Keith Wurtz

Bio: Keith Wurtz is an academic researcher. The author has contributed to research in topics: Logistic regression & Regression analysis. The author has an hindex of 1, co-authored 1 publications receiving 1 citations.

A Methodology for Generating Placement Rules that Utilizes Logistic Regression.

Abstract: The purpose of this article is to provide the necessary tools for institutional researchers to conduct a logistic regression analysis and interpret the results Aspects of the logistic regression procedure that are necessary to evaluate models are presented and discussed with an emphasis on cutoff values and choosing the appropriate number of candidate predictor variables In order to demonstrate the process of conducting a logistic regression analysis, models are generated using educational background measures (eg, last grade in English, high school grade point average, etc) to predict the dichotomous outcome of success (ie, grade of A, B, C, or CR) in an English course At the same time, the information presented here can be applied to other logistic regression studies Topics covered include setting up the database, dummy coding, data reduction, multicollinearity, missing cases, setting the cutoff value, interpretation of the results, selecting a model, and the interpretation of odds ratios when they are negative Introduction Logistic regression is being used more frequently by institutional researchers conducting educational research across the country (Peng, Lee, & Ingersoll, 2002; Desjardins, 2001) For example, the California Community College Chancellor's Office used logistic regression to identify predictor variables for indicators in the Accountability Reporting for Community Colleges (ARCC) In addition, logistic regression has been used to predict retention of Hispanic/Latino students (Wilson & Hughes, 2006), and to predict college attrition (Zvoch, 2006) Equally important, according to Morest and Jenkins (2007), community colleges are beginning to demand higher level statistical analysis to help with the decision-making process and to identify methods that can help increase student outcomes Logistic regression has many advantages over other similar procedures like multiple regression and discriminant analysis (Tabachnick & Fidell, 2007) The advantages of logistic regression are that the candidate predictor variables do not have to be normally distributed, linearly related, or have equal variances (Mertler & Vannatta, 2005) In addition, the candidate predictor variables can be continuous, discrete, and/ or dichotomous Even with all of these advantages, there are some limitations to logistic regression For instance, the ratio of the number of candidate predictor variables to the number of cases can be problematic if there are too few cases (Mertler & Vannatta, 2005) This limitation and others are discussed in the sections that follow Logistic regression has the potential to be a powerful tool in predicting dichotomous outcomes (Peng et al, 2002) A dichotomous outcome is one that can be coded into one of two mutually exclusive categories, eg, students who earn grades on record (or grades) as compared to those who do not or students who persist from fall to fall versus those who do not As an illustration, a research study with a dichotomous outcome might involve identifying educational background characteristics to inform decision makers during enrollment management planning Conducting logistic regression requires that researchers are at a minimum familiar with both evaluating the validity of a model and the steps to conducting a logistic regression analysis The methodology used to predict a dichotomous outcome is extremely important because of the impact that the models have on students For example, if the model does not do well at predicting success, students may take courses that they don't need resulting in increased education costs as well as additional time to reach their goals Many community college researchers are responsible for validating third-party assessments like College Board's ACCUPLACER This paper discusses the process of using logistic regression to predict success in an English course based on educational background measures and assessment scores …

Effects of Learning Communities on Community College Students' Success: A Meta-Analysis

Abstract: Effects of Learning Communities on Community College Students’ Success: A Meta-Analysis by Keith Wurtz M.A., CSU, Fullerton, 1997 B.S., California Polytechnic University, Pomona, 1990 Dissertation Submitted in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy