Reference no: EM133504537
Case Study: The data set stored in the file "hw2data.txt" is from a case control study where the objective was to assess the protective effect of omega-3 fatty acids absorbed by the red cells on Primary Cardiac Arrest. There are six variables in the data set: Case/Control status (1: Primary Cardiac Arrest, 0: Control), Age (in years), Gender (1: Female, 0: Male), total kcal spent in physical activity, DHAEPA (dietary intake of omega-3 fatty acids estimated from the survey reports of fish intake from the spouses of the cases and controls) and REDTOT a measure of omega-3 fatty acids absorbed in the red cell membrane. The covariates thought to be confounders are age, gender and physical activity. Unfortunately, there is considerable amount missing values in REDTOT.
Question 1. How will you assess whether there is any evidence against missing completely at random mechanism for REDTOT?
Question 2. Develop adjustment cell weights to compensate for potential selection bias in complete case analysis.
Question 3. Develop a response propensity model for REDTOT and develop weights based on the estimated response propensity.
Question 4. Develop a regression model for predicting REDTOT and use it to develop adjustment weights.
Question 5. Obtain complete-case, and the three weighted estimates (weights developed in (2), (3) and (4)) of the regression coefficient for REDTOT in a logistic regression with the case-control status as the dependent variable and age, gender, total kcal and redtot as independent variables.
Question 6. The standard error of the estimated regression coefficients needs to be obtained using survey analysis software. Review Proc Surveylogistic section of the SAS manual or use R routines developed by Thomas Lumley (available in the R library) to obtain the standard error the three weighted estimates