Our book is available online! Just go to that page and scroll down to the “ROS online PDF” link. It’s better to link to the webpage than the pdf, so that you and others will get the latest corrections and see the extra online material.
Personally I prefer the paper copy as it’s easier to flip through and find things, but a pdf copy is helpful too, sometimes.
As a reminder, here’s the fun version of the table of contents:
• Part 1:
– Chapter 1: Prediction as a unifying theme in statistics and causal inference.
– Chapter 2: Data collection and visualization are important.
– Chapter 3: Here’s the math you actually need to know.
– Chapter 4: Time to unlearn what you thought you knew about statistics.
– Chapter 5: You don’t understand your model until you can simulate from it.
• Part 2:
– Chapter 6: Let’s think deeply about regression.
– Chapter 7: You can’t just do regression, you have to understand regression.
– Chapter 8: Least squares and all that.
– Chapter 9: Let’s be clear about our uncertainty and about our prior knowledge.
– Chapter 10: You don’t just fit models, you build models.
– Chapter 11: Can you convince me to trust your model?
– Chapter 12: Only fools work on the raw scale.
• Part 3:
– Chapter 13: Modeling probabilities.
– Chapter 14: Logistic regression pro tips.
– Chapter 15: Building models from the inside out.
• Part 4:
– Chapter 16: To understand the past, you must first know the future.
– Chapter 17: Enough about your data. Tell me about the population.
• Part 5:
– Chapter 18: How can flipping a coin help you estimate causal effects?
– Chapter 19: Using correlation and assumptions to infer causation.
– Chapter 20: Causal inference is just a kind of prediction.
– Chapter 21: More assumptions, more problems.
• Part 6:
– Chapter 22: Who’s got next?
• Appendixes:
– Appendix A: R quick start.
– Appendix B: These are our favorite workflow tips; what are yours?
And here’s the list of what you should be able to do after reading and working through the book:
• Part 1: Review key tools and concepts in mathematics, statistics, and computing.
– Chapter 1: Have a sense of the goals and challenges of regression.
– Chapter 2: Explore data and be aware of issues of measurement and adjustment.
– Chapter 3: Graph a straight line and know some basic mathematical tools and probability distributions.
– Chapter 4: Understand statistical estimation and uncertainty assessment, along with the problems of hypothesis testing in applied statistics.
– Chapter 5: Simulate probability models and uncertainty about inferences and predictions.
• Part 2: Build linear regression models, use them in real problems, and evaluate their assumptions and fit to data.
– Chapter 6: Distinguish between descriptive and causal interpretations of regression, understanding these in historical context.
– Chapter 7: Understand and work with simple linear regression with one predictor.
– Chapter 8: Gain a conceptual understanding of least squares fitting and be able to perform these fits on the computer.
– Chapter 9: Perform and understand probabilistic prediction and simple Bayesian information aggregation, and be introduced to prior distributions and Bayesian inference.
– Chapter 10: Build, fit, and understand linear models with multiple predictors.
– Chapter 11: Understand the relative importance of different assumptions of regression models and be able to check models and evaluate their fit to data.
– Chapter 12: Apply linear regression more effectively by transforming and combining predictors.
• Part 3: Build and work with logistic regression and generalized linear models.
– Chapter 13: Fit, understand, and display logistic regression models for binary data.
– Chapter 14: Build, understand, and evaluate logistic regressions with interactions and other complexities.
– Chapter 15: Fit, understand, and display generalized linear models, including the Poisson and negative binomial regression, ordered logistic regression, and other models.
• Part 4: Design studies and use data more effectively in applied settings.
– Chapter 16: Use probability theory and simulation to guide data-collection decisions, without falling into the trap of demanding unrealistic levels of certainty.
– Chapter 17: Use poststratification to generalize from sample to population, and use regression models to impute missing data.
• Part 5: Implement and understand basic statistical designs and analyses for causal inference.
– Chapter 18: Understand assumptions underlying causal inference with a focus on randomized experiments.
– Chapter 19: Perform causal inference in simple settings using regressions to estimate treatment effects and interactions.
– Chapter 20: Understand the challenges of causal inference from observational data and statistical tools for adjusting for differences between treatment and control groups.
– Chapter 21: Understand the assumptions underlying more advanced methods that use auxiliary variables or particular data structures to identify causal effects, and be able to fit these models to data.
• Part 6: Become aware of more advanced regression models.
– Chapter 22: Get a sense of the directions in which linear and generalized linear models can be extended to attack various classes of applied problems.
• Appendixes:
– Appendix A: Get started in the statistical software R, with a focus on data manipulation, statistical graphics, and fitting and using regressions.
– Appendix B: Become aware of some important ideas in regression workflow.
Enjoy! And, again, here’s the link.
Thank you for putting the content online for free!
Arne:
Thank you to our many sponsors who have supported this work over the past few decades!
Ditto. That is thoughtful gesture, Andrew.
Thank You.
I hope you consider featuring co-edited introductory statistics textbook.
Ah merci mille fois ! Thank You Andrew and all the people involved in the book and in this decision.
I buy a paper version of ROS (and a book of J. Pearl) on Amazon but the parcel never touch home and no-one know where it was delivered. I hope the god elected person enjoy the two book …
This is awesome! Thank you so much for putting this online free! I never want to publish behind a paywall but I often feel forced into it by journals/book publishers. If you don’t mind answering, I’d be grateful to know how you got permission from Cambridge University Press to post it freely? Is this particular publisher typically okay with that, or did you need to negotiate? If so, any advice on how to do that successfully?
Thank you again so much!!!
This is great. I already own the paper copy, but I have recommended it to people and having a free digital version is really handy. May I ask when you currently expect your multilevel textbook to be released?
+10
Andrew:
I’m curious about why it took CUP so long, given that It has been available for free since 2020 — along with most other published books: **
John
John:
I have no idea what that site is, but I doubt it has the latest version of our books. I recommend you Regression and Other Stories from here and Bayesian Data Analysis from here.
Thanks Andrew!
The approach you use in the book is very refreshing and I’ll be sure to link this pdf to friends and colleagues.
I’m from economics and I found myself frequently getting upset that students think that econometrics is what is presented in Mostly Harmless Econometrics. While it’s a fine book in itself, it represents only a very, very small subset of what we should think about when discussing statistics and Econometrics (it doesn’t help that the authors think their book is the right way to teach Econometrics from the scratch…)
Would be really curious to hear what you think the pros and cons of MHE are. I was handed MHE once (by an economist, of course) and I skimmed through it — some useful stuff on causal inference, but I admit I didn’t go through the whole thing in detail, although it’s on my to read list.
Hey AnonStat,
While I definitely don’t think MHE is a bad book per se, I do have a few problems with the way people recommend it. My main criticism is similar to what Francis Diebold (an influential economist working on time series) has said.
” There’s a lot to like about MHE. It’s an insightful and fun treatment of micro-econometric regression-based causal effect estimation — basically how to (try to) tease causal information from least-squares regressions fit to observational micro data. It’s filled with wisdom, exploring many subtleties and nuances. In many ways it’s written not for students at age 23, but rather for seasoned researchers at age 53. And it tells its story in a marvelously engaging conversational style. ”
“But there’s also a lot not to like about MHE. The problem isn’t what it includes, but rather what it excludes. Starting with its title and continuing throughout, MHE promotes its corner of applied econometrics as all of applied econometrics, or at least all of the “mostly harmless” part (whatever that means). Hence it effectively condemns much of the rest as “harmful,” and sentences it to death by neglect. It gives the silent treatment, for example, to anything structural — whether micro-econometric or macro-econometric — and anything involving time series. And in the rare instances when silence is briefly broken, we’re treated to gems like “serial correlation [until recently was] Somebody Else’s Problem, specifically the unfortunate souls who make their living out of time series data (macroeconomists, for example)” (pp. 315-316).”
And to summarize I highly agree with this last part also:
“All told, Mostly Harmless Econometrics: An Empiricist’s Companion is neither “mostly harmless” nor an “empiricist’s companion.” Rather, it’s a companion for a highly-specialized group of applied non-structural micro-econometricians hoping to estimate causal effects using non-experimental data and largely-static, linear, regression-based methods. It’s a novel treatment of that sub-sub-sub-area of applied econometrics, but pretending to be anything more is most definitely harmful, particularly to students, who have no way to recognize the charade as a charade.”
(link here: https://fxdiebold.blogspot.com/2015/01/mostly-harmless-econometrics.html)
If one is interested in causal inference I think other books would be better suitable. There are the Pearl books for example. This one, recently published and free online, seems to be a very good addition to causal inference books (https://theeffectbook.net/)
Conditional:
I liked Mostly Harmless Economics; here’s my review. I agree with Diebold on one point, though; as I wrote in the very second paragraph of my review,
I continued, though, with some positive words:
So I think we should forgive Mostly Harmless Economics its misleading title and celebrate its positive content.
I also clicked on the link at the end of your post. I agree that the book there looks good. In its style, it reminds me a lot of Mostly Harmless Econometrics. I have mixed feelings about the content, but that would be true of any book on this topic. Compared to Regression and Other Stories and Mostly Harmless Econometrics, this new book is focused on design and concept, not analysis, and that’s a good thing. I’d say the closest existing book is the recent methods text by Ethan Bueno de Mesquita and Anthony Fowler.
> recent methods text by Ethan Bueno de Mesquita and Anthony Fowler.
Interesting! Thanks for the reference.
> forgive Mostly Harmless Economics its misleading title
It’s hard to take it so literally as to be misled though!
I have the book on my desk (after reading it has retired to being part of my laptop stand) and I see examples on impact of education on earnings, the draft on earnings, and some sort of election thing.
Income! Education! Wars! Elections! Big stuff. No way you can actually do econometrics and not accidentally involve yourself in some harm — it’s sorta baked into the cake.
If the authors are super pushing that their book is the Only Way then I guess that’s probably not right, but I didn’t pick that up from reading the book. What I got was, we’re gonna talk you through three methods for causal inference that we like — here goes!
Andrew:
Thanks for the suggestion of this book from Ethan Bueno de Mesquita and Anthony Fowler.
Admittedly, my criticism of MHE may sound a bit harsh because I’m in fact taking other factors into account that are not necessarily the “book’s fault”. More precisely, I think a lot of students (at least from what I’ve seen in some departments) are being exposed maybe *too early* (and perhaps too strongly to the book). A lot of professors talk a lot about the book in a first econometrics course, and I think that leaves a very limited impression of what types of problems we encounter in econometrics and statistics in general.
All in all, it would be unfair to incorporate this into my criticism because, well, the book is just a book and I’m taking into account the behavior of instructors. However, the authors themselves seem to actively believe that econometrics should be taught according to the approach in their book, which I strongly disagree with.
(I’m again citing partially Diebold here: https://fxdiebold.blogspot.com/2017/02/econometrics-angrist-and-pischke-are-at.html ;
and this is the paper he is referencing: https://www.nber.org/papers/w23144?utm_campaign=ntw&utm_medium=email&utm_source=ntw)
I’m probably not alone in willingness to pay for part two of RAOS, hopefully using brms for examples… Is there a sense of timeline as to when part two will hit the bookshelves?