The Relative Advantage of One-Way to Repeated-Measures ANOVA With High Student Turnover

• In Event: Marriott Roundtable Session 7
  In Roundtable Session: Applied Methods and Measurement in Education Research

Abstract

Repeated-measures ANOVA (RMA) is appropriate for studies with various measurement points, but loses power with high missing cases. Methods of handling missingness lead to misinterpretation of results due to lack of statistical power (e.g., listwise deletion) or increased Type-I errors (e.g., mean substitution). Even possible appropriate strategies (i.e., multiple imputation) require extensive statistical knowledge. Because of this, teachers wanting to verify their own interventions struggle finding a favorable method. Therefore, the study determines if using one-way ANOVA (OWA) instead of RMA leads to greater statistical power without artificially increasing Type-I errors. Results indicate that OWA outperforms RMA across all sample sizes when data are uncorrelated and missingness is high. Implications and future directions for research in this topic are discussed.

Authors

• Gabriela A. Aquino Aguilar, St. Mary's University - San Antonio

• Rick Sperling, St. Mary's University College