Information Pertaining to Review for Final Exam


Note: You should plan to take the final exam at the time it is scheduled for --- it will not be easy for you to convince me to give you the exam at any other time. I plan to use the time between the last lecture and the exam to create the exam, and so I don't anticipate that it will be possible to take the exam early.

The final exam is an open book(s) / open notes exam --- you can use whatever books and other printed or written material that you want to bring with you. I recommend that you also bring a basic scientific calculator. There should be no real need for a computer, but you can bring one if you wish to do so.

Extra Office Hours

I will hold extra office hours in Room 222 of Thompson Hall (so same building but not same room as where we have class) at the following times: (Note: In Thompson Hall there is a 222A and a 222B, in addition to 222, but they aren't all next to each other.)

During the first 85 to 90 minutes of each 3 hour time slot I'll lecture about some of the material that's in the notes, but that I can't seem to cram into the regular lectures. (This is basically to benefit you with regard to completing my discussion of applied statistical techniques that I'd like for you to know about. If you're trying to do the Bonus HW problems and are unsure about something, then maybe I'll cover it when I lecture. The final exam is designed so that you can get a perfect score w/o having heard my final 90 minutes worth of material.) After a short break, I'll be glad to answer reasonable questions you may have pertaining to the final exam.

What to Study

I am planning to give you 6 multiple choice and/or true/false questions worth 2.5 points each, and 10 other parts (where you are expected to show your work and produce answers) worth 5 points each. I'll take your best 7 out of 10 scores from the 5 point parts to get a score out of 35 that I'll add to your score from the multiple choice questions --- making the exam worth 50 points in all. Your exam will be easier than most of the previous STAT 554 final exams.

I'm planning an exam to match well with this study guide, which points out material that will be important for your final exam. In addition to understanding the material referred to on the study guide, I suggest that you make sure that you understand the meaning of terms that I used often throughout the semester: terms like power, homoscedasticity, heteroscedasticity, conservative test, and anticonservative test. Also, make sure that you can get a p-value, or make a statement about the p-value in cases where an exact p-value isn't practical, using the tables for the sign test, the Wilcoxon rank sum test, Friedman's test, and the tables of critical values for chi-square distributions (which could be used for the Krukal-Wallis test and Pearson's chi-square goodness-of-fit test (including the version of the test that can be used to test for lack of independence using a contingency table)). Also, be prepared to get an exact p-value for Fisher's exact test, and the Wald-Wolfowitz runs test. Your exam will have at least two very straightforward parts where you are expected to do a standard test on a smallish amount of data. (And the tables may be useful for one or more other parts of the exam as well. I have one part where you will need to make use of the tables for an exact p-value from a Wilcoxon-Mann-Whitney test, but you won't be asked to do such a test. (Rather than try to guess what I mean by this, just be prepared to do an exact Wilcoxon-Mann-Whitney test, and you should do okay with what I have planned for you.)) I'll also inform you that you will be requested to produce at least one confidence interval for a parameter, and you'll be asked to give at least one point estimate of a variance.

I distributed two previously used final exams. (Some comments about these old exams, and the answers for one of them, can be found on the study guide. The answers to the other exam are supplied in the handout containing the two exams.) As for the problems on the old exams that I distributed, the most pertinent ones may be 2, 3, 4, 5, and 9 from the Fall 1997 exam, and 1, 2, 4, 5, 6, and 8 from the Spring 1994 exam. If you can do these problems, you'll get some practice with concepts and techniques that may be useful on your exam. But don't spend too much time on these problems and fail to spend adequate time studying the other material that I've indicated is important --- for your exam you'll need to know more than just how to do these problems (and some aspects of some of these problems are not at all strongly related to your exam).

Unlike some other exams I have given in the past, your exam will not be extremely hard. Since I will only count your best 7 of 10 parts from the last portion of the exam where you work problems, you can screw up and not learn all of what you are supposed to learn and still get a very high score.