Mediating and moderating effects in job design

Journal of Management, Dec, 1992 by Gary Johns, Jia Lin Xie, Yongqing Fang

The results for security and social satisfaction ran counter to theoretical predictions, in that there was some tendency for managers who were lower on these facets to report that higher scope jobs induced more positive psychological states. Also, the relationship between experienced meaningfulness and self-rated performance was most negative among those who reported the most social satisfaction. The variance increments for these tests were in the 1-2% range, and the t's ranged from 1.97 to 2.90.

Significant moderator effects for pay satisfaction were confined to the relationship between the psychological states and turnover cognitions. As predicted by the theory underlying the JCM, those who were most satisfied with pay thought less about quitting when they were experiencing favorable job-related psychological states. Turnover cognitions were essentially unrelated to psychological states among those who were less satisfied with their pay. Again, variance increments were in the 1-2% range, with t's ranging from 2.02 to 2.42.

The moderator effects for supervisory satisfaction were the strongest, the most pervasive, and the most complicated in pattern. In essence, supervisory satisfaction negatively moderated the relationship between job scope and the psychological states, again running counter to JCM theory. That is, managers who reported less satisfaction with supervision were more likely to describe high MPS jobs as inducing positive psychological states. A similar moderator effect was observed for the relationship between MPS and the outcome variable of general satisfaction. Variance increments for these interactions averaged 2%. and t's ranged from 2.36 to 3.95. On the other hand, an opposite moderating effect existed between the psychological states and performance (and, in one instance, turnover cognitions). That is, managers who were most satisfied with their bosses were most likely to rate their own performance highly when they reported favorable psychological states. Again, variance increments averaged 2%, with t's ranging from 2.05 to 2.83.

Instrumentalities. The moderator effects for pay and promotion instrumentality are shown in the bottom portion of Figure 2. For pay instrumentality, significant effects can be observed in both the "front" and the "back" parts of the JCM. All of these moderator effects except those involving self-ratings of performance are negative effects. That is, lower perceived connections between pay and performance result in more positive relationships between MPS and the states and between the states and the outcomes. However, in both cases where pay instrumentality interacts with psychological states to predict performance the moderator effect is the opposite. That is, managers tended to respond to favorable psychological states with enhanced performance if pay was seen to be contingent on performance.

Significant moderator effects for promotion instrumentality were confined almost exclusively to the links between the psychological states and outcome variables. Increased knowledge of results boosted general and growth satisfaction and reduced turnover cognitions when promotions were not perceived to be related to performance. When promotions were seen as performance-contingent, meaningful and responsible work were both positively related to self-ratings of performance. The last interaction was the strongest in the study in terms of variance explained (|delta~|R.sup.2~ = .05, t = 3.88, p |is less than~ .001). Otherwise, variance increments for significant instrumentality interactions averaged about 2.5%, with t's ranging from 1.95 to 3.38.