Longitudinal data enable CALDER’s researchers to study the impact of various policies on individuals and schools over time. Because census files do not confine researchers to a sample selected for a specific research question, the Center’s experts can estimate the effects of policies as they emerge and target their analyses to the population(s) intended to be affected by the policy.
The Center’s research questions are pursued using state-of-the art analytic techniques, including:
- Panel data models
- Hazard models
- Regression discontinuity
- Qualitative studies to help interpret quantitative results.
Experimental data is the gold standard of statistical inference, but it is hard to come by, in part because of high costs (both pecuniary and political). As a consequence, experiments often rely on relatively small samples in limited settings making it difficult to propose general policies. The statistical techniques, including the quasi-experimental methods, listed above offer the potential of identifying the treatment effect reliably in a non-experimental setting and enhance external validity with large samples and direct observation of sub-populations.