Logistic regression is usually used to predict a dichotomous outcome variable on the basis of a set of continuous, dichotomous, ordinal, or categorical predictor variables. Using logistic regression can be extremely useful when analyzing complex relationships among different predictor variables, and when (some of) the predictors are not normally distributed or if the relationship between them is not linear. This paper assessed a specific database to examine whether age, socioeconomic status, teaching approach (computer versus traditional), and amount of reading predicted reading level among college students. In Chapter I, the paper presents the formulated research questions, followed by the descriptive statistics, and the results of the Chi-Square tests. In Chapter II, the results of univariate and multivariate logistic regression analyses are reported with a comparison between univariate versus multivariate results. These sections particularly discuss the Wald tests, p-values, logistic regression equations, goodness-of-fit tests, and the predictive accuracy of the logistic models. Throughout the paper, the main assumptions of performed analyses are checked and discussed carefully. All statistical tests are two-tailed, and p<.05 is considered significant. The final section, Chapter III, concluded that college students who were identified as good readers in the examined database were likely to read more and appeared, in this specific sample, to take a traditional approach to teaching.