Estimating Variation In Treatment Effects in Sociological Experiments Margo Katherine Wilke Undergraduate Research Internship Program Spring 2024 Accepted Sociological research methods, applications to areas including gender, violence, crime, inequality, race, Classical experiments, whether for social policy or for medicine, determine the effects of “treatments” by comparing the mean outcomes of a treatement groups to those of an untreated group. That provides good information about “mean” treatment effects, but no information about potentially wide variation in the effects on different indlviduals, We (Drs. Scott Feld and Trent Mize) are developing and applying new methods for revealing those variations. In this research, we plan to try out this new method by reanalyzing data from national survey experiments that were previously used in publications that simply reported mean effects. We hope to be able to show that some experimental treatments actually do provide relatively consistent effects on most people, but that others have widely variable effects, where the mean provides a misleading summary of the extent to which some people change in each direction Scott L Feld We intend to apply our approach to previous studies based upon archived data provided by a multi-year National Science Foundation funded project on Time-Sharing Experiments in the Social Sciences (TESS) These studies involve a broad range of topics for survey experiments (e.g. regarding crime, race, gender, violence, etc.) We are seeking a student who will review many of the publications from these data, help select those that might be of particular interest, and reanalyze those data using our new approach. The student will need to be able to download and review research publications, download and analyze data using any standard stastistical software (e.g. SPSS, STATA, R, or others) to create and manipulate data files and perform basic statistical analyses (e.g. linear regression on various different samples (e.g. men, women, young, etc.) with our instruction, guidance and supervision. The student must be interested in participating in the process of discovery in current social science research on topics such as gender, crime, violence, inequality, religion. Experience with using some standard statistical softward to perform standard statistical analyses on real data would be helpful. Some familiarity and comfort with the handling and analysis of data is necessary. Please note: The Margo Katherine Wilke Undergraduate Research Internship (Wilke) program is designed to support Purdue College of Liberal Arts undergraduates. As a result, students without a CLA major or minor may not be considered for this program. 3 0 (estimated)