ABSTRACT
Both researchers and practitioners agree that having highly engaged employees results in individuals and organizations reaping various positive consequences related to performance and absenteeism. However, available research syntheses date from the early years of this line of research, thus cover only a small fraction (under 10%) of the available studies. The present meta-analysis updates the results of work engagement and its three facets (vigour, dedication, and absorption) on task performance and includes a substantial number of studies on absenteeism with separate analyses of longitudinal studies. A total of 179 unique correlations representing an overall sample size of N = 139,182 was examined, confirming and enhancing a positive association between work engagement and task performance (ρ = .483) and a negative association between work engagement and absenteeism (ρ = −.171). The three facets of engagement had similar effects on performance, but only vigour and dedication correlated significantly negatively with absenteeism. Work engagement is linked positively to future task performance and negatively to future absenteeism. The influence of several methodical moderators is examined. Finally, we discuss how these findings can inform research and practice in order to contribute to a more effective and healthy work environment for employees.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1. It is noteworthy that earlier versions of the JD-R model exist that do not include job performance as an integral part; however, the current version we are referring to explicitly incorporates job performance as a main dependent variable
2. To prevent any confusion with the used statistical symbols, we would like to note that we use the following notation throughout this work: r = uncorrected estimated correlation coefficient; ρ = estimated correlation coefficient corrected for attenuation.
3. Lacking a clean test procedure, we compared two meta-analyses with different numbers of included studies using the estimated confidence intervals. This does not allow any statement on the significance of the difference.