Exercises

These exercises are intended to get you familiar with the concepts underlying the various statistical approaches we introduced in the book and build your skills in taking a design, interrogating the data, and drawing some conclusions.

You won’t find standard revisionary questions, such as “What are the assumptions underlying x?”, or “Why would you use method x over method y?”. These exercises will build your understanding of model structure, assessing assumptions, making decisions about how models fit and how to describe effects, but we think it better when you need to answer these questions while looking at data used to answer biological questions.

You will find more of the kinds of worked examples we used throughout the book - real world data analysis situations, taken from published papers. We’ll give you a bit of context for the data, and then ask you a series of questions that will result in your translating the question to a statistical model, identifying the best way to fit that model, assessing assumptions, and describing the patterns that you infer from the data analysis.

For these data sets,

• You’ll find the full reference, including a link to the paper, at the end of each chapter’s exercises.
• We’ve used papers that are Open Access, as far as we know.
• The papers have been chosen because the authors followed the good practice of making the raw data available publicly. They are setting a great example.
• Getting the data.
  • In the ideal cases, you can just download the raw data and import it to R. For us, the ideal case means a csv file (or equivalent), with one row of headers, then the data. The variables have simple, short names, without spaces or other things that make us use quotes everywhere in our code.
  • In other cases, the files aren’t quite as straightforward, and we’ve given you a bit of R code to import the file as a well-behaved dataframe. That step might include skipping rows or columns, renaming variables, etc.
  • In some cases, we’ve found it simpler to create our own data file and give you a link to it.
• In the worked examples, we showed you an approach that matched what the original paper’s authors did. We did this to allow you to see how the authors put their data analysis into a larger story. In the exercises, we haven’t been constrained that way - we’ve seen opportunities to make the analysis simpler or more complex, we’ve asked you to use more than one approach to a data set, and so on. As a result, you shouldn’t worry about your results exactly matching what’s in a paper.

R-wrangling

We’ve not provided the code for these exercises, other than to help you with getting hold of the data. The code you’ll want will depend on the solutions you identify. However,…

• The worked examples provide a “bank” of code for particular tasks, and you may well have identified ways to improve or streamline that code

• Instructors have access to suggested solutions to the exercises, which include the code. If you’re working through these exercises in a way linked to a class, ask your instructor about code and solutions. We’ve run those solutions in R 4.3.1.

• We’ve often kept a bit of code, particularly when opening a file we’ve produced locally or to tidy up a data set, and we’ll show you the first few lines of the file as an illustration.

The exercises are organized by chapter.

• Chapter 1. Relax and enjoy the scenery. This chapter sets the scene for the rest.

• Chapter 2. This chapter is revisionary, and we’d use exams to check students’ understanding. We tend to make these questions a gentle start to the inquisition.

• Chapter 3. This chapter covers some basic design principles, which we pick up in some other exercises and assessment exercises.

• Chapter 4

• Chapter 5

• Chapter 6

• Chapter 7

• Chapter 8

• Chapter 9

• Chapter 10

• Chapter 11

• Chapter 12

• Chapter 13

• Chapter 14. This chapter introduces multivariate analyses, which are expanded in the following two chapters.

• Chapter 15

• Chapter 16

• Chapter 17

Group Exercises

• Specifying linear models

• Critiquing published data analyses

• See also Chapter 17