An empirical examination of the impact of performance attributions and job satisfaction on turnover intentions.

Huning, Tobias M. ; Thomson, Neal F.

INTRODUCTION

"Voluntary turnover" has been one of the most salient topics in management research for at least the last half century (March & Simon, 1958, Hom & Kinicki, 2001). Every year, companies spend significant sums of money replacing employees who voluntarily separate from their organizations. The costs associated with voluntary employee turnover include disruptions of work, loss of knowledge, skills, and organizational memory (Griffeth & Hom, 2001). A key goal for many organizations is to effectively manage voluntary turnover of employees that is caused by dissatisfaction with their jobs or employers. Extant research recognizes that attitudes and intentions explain around 5% and 15% of the turnover variance respectively (Griffeth, Hom, & Gaertner, 2000, Hom & Griffeth, 1995). Since one of the key determinants of turnover is the intention to turnover, a key question becomes "what causes an employee to decide that they want to leave?" The study that follows proposes that a key factor in this process is the style of attributions used by employees to explain their performance successes and failures. To be sure, we test a model which postulates that employees quit their jobs based on attributions they make regarding their performance.

ATTRIBUTION THEORY

Attribution theory has its roots in Heider's (1958) description of the "naive psychologist" who attempts to find causal explanations for events and human behaviors. Several models have been developed from this idea, which attempt to explain the process by which these attributions are made both in the case of self attribution (e.g. Weiner, 1974; Abramson, Seligman & Teasdale, 1978) and social attributions or attributions made regarding the behaviors and outcomes of others (e.g. Kelley, 1973, Thomson and Martinko, 2004).

Weiner (1974), in his development of the achievement motivation model of attributions, classified causal attributions across two dimensions; the locus of causality, and the stability of the cause. The first, locus of causality, originally proposed by Rotter (1966), is the degree to which the attributed cause is internal to the person, or part of the external environment. Internal attributions might include factors such as low intelligence, or lack of attention. External attributions could include weather conditions, or task difficulty. A second dimension, stability, refers to the degree to which the cause remains constant over time. The example of low intelligence would be stable, where the example of lack of attentiveness, would be unstable. Weiner (1979) and Zuckerman and Feldman (1984) added the dimension of controllability to the achievement motivation model. This dimension focused on whether the cause of an event or behavior is controllable or uncontrollable.

McAuley, Duncan and Russell (1992) expanded the concept of controllability by proposing dual dimensions of personal and external control. For personal control, the attributor indicates that he or she either can or cannot personally control the outcome of the event. The external control dimension measures the degree to which the attributor sees the situation as being controllable by anyone else, such as a supervisor or co-worker. As Vielva and Iraurgi, (2002) point out, a response indicating external control, is different than a response indicating uncontrollability. This paper proposes that type of attribution made by an employee across these dimensions is likely to impact an employee's satisfaction with their job, as well as the likelihood that they will decide that they want to leave their position.

JOB SATISFACTION

Job satisfaction is the most studied variable in organizations. Job satisfaction has been defined as a pleasurable emotional state the results from the appraisal of one's job (Locke, 1976). In other words, job satisfaction describes an affective reaction to one's job as well as attitudes toward the job. This in turn suggests that job satisfaction is formed from affect, cognition, and ultimately will result in satisfaction contingent job-related behaviors. Some of the most commonly studied outcomes of job satisfaction are organizational citizenship behaviors, absenteeism and turnover (Organ & Ryan, 1995; Wegge, Schmidt, Parkes, & van Dick, 2007; Saari & Judge, 2004). Job satisfaction is the central variable in among the central theoretical and empirical contributions in employee turnover.

TURNOVER INTENTIONS

Voluntary turnover refers to an employee voluntarily leaving and organization. Early approaches such as March and Simon's (1958) contributions and inducements model have identified that job satisfaction determines the perceived desirability of movement, which ultimately determines whether an individual quits the job or not. In March and Simon's model job satisfaction is driven by the match between the job and the self-image, the match between the job and other roles, as well as the predictability of future relationships inside the organization. Additionally, based on the aforementioned dissatisfaction, quitting is contingent on an evaluation of the expected utility of the perceived alternatives.

Furthermore, Mobley (1977) suggested that job satisfaction follows and evaluation of one's existing job, which then triggers a sequences of cognitive and behavioral process leading to the quit/stay decision. It is however, essential to note that Mobley anchors his theory on the experience of job satisfaction-dissatisfaction.

Recent theorizing has included the role of job performance into the employee withdrawal process. Allen and Griffeth (2001) hypothesized and found evidence for the moderating effect of the ease of visibility on the relationship between performance and perceived ease of movement. They also found that rewards moderate the relationship between performance and the desirability of movement. It is important to note that not all voluntary turnover is bad. In fact, it would be desirable to most organizations for weak performers to quit.

Collectively, this leads us to ask whether the quit decision of employees is contingent on their implicit theories about the causes of their performance. In other words, we ask if the attributions employees make regarding their performance determined their levels of job satisfaction and ultimately their intentions to quit or remain with their respective organizations. In the following section we present specific research hypotheses grounded in attribution theory and based on a rich body of knowledge on voluntary turnover.

HYPOTHESES

Past studies have looked at the role of attributions in job satisfaction (McCormick, 1997, Norris and Niebuhr, 1984). Of specific relevance to this study, Norris and Niebuhr (1984) found that individuals who tended to attribute their performance to internal causes also had higher job satisfaction. Based on their findings, we hypothesize the following:

H1: Locus of causality will be related to job satisfaction with internal attributions leading to higher job satisfaction and external attributions leading to lower job satisfaction.

Additionally, there are numerous studies examining the role of job satisfaction on turnover intentions. Tett and Meyer (2006) provide a meta-analytical examination of past findings in this area and conclude that job satisfaction is very strongly related to turnover intentions, having a greater effect than organizational commitment. Therefore, based on their meta-analytical examination of 155 studies in the area we hypothesize:

H2: Job Satisfaction will be negatively related to turnover intentions.

A recent study by Harvey, Harris and Martinko (2008) examined the role of attributions as predictors of job satisfaction, stress and turnover intentions. This was one of the first studies to examine the roles of these variables simultaneously, and specifically to include attributions. While their focus was specifically on hostile attributions, the findings relate to our study as well. They found a relationship between hostile attribution styles and turnover intentions. Hostile attribution style is explained as "blaming others when things go wrong in their lives." (Harvey, Harris and Martinko 2008) This relates to the CDSII dimensions as follows: Blaming others is external LOC, but also high external control and low personal control. Hostile attributions generally also indicate a bias toward high stability, as the "offender" is likely not to change. Based on their finding of a relationship between attributions and turnover intentions, we hypothesize:

H3: External LOC will be related to higher turnover intention

H4. High stability will be related to turnover intentions

H5. High external control will be related to higher turnover intention

H6: Low internal control will be related to higher turnover intention.

METHOD AND SAMPLE

Participants were 363 students at a regional state university located in the southeastern United States. The sample consisted of graduate and undergraduate students at the university's college of business. We distributed a survey instrument together with a cover letter and consent form. We asked the participants to read the cover letter and sign the consent form, provided they chose to participate. The cover letter explained the study and reiterated the fact that participation was voluntary. We explained that incentives were (or were not) provided at the discretion of the respective course instructor. The participants were also informed that they could discontinue the survey at any time without penalty or loss of reward that they were otherwise entitled to receive. We instructed the participants that they were to treat these questions as they relate to the jobs the currently hold, a job they have held in the past in case they currently did not work, or if they have never worked to treat being a student as their current job. The survey contained the measurement scales as well as questions on demographics of the participants. The participants took the survey during their respective class periods. 99% of the participants returned a usable survey.

About fifty-one percent (51.2%) of the participants were female, 47.1% were male; 1.7% did not respond to this question. The average age was between 23 and 25 years of age with 9.9% of the sample age 35 or older. 56.5% were white (non-Hispanic), 30% African-American, 4.7% Hispanic, 3.6% Asian, .6% Native American, and 2.2% specified as "other", 43.3% responded that they had high school diplomas, 11.8% indicated they had associate degrees, 38.9% stated they held a bachelors degree, and 3.9% stated that they had master's degrees. .3% suggested they had doctorates. The average work experience of this sample was 6 years and 5 months. 92.4% of the respondents had at least one year of work experience, 83.6% reported work experience of at least 2 years, 46.4% reported 5 years or more, and 14.8% expressed that they had worked for at least 10 years. We believe that this demographic composition of the sample makes a strong argument for the generalizability of the sample to an average "working" population. The average participant also maintained a 2.9 GPA.

MEASURES

Attributions

For the measurement of performance attributions, we used the Causal Dimension Scale II (CDS II), developed by McAuley, Duncan and Russell (1992). The CDS II consists of a 12 questions, which make up 4 scales, with three items per scale, which evaluated the attributional dimensions of (1) locus of causality, (2) external control, (3) stability, and (4) personal control. Reliabilities using the CDS II are generally reported to be high (McAuley, Duncan and Russell, 1992). The reliabilities of the scales in our sample are as follows: Locus of causality [alpha] = .74, external control [alpha] =.7, stability [alpha] = .6, and personal control [alpha] = .83

Job satisfaction

Job satisfaction was measured with 3 items from Hoppock (1935). Respondents rated the items on a 5-point Likert-type response scale (1 = "Strongly disagree"; 5 = "Strongly agree"). A sample item is, "All in all, I am satisfied with my job." This scale produced a coefficient alpha of .89.

Turnover intentions

Turnover intentions were measured with three items adapted from the scale developed by Hom and Griffeth (1991). The items were rated on a 5-point Likert-type scale ranging (1 = "Definitely not"; 5 = "Definitely yes"). The scale produced a coefficient alpha of .92.

ANALYSIS

We conducted a series of regression analyses to examine the relationships between attribution styles, job satisfaction, and turnover intentions using SPSS. Means, standard deviations, reliabilities, and correlations are reported in Table 1.

The initial results show that, as hypothesized in H1, based on the findings of Norris and Niebuhr (1984) locus of causality is significantly related to job satisfaction at p=.01. The standardized path coefficient for the relationship between locus of causality and job satisfaction was estimated to be [beta] = .20. The remaining dimensions, external control, personal control, and stability were not statistically significant with respect to job satisfaction. However, external control could be described as marginally significant at p=.08 with a standardized path coefficient of [beta] = -.10.

Further, we tested whether the attribution dimensions and job satisfaction were significantly related to turnover intentions. Supporting H3 and H4, the attribution dimensions locus of causality and stability were statistically significant at p=.05 and p=.02 respectively. External control and personal control failed to meet the significance threshold. The standardized coefficients were -.15 for locus of causality and -.14 for stability. This provides some preliminary evidence to the relationship between attributions and turnover intentions.

However, as indicated by H2, we were also interested to determine whether job satisfaction mediated between attributions and turnover intentions. Therefore, we included job satisfaction in the regression analyses and found that locus of stability was no longer statistically significantly related with turnover intentions. This led us to believe that the relationship between locus of causality attributions and turnover intentions is fully mediated by job satisfaction. This result was confirmed with a Sobel-test indicating a one-tailed probability of p <.01. Further, the results of a Sobel test indicated that the relationship between stability attributions and turnover intentions was partially mediated by job satisfaction indicated by the one-tailed probability of p=.01.

H5 and H6 were not supported. There was no significant relationship found between either internal control or external control and turnover intentions.

DISCUSSION

While past studies have clearly delineated the relationship between job satisfaction and turnover intentions (Tett and Meyer, 2006), very few have looked at attribution styles, and job satisfaction simultaneously as predictors of turnover intentions (Harvey, Harris and Martinko, 2008). This study builds on their findings, which tied hostile attributions to job satisfaction and turnover intentions by looking at more general patterns of attribution styles and relating them to job satisfaction and turnover intentions.

Not surprisingly, we found that job satisfaction was a strong predictor of turnover intentions. We also found that attributional tendencies or styles are a significant influence on job satisfaction. It was interesting to find that while the tendency toward stability attributions had a direct positive effect on turnover intentions, even when job satisfaction was included in the model, the impact of locus of causality attributions appears to be fully mediated through the variable of job satisfaction.

The failure to find the relationships predicted in H5 and H6 suggest an interesting interpretation of these findings. If neither high internal control nor high external control influenced turnover intention, then the remaining conclusion is that uncontrollable causes for performance related failures increase the intent to turnover. In other words, having your performance related outcomes depend on chance, luck or the whim of weather are more likely to cause you to want to leave your job than having your outcomes based on another person such as a supervisor.

This finding poses an interesting contrast to the findings of Harvey, Harris and Martinko, (2008). While hostile attributions would typically imply blaming the supervisor or other co-worker, these findings suggest that voluntary turnover is more likely to be caused by feelings of uncontrollability than feelings that another person controls the outcome.

BIBLIOGRAPHY

Abramson, Lyn Y.; Seligman, Martin E.P.; Teasdale, John D. (1978) Learned Helplessness in Humans: Critique and Reformulation. Journal of Abnormal Psychology, 87, 49-74.

Allen, D.G. & Griffeth, R.W. Test of a mediate performance-turnover relationship Highlighting the moderating roles of visibility and reward contingency. Journal of Applied Psychology, 2001, 86(5), 1014-1021.

Griffeth, R.W., Hom, P.W., & Gaertner, S. A meta-analysis of antecedents and correlates of employee turnover:Update, moderator tests, and research implications for the next millennium. Journal of Management, 2000, 26, 463-488.

Griffeth, R.W. & Hom, P.W. (2001). Retaining Valued Employees, Thousand Oaks: Sage.

Harvey, Paul ; Harris, Kenneth J. and Martinko, Mark J. (2008) The Mediated Influence of Hostile Attributional Style on Turnover Intentions, Journal of Business and Psychology, 22, 333-343.

Heider, F. (1958) The Psychology of Interpersonal Relations. New York, Wiley.

Hom, P.W., & Griffeth, R.W. Structural equations modeling a test of turnover theory: Cross-sectional and longitudinal analyses. Journal of Applied Psychology, 1991, 76, 350-366.

Hom, P.W. & Griffeth R. (1995). Employee Turnover. South-Western College Publishing, Ohio.

Hom, P., & Kinicki, A. (2001). Towards a greater understanding of how dissatisfaction drives employee turnover. Academy of Management Journal, 44, 975-987.

Kelley, H.H. (1973) The Process of Causal Attribution. American Psychologist, 28, 107-128.

Locke, 1976 cited in Brief, A. P., & Weiss, H. M. (2001). Organizational behavior: affect in the workplace. Annual Review of Psychology, 53, 279-307, p. 282

March, J.G. & Simon, H.A. Organizations. New York: John Wiley, 1958.

McAuley, E. ; Duncan, T. E. & Russell, D. W. (1992) Measuring causal attributions: The revised causal dimension scale (CDSII). Personality & Social Psychology Bulletin, 18, 566-573.

McCormick, John (1997) An attribution model of teachers' occupational stress and job satisfaction in a large educational system. Work & Stress, 11(1) 17-32.

Mobley, W.H. Immediate linkages in the relationship between job satisfaction and employee turnover, Journal of Applied Psychology, 1977, 62, 237-240.

Norris, Dwight R. and Niebuhr, Robert E. (1984) Attributional influences on the job performance-job satisfaction relationship. Academy of Management Journal, 27(2), 424-431.

Organ, D. W., & Ryan, K. (1995). A meta-analytic review of attitudinal and dispositional predictors of organizational citizenship behavior. Personnel Psychology, 48, 775-802.

Rotter, J.B. (1966) Generalized Expectations for Internal versus External Control of Reinforcement. Psychological Monographs, 801-828.

Saari, L. M., & Judge, T. A. (2004). Employee attitudes and job satisfaction. Human Resource Management, 43, 395407.

Tett, Robert P. and Meyer, John P. (2006) Job Satisfaction, Organizational Commitment, Turnover Intention, and Turnover: Path Analyses Based on Meta-Analytic Findings. Personnel Psychology, 48(2), 259-293.

Vielva, I. and Iraurgi I. (2002) Control perception and coping behavior in abstinence in alcoholics. Salud y Drogas 2(1), 29-42.

Wegge, J., Schmidt, K., Parkes, C., & van Dick, K. (2007). 'Taking a sickie': Job satisfaction and job involvement as interactive predictors of absenteeism in a public organization. Journal of Occupational and Organizational Psychology, 80, 77-89

Weiner, B. (1974) (ed.) Achievement Motivation and Attribution Theory, Morristown, NJ: General Learning Press.

Weiner, Bernard (1979) A theory of motivation for some classroom experiences. Journal of Educational Psychology, 71, 3-25.

Zuckerman, M. and Feldman, L (1984) Actions and occurrences in attribution theory. Journal of Personality and Social Psychology, 3, 541-550.

Tobias M. Huning, Columbus State University

Neal F. Thomson, Columbus State University

Table 1: Means, Standard Deviations, Reliabilities, and Correlations
among all Variables

                      Mean       SD         1          2

Locus of Causality    6.3        1.46       (0.74)
External Control      5.1        1.53       -.10       (0.70)
Stability             5.3        1.49       .44 **     .17 **
Personal Control      2.4        1.46       .68 **     -.26 **
Job Satisfaction      3.8        0.48       .18 **     -.13 *
Turnover Intention    3.1        0.89       -.14 **    .01

                      3          4          5          6

Locus of Causality
External Control
Stability             (0.62)
Personal Control      .35 **     (0.83)
Job Satisfaction      .07        .17 **     (0.89)
Turnover Intention    -.18 **    -.1        -.58 **    (0.92)

Note: Reliabilities (Cronbach's Alphas) are given in parentheses.

** Correlation is significant at the .01 level.

* Correlation is significant at the .05 Level.

Table 2: Regression results:

Regression results with Job Satisfaction as Dependent Variable

                      [beta] *    s.e.    t-value    p-value

Locus of Causality       0.2      0.37      2.54       0.01
External Control        -0.10     0.04     -1.76       0.08
Stability               0.01      0.04      0.12       0.91
Personal Control        -0.02     0.06      -.23       0.82

* Standardized path coefficient

Regression results with Turnover Intention as Dependent Variable

                      [beta] *    s.e.    t-value    p-value

Locus of Causality      -.15      0.07      -.19       0.05
External Control        0.03      0.05      0.54       0.59
Stability               -.14      0.05     -2.30       0.02
Personal Control        0.06      0.07      0.79       0.43

* Standardized path coefficient

Regression Results including Job Satisfaction as Mediator

                      [beta] *    s.e.    t-value    p-value

Locus of Causality      -.04      0.06      -.64       0.53
External Control        -.02      0.04      -.46       0.64
Stability               -.14      0.05     -2.65       0.01
Personal Control        0.05      0.06     0.786       0.43
Job Satisfaction        -.54      0.06     -11.52      0.00

* Standardized path coefficient