Abstract:
The main aim of this study was to evaluate the employment outcomes of participants in a Ministry of Social Development (MSD) programme, In-Work Support for Maori and Pacific peoples (IWS). A second aim of the study was to assess the contribution of several different outcome measures to the evaluation of programme participants' employment outcomes. The third aim was to test the suitability of a non-equivalent groups design for this study.
A non-equivalent groups design was used to compare the outcomes of the participant group to the outcomes of the group who were eligible to participate in the programme, but did not (the comparison group). The study used data extracted from MSD administrative databases. Data was collected about the demographic characteristics and the benefit receipt trends of IWS participants and the comparison group.
The analysis showed that IWS participants achieved more favourable outcomes than the comparison group; participants remained in employment for longer continuous periods after leaving an income-tested social welfare benefit and spent a greater amount time in employment during the first three months after leaving.
The non-equivalent groups design was found to have some limitations in the current study. The participant and comparison groups differed on some characteristics, which meant that the outcomes of the groups were not easily comparable. Therefore, the use of an alternative method of constructing a comparison group, such as propensity matching, is recommended for future studies of in-work support programmes.
The demographic differences between the two groups suggested that participants might have been more disadvantaged in the labour market than the non-participant comparison group, and could be expected to achieve less favourable outcomes than the comparison group. This suggested that the outcome analysis provided a conservative estimate of programme participants' outcomes.
The outcome measures used in the study were all found to offer useful information about different aspects of the outcomes of programme participants. The use of a range of measures, as included in this study, would be recommended for similar future research.
