Fall 2013 Stat6800 Final Exam

Web Address of This Page

http://www.stat.wmich.edu/wang/680/2013FinalExam

Problem Description

XYZ Pharmaceutical, Inc. conducted a clinical study. A portion of the coutcomes recorded two types of adverse events (AE, Types I or II) with serverity (Mild, Moderate, and Severe) that occurred in the study dated by Ady (Analysis Day). These information for the six sites (A, B, C, D, E, and F) of 200 subjects each were recorded in ae.txt. Subjects were labelled as 0001, ..., 1200.

Through the entire study, if a subject has at least 2 occurrences of Severe Type I AE, then he/she is flagged with a "Yes" for critically critical. In addition, if a subject has either 3 or more occurrences of Moderate Type II AE, or at least 2 occurrences of Severe Type II AE, then he/she is flagged with a "Yes". Thus, the potential combinations of flags for two types (named ccfl in SAS codes below) are (No, No), (No, Yes), (Yes, No), and (Yes, Yes). A subject having ccfl of (No, Yes) has No for Type I and Yes for Type II.

It is important to know the frequency counts of subjects for the four possible values of ccfl within site. An example report for site F (site-F.txt) was created by (the correct) SAS program.

Your WIN will set a random seed so that the raw ae data of a randomly selected site (from sites A to E) will be assigned to you to produce a summary report for that site.

You are provided a SAS Codes (FinalExam.sas) in which many coding errors occur. These errors could be

You are to correct all errors. The SAS log generated from running your SAS codes is expected to be error free.

To help test your (corrected) SAS codes, uncomment the line (first statement after the data _null_ step)

    %*let _site = F;
   
to
    %let _site = F;
   
and compare your SAS output with the example report for site F (site-F.txt).
Rember to comment this line again when you are done testing. Name your corrected SAS codes as yourLastName_Final.sas. Clearly mark all errors in the provided SAS codes. You can place the buggy code chunks (i.e., a code chunk=a data step or a procedure step or a standalone statement) in a SAS comments and identify errors in it.

Program Submission

You are to submit SAS codes only (as an attachment) through email and get a confirmation of receipt from the instructor before you leave. My email address is

jung-chao.wang@wmich.edu