Sociology 229A: Event History Analysis

Short Assignment #5: EHA Outliers

Due: Start of class (2:00) May 13

This assignment requires a dataset on the course website entitled “Assignment 3 Environment.dta” and an accompanying do-file.

Regarding the data: The unit of analysis is the nation-state. The outcome of interest is first enactment of major pro-environmental legislation.

1. Download the dataset

2. Run a Cox Model

3. Compute martingale residuals, deviance residuals, efficient score residuals, schoenfeld residuals, and scaled schoenfeld residuals.

4. Make visual plots of residuals versus caseid for:

a. Deviance score residuals

b. Schoenfeld (or Scaled Schoenfeld) residuals for each X variable

Note: obtain country names on the plot, add the option: “mlabel(newname2)”

5. See if you can identify any outliers. If so, drop the outlier(s) and re-estimate the Cox Model. (If you can’t find any real outliers, practice removing a case and re-estimating the model anyway…)

Hand in the following:

1. At least 2 of your residual plots (noting outliers, or lack thereof).
2. Cox model results before & after removal of cases. Note exactly which cases were removed. Note major changes in the results.

Optional (advanced) exercise: Use the efficient score residuals to compute dfbetas. Make relevant plots to identify outliers.