Logistic Regression via Excel Spreadsheets: Mechanics, Model Selection, and Relative Predictor Importance

Abstract

Logistic regression is one of the most fundamental tools in predictive analytics. Graduate business analytics students are often familiarized with implementation of logistic regression using Python, R, SPSS, or other software packages. However, an understanding of the underlying maximum likelihood model and the mechanics of estimation are often lacking. This paper describes two Excel workbooks that can be used to enhance conceptual understanding of logistic regression in several respects: (i) by providing a clear formulation and solution of the maximum likelihood estimation problem; (ii) by showing the process for testing the significance of logistic regression coefficients; (iii) by demonstrating different methods for model selection to avoid overfitting, specifically, all possible subsets ordinary least squares regression and l1-regularized logistic regression (lasso); and (iv) by illustrating the measurement of relative predictor importance using all possible subsets.

Teaching binary logistic regression modeling in an introductory business analytics course

“…Importantly, these studies show that teaching logistic regression helps reinforce linear regression concepts (Morrell & Auer, 2007) and enhance critical and analytical thinking (Brusco, 2022; Li et al., 2018). In contrast to linear regression, binary logistic regression offers unique opportunities for students to learn about out‐of‐sample testing and how to evaluate models using classification performance metrics (accuracy, sensitivity, and specificity).…”

“…Literature has shown that logistic regression is among the most popular methodological tools in predictive analytics (Brusco, 2022). Given that more introductory IBA and statistics books are including logistic regression (Morrell & Auer, 2007), one could expect that logistic regression has been taught widely in undergraduate IBA courses.…”

“…Most studies provide insights on teaching logistic regression in advanced undergraduate courses or post-graduate courses. For example, Li et al (2018) highlight the inclusion of logistic regression in an introductory business statistics course for post-graduate students, and Brusco (2022) describes the use of Excel templates to enhance conceptual understanding of binary logistic regression for graduate students. Morrell and Auer (2007) illustrate classroom activities related to logistic regression in an advanced applied statistics course.…”

“…Interestingly, Al-Haddad et al (2019) report that students who learned Excel alongside IBA performed better than students learning only Excel skills in an information technology course. Also, literature shows that purpose-built Excel templates or workbooks can play a vital role in teaching logistic regression or Bayesian methods in more advanced courses (Brusco, 2022;Johnson & Berenson, 2019). Guided by these studies, we decided to use Excel in our IBA courses.…”

Prediction of Graphitization Behavior during Long-Term Creep in Carbon Steels

“…Each partial regression coefficient was determined for both elongated graphite and spherical graphite to minimize the absolute value of the sum of the log-likelihood L(c) of all data points by fitting with the solver in Microsoft Excel. 23) Here, L(c) is given by…”