Developing Software-Based Statistical Models for Educational Incentives in Middle School Classrooms

Abstract

There is a need for more statistical, computerized representations in studies via fixed effects and mixed effects models. This article gives meta-analytic examples of (a) adequate literary, statistical, and conceptual coverage of token reinforcement as defined within educational interventions and (b) practical mixed-effects modeling that is relevant for determining how treatment effect size fits with other characteristics in literature on incentives. The findings from the meta-analytic modeling indicate that sample size, grouping options, timing, study type, and treatment effect size variation have significant influences on the practical significance (effectiveness) of incentives with middle school students. Accounting for these variables helps stakeholders in education develop supports that offer more standardization, versatility, and appeal to students as a whole. A variety of treatment effects for reinforcers may exist, but the overall effect of reinforcement can be positive. This article is recommended for those interested in developing better instructional practices for students, regardless of academic abilities.

Keywords: Modeling, Reinforcement, Instruction, Education, Supports