Data Analysis 765 Syllabus

Spring 2005

Instructor: Ken Petren *email: ken.petren@uc.edu (6-9719)

Office Hours: by appointment: 802 Rieveschl

TA: Andrew Osterberg. 6-9715 osterbar@email.uc.edu

Lecture & Lab:

2-5 Tues Thurs. (required) Computer Lab: Room 622 Rieveschl

GOALS:

To make you better able to understand, analyze and interpret data.

The course will emphasize:

- Basic statistical foundations and framework (biological context)

- Experimental design

- Hypothesis formulation and testing

- Interpretation and presentation of results

My goal is to provide you with a conceptual framework to facilitate your career development. I will try to make you aware of limitations and assumptions you will encounter in analyzing data. As a graduate student, you must supply the motivation to get the most out of this course. For instance, you must be prepared to delve further into the reference texts and other literature on your own when you are designing experiments and analyzing your own data.

RESOURCES

Computer programs: SYSTAT 10.2 (PC); Resampling Stats.

Text: Zar, J. H., Biostatistical Analysis (strongly recommended)

Computer Lab: You have access any time other classes are not using the lab.

There will be a fair amount of traffic 9-12 and 1-4 for Ecology 303.

We will post a schedule soon. See me for other access to SYSTAT

GRADING: your final grade will be based on 100 points. 90-100 points will earn you an A; 80-90 a B and so on, however grades may be scaled up for the entire class based on group and individual effort as judged by the instructor.

Problem Sets 75% (see tips below)

Weekly problem sets will be graded for logic, presentation and interpretation. Points will be deducted for cutting and pasting raw output, poor graphical presentation, poor grammar, and inaccurate and unjustified analyses. Please see the Guide to Problem Sets.

** The last problem set is worth twice as much as others, and in it

you will focus on a problem of your choice.

Participation 25%

-Paper presentation. Objective is not to summarize the paper,

but to ask insightful DATA ANALYSIS questions

of your comrades.

-Paper discussion. You must read the paper twice, and you must be

prepared to participate.

-Asking / answering questions in Lecture and Lab.

PROBLEM SETS: Think of problem sets as miniature papers. Each should have a 1-2 sentence statement of the question (In your own words!), hypotheses, analysis interpretation and broader conclusions. Answers to problem sets must be typed, combined with graphics into a single MS WORD file and turned in via email. You must summarize numeric output from SYSTAT (do not paste SYSTAT output tables directly). The use of your own summary tables is strongly encouraged where appropriate (e.g. for results from 5-7 similar t-tests). SYSTAT graphs can be pasted directly into MS Powerpoint for quick manipulation, labeling and drawing, and these figures can be pasted into MS WORD. See the Guide to Problem Sets for help.

ASK FOR HELP!: Please do not hesitate to ask for help if you need it on any aspect of problem sets or understanding the material. However, for help with analyses,. you MUST at least try to solve your problem before asking for help! Keep in mind that my door is open for statistical advice down the road IF you have made an effort to do as much as you can on your own in this class.

ATTENDANCE is mandatory. You must make arrangements BEFOREHAND if you will miss all or part of a class, or you will be docked 10 percentage points on your final grade. Scientific reasons are acceptable for missing class (e.g. going to a conference, data that must be collected during a specific time). Social events are not acceptable excuses.

PLAGARISM: Group participation is encouraged, but everybody should conduct the analysis at their own computer. Plagiarism will not be tolerated. This means that WRITTEN REPORTS MUST BE IN YOUR OWN WORDS, and the work of others must be properly cited. The penalty for plagiarism is a zero on that assignment.

COURSE OUTLINE (DA 765): (Subject to Revision)

Week 1 (3/29)

Philosophy of Data Analysis

Types of Data / Definitions

Descriptive statistics

Distributions & Variation

The normal distribution

The central limit theorem

Week 2 (4/5)

Probability and chance

Resampling and the bootstrap.

Random variables

Introduction to hypothesis testing

Week 3 (4/12)

Hypothesis testing

Area under the curve, alpha, Type I and II errors,

Comparing the means of two groups

t-distribution

Introduction to statistical power

Week 4 (4/19)

Data transformations I

Log, square root, arcsin

Graphical representation of data

Correlation

Assumptions and uses

Week 5 (4/26)

Regression

Linear, IV/DV assumptions.

Relationship to ANOVA and the GLM

Multiple regression

Scaling relationships

Week 6 (5/3)

Categorical variables

Goodness of fit tests

Chi square, K-S test

Contingency tables

Nonparametric statistics

strategies of use

Week 7 (5/10)

ANOVA

Introduction to general linear models

F-statistics calculations and assumptions

Introduction to experimental design

Independence, confounding factors and pseudoreplication

Discussion: paper TBA (e.g. Hurlbert pseudoreplication)

Week 8 (5/17)

Higher-order ANOVA,

Repeated measures

fixed vs. random effects

Discussion: paper TBA (e.g. Petren Case Ecology paper)

Week 9 (5/24)

ANCOVA

Power analysis

Review of basic concepts and examples

Discussion: paper TBA (e.g. Jayne lizard tracks paper).

Week 10 (5/31)

Designed based on class composition.

- Phylogenetic inference? Character evolution?

Independent contrasts?

- A brief introduction to basic multivariate analyses (PCA?)

Discussion: paper TBA (e.g. Hamilton Zuk experiment paper).