Logistic regression modelling: procedures and pitfalls in developing and interpreting prediction models

Abstract

This study sheds light on the most common issues related to applying logistic
regression in prediction models for company growth. The purpose of the paper is 1) to provide a detailed demonstration of the steps in developing a growth prediction model based on logistic regression analysis, 2) to discuss common pitfalls and methodological errors in developing a model, and 3) to provide solutions and possible ways of overcoming these issues. Special attention is devoted to the question of satisfying logistic regression assumptions, selecting and defining dependent and independent variables, using classification tables and ROC curves, for reporting model strength, interpreting odds ratios as effect
measures and evaluating performance of the prediction model. Development of a logistic regression model in this paper focuses on a prediction model of company growth. The analysis is based on predominantly financial data from a sample of
1471 small and medium-sized Croatian companies active between 2009 and 2014. The financial data is presented in the form of financial ratios divided into nine main groups depicting following areas of business: liquidity, leverage, activity, profitability, research and development, investing and export. The growth prediction model indicates aspects of a business critical for achieving high growth. In that respect, the contribution of this paper is twofold. First, methodological, in terms of pointing out pitfalls and potential solutions in logistic regression modelling, and secondly, theoretical, in terms of identifying factors responsible for high growth of small and medium-sized companies.