Skip nav

Thumbnail of data collection design model. Click to open. Click to open the data collection design model.

The most sophisticated evaluation studies use an experimental design, considered to be the gold standard. The hallmark of this design is randomization. Randomization reduces the likelihood that study groups will differ from one another before the intervention, which can be a potential confound. However, randomization by itself cannot guarantee the absence of group differences. To confirm that the randomization was effective, you must test the groups after randomization for demographic differences or any other preexisting differences that may affect outcomes. This step helps to ensure the groups are comparable.

Data collection in an experimental study resembles that of quasi-experimental studies, but with the addition of random assignment. The image at right illustrates what data collection looks like in an experimental design.

Select "Next" to continue.