Basics

Tests for nominal variables

Descriptive statistics

Tests for one measurement variable

Tests for multiple measurement variables

Multiple tests

Miscellany

BISC 667, Research Methods in Biology, Fall 2007

Section 010

Tuesdays and Thursdays, 11 a.m.-12:15 p.m.

051 McKinly

Instructor: John McDonald
322 Wolf Hall (office)
026B Wolf Hall (lab)
E-mail: mcdonald@udel.edu
Phone: 831-2007 (I rarely check messages, so e-mail is better)
Class web page: http://udel.edu/~mcdonald/statsyllabus.html


Note: Those homework assignments that are not in their final version are shown in italics; don't start working on a homework assignment until the final version is posted, as indicated by regular type. Study guides in italics may also be revised, so don't print them out until their link is in regular type.

Date Day Lecture topic Homework due
Aug. 28 Tuesday Introduction;
Kinds of biological data
 
Aug. 30 Thursday Probability  
Sept. 4 Tuesday Hypothesis testing;
Exact binomial test
HW 1
Sept. 6 Thursday Power analysis;
Random sampling
 
Sept. 11 Tuesday Chi-square test,
G-test
and Randomization tests for goodness-of-fit
HW 2
Sept. 13 Thursday Chi-square test,
G-test
and Fisher's exact test of independence
 
Sept. 18 Tuesday Small numbers in chi-square and G-tests;
Repeated G-tests of goodness-of-fit
HW 3
Sept. 20 Thursday Review for the exam  
Sept. 25 Tuesday First exam (see the study guide)  
Sept. 27 Thursday Go over exam answers;
Descriptive statistics: Central tendency
 
Oct. 2 Tuesday Descriptive statistics: dispersion,
standard error and
confidence limits
HW 4
Oct. 4 Thursday Student's t-test;
Introduction to anova;
Single-classification anova: Model I vs. Model II;
testing homogeneity of means
 
Oct. 9 Tuesday Planned comparisons among means;
unplanned comparisons among means;
estimating added variance components.
HW 5
Oct. 11 Thursday Assumptions of anova:
Normality;
homoscedasticity
 
Oct. 16 Tuesday Data transformations;
Kruskal–Wallis test
 
Oct. 18 Thursday Nested anova HW 6
Oct. 23 Tuesday Two-way anova;
paired t-test
HW 7
Oct. 25 Thursday Wilcoxon signed-rank test;
sign test
 
Oct. 30 Tuesday Second midterm (see the study guide)    
Nov. 1 Thursday Linear regression;
correlation
 
Nov. 6 Tuesday Spearman rank correlation HW 8
Nov. 8 Thursday Polynomial regression  
Nov. 13 Tuesday Analysis of covariance HW 9
Nov. 15 Thursday Multiple regression  
Nov. 20 Tuesday Logistic regression HW 10
Nov. 22 Thursday Thanksgiving, no class  
Nov. 27 Tuesday Multiple comparisons  
Nov. 29 Thursday Meta-analysis HW 11
Dec. 4 Tuesday Review for exam, evaluations  
Dec. 13 Thursday
1-3 p.m.
105 Sharp Lab
Final exam (see the study guide
and the practice questions
and the second set of practice questions.)
 

Purpose of the course

This course is designed for biologists who want to apply appropriate statistical tests to their data, and who want to understand the statistical tests that other biologists have used. We will therefore spend little time on the mathematical basis of the statistical tests, focusing instead on how to choose the appropriate test for a given data set, how to design experiments to make them more suitable for statistical analysis, and how to interpret the results of statistical tests. While it would be impossible to cover every statistical test ever used by biologists in a single course, we will cover many of those techniques that are commonly used.

At the end of the course, you should be able to determine the correct statistical technique to apply to many biological experiments, and you should be able to apply each technique and interpret the results. You should also be able to recognize experimental designs for which you have not learned the appropriate statistical test, and you should be able to ask intelligent questions when consulting with a statistician about such experiments.

Office hours

There are no fixed office hours. You can make an appointment by talking to me before or after class or by e-mailing me. If you have questions while studying or doing the homework assignments, feel free to e-mail me or drop by my office.

Textbook

We will not use a textbook this fall. All of the information you need will be delivered in class or available on the class web pages. On the class web pages, I will reference the sources of my information, primarily Biometry by R.R. Sokal and F.J. Rohlf (Third Edition, 1995, W.H. Freeman and Co.) and Biostatistical Analysis by J.H. Zar (Fourth Edition, 1999, Prentice Hall). I provide these references in case you need to look up more information about a particular test for a publication or grant proposal. I do not recommend using these texts if you need help understanding the material required for class. If something isn't clear to you from the lectures and web pages, please see me for help.

Attendance

Attendance is not mandatory, and you will not be graded on your in-class participation. However, I think it will be a lot easier to learn the material if you attend class. Homework assignments and class handouts will be available from the class web page (http://udel.edu/~mcdonald/statsyllabus.html). If you are absent when a homework assignment is due, please try to turn in the assignment before the next class. Do not e-mail me your homework assignments.

Grading

Your grade will be based on the following:

15% First midterm exam
30% Second midterm exam
40% Final exam
15% Homework assignments

The grades will not be curved. There are no extra credit projects. The points from the homework and exams will be combined and converted to letter grades as follows:

A 93-100; A- 90-92.9; B+ 87-89.9; B 83-86.9; B- 80-82.9; C+ 77-79.9; C 73-76.9; C- 70-72.9; D+ 67-69.9; D 63-66.9; D- 60-62.9; F 0-59.9.

Exams

The exams will be cumulative. The main emphasis of the exams will be testing your knowledge of what the appropriate statistical test is to use in a particular situation and how to interpret the results. Any equations or tables required for the exam will be provided. You will not be allowed to use reference books or notes during the exams, and you will not need a calculator. If your native language is not English, you may use a dictionary.

Homework assignments

A homework assignment is due on most Tuesdays. They will be available on the class web page a couple of weeks before they are due. I will not hand out printed copies of the assignments in class.

The homework assignments make up a relatively small portion of your grade, and they will be loosely graded; if you make a sincere effort on all parts of an assignment, you will probably get full credit. Despite this, I suggest that you put a lot of effort into the homework. Many of the skills I hope you learn from this class can only be learned by doing, not by listening to me talking, so your statistical education will be incomplete (and you will struggle on the exams) if you do not do the homework to the best of your ability. Please drop by my office/lab or e-mail me if you need help on the homework.

Software

We are going to use Excel spreadsheets for most of our data analysis. The web page for almost every statistical test will include a spreadsheet that you can download and use with the Windows or Macintosh versions of Excel. I think these spreadsheets should also work using the free program Calc, part of the OpenOffice.org suite of programs, although I haven't tested them yet.

I like using Excel for statistics because most of you already know how to use it, and because it makes it easy to combine graphics with the statistical test. If you're interested in a field that only uses fairly basic statistics, such as molecular or cell biology, the tests that you can do with Excel are likely to be all you'll ever need.

Wherever possible, I've linked to someone else's web page that will perform each statistical test, and you may use them for the homework if you prefer. I found most of these web pages using John Pezzullo's excellent list of Interactive Statistical Calculation Pages, which is a good place to look for information about tests that are not discussed in this course. For most tests, the web page version is not as easy to use as a spreadsheet, and many web pages only handle limited sample sizes. Some web pages, however, do tests that would be difficult to set up on a spreadsheet.

Each of my web pages about a statistical test will also include instructions for performing the statistical test using SAS. You do not need to learn SAS for this course. SAS is a powerful but user-unfriendly statistical package that can do statistical analyses that are far beyond the abilities of Excel. If you are interested in a research area that makes heavy use of sophisticated statistics (such as epidemiology, ecology, or quantitative genetics), you should learn SAS, and a good way to start would be by doing all the homework assignments in SAS.

If you already know another statistical program, such as SPSS, you may use it on the homework assignments, but don't ask me for help.


Return to John McDonald's home page

This page was last revised October 12, 2007. Its URL is http://udel.edu/~mcdonald/statsyllabus.html