Finding Stuff in This Book

0.9 The Effect Is Big

Copy link

As you may have already noticed, especially if you have the physical version, The Effect is a big fat book with a lot of stuff in it. And sure, there are chapter and section titles, and an index in the back, but it can still be tricky to find the topics you’re looking for, especially if you might know that topic by a different name.

Accordingly, a common experience for me is to hear from an instructor using the book that they sure wish I was able to include topic X or feature Y. But I did include topic X and feature Y! Just… maybe in a chapter they skipped over, or kept online.77 More than once, the request for topic X has been paired with a suggestion that chapter Z isn’t something they’d assign in their class and could be cut from the book… and chapter Z is the chapter with topic X! Of course, there are also plenty of topics that are actually absent. Can’t include everything! You won’t find guidance on qualitative research, or (aside from the power analysis stuff in Chapter 15) on running field experiments or randomized controlled experiments. Those things are great, but just not the focus of the book.

In this chapter I’ll point out the places you can find different topics or materials in a way that’s maybe more accessible than an index but more detailed than the chapter titles.

0.10 Additional Online Materials

Copy link

The first place I’ll start is actually outside of the book entirely. This book has a set of online materials available for instructors or anyone else who finds them useful. All the materials can be found linked on the first page of the book’s website at theeffectbook.net.

These online materials include:

• I want practice problems. There is a homework assignment available for each chapter, including both conceptual and technical questions that you can use for practice or discussion.
• I want more real-world research and more code examples. In addition to the conceptual materials, there is a second set of homework assignments centered around coding and technical application. Most of these assignments are constructed around walking through an actual published academic paper, allowing you (or your students) to apply the tools in real-world contexts.
• I want video material. There are more than 70 videos covering every chapter and multiple topics, all available for free and linked on the book website.
• I want material on introductory coding and data cleaning. The website contains introductions to both. Some of this material is by me, and some of it is linked out to others.
• I want animations. Animations demonstrating many of the causal inference methods covered in this book are available on my other website here.

0.11 Topics in the Book

Copy link

You can already see the topics listed in the chapter titles. I don’t need to tell you where to look to find information about difference-in-differences. It’s in the chapter called “Difference-in-differences.”88 OK, sometimes I do get people asking me where the material is on the newfangled crop of difference-in-differences estimators. It’s in there, I promise! This section is more about the stuff that is definitely in the book but perhaps in a less obvious way.

I want an introduction to basic statistical analysis. I can forgive you for not spotting this in the chapter list, since I gave them titles that err on the side of describing what you’re actually doing rather than matching standard terminology. But there are two whole chapters on this! Chapters 3 and 4 cover descriptive statistics for single variables and for two-variable relationships, respectively.

I want material about hypothesis testing and p-values. This topic gets covered in Chapter 3, with plenty more as it relates to linear regression in Chapter 13, especially the “Hypothesis testing in OLS” section (which also includes discussions of things like p-hacking and how to interpret significance).

I want material on robustness tests and evaluating model assumptions. This material is both spread out over the book and found concentrated in some key places. In basically every chapter in The Toolbox (Part II of the book), there are robustness tests and ways to evaluate model assumptions baked into the chapter as they become relevant. In a more concentrated form, you can find discussion of robustness tests more generally (including different approaches to testing robustness) in Chapters 11 and 21. Also check out Chapter 15 for ways of evaluating a method’s sensitivity to violations of its assumptions.

I want help reading an actual paper. This is a topic that’s really dispersed throughout the book. But “read the whole book” isn’t super helpful. A good place to start, though, is the walkthrough of how to read a regression table in Chapter 13.

I want material on machine learning. Machine learning isn’t really the central approach taken in this book.99 It’s not really the primary approach taken in most causal inference, at least not for now, so I hope you’ll forgive me. So it’s in here, but don’t expect this to be an in-depth machine learning source. Chapter 13 introduces regularized regression near the end. Many of the other toolbox chapters, for example Chapter 14, have hints on how machine learning can be incorporated into a given method. But the place you’ll find the most machine learning is in Chapter 22, which introduces the most broadly-applicable machine learning method for causal inference, double machine learning, also known as double debiasing, as well as some other machine learning-based methods for performing causal inference, like causal forests.

I want material to help put together my own research project, or critique an existing study. Your best bet here is the first half of the book! Rather than skipping to the Toolbox chapter with the method you think is appropriate, start instead with Chapters 1, 2, and 5 through 8.

I want more coding. The chapter that will best let you flex your coding muscles is Chapter 15.

I want Bayesian statistics or time-series analysis. These are probably the most frequent requests I get for things that I will definitely admit the book doesn’t cover in much depth. You’re probably best off looking elsewhere. For Bayesian statistics there’s a minor introduction to random effects and hierarchical modeling in Chapter 16. And while panel data (multiple time series all estimated together) stuff is well covered, tools for analysis of a single time series, like ARIMA modeling, gets only a brief introduction in Chapter 17.