Emphasizing the inductive nature of statistical thinking, Environmental and Ecological Statistics with R connects applied statistics to the environmental and ecological fields. It follows the general approach to solving a statistical modeling problem, covering model specification, parameter estimation, and model evaluation. The author uses many examples to illustrate the statistical models and presents R implementations of the models.
The book first builds a foundation for conducting a simple data analysis task, such as exploratory data analysis and fitting linear regression models. It then focuses on statistical modeling, including linear and nonlinear models, classification and regression tree, and the generalized linear model. The text also discusses the use of simulation for model checking, provides tools for a critical assessment of the developed model, and explores multilevel regression models, which are a class of models that can have a broad impact in environmental and ecological data analysis.
Based on courses taught by the author at Duke University, this book focuses on statistical modeling and data analysis for environmental and ecological problems. By guiding readers through the processes of scientific problem solving and statistical model development, it eases the transition from scientific hypothesis to statistical model.
The Everglades Example
What Is R?
Getting Started with R
The R Commander
The Normality Assumption
The Independence Assumption
The Constant Variance Assumption
Exploratory Data Analysis
From Graphs to Statistical Thinking
Estimation of Population Mean and Confidence Interval
A General Procedure
Nonparametric Methods for Hypothesis Testing
Significance Level alpha, Power 1 - beta, and p-Value
One-Way Analysis of Variance
ANOVA as a Linear Model
Simple and Multiple Linear Regression Models
General Considerations in Building a Predictive Model
Uncertainty in Model Predictions
Smoothing and Additive Models
Classification and Regression Tree
The Willamette River Example
Generalized Linear Model
Seed Predation by Rodents: A Second Example of Logistic Regression
Poisson Regression Model
Generalized Additive Models
ADVANCED STATISTICAL MODELING
Simulation for Model Checking and Statistical Inference
Summarizing Linear and Nonlinear Regression Using Simulation
Simulation Based on Resampling
Multilevel Structure and Exchangeability
Multilevel Linear Regression
Generalized Multilevel Models
Song S. Qian is an associate research professor in the Nicholas School of the Environment at Duke University. Dr. Qian’s research consists of adaptive management strategies for watershed TMDL, GIS-assisted watershed modeling, water quality assessments, modeling marine mammal habitats, environmental sampling design, and more.
My favorite feature of this book is the large number of datasets (available on Qian’s web page) … Overall, I liked the book. I expect that I will be pulling examples from it when I teach methods courses to students in the sciences. … it does contain many interesting and intriguing examples, and good examples of R code. So I can and do recommend it as a helpful resource …
—Jane L. Harvill, The American Statistician, November 2011
Qian effectively blends fundamentals of scientific methods with statistical thinking, modeling, computing, and inference. … the text is well formatted with liberal use of illustrative portions of R code … It is clear that Qian has taken great care in developing this book and has succeeded in meeting his stated purpose. The book reflects Qian’s insights into teaching environmental and ecological modeling developed over many years in applied statistics and as an educator in applied sciences. …
—Biometrics, June 2011
his book gives a data-oriented introduction to statistical modeling of environmental and ecological phenomena. It is a beautiful scientific guideline for a computer-based model building and evaluation process. … This introductory book gives a diversified overview of modern applied statistics while always following an inductive, data-based approach. Numerous data sets and R scripts, all available online, help to understand even the subtle differences between various models. Here, the reader profits from the obvious practical experience of the author. Meaningful graphics and R code/output embedded in the text support the conclusions drawn and facilitate the application to own data sets. … Students and researchers of environmental sciences with basic knowledge in statistics will find this book valuable as both a work of reference and an introductory guide to statistical modeling with R.
—Sebastian Engelke and Martin Schlather, Biometrical Journal, 2011