Skip to main content

Welcome, the Hub connects all projects


Testing Causal Hypotheses Using Longitudinal Survey Data: A Modest Proposal for Modest Improvement


"One-shot surveys were developed to describe populations and longitudinal surveys were developed to describe changes in populations. Relative to the range of all questions that the community of educational researchers asks, the scientific function of surveys is limited, just as the scientific function of other methods is also limited. Surveys were not developed to describe or explain causal connections between interventions and outcomes. There are better methods for this, such as experiments, but they are generally weaker than the survey when it comes to population description. It is the research purpose that should determine method choice. But since science values many different kinds of purpose, no method can be superior to all the others for all the relevant purposes to which research is put."