientiousness X Organizational-Politics Perceptions) at the final step. As shown in Table 2, the cross-product term of PCImeasured conscientiousness and organizational politics explained a significant incremental portion of variance (A/?2 = .01, p < .01) in job performance over and above the main effects of the demographic control variables, GMA, education, conscientiousness, and politics. We repeated the analyses on Sample 3 data using the OPQ measure of conscientiousness. OPQ-measured conscientiousness was unrelated to performance in Sample 3 (r = .01, ns). However, the change in R2 of the cross-product term of OPQ-measured conscientiousness and organizational politics from Sample 3 was statistically significant, adjusted R2 = .05, Afl2 = .02, F(9, 193) = 5.00, p < .03. These effects sizes (AR2) are within the typical range (i.e., A/?2 = .01-.03) for moderator effects in nonexperimental studies (Champoux & Peters, 1987; Chaplin, 1991) and indicate that perceptions of politics had an effect on job performance across the entire range of conscientiousness scores. We plotted the prediction-of-performance scores at the mean as well as at high and low levels of politics (1.0 and —1.0 standard deviations from the mean; Stone & Hollenbeck, 1989). Figure 1 reveals the interactive effects. Perceptions of high levels of organizational politics may have been particularly detrimental to lowconscientious workers, as they had lower performance scores than high-conscientious workers. Considering Cohen's (1988) categories of effect sizes (.20 = small, .50 = medium, and .80 = large), politics perceptions had at least a small effect on job performance among workers with conscientiousness scores at or below 2.64 (z = — .26). The impact of conscientiousness scores on job-per-

475

formance scores was dependent on the level of politics perceptions: Conscientiousness was related to job performance among workers perceiving organizational politics to be at or above average levels. In contrast, individual differences in conscientiousness had little, if any, relationship to performance among workers seeing low levels of organizational politics. The graph of the interaction using the OPQ measure of conscientiousness collected in Sample 3 (not shown) was similar in form.

Discussion Our test of Hypothesis 1 yielded mixed results. When measured by the PCI, conscientiousness was significantly related to performance. When measured by the OPQ, it was not. Perhaps different operationalizations of conscientiousness have contributed to the inconsistency of previous findings. However, our results confirmed Hypothesis 2 regardless of the conscientiousness measure used. Specifically, conscientiousness explained variance in supervisor-rated job performance for workers perceiving moderate to high levels of organizational politics. When perceiving moderate to high levels of politics, workers may have relied on idiosyncratic preferences for getting the work done and promoting their careers. In such cases, conscientious workers may have been more successful than those low in conscientiousness. However, among workers perceiving low levels of organizational politics, conscientiousness explained little, if any, variance in performance. These workers may have received sufficient guidelines and support for getting the work done such that the paths to achieving expectations were clear and needed few individually mapped detours (Ferris et al., 1989). This study has several strengths. First, congruent with our theoretically derived hypothesis, we found a consistent conscientiousness-politics interaction across four samples, providing some evidence of replication as advocated by Golding (1975). Moreover, the form of the interaction was similar when two different measures of conscientiousness were used. Second, we found the interactive effect even after controlling for demographic variables previously shown to be related to job performance, namely, sex, age, tenure, education, ethnic minority status, and GMA. Third, variations in our data collection procedures helped reduce the likelihood that common method variance was a prob-

Table 1 Intercorrelation Matrix and Descriptive Statistics Variable

M

SD

1

2

3

4

5

6

7

8

9

1. Performance" 2. Perceptions of politics 3. Conscientiousness 4. Sex

-0.01 2.84

0.99 0.52 0.21 0.54 0.49 0.97 4.86 1.82 11.09

_ -.11* .16*** .05 -.14*** .18*** .16*** .05 .12***

(.71) -.17*** .06 .10** -.17*** -.04 .03 -.02

(.77) .10** -.09** .06 .06 .04 .16***

— .03 -.09** .05 -.24*** .01

— -.33*** -.14*** -.02 -.18***

— .05 .19*** .11**

— -.04 .42***

— .11**



5. Ethnic minority 6. General mental ability 7. Tenure 8. Education 9. Age

2.69 1.66 1.41 0.06 4.40 1.82 32.09

Note. Internal reliability estimates (a) are presented in the diagonal. * The different performance measures were converted to z scores for the combined sample; Sample 1 a = .93; Sample 2 a = .75; Sample 3 a = .82; Sample 4 a = .75. *p

Suggest Documents