Kinds of Research Logic John L. Couper, Ph.D. (from G. F. Hegel) a) Descriptive: describe components and relationships without explaining processes atheoretical start from but refine commonsense labels and interactions can only consider correlation, not causality * communication example: mood changes are caused by violent programming b) Abductive: find the best hypothesis that explains a body of data atheoretical but can help to develop explanatory theory select through comparison and elimination of possible explanations confirms applicability for explaining a phenomenon empirical: based on observation rather than concepts * communication example: observe several reasons for mood change, and find which best fits the observed process c) Deductive: given initial assumptions, construct rules to confirm/disconfirm hypotheses theoretical (because methodically testing pre-existing theory) based on deductive logic start from expected relationships and rules of change mixes empirical and theoretical bases * communication example: nerve stimulation and inhibition shown to lead to mood change (while controlling for other variables), which suggests a direct link between mood and sensory stimulation d) Inductive: find patterns that emerge within a mass of data uses observation to develop theory with few a priori assumptions the researcher methodically and honestly identifies patterns these patterns are used to develop a more-abstract idea or model (i.e., theory) empirical: rejects external, assumed logic and relationships * communication example: watch people watching television to discover patterns of mood change that might or might not lead to theory You can design a study that uses these kinds of research logic  separately: for specific for simpler design and analysis  together: e.g., combine abduction and description to develop testable theories  sequentially: use one of them for data and/or understanding that will improve another one