An Entropy-Based Measure for Assessing Fuzziness in Logistic Regression

• In Event: Poster Session 7
  In Poster Session: Division D Section 2 Poster Session

Abstract

Methodologists recommend that applied researchers use a combination of data-model fit measures. In logistic regression, a small number of data-model fit indices currently exist and are criticized for being sensitive to sample size, may lack power, and being ineffective at describing the accuracy of classification. The current study introduces a measure of entropy that can be used to assess the quality of logistic-regression models. Previously, this measure has been utilized to quantify how well individuals are classified into classes in latent class analysis. The current study shows how this measure can also be used in logistic regression to examine the quality of classification. Entropy can and should be used in conjunction with other measures of model fit in logistic regression.

Authors

• Brandi A. Weiss, The George Washington University

• William R. Dardick, The George Washington University