Chao-Ying Joanne Peng

An Introduction to Logistic Regression Analysis and Reporting

Abstract The purpose of this article is to provide researchers, editors, and readers with a set of guidelines for what to expect in an article using logistic regression techniques. Tables, figures, and charts that should be included to comprehensively assess the results and assumptions to be verified are discussed. This article demonstrates the preferred pattern for the application of logistic methods with an illustration of logistic regression applied to a data set in testing a research hypothesis. Recommendations are also offered for appropriate reporting formats of logistic regression results and the minimum observation-to-predictor ratio. The authors evaluated the use and interpretation of logistic regression presented in 8 articles published in The Journal of Educational Research between 1990 and 2000. They found that all 8 studies met or exceeded recommended criteria.

Principled missing data methods for researchers

The impact of missing data on quantitative research can be serious, leading to biased estimates of parameters, loss of information, decreased statistical power, increased standard errors, and weakened generalizability of findings. In this paper, we discussed and demonstrated three principled missing data methods: multiple imputation, full information maximum likelihood, and expectation-maximization algorithm, applied to a real-world data set. Results were contrasted with those obtained from the complete data set and from the listwise deletion method. The relative merits of each method are noted, along with common features they share. The paper concludes with an emphasis on the importance of statistical assumptions, and recommendations for researchers. Quality of research will be enhanced if (a) researchers explicitly acknowledge missing data problems and the conditions under which they occurred, (b) principled methods are employed to handle missing data, and (c) the appropriate treatment of missing data is incorporated into review standards of manuscripts submitted for publication.

THE USE AND INTERPRETATION OF LOGISTIC REGRESSION IN HIGHER EDUCATION JOURNALS: 1988-1999

This article examines the use and interpretation of logistic regression in three leading higher education research journals from 1988 to 1999. The journals were selected because of their emphasis on research, relevance to higher education issues, broad coverage of research topics, and reputable editorial policies. The term “logistic regression” encompasses logit modeling, probit modeling, and tobit modeling and the significance tests of their estimates. A total of 52 articles were identified as using logistic regression. Our review uncovered an increasingly sophisticated use of logistic regression for a wide range of topics. At the same time, there continues to be confusion over terminology. The sample sizes used did not always achieve a desired level of stability in the parameters estimated. Discussion of results in terms of delta-Ps and marginal probabilities was not always cautionary, according to definitions. The review is concluded with recommendations for journal editors and researchers in formulating appropriate editorial policies and practice for applying the versatile logistic regression technique and in communicating its results with readers of higher education research.

Logistic Regression Analysis and Reporting: A Primer

Logistic regression, being well suited for analyzing dichotomous outcomes, has been increasingly applied in social science research. That potential expanded usage demands that researchers, editors, and readers be coached in terms of what to expect in an article that used the logistic regression technique: What tables should be included? What assumptions tested? What figures or charts should be expected? In this article we seek to answer these questions with an illustration of logistic regression applied to a real world data set. Results were evaluated and diagnosed in terms of the overall test of the model, interpretability and statistical significance of each predictor, goodness-of-fit statistics, predictive power, accuracy of prediction, and identification of potential outliers. Guidelines are offered for modeling strategies and reporting standards in logistic regression. Furthermore, 6 statistical packages were employed to perform logistic regression. Their strengths and weaknesses were noted in terms ...

Being Labeled Gifted or Talented: Meanings and Effects Perceived by Students in Special Programs

Questionnaires were administered to 295 students in summer programs for 9-15 year old academically gifted and artistically talented students, asking for their conceptions of giftedness and talent and their perceptions of the causes and consequences of being so identified. Findings suggested that their stereotypes regarding gifted and talented students were highly positive. These students reported that giftedness can be attained by hard work, that they are not very different from others, and that others treat them well. A minority reported negative reactions from peers.

Modeling categorical variables by logistic regression.

OBJECTIVE To demonstrate the use of logistic regression in health care research. METHOD Forward and backward stepwise logistic regression algorithms were systematically applied to a real-world data set comprising 301 cancer patients and a set of explanatory variables. RESULTS Four variables were identified as effective predictors of pain reporting by cancer patients during chemotherapy: fatigue, depression, severity of colds or viral infections, and insomnia. The 4-predictor model was validated by (a) significance tests of regression coefficients at p<0.05, (b) significant improvement of this model over competing models, and (c) goodness of fit indices. CONCLUSIONS Logistic regression is useful for health-related research in which outcomes of interest are often categorical.

Comparison of Two Approaches for Handling Missing Covariates in Logistic Regression

For the past 25 years, methodological advances have been made in missing data treatment. Most published work has focused on missing data in dependent variables under various conditions. The present study seeks to fill the void by comparing two approaches for handling missing data in categorical covariates in logistic regression: the expectation-maximization (EM) method of weights and multiple imputation (MI). Sample data are drawn randomly from a population with known characteristics. Missing data on covariates are simulated under two conditions: missing completely at random and missing at random with different missing rates. A logistic regression model was fit to each sample using either the EM or MI approach. The performance of these two approaches is compared on four criteria: bias, efficiency, coverage, and rejection rate. Results generally favored MI over EM. Practical issues such as implementation, inclusion of continuous covariates, and interactions between covariates are discussed.

Beyond Cohen's d: Alternative Effect Size Measures for Between-Subject Designs

Given the long history of discussion of issues surrounding statistical testing and effect size indices and various attempts by the American Psychological Association and by the American Educational Research Association to encourage the reporting of effect size, most journals in education and psychology have witnessed an increase in effect size reporting since 1999. Yet, effect size was often reported in three indices, namely, the unadjusted R 2, Cohens d, and η2 with a simple labeling of small, medium, or large, according to Cohens (1969) criteria. In this article, the authors present several alternatives to Cohens d to help researchers conceptualize effect size beyond standardized mean differences for between-subject designs with two groups. The alternative effect size estimators are organized into a typology and are empirically contrasted with Cohens d in terms of purposes, usages, statistical properties, interpretability, and the potential for meta-analysis. Several sound alternatives are identified to supplement the reporting of Cohens d. The article concludes with a discussion of the choice of standardizers, the importance of assumptions, and the possibility of extending sound alternative effect size indices to other research contexts.

Leisure-time physical activity and obesity in Black adults in Indianapolis.

OBJECTIVE To examine the relationship between the total volume of leisure-time physical activity (LTPA) and obesity among African American adults in Indianapolis. METHODS Logistic regression analysis with 649 African American adults. RESULTS The data show an inverse graded relationship between the total volume of LTPA and obesity for African American women, but not for men. CONCLUSIONS African American women who accumulate a high volume of LTPA (ie, 300 minutes or more per week) are less likely to be obese. Further research is needed to investigate the gender difference in the effect of LTPA on obesity.

Social-ecological factors of leisure-time physical activity in black adults.

OBJECTIVES To develop and test an explicative model of leisure-time physical activity (LTPA), including 6 selected contributory factors: self-efficacy, self-regulation, social support, perceived physical environment, outcome-expectancy value, and policy beliefs. METHODS A social-ecological model of LTPA using the structural equation modeling technique was estimated in a regional, church-going sample of 649 African Americans. RESULTS The results indicated this model is good fit to the data. LTPA was associated with self-regulation and gender directly (P<.05) and social support, self-efficacy, perceived access to LTPA facilities, and positive outcome-expectancy value indirectly (P<.05). CONCLUSIONS Multitiered interventions considering cultural relevance are recommended to improve LTPA engagement.