ELME is a summer educational program at the Kellogg Biological Station devoted to Enhancing Linkages between Mathematics and Ecology.
Research often involves addressing hypotheses with data that are non-linear, non-normal, hierarchical or otherwise complex. This year’s ELME program focuses on introducing graduate students (or advanced undergraduates) to modern statistical methods that make it possible to link theory with empirical data in ecological and evolutionary research. The program will consist of three, intensive week-long courses.
Each course features a hands-on environment; students will learn the basics in a lecture setting and while cementing their knowledge with independent and collaborative lab projects using R. An emphasis is placed on applying course concepts to students' own research and data. Weekly topics are complementary, but each course is independent, so students can take any combination (one, two or all three) of these courses. Class runs 9 am - 5 pm, M-F and may involve evening projects, so residence is encouraged.
Week 1: June 2-6
This course explores maximum likelihood estimation (MLE) as a tool for model fitting and parameter estimation, and a powerful alternative to least squares methods. We’ll cover how MLE can be used to fit standard models (such as linear regressions and ANOVAs), as well as its application to non-linear, non-normal data. We’ll also address topics such as model selection and multi-model inference that allow us to choose between competing models and test hypotheses. Advanced topics will include time series analysis and working with zero-inflated data. Throughout the course, lab activities and projects will be based on real ecological data, including student’s own data.
Week 2: June 9-13
This course will address the who, what, why, and how of Bayesian statistics. Who was Bayes and what was his insight? What are Bayesian statistical approaches and how do they differ from more traditional methods? Why should or shouldn’t one use a Bayesian approach? How does it work? This is a hands-on course. By the end of the week, you will be conducting Bayesian analyses of real ecological data, ideally your own.
Week 3: June 16-20
We will cover the theory and application of data analysis using causal network models. Topics will include the logic of causal networks; linear gaussian models; the analysis of latent variable models; and non-linear, non-gaussian models. The course will consist of lectures, lab activities and student projects. Bring data!
Hours: Mon-Fri 9 am - 5 pm
Target audience: 12-18 graduate students and postdocs; exceptional undergraduates will be considered
Prerequisites: Previous experience with statistics including familiarity with generalized linear modeling (eg, ANOVA, linear regression, logistic regression) is recommended. Programming experience (any language, but particularly with the statistical programming language R) is beneficial.
Format: A mixture of lecture, and guided computer labs.
Admissions: favors, but is not limited to, students who can enroll through MSU or a participating CIC institution. Additionally, scholarship support is available to assist with tuition and expenses.
Click here for more information on Admission
Click here for more information on Scholarships
Contact email@example.com with questions.
|Last Updated on Friday, 21 February 2014 15:03|