Mixed Models for Biologists: A practical crash course

Mixed models are a powerful, flexible tool that empower biologists to control for various sources of non-independence ("pseudoreplication") without losing (hardly any) statistical power to detect a significant pattern in their variable of interest. In this four-day course, students, who must already be familiar with linear models and GLMs in R, will learn how to code and implement the next natural step: mixed models. Students will learn the difference between fixed effects and random effects, will learn how to ditch clunky transformations and instead embrace the native distribution of the data, and will be forced to think critically about experimental design. Towards the end of the course, we will also introduce Bayesian mixed models, which offer further flexibility in modelling. Students will also analyze their own data, so this course is particularly well suited to those with data in hand (datasets will also be provided for those without their own data). This course is practical, hands-on and workshop style with very little mathematics per se.

9:45-10:00 | Introductions & icebreaker

Lecture: Overview of linear models

Lecture: Overview of GLMs

Lecture: Distributions

Lunch

Afternoon: practical examples and working through GLM code

Lecture: Overview of mixed models

11:00 | Coffee break

11:15-11:45 | Q & A session as needed

11:45-12:30 | Work through first GLMM example

12:30-13:00 | Q & A on distributions and miscellaneous topics

Lunch

Go through annotated code

14:00-15:00 | Seed example + Q & A

15:15 | Break

15:30-16:30 | Hunting example + Q & A

16:30 | Break

16:35-17:45 | Count data example (Diogo Ferreira) + Q & A

9:45-10:30 | Group discussion: student example one

10:30-11:00 | Group discussion: student example two – define questions and set up global models on whiteboard

11:00 | Coffee break

11:15-13:00 | Break-out groups of three students: define questions and construct global models on paper

Lunch

14:00-14:30 | Short lecture: additional considerations when modelling

14:30-17:45 | Students work through their own data

Students work through their own data

4 days - 24 hours - Schedule: 09:30 - 12:30 & 14:00 - 17:00

Coordination

Luke L. Powell | BIOPOLIS-CIBIO

Instructors

Carolina Ocampo-Ariza

Diogo Ferreira | BIOPOLIS-CIBIO

Panagiotis Nikolaou

Tiago Silva

This course builds on familiarity with linear models and a basic working knowledge of generalized linear models (GLMs) in R. Participants should be comfortable with the main concepts of linear modeling. If you’re unsure whether the course is the right fit for your experience, please feel free to email the instructor.

The course will be open to a maximum number of 20 participants.

75% of available student slots are reserved for BIODIV students.

Priority will be given to:

• 1st year and other PhD students attending the BIODIV Doctoral Program;

• PhD students attending other courses;

• Other post-graduate students and researchers.

(A minimum number of participants is required for the course to take place)

To apply fill the FORM

A copy of your submitted form will be sent to the email address you provided. This email is your confirmation that the application was successfully submitted.

If you don't receive this email, your application was not successfully submitted.

We will email you the results within 10 working days after the application deadline.

80€ (students) | 150€ (other participants).

BIOPOLIS/ CIBIO members will have an additional discount of 20%.

Participation is free of charge for BIODIV Students (M BGE & PhD) & CIBIO's TwinLabs members.

(If applicable, payment will be made to Associação BIOPOLIS - VAT No. 516033727. Information on the payment amount and the payment deadline will be sent together with the results notification, and a proforma invoice may be issued in due time).

Registration fees do not include accommodation or meals.

The selected BIODIV (FCUP) students will have their travel (from Porto) and accommodation expenses covered.

No ECTS credits are awarded for attending the course. Participants receive a certificate of attendance without quantitative evaluation.

Education & Training Unit (Post graduation office)

+351 252 660 425