The Care and Feeding of Survey Data

 

Hans D. Lemke

Tested Studies in Laboratory Teaching, 2017, Volume 38

Poster file:

Abstract

Surveys containing Likert scales are commonly used to assess opinions in educational research. There has been significant debate concerning the handling of data generated from these instruments. The primary source of this confusion is the difference between Likert-style items (LSI) and true Likert scales. The goals of this paper are to clarify the differences between them and provide guidelines for their analysis and presentation. Data generated from LSI are ordinal in nature and should be analyzed using appropriate nonparametric statistical tests. Likert scales require extensive testing and validation of the instrument and analysis may be performed with parametric or nonparametric statistics, depending on the structure of the data. These guidelines provide the tools for researchers to properly analyze and present their data and practitioners to evaluate and interpret published results.

Keywords:  statistical analysis, survey, Likert scales, Likert-style items, ordinal data

University of Houston (2016)