Statistical Modeling for Conservation

Target Audience:  

Biologists who are responsible for monitoring or conservation design projects and other population-habitat related studies.  

Summary and Objectives:  

Statistical models are used to understand the mechanisms behind the patterns we see in nature. Statistical models are particularly good at this task because they embrace uncertainty in the environment, which provides a range of possibilities for conservation decision making. This course focuses on modeling trends and population-habitat relationships using "R" software. Emphasis is placed on model development, interpretation, understanding assumptions and evaluation of competing models. Statistical models explored in this class include simple and multiple linear regression, linear mixed models, generalized linear models such as Logistic, Poisson, and Negative Binomial regression and occupancy modeling. Other topics include variable selection and screening, model comparison techniques using AIC, BIC and cross-validation. Students are encouraged to bring a project or dataset to class for one-on-one consultation and for examples that may be integrated into the class. This coursework prepares students for CSP4220 Species Distribution Modeling for Conservation and CSP4230 Design and Analysis of Biological Monitoring.  

Upon completion of this course, participants will be able to:  

• Develop simple linear and multiple linear regression models to examine relationships and predict abundance, density, etc.  
• Use generalized linear models (e.g., logistic regression) to develop predictive species distributions.  
• Use linear and generalized linear mixed models to evaluate trends in monitoring studies.  
• Perform variable selection and screening, and compare models using AIC, BIC, and cross-validation techniques.  
• Use occupancy modeling for understanding changes in the proportion of sites occupied by a species when there is imperfect detectability. 
• Evaluate multicollinearity and diagnose outliers. 
• Gain exposure to integrating statistical models with GIS technology for the purposes of creating resource selection/species distribution maps.  

Prerequisite:  

Making Sense of Biological Data with R or equivalent experience with R and background in statistical concepts. Consult with the course leader with requests for bypassing course prerequisites. 

Competency Addressed:  

Data Management- Awareness, Modeling and Simulation - Basic, Research and Statistics - Basic, Data Interpretation - Basic