A Simple and Transparent Alternative to Logistic Regression

Authors

  • James W. Grice Oklahoma State University
  • Lisa D. Cota Oklahoma State University
  • Paul T. Barrett Cognadev UK Ltd.
  • Karl L. Wuensch East Carolina University
  • Michael Poteat East Carolina University

DOI:

https://doi.org/10.14738/assrj.37.2125

Abstract

Observation oriented modeling was compared to logistic regression in the re-analysis of data from a published study. In the original study students’ decisions to stop or continue controversial animal research was predicted from gender, the type of research being conducted, and measures of ethical attitudes. With regard to classification accuracy, results from the observation oriented modeling analyses compared favorably to results obtained from logistic regression and two other methods of classification. Prediction profiles created from the observation oriented modeling analyses were moreover parsimonious and transparent, and the analyses themselves were free of assumptions. Unique features of the observation oriented modeling approach were discussed.

Downloads

Published

2016-07-30

How to Cite

Grice, J. W., Cota, L. D., Barrett, P. T., Wuensch, K. L., & Poteat, M. (2016). A Simple and Transparent Alternative to Logistic Regression. Advances in Social Sciences Research Journal, 3(7). https://doi.org/10.14738/assrj.37.2125