Logistic   Regression

Logistic   Regression

Many variables in health research are binary, which means something composed of two parts, such as, having or not having a condition.  The following are examples of binary variables: having or nor having a myocardial infarction (heart attack),  living or dying, being exposed or nor exposed to a risk factor, smoking or not smoking, obese or not obese, taking or not taking a medication, etc.  When a binary variable is predicted from one or more variables, linear regression is not appropriate.  Instead, logistic regression is employed.  Other types of regression can also be used, such as Cox regression when time to event is also modeled, or Poisson regression for frequency outcomes, but logistic regression is one of the most widely used and can serve as a model for understanding  the prediction of discrete outcomes.  For this discussion you are asked to search the Ashford University Library for a scholarly, peer-reviewed journal article describing a study that utilized logistic regression.  Share with your peers the study design, the participants and the research question/s.  What descriptive analyses were performed and what did they reveal?  What was the dependent variable?  What were the independent variables?  What were the main results and the main conclusions?  What implications do the results have and what directions do you see for further research?  Cite your study using APA format as outlined in the Ashford Writing Center.