The mediating role of job characteristics in job redesign interventions: A serendipitous quasi-experiment DAVID J. HOLMAN*, CAROLYN M. AXTELL, CHRISTINE A. SPRIGG, PETER TOTTERDELL AND TOBY D. WALL Institute of Work Psychology, University of Shefﬁeld, Shefﬁeld, U.K.
The aim of this paper is to examine the mediating role played by ﬁve key job characteristics in the relationship between employee participation in a job redesign intervention and employee well-being. In studies of job redesign interventions, it has been assumed that any effects of employee participation in job redesign on well-being are a result of changes in job characteristics rather than participation in change per se. It is therefore important to statistically test for mediation in job redesign intervention studies to help establish that the change in job characteristics is the mechanism through which job redesign interventions work. However, this has rarely been tested directly, either because data to allow tests of mediation have not been collected (e.g. assessments of job characteristics) or because data have been collected but mediation has not been tested using accepted procedures. This makes it unclear whether changes in job characteristics explain the effects. Results from multilevel analyses of a longitudinal 9-month long serendipitous quasi-experimental participative job redesign intervention showed that changes in job control, participation, skill utilization and feedback, but not task obstacles, were sufﬁcient to account for the relationship between the intervention and employee well-being. Copyright # 2009 John Wiley & Sons, Ltd.
There is a long history of using job redesign interventions to change job characteristics in the expectation that this will improve employee well-being (Parker & Wall, 1999). It is generally assumed that the change in job characteristics is the mechanism through which job redesign interventions affect employee well-being. Yet this is rarely tested; either because data to allow tests of mediation have not been collected (e.g. assessments of job characteristics) or because data have been collected but mediation has not been tested using accepted procedures (Baron & Kenny, 1986). To date, studies in which tests of mediation have been made have focused exclusively on changing one job characteristic, job control (Bond & Bunce, 2001; Logan & Ganster, 2005). However, many job redesign interventions aim to change multiple job characteristics on the assumption that this will produce a greater impact on * Correspondence to: David J. Holman, Institute of Work Psychology, University of Shefﬁeld, Shefﬁeld, S10 2TN, U.K. E-mail: d.holman@shefﬁeld.ac.uk
Copyright # 2009 John Wiley & Sons, Ltd.
Received 2 May 2008 Revised 6 May 2009 Accepted 15 May 2009
THE MEDIATING ROLE OF JOB CHARACTERISTICS
well-being (Semmer, 2003) but the efﬁcacy of so doing has not been tested using required mediation procedures. Indeed, analysis might reveal that a change in one job characteristic accounts for all the change in well-being, implying that changing multiple job characteristics may not be an efﬁcient and parsimonious approach. There are, therefore, two reasons why it is important to test for mediation in job redesign interventions that aim to change multiple job characteristics. First, it is important to establish that a change in job characteristics is the mechanism through which the job redesign intervention affects well-being, and that any change in well-being is not just a direct result of being involved in an intervention per se. This is particularly important in participative job redesign interventions, in which employees are involved in suggesting and implementing change, as improvements in well-being may result from the...