Perspectives on Randomization and Readiness for Change in a Workplace Intervention Study

Randomization is often recommended above self-selection when allocating participants into intervention or control groups. One source of confounding in non-randomized studies is the participants’ attitudes towards the intervention. Because randomized workplace interventions are not always feasible, it is important to investigate differences between study groups in readiness for change. To meet this aim, we used data from an intervention study of the effects of work-time control. The study design entailed both self-selection (i.e. non-random) and random allocation into intervention and control groups. Some team leaders rejected randomization because they considered it to be fairest to increase work-time control among employees in most need. Others accepted randomization arguing that it was fairer to allocate a potential benefit by random. We found no difference in readiness for changes when comparing the self-selected intervention and control groups. In contrast, the randomized intervention group reported higher readiness for change when compared with both the randomized control group and the self-selected intervention group. This suggests that self-selection into intervention and control groups may reflect the local leaders’ rather than the employees’ readiness for changes and that randomization may influence the participants’ attitude towards the intervention perhaps by evoking an experience of ‘winning or losing in the lottery’