Ph.D. Comprehensive Exam in Health Systems

Question 1 – Seeds: Random Effect Logistic Regression Logistic regression is a commonly overlooked but powerful modeling tool used to predict a binary condition by using predictors that can be either categorical or quantitative. For example, one could measure the probability that a basketball player will successfully shoot a free throw based on factors such as historical performance, age, weight, gender, time of year, conditioning, and playing experience. Although there are many similarities between logistic and linear regression, we cannot simply apply linear regression directly to a problem with a binary outcome because one of the assumptions of linear regression is that the relationship between variables is linear. Transforming data using logarithmic transformation is one way to solve this problem using linear regression, and in fact, the logistic regression model is based on the idea of expressing the multiple linear regression equation in logarithmic terms. [1] In this question, we first examine applicability of logistic regression on a simple example that looks at how to predict germination of seeds depending on seed and type of root extract. Next, we import data to SPSS and perform relevant statistical summaries and analyses of the data. Finally, we look at how a similar model can be used in the health sector to measure and predict bed occupancy rates at hospitals. (i) Posterior Estimators and Simulated Variances First, the data were imported from BUGS example Seeds, available free at germination numbers (number that germinated and number planted) on each of 21 plates that are arranged according to a 2 by 2 factorial layout by seed and type of root extract. D. Huang Dr. Vidakovic - 3 These data are summarized below in Table 1-1. Root Type S