Statistical Modeling for Ecology and Conservation
A five-day intensive workshop held immediately before fall term. Builds the probability, GLM, mixed-model, maximum-likelihood, and Bayesian-modeling foundation that ecologists and conservation scientists need to analyze their own data.
Course information
| Course name | Statistical Modeling for Ecology and Conservation |
|---|---|
| Course number | FW 536 |
| Credits | 3 |
| Meeting times | Five days prior to start of fall term classes (typically Friday through Tuesday, including the weekend). 9 am – 12 pm and 1 pm – 5 pm, with a midday break for lunch. |
| Prerequisites | None. Familiarity with the R programming language and prior introductory statistics courses are strongly recommended. |
| Instructor | Taal Levi, Department of Fisheries, Wildlife, and Conservation Sciences |
Catalog course description
Explores statistical modeling and practical analysis of data typical of ecology, conservation, and natural resources. Facilitates deep interpretation and application by covering probability, continuous and discrete probability distributions and their use for ecology, generalized linear models and their interpretation, mixed effects models for ecological data, model selection, and applications of maximum likelihood and Bayesian statistics for common hierarchical modeling problems in ecology. Includes hands-on exercises in the R programming language and the BUGS language implemented in Nimble and JAGS.
Daily schedule
Each day has a morning lecture + morning lab and an afternoon lecture + afternoon lab. The morning lab is the assignment that goes with the morning lecture, and the afternoon lab is similar. Each lab is worth 20 points.
| Day | Session | Lecture content | Lab activity |
|---|---|---|---|
| Day 1 | Morning | Probability theory; random variables; probability density functions; cumulative distribution functions; expectation; variance. | Probability problem set (Day 1 morning lab — 20 pts) Pre-course exam at start of day |
| Afternoon | Discrete distributions: Bernoulli, Binomial, Poisson, Negative Binomial; their use for ecological data. | Discrete distributions problem set (Day 1 afternoon lab — 20 pts) | |
| Day 2 | Morning | Continuous distributions: Normal, t, Exponential, Uniform, Lognormal, Gamma, Beta; their use for ecological data. | Continuous distributions problem set (Day 2 morning lab — 20 pts) |
| Afternoon | Linear models and Generalized Linear Models I — using and interpreting GLMs for ecological problems; the log and logit links as biology. | Linear models problem set (Day 2 afternoon lab — 20 pts) | |
| Day 3 | Morning | Generalized Linear Models II / Hierarchical models — mixed effects models for ecological data. | Hierarchical modeling problem set (Day 3 morning lab — 20 pts) |
| Afternoon | Hypothesis tests and model selection in ecology — LRT, AIC, AICc, BIC, cross-validation. | Model selection problem set (Day 3 afternoon lab — 20 pts) | |
| Day 4 | Morning | Maximum likelihood for nonlinear models in ecology — grid search, numerical optimization, likelihood profiles. | Maximum likelihood problem set (Day 4 morning lab — 20 pts) |
| Afternoon | Introduction to Bayesian inference — likelihoods, priors, posteriors, conjugacy, MCMC; introduction to JAGS and Nimble. | Bayesian modeling problem set 1 (Day 4 afternoon lab — 20 pts) | |
| Day 5 | Morning | Bayesian hierarchical modeling — priors, random effects structures, latent variables, occupancy, N-mixture. | Bayesian modeling problem set 2 (Day 5 morning lab — 20 pts) |
| Afternoon | Bayesian hierarchical modeling, continued; wrap-up and pathways for continued learning. | Open lab — finish the Bayesian problem set and catch up on any earlier problem sets |
Course-specific measurable student learning outcomes
By the end of this course, students will be able to:
- Determine the appropriate probability distributions for distinct data types.
- Interpret and implement generalized linear models (GLMs) to analyze ecological and conservation data.
- Construct and evaluate generalized linear mixed-effects models (GLMMs) to address hierarchical ecological data.
- Implement model selection appropriately to evaluate multiple hypotheses for ecological data.
- Construct and evaluate nonlinear and hierarchical models common in the ecological and wildlife sciences in both maximum likelihood and Bayesian frameworks using JAGS and Nimble.
Evaluation of student performance
Your course grade is based on a pre-course exam and nine problem-set assignments (one per morning or afternoon session except Day 5 afternoon, which is open lab time to finish any unfinished problem sets).
| Component | Each | Count | Total |
|---|---|---|---|
| Pre-course exam | 50 pts | 1 | 50 pts |
| Problem sets (one per morning/afternoon lab) | 20 pts | 9 | 180 pts |
| Course total | 230 pts |
Grading scale
| A | A− | B+ | B | B− | C+ | C | C− | D+ | D | D− | F |
|---|---|---|---|---|---|---|---|---|---|---|---|
| ≥ 92 | 90–91.9 | 88–89.9 | 82–87.9 | 80–81.9 | 78–79.9 | 72–77.9 | 70–71.9 | 68–69.9 | 62–67.9 | 60–61.9 | < 60 |
Grades are based on your percentage of the 230 points above. Each band runs from its lower number up to (but not including) the next band's — so exactly 92 is an A, and exactly 90 is an A−.
Submission format
All assignments are submitted as R Markdown HTML outputs uploaded to Canvas. You can embed photos of hand-written derivations or diagrams directly into the R Markdown — many probability problems are easier to do on paper, and a photo of clean work plus a typed interpretation is acceptable for full credit.
An R Markdown tutorial on how to install R Markdown, knit to HTML, and embed photos is the first technical activity of Day 1 — please complete the install steps before the workshop begins so we don't lose time on setup.
Per-problem-set workflow
- Open the problem set in your browser to read the questions.
- Open the matching R Markdown template (provided alongside each problem set in the day's folder).
- Add your code, prose, and (if needed) photos of written work.
- Knit to HTML.
- Upload the resulting
.htmlfile to Canvas by the deadline below.
Deadlines and late work
The short version
Each problem set is due before the next morning session. If you need more time, just ask — extra time is readily granted.
- A session's problem set (morning or afternoon) is due before the start of the next morning session. Both of Day 1's sets are therefore due before Day 2 begins, and so on. You have the evening to finish; nothing is due at lunch.
- Day 5's problem set, and anything still outstanding from earlier days, are due after the workshop ends — I will confirm the date in class. Day 5 afternoon is open lab time set aside specifically for catching up, and I will be there to help.
- Extensions are granted on request, and you do not need to justify the request. This is a one-week intensive; falling behind is normal and is not a character flaw. Email me, or just say so in class, and we will agree on a new date. There is no late penalty for work covered by an extension.
- Work submitted late without an extension is still accepted and still graded. Please talk to me if you find yourself in that position, so I can help rather than just mark it.
This policy is deliberately generous because the grade is based on effort rather than aptitude. A complete, genuine attempt handed in late is worth far more to you — and scores far better — than a rushed or empty one handed in on time.
For each lab session, you'll see two problem-set documents:
- One practice problem set with visible worked answers — this is what I will walk through during lab time.
- One assessed problem set — a single dataset-driven set of problems. Its answer key lives in a separate instructor-only file and is released after grading.
Pre-course exam
The course includes a short pre-course exam given at the start of Day 1 morning. It touches on all course content — probability, distributions, GLMs, mixed models, model selection, MLE, and Bayesian inference — but is meant as a diagnostic, not a test of what you already know.
Its purpose is to tell the instructor where the class is starting from, so the workshop can be paced and supplemented appropriately. You will not be penalized for low performance — its scoring is intentionally generous, because you are not expected to know all of the material going in. Come as you are; the pre-course exam simply sets a baseline.
Learning resources
Students must have a laptop with the following installed and working:
- R and RStudio.
- JAGS.
- A C++ compiler for Nimble — see installing Nimble.
Once JAGS is installed, install and load the R packages rjags, R2jags, and jagsUI to confirm that JAGS and R are interfacing. Install and load the nimble package and confirm Nimble is interfacing with R.
For R Markdown, install the rmarkdown and knitr packages (RStudio will prompt you the first time you create an Rmd file). Walking through the R Markdown tutorial is the first activity of Day 1.
Highly recommended (not required) books:
- Hobbs & Hooten (2015). Bayesian Models: A Statistical Primer for Ecologists. Princeton.
- Kéry & Schaub (2012). Bayesian Population Analysis using WinBUGS. Academic Press.
- Bolker (2008). Ecological Models and Data in R. Princeton.
- McElreath (2020). Statistical Rethinking. CRC Press.
Course policies
Artificial Intelligence (AI) Usage
The use of artificial intelligence is encouraged to help you learn. Tools such as large language models can be excellent tutors — use them to explain concepts, debug your R code, check your reasoning, and explore "what if" questions. The one expectation is that you use AI to understand the material, not to bypass understanding it: write up your own reasoning in your own words, and be ready to explain any answer you submit.
Academic calendar
All students are subject to the registration and refund deadlines as stated in the OSU Academic Calendar.
Statement regarding students with disabilities
Oregon State University is committed to making learning experiences accessible to all, which includes consideration of the course design, course learning outcomes, and barriers experienced by the disabled student. Disability Access Services (DAS) staff and students work together to determine reasonable accommodations. Students connected with DAS who have questions about their accommodations should reach out to their DAS Adviser. Students who have not yet completed the process for setting up accommodations should contact DAS at 541-737-4098, visit the DAS website, or email disability.services@oregonstate.edu to explore accommodations. Students and faculty members are encouraged to collaborate around the implementation of accommodations.
Student conduct expectations
Student Bill of Rights
OSU has twelve established student rights. They include due process in all university disciplinary processes, an equal opportunity to learn, and grading in accordance with the course syllabus: https://asosu.oregonstate.edu/advocacy/rights.
Reach out for success
University students encounter setbacks from time to time. If you encounter difficulties and need assistance, it's important to reach out. Consider discussing the situation with an instructor or academic advisor. Learn about resources that assist with wellness and academic success at oregonstate.edu/ReachOut. If you are in immediate crisis, please call or text the Suicide & Crisis Lifeline at 988.