As with any research design that seeks to elucidate causality, evaluative research requires a protocol that outlines in detail the causal model under evaluation, the variables of the study, the units of observation, the measurement instruments used and the means deployed to control threats to the validity of the study. There are numerous factors present in the analysis of any intervention or organization that can threaten the validity of a study, so the researcher must implement additional control measures. Two types of validity:

INTERNAL VALIDITY: the study must be designed so the researcher can distinguish between changes produced by the intervention under evaluation and changes produced by other causes.

EXTERNAL VALIDITY: A study that is externally valid is one that shows that the results of the intervention are applicable to other people, in other places, at other times.