Probability and Statistical Inference

Instructions for the Final Exam

Relevant Topics

The final exam covers all topics we have learned in class. The
exam focuses about 45% on material from Midterm 1, about 45% on
material from Midterm 2, and about 10% on topics after Midterm 2.

The final exam is designed to test your understanding of statistical thinking and methods. You will not be asked to perform laborious calculations. For example, you don't have to calculate correlations or standard deviations of a list of numbers by hand. You will be asked to manipulate probability distributions, compute expected values and variances for discrete and continuous random variables, and compute covariances for two discrete random variables. You will be asked to use the law of iterated expected values. You will be asked to compute variances of linear combinations of random variables. You will be asked to compute maximum likelihood estimates for simple problems. Regression and Bayesian inference will be covered on the exam.

You may be given JMP output and asked to interpret it. For example, you may be given some problems with two or more sets of JMP output reporting the results of different analyses, and you have to choose the correct analysis. You do not need to memorize JMP commands or click-sequences for the exam.

You also will be asked some questions that are not attached to specific data. For example, you may be asked to design a particular study or to comment on the veracity of statistical statements.

Logistics

The exam is on **Wednesday, May 2,
7:00
PM -- 10:00 PM.**

During the exam, you are allowed to use two sheets of 8 by 11 paper filled with any information you want to put on them. Both sides of the sheets can be used. You cannot use any other material during the exam.

Bring a calculator. Bring your tables for the normal curve, the t-curve, and the chi-squared curve.

The exam is designed to take about 2.25 hours, although you have the full three hours to complete it.

Suggestions for Studying

A useful way to study for these exams is to do the problems from the supplement and from the text. Often, the hardest and most crucial part of statistical analyses is deciding which inferential technique to use. Thus, you may find it helpful to ask a friend to select problems from an introductory text book, have them cover up any instructions like "use a one-sample t-test to...", and then decide which procedure to use. This simulates real-life, and it also simulates the exam. You also should redo the probability problems form the supplement. You might alter the numbers slightly and check your answers with friends. Another great way to study is to write your own questions, then give them to study partners to take. This is especially good for the probability and sampling distribution problems and for deciding what analyses to use.

Below is a link to a whole bunch of practice problems. There
are more practice problems than will be on the final exam.
Some topics are not covered in the problems. They still may be
covered on the exam.