Researchers often want to understand how well their findings from a postsecondary randomized controlled trial generalize to a broader population of colleges. When that is the goal, it is important to assess how well the study sample of colleges represents the intended inference population of colleges.

The Generalizer is a valuable tool for this purpose. It guides researchers through two steps:

1. Defining the inference population. What types of colleges do researchers want to know if their study findings generalize to? The tool will apply this definition to the population of colleges in each state and in the United States overall.
2. Selecting variables that may explain cross-college variation in intervention effects. If intervention effects are constant, the evaluation results generalize broadly. If effects vary, results may only generalize to colleges that are similar in composition to the ones in the study. The tool guides users through the process of choosing a set of possible moderators for comparing the sample and inference populations.

Once these steps are complete, The Generalizer compares the selected characteristics of a user’s study sample to the inference population, both nationally and by state. Doing so helps users assess the representativeness of their samples and make more informed judgments about how broadly their findings may apply.

In addition, this report—commissioned by the Institute of Education Sciences (IES)—provides an excellent overview of the steps involved in assessing the similarity between a study sample and the inference population (among other things).^[1]