Positive Work Affect Under Heighten Job Demands: Comparing the Motivational Effects of Two Conceptualizations of Instrumental Feedback

Job strain contributes greatly to turnover, substance abuse, work-related accidents, and family dysfunction(Cooper, Dewe, & O'Driscoll, 2001). Monetarily, the construct costs society billions due to reduced work productivity (Levi, 2002) and escalating health care expenditures (Rice, 1999). Not surprisingly, workplace health has increasingly captured the interest of organizational and social work scholars (Macik-Frey, Quick, & Nelson, 2007; Preston, 2007) with a notable amount of research focused on job strain (Cooper et al., 2001). Within this body of literature, recent studies have identified the strain buffering effects of instrumental or goal-related feedback (Vaananen et al., 2003).

Action regulation theory (ART), posits work action as a feedback-dependent process that encompasses the development and hierarchical arrangement of goals; information search and collection; strategy formation, execution, monitoring; and goal-performance evaluation (Frese & Zapf, 1994). The theory, similar to social cognitive theory (Latham, 2007), asserts that progress toward goal attainment (goal-performance evaluation), attributable to one's abilities or actions (i.e., job control), fosters positive affective states (motivated job mastery and learning). Although both instrumental feedback from supervisors and (e.g., co-workers) address goal-performance gaps, research has clearly demonstrated that feedback is most impactful when it is frequent, credible, accurate, and timely (Ashford, Blatt, & VandeWalle, 2003). To date, however, no study has compared the two sources of instrumental feedback. Based on ART, four hypotheses were advanced:

1) a significant main effect for instrument feedback (supervisor); 2) a significant main effect for instrumental feedback (other sources); 3) a significant job demands-control-feedback (supervisor) interaction, and 4) a nonsignificant job demands-control-feedback (other sources) three-way interaction.

To test the four hypotheses, over 1000 employees from three human service agencies in the city of New York were surveyed. Seven hundred and one completed surveys were returned for a response rate of 69%. Survey procedures followed methods outlined by Dilman (1978). No violations of OLS regression were found. Outliers beyond a Cook's distance and center leverage value of 0.09 and 0.20 were removed. The internal consistency of the survey's measures ranged from 0.83 to 0.92. Confirmatory Factor Analysis (CFA) was performed and corresponding items loaded heavily onto their respective factor from 0.55 to 0.89. Discriminant validity was evaluated and fit indices pointing towards good model fit. All alternative models were examined and all failed to achieve acceptable fit.

Hierarchical multiple regression was used to test all hypothesized interactions. After including control variables, support was noted for hypothesis 2, 3, and 4. As predicted, both the main effect for instrumental feedback (other sources) (β = 0.14, p > 0.05) and the three-way demands-control-feedback (supervisory) (β = -0.12, p > 0.05) interaction on positive work affect were significant, the demands-control-feedback (other sources) interaction was not. The 0.10 percent increase in adjusted R2 accounted for by the significant interaction was five times greater than the median effect size (R2 = .002) found by Aguinis, Beaty, Boik, and Pierce (2005). As such, this study extends the JD-C literature by further clarifying instrumental feedback's motivational effects. Implications for both social work management theory and practice are extensively discussed.