An Exploration of the Role of Likert Items in School-Based Survey Analysis

• In Event: 35.073 - Innovations in Survey Research Methodology

Abstract

This paper explores Likert item treatment in school-based surveys. It is common to treat Likert items on education surveys as interval data and perform arithmetic analyses reserved for interval data on this ordinal data. To explore the treatment of Likert items in school-based surveys construct analysis is performed using principle component extraction (PCA) and data is analyzed using a stratified sample t-test and ordinal regression. There are important distinctions in understanding the results of the two approaches used for this analysis. The t-test approach examines the difference in means. Conversely, the ordinal regression approach treats the data as ordinal. The ordinal regression approach yielded results that are largely consistent with the results from the stratified sample approach.

Authors

• Liz Bergeron, New Tech Network

• Max Dugan-Knight

• Sohail Kamdar

• Jen Vorse Wilka

• Beth Boesche-Taylor, New Tech Network