EEB 480. Model-based Statistical Inference for Ecology.
This course is an introduction to the modern theory and practice of scientific data analysis using both standard and innovative approaches. The unifying concepts are those of probability model, information, and inference. Students will learn and use the basic principles of model formulation, estimation, interpretation, criticism, and refinement. The course will make use of lectures, readings, and computer exercises in the R statistical computing environment. Students will obtain hands-on experience in data analysis using data provided by the instructor and students. In particular, students with scientific questions of their own and data sets to analyze will have a chance to work on these in the course. Students will develop and practice good habits in the organization, performance, and presentation of data and data analysis. Although examples will be for the most part drawn from Ecology, students from other disciplines, including Evolutionary Biology, Public Health, and Natural Resources, will learn valuable technique.
Students
will learn and use the basic principles of model formulation,
estimation, interpretation, criticism, and refinement. The course
will make use of lectures, readings, and computer exercises in the
R statistical computing environment. Students will obtain hands-on
experience in data analysis using data provided by the instructor
and students. In particular, students with scientific questions of
their own and data sets to analyze will have a chance to work on
these in the course. Students will develop and practice good habits
in the organization, performance, and presentation of data and data
analysis. Although examples will be for the most part drawn from
Ecology, students from other disciplines, including Evolutionary
Biology, Public Health, and Natural Resources, will learn valuable
technique.Students completing the course will have gained
- facility with the statistical analyses most commonly needed in ecology,
- good habits of data organization, visualization, and analysis, which will facilitate their ability to perform sound, transparent, and reproducible research,
- a deep understanding of the fundamental theory of statistical inference, which will allow the design of new analyses customized to specific questions and data, and
- the background needed to understand and criticize models and statistical inference in the scientific literature.
Course topics include:
- theory of scientific inference
- data visualization
- exploratory data analysis
- literate and reproducible programming
- review of probability theory
- general linear models
- generalized linear models
- nonlinear regression models
- stochastic simulation
- likelihood theory
- maximum-likelihood inference
- Bayesian inference
- hierarchical/mixed-effects models
- dynamic models
Additional topics that may be covered, according to opportunity and interest, include:
- phylogenetic comparative analysis
- the bootstrap and other resampling schemes
- spatial models
Prerequisites: An undergraduate-level grounding in calculus and statistics. Students unfamiliar with numerical computation in any language on any platform should consult with the instructor before registering for the course.
