Stat 2000: Preparing for your midterm exam and the Cheng Multiple-Choice book

Published: Sun, 10/26/14

Here are some suggestions to help as you prepare for your midterm exam.

Make sure that you have gone through the two sample exams they have provided you thoroughly.  They are the best indication of what your exam may be like.
 
If you have purchased the Smiley Cheng Multiple-Choice Problems Set for Basic Statistical Analysis I (Stat 2000) available in the Statistics section of the UM Book Store, you can certainly use the old midterm exams in that book for additional practice. 

Note that the course has undergone changes in topics and philosophy over the years, so some questions in the old midterms must be omitted (you will cover those topics later in the course).  Also some topics (probability) are not adequately represented since that has been only added to the course in recent years.

I suggest you start with the most recent exams and work your way backwards.  The more recent exams are probably more indicative of what your exam will be like.  The exams from the 90s are probably too easy, as the midterm has definitely gotten harder over the years. 

Do realize that these exams were only 1.5 hours long, not 2 hours like yours, so they are shorter.

The solutions to these term tests are in Appendix C of my book starting on page C-1.  (But you have to have the Smiley Cheng book to see the questions.) 
Some facts to be aware of if you are using my answer keys to help solve the old exams in Cheng:
  • Please understand many of these solutions were written several years ago, and the course has undergone many changes during that time.  Specifically, you may see me referring to different statistical tables (Cheng/Fu Tables 1, 2, etc.).  You are expected to use the tables in Moore & McCabe to answer all of the questions.  (In the case of Poisson, of course, you have no tables, so you will have to use the Poisson probability formula.)
  • In all of the exams prior to 2007-2008, I use the conservative method for the two-sample t test when the pooled method is not called for.  This is essentially the generalized method, but where we use the smaller of df = n1 – 1 and df =  n2 – 1 for our degrees of freedom rather than that horrible formula we use nowadays.  Back then, the course used the conservative method rather than the generalized method.  You would have to use that really complicated df formula given as #1 on the formula sheet these days (but, then again, they will probably compute the df for you in that case).
  • You may find me using the conservative t-test in some of the exams from the 90’s even though the standard deviations suggest that you should use the pooled t-test (for example, question 12 from 98/99 Term 1).  This is because they were using different criteria back then.  On your exam, use the “Rule of Thumb” discussed in Lesson 4 to determine if the pooled t-test should be used (unless your prof has told you to do otherwise).
  • The 1997/98 Term 1 Midterm Exam includes nonparametric test questions that you will not do until after the midterm, if at all.  Specifically, omit the following questions: #12 up to and including #16 and omit Part B #2(d).  The 1997/98 Term 2 Midterm Exam is almost pointless to look at.  Specifically, omit the following questions: #7 up to and including #16, #18, and #20.