Metodo

International Studies in Phenomenology and Philosophy

Book | Chapter

178544

Correlation to logistic regression illustrated with a victimization-sexual orientation study

Creaig A. Dunton Mark Beaulieu

pp. 173-198

Abstract

Dunton and Beaulieu hold a positivist ideology. In their chapter, they explain a common positivist technique: correlation. They go on to discuss regression and a specialty technique: logistic regression. Correlation and regression are generally deductive within-group unit of analysis strategies, since factors of interest are measured as predictors of the dependent variable. The factors and dependent variable of interest in the unit of analysis are established through a scholarly literature review. As with all true positivistic ideologies, hypotheses are developed to test the unit of analysis. A unique aspect of their example was the ex post facto use of logistic regression on existing data. Using correlation and regression is not considered mixed methods or multi-methods because researchers with a positivist ideology generally use correlation first to show evidence of the hypothesized relations between factors or between factors and the dependent variable, otherwise it may not be feasible to continue the analysis. Logistic regression has specific assumptions that must be met in order to be applied, and they discuss this.

Publication details

Published in:

D Strang Kenneth (2015) The Palgrave handbook of research design in business and management. Dordrecht, Springer.

Pages: 173-198

DOI: 10.1057/9781137484956_12

Full citation:

Dunton Creaig A., Beaulieu Mark (2015) „Correlation to logistic regression illustrated with a victimization-sexual orientation study“, In: K. D Strang (ed.), The Palgrave handbook of research design in business and management, Dordrecht, Springer, 173–198.