Pattern Matching, Validity, and Conceptualization in Program Evaluation

Abstract

All social research is based on pattern matching ideas. A pattern match involves a correspondence between a theoretical or conceptual expectation pattern and an observed or measured pattern. Two quasi-experimental designs-the nonequivalent dependent variable design and the reversed treatment design—illustrate pattern matching logic well. In program evaluation three pattern matches are important: the program pattern match that assesses program implementation; the measurement pattern match that assesses the validity of the measures; and the effect pattern match that assesses the causal hypothesis Conceptualization methods are needed to facilitate the articulation of rich theoretical patterns. An example of a conceptualization study is presented and the utility of conceptualization methods for pattern-matching research is discussed.