Testing Individual vs Group Mean Differences in Social Science Research

A true experimental design requires random selection and random assignment of subjects to control and experimental groups. A hypothesized statistically significant mean difference in the dependent variable between these two groups is typically specified. This methodology is also referred to as a randomized clinical trial when testing for group differences. Individual differences are generally not considered, rather the focus is on the average control group and experimental group difference. This article offers another approach that illustrates testing for individual differences over time. h e true experimental design conducts a test of control group versus experimental group average dependent variable difference using analysis of variance statistical tests (Maxwell & Delaney, 2004). This methodology is also referred to as a randomized clinical trial when testing for group mean differences (Machin & Fayers, 2010). Oftentimes a true experimental design is not possible, so the researcher uses a quasi-experimental design. A quasi-experimental design uses a comparison group rather than a control group. The typical quasi-experimental design considers a pre-test measure, followed by treatment, and then a similar post-test measure for the subjects in the comparison group and the experimental group. In the statistical analysis, individual post-test measure differences are adjusted for individual pre-test measure differences to control for bias. This adjustment is referred to as analysis of covariance and expressed in the general linear model as: are