TYPES OF EVALUATION DESIGNS
The evaluation design you choose will mainly depend on the question(s) that you wish to answer using your evaluation (and, of course, the resources you have). By choosing an appropriate evaluation design, you will be able to get clear and reliable answers to your evaluation questions.
Most program evaluations seek to answer one or more of the following questions:
There are three general types of designs that you can utilise for your evaluations:
Most program evaluations seek to answer one or more of the following questions:
- Were there any changes in the outcome measures?
- Did your program contribute to these outcomes?
- Which specific part or feature in your program led to these changes?
There are three general types of designs that you can utilise for your evaluations:
- Experimental design: Participants are randomly assigned to the intervention or the comparison group, and data is collected both before and after the intervention. The changes in the intervention group are then compared to those in the comparison group. This is the most rigorous design and often considered the 'gold standard', but in some situations it may simply not be feasible.
- Quasi-experimental design: This usually involves two groups, an intervention group and a comparison group, but participants are not randomly assigned to the groups. As it cannot be assumed that the two groups are equal at baseline, any differences between the two groups must be considered when conducting the analysis. For this reason, the comparison group should ideally be matched to those in the intervention group on relevant characteristics, e.g, age, gender, type of dementia etc., as much as possible.
- Non-experimental design: This is the weakest study design where you have only the intervention group. This design makes it difficult to establish the relationship between changes in outcome measures to the program because evaluators do not know what would happen if people did not take part in the program. While this is considered 'weak' in terms of establishing cause and effect, this may be the only feasible option with the resources available.