Normality Versus Non-Normality: Enhancing Understanding of Engineering Interventions' Impact for Multilevel Models With Variance Heterogeneity

• In Event: 53.082 - Poster Session 10
  In Poster Session: 53.082-3 - Special Topics in Engineering and Computer Science: Engineering Design, Computational Thinking, and Assessment

Abstract

Variance heterogeneity, typically treated as a nuisance, is a common feature of education data when treatment differences expressed through means are present. A deeper understanding of a treatment’s impact is possible by identifying predictors of variances of the outcome variable, which can enhance understanding of who a treatment does and does not benefit in ways that can inform and improve the treatment. Normal and non-normal distribution modeling theories are used in this study to illustrate two methods of modeling variance heterogeneity for data from a study of the impact of an engineering design-based STEM curriculum on student achievement with a focus on two-level (multilevel) models.

Authors

• Yadira Peralta, University of Minnesota - Twin Cities

• Michael R. Harwell, University of Minnesota

• Selcen Guzey, Purdue University

• Mario Moreno, University of Minnesota

• Tamara Jo Moore, Purdue University