Mixture Experiments in Social Science: Ideal Methodology for Testing Criminological Theory Within and Between

David N. Khey, University of Florida

ABSTRACT
New methodologies for examining the performance of criminological theory are presented in this project. Mixture experiments, a broad category of statistical modeling that is predominately used within industry for optimizing product, can hypothetically analye the performance of measures of criminological theory as well as having the power to compare theories' performances against each other. The analysis generates a visual output that allows researchers to easily diagnose the strengths and weaknesses within and between ciminological theory's measures. Social learning theory is analyzed individually and social learning, social bonding, and deterrence theories together are presented in this manner as illustration of this technique. The data used within this project includes self report measures that have been exemplified in prior research as accepted measures for testing these theories. The advantages of this technique are: it is relatively simple using basic statistical modeling techniques, it yields easily interpretive outputs, and it can diagnose strengths and weaknesses within and between theories with extreme precision. These analyses can provide insight into the drafting of new theory and integrating theories with optimal results.

(Return to Program Resources)

Updated 05/20/2006