Weekly Schedule: Wednesdays from 1-4
Office Hours:Prof. Byrnes will hold office hours Tuesday and Thursday from 2-4. If you need to schedule a different time,
email him.
Overview & Objectives:
In Special Topics in Biological Data Analysis, we will take a semester and explore a particular topic or concept in data analysis in detail. Students will come to the course either with a project in mind that requires this technique, or, within one week of beginning the course will start a new collaborative project using the technique.
You will:
1) Develop an understanding of the role of meta-analysis in the scholarly literature.
2) Learn to implement meta-analysis for one's own work.
3) Conduct a publication quality meta-analysis that contributes to the development of students' research objectives.
Prerequisites:
Previous experience with statistical data analysis at an advanced level or permission of the instructor.
Experience with mixed models helpful.
Students who are not familiar with R should work through at least one of the tutorials at
http://pairach.com/2012/02/26/r-tutorials-from-universities-around-the-world/
OR (highly recommended) through the introductory modules in the swirl package at
http://swirlstats.com/ before taking the course. Advanced undergraduates welcome at instructor's discretion.
Also, you should install
R and
R-studio on your laptop before the first day of class. In addition, please install the
metafor,
devtools,
rmeta,
car,
plyr, and
ggplot2 packages. Please bring your latop to all classes.
If students wish to work in other languages, they are more than welcome to do so.
Required Texts:
While the topic covered in this course is general to all sciences, we'll use a textbook that was produced by a group of ecologists and evolutionary biologists. Despite it's title, the text is a general book for many disciplinary audiences with some additional material at the start and finish that are specific to EEB.
Handbook of Meta-Analysis in Ecology and Evolution. 2013. Edited by Julia Koricheva, Jessica Gurevitch, Kerrie Mengersen. Princeton University Press.
http://www.amazon.com/Handbook-Meta-analysis-Ecology-Evolution-ebook/dp/B00CC3N17C/
Recommended Texts:
These two books are some of the standards in the field. We'll have them available as reference as needed and some chapters may be used as readings.
Introduction to Meta-Analysis. 2009. Michael Borenstein, Larry Hedges, Julian Higgins, and Hannah R. Rothstein. Wiley.
http://www.amazon.com/Introduction-Meta-Analysis-Statistics-Practice-Borenstein/dp/0470057246/
The Handbook of Research and Synthesis and Meta-Analysis, 2nd Edition. Edited by Harris Cooper, Larry Hedges, and Jeffry Valentine. Russel Sage Foundation.
http://www.amazon.com/The-Handbook-Research-Synthesis-Meta-Analysis/dp/0871541637/
Grading:
Students will be graded on presentations, class participation, and the final paper. Students will each present at one to two chapters during the course. This presentation will be worth 30% of their grade. Participation in classroom discussions about the material will be worth 20%. The final paper (written up either singly or in groups) and its concomitant public presentation will be worth 50% of the grade. Strong evidence of intent to submit for publication will give students an extra 10%.
Teaching approach
The course will be a mix of lecture from both students and the professor as well as discussion about the material and how it relates to the projects conducted by students.
Final Paper:
For the final paper, students will be asked to define a question answerable by meta-analysis and write a paper that addresses it by the end of the semester. This work should be publication quality, contributing to the students' intellectual advancement. Students can do the project in groups or by themselves (e.g., there may be one class project, two different projects dividing the class in half, everyone can do their own project, etc.). Groups are encouraged, as data gathering can take some time, and many hands make light work. In addition, different members of the group may be able to better lend their expertise to different parts of the project (e.g., data gathering, analysis, writing). It is the students' choice as to how work is partitioned within a group. Every week we will take some class time to discuss project progress, particularly in the context of the topic of the week.
By the end of week two, students must have approval of the topic by the course instructor. Data gathering should be complete by the middle of the semester, and students will be asked to present a mid-semester update. Students will present the final results in a talk open to the school at the end of the semester.
Course Content:
Students presenting chapters will be chosen at the beginning of the semester. Additional papers will be assigned as needed.
Date
|
Topic
|
Chapters & Papers
|
Assignment Due
|
Slides |
Code |
Jan 29th
|
What is
Meta-Analysis, Project Discussions
|
1-2,25
(yes, read these before day 1! it's a long break!)
|
Bring potential meta-analysis research ideas
|
Intro Lecture
|
|
Feb 5th
|
Finding and
Gathering Data
|
3-4,20
|
Discuss and
finalize research project questions
|
Starting a Meta-Analysis Lecture
|
|
Feb 12th
|
Assessing Data
Quality, Data Extraction
|
5,23
|
Create search
strategies and bring an initial prospective literature list
|
Data Extraction Lecture
Discussion Questions
|
Web Plot Digitizer |
Feb 19th
|
Metrics of
Effect
|
6-7
|
Randomly
sample prospective studies to evaluate feasibility, refine study list
|
Effect Size Metrics |
Calculating Effect Sizes in R
Marine BEF Data
Code for Estimating Effect Sizes |
Feb 26th
|
Publication
Bias
|
14
|
Present metrics that will be used for evaluation, demonstrate feasibility
|
Bias Lecture |
Checking for Bias in R Presentation
Code for Bias Checking |
March 5th
|
Missing Data
|
13
|
Gather data
|
Missing Data Pt. 1
Missing Data Pt. 2
|
|
March 12th
|
Introduction
to Visualization & Statistical Inference in Meta-Analysis
|
8,12,21
|
Gather data
|
Fixed and Random Models in Meta-Analysis |
Lepidoptera data
Code for Fixed and Random Lepidoptera Analysis |
March 26th
|
Moment Based Approaches
|
9
|
Gather data
|
Mixed Models and Meta-Regression in Meta-Analysis |
Code for Mixed and Meta-Regression Lepidoptera Analysis
Code to Load Marine Meta-analysis Data |
April 2nd
|
Maximum Likelihood Based Approaches
|
10
|
Perform one to two analyses on data gathered to date
|
Likelihood & Hierarchical Models in Meta-Analysis |
Code for Hierarchical Meta-Analytic Models
rma v. lm and lme |
April 9th
|
Bayesian Approaches
|
11
|
|
An Overview of Bayesian Techniques in Meta-Analysis |
Simple Bayesian Meta-analysis with MCMCglmm
Code for Simple Bayesian Meta-Analysis |
April 16th
|
Temporal & Spatial Trends in Data
|
15, 19
|
Compare
multiple forms of analysis with gathered data. Ask, which is appropriate
|
|
Cumulative Meta-Analysis in R
R code for cumulative meta-analysis |
April 23rd
|
Power Analysis & Other forms of Validation
|
22
|
Evaluate power
of analyses
|
Power Analysis in Meta-Analysis |
R code for power analysis by simulation
|
April 30th
|
Non-Independent
Data
|
16
|
|
Modeling Non-Independence in Meta-Analysis |
R code for modeling non-independent data
robumeta
A meta-sandwich
Another meta-sandwich
Meta-sandwich with extra mustard
|
May 7th
|
Phylogenetic
Non-Independence
|
17
|
|
|
|
May 14th
|
Dealing with
Raw Data
|
18
|
Present
results corrected and not corrected for non-independence
|
|
|
May 21st
|
Final
Presentations!
|
|
Invite guests
|
|
|
May 28th
|
Final Projects
Due
|
|
|
|
|