Ecology 145—Lecture Outlines

Lecture 1 (Wednesday, January 11): Course overview, types of probability models, probability basics

Lecture 2 (Friday, January 13): Bernoulli distribution, binomial distribution, Poisson distribution

Lecture 3–R Lab Session 1 (Tuesday, January 17): Introduction to R, plots and linear regression in R, residual analysis

Lecture 4 (Wednesday, January 18): Probability mass function for the Poisson distribution, negative binomial distribution

Lecture 5 (Friday, January 20): Negative binomial distribution in ecology, dispersion, link between Poisson and negative binomial, nonhomogeneous Poisson process

Lecture 6 (Monday, January 23): Negative binomial distribution as a model of heterogeneity, gamma distribution

Lecture 7 –R Lab Session 2 (Tuesday, January 24): Modeling mean-variance relations in data, probability functions in R, the influence of θ and μ on the negative binomial probability mass function

Lecture 8 (Wednesday, January 25): Wrap-up on distributions, additional motivations for the negative binomial, gamma and lognormal distributions

Lecture 9 (Friday, January 27): Parameters and statistics, properties of estimators

Lecture 10 (Monday, January 30): Maximum likelihood estimation

Lecture 11-R Lab Session 3 (Tuesday, January 31): Maximum likelihood estimation using R—loglikelihoods with one parameter

Lecture 12 (Wednesday, February 1): Properties of maximum likelihood estimators

Lecture 13 (Friday, February 3): Information as curvature, likelihood ratio test, Wald test

Lecture 14 (Monday, February 6): G² and Χ² goodness of fit tests

Lecture 15-R Lab Session 4 (Tuesday, February 7): Maximum likelihood estimation using R—loglikelihoods with two parameters

Lecture 16 (Wednesday, February 8): Information-theoretic methods for model selection. Introduction to AIC.

Lecture 17 (Friday, February 10): Using AIC to select models--the Burnham & Anderson protocol

Lecture 18 (Monday, February 13): AIC for transformed response models, zero-inflated models and conditional models for data with extra zeros

Lecture 19-R Lab Session 5 (Tuesday, February 14): Using AIC to screen models using the Burnham & Anderson protocol. Comparing Poisson, negative binomial (HW), log-transformed, and ZIP models for count data.

Lecture 20 (Wednesday, February 15): Introduction to generalized linear models (GLIMs)

Lecture 21 (Friday, February 17): Introduction to generalized linear models (GLIMs)—Part 2

Lecture 22 (Monday, February 20): Further details on generalized linear models—other link functions and the deviance

Lecture 23-R Lab Session 6 (Tuesday, February 21): Generalized linear models in R. Revisting the Crawley slug data set. Plant species-area curves and the Galapagos Islands

Lecture 24 (Wednesday, February 22): Offsets and residuals in generalized linear models. The systematic component and the use of regression models in science

Lecture 25 (Friday, February 24): Uses of regression models. Incorporating categorical variables in multiple regression models

Lecture 26 (Monday, February 27): The coding of categorical variables in regression models

Lecture 27-R Lab Session 7 (Tuesday, February 28): Negative binomial regression, residual analysis for assessing model form, spatial correlation in regression residuals

Lecture 28 (Wednesday, March 1): Categorical variables with three levels. Analysis of covariance.

Lecture 29 (Friday, March 3): Analysis of covariance example—energy expenditure by volant vertebrates

Lecture 30 (Monday, March 6): Testing linear combinations of contrasts. Orthogonal contrasts

Lecture 31-R Lab Session 8 (Tuesday, March 7): Analysis of covariance. Categorical variables in regression.

Lecture 32 (Wednesday, March 8): Contrasting the "new statistics" with the "classic statistics"

Lecture 33 (Friday, March 10): Contrasting the "new statistics" with the "classic statistics" (continued)

Lecture 34 (Monday, March 20): Interpreting logistic regression coefficients, grouped versus ungrouped binary data

Lecture 35-R Lab Session 9 (Tuesday, March 21): Fitting logistic regression models in R; goodness of fit tests

Lecture 36 (Wednesday, March 22): Goodness of fit tests for logistic regression; confusion matrix

Lecture 37 (Friday, March 24): Sensitivity, specificity, and ROC curves as tools for model calilbration

Lecture 38 (Monday, March 27): Recognizing structure in data sets

Lecture 39-R Lab Session 10 (Tuesday, March 28): Fitting habitat suitability models, ROC curves, and cross-validation

Lecture 40 (Wednesday, March 29): Introduction to the multilevel model

Lecture 41 (Friday, March 31): Mulitilevel model terminology; modeling heterogeneity also models the correlation

Lecture 42 (Monday, April 3): Strategies for model building with multilevel models

Lecture 43-R Lab Session 11 (Tuesday, April 4): The nlme package and multilevel models in R

Lecture 44-R Lab Session 11b (Wednesday, April 5): Adding a level-2 predictor to a multilevel model

Lecture 45 (Friday, April 7): Subject-specific and population-averaged models. Converting between multilevel, composite, and R notation for mixed effect models

Lecture 46 (Monday, April 10): Matrix formulation of the multilevel model, correlation structures, methods of estimation

Lecture 47-R Lab Session 12 (Tuesday, April 11): Diagnostic plots of level-1 and level-2 residuals. Specifying a residual correlation structure.

Lecture 48-R Lab Session 12b (Wednesday, April 12): Multilevel models with multiple level-1 predictors. Multilevel models with three levels.

Lecture 49 (Monday, April 17): Alternatives to mixed models for structured data sets—generalized least squares and generalized estimating equations

Lecture 50-R Lab Session 13 (Tuesday, April 18): Nonlinear mixed models—fitting a Gompertz model to growth data

Lecture 51-R Lab Session 13b (Wednesday, April 19): Assessing the fit of a nonlinear mixed model

Lecture 52 (Friday, April 21): Introduction to Bayesian analysis

Lecture 53 (Monday, April 24): Markov chain Monte Carlo

Lecture 54-R Lab Session 14 (Tuesday, April 25): Bayesian modeling and the use of WinBUGS

Lecture 55-R Lab Session 14b (Wednesday, April 26): Specifying a multilevel model in WinBUGS

Lecture 56 (Friday, April 28): Discussion of final exam and course evaluations

Course Home Page