STAT2012 Statistical Tests

General Information

This page contains information on the Intermediate Unit of Study STAT2012: Statistical Tests.

This unit is offered in Semester 2.

The lecturer for this unit of study is Jennifer Chan.

For further information on Intermediate Mathematics and Statistics, refer to the Intermediate Handbook.

You may also view the Faculty Handbook entry for STAT2012 from the central units of study database.

Credit point value: 6CP.
Classes per week: Three lectures, one tutorial and one computer laboratory session.

Links to online resources for this unit of study will be added here as the material becomes available.

Please note that assignment one is due on this Wed. You can type or hand-write your ass. but you should provide working steps. You are encouraged to include R output for checking your answers. You will get bonus mark for this. You may hand the ass. to me during lecture or put it into the ass. box just outside my office in Carslaw 817.
Computer practical: Q1 t-test is more powerful (higher prob. of rejecting H0 if H0 is true) but NOT necessarily more accurate or better in general! If the normality assumption is not valid, it may give false result. Also we say sign test is less powerful NOT more reliable than t-test. With less assumption, sign test and Wilcoxon test are more general but if normality assumptions applies, t-test should be used.
Computer practical: Q1 Remember to write binom.test(x,n,0.5,alt="greater,0.95). Some miss out 0.5 and get a large p-value.
Computer practical: Q2 you should say Wilcoxon sign rank test is more powerful than sign test but Wilcoxon sign rank test is NOT the same as t-test even though they give similar p-values.
We are up to P.96 of the lecture note, just finished 2-sample z-test.
Previous reminders.

Last revised 19/08/08

All rooms are in the Carslaw building unless otherwise indicated.