Research Foundations: Theory, Concepts, Measures

This chapter introduces three major concepts in conducting social science research: theory, concepts, and measures. Together, the three make up the basic logical infrastructure for conducting scientific research in political science. Theories, which are simplified explanations of the variation that occurs in the world, drive research in political science, particularly basic research. Research can either be theory-building, in which case it uses factual information about the world to generate a theory that can be further explored, or theory-testing, which takes a theory developed from earlier research and tests it. The strongest theories are parsimonious, which means simple and having fewer parts; generalizable, meaning that they can apply to a wide variety of circumstances; and falsifiable, meaning that they are capable of being proven wrong.

A theory is not framed in terms of actual events but in terms of concepts, which are definitions that help us classify and understand occurrences. A theory must contain at least two concepts: an independent variable or cause, and a dependent variable, or an effect. Conceptualization, or the process of defining a concept for the purposes of a research project, helps determine how a project is constructed and how other researchers will understand it. Once concepts are defined, they then must be operationalized, turned into concrete measurements that can be used to collect data. Measurement is the process of determining the values of the variables that represent concepts in order to look for relationships between them. Measurement requires precision, and it also requires understanding what type of a variable you are working on — a categorical, continuous, interval, or ratio variable.

Concepts are then used to construct hypotheses, which describe causal relationships between concepts. The criteria for arguing that a causal relationship is correlation, temporal order, absence of confounding variables, a plausible causal mechanism, and consistency. For a strong causal relationship, all five should be present.