What do you think?
Rate this book
Understanding Statistics: An Introduction
The modern world is brimming with statistical information--information relevant to our personal health and safety, the weather, or the robustness of the national or global economy, to name just a few examples. But don't statistics lie?
Well, no--people lie, and sometimes they use statistical language to do it. Knowing when you're being hoodwinked requires a degree of statistical literacy, but most people don't learn how to interpret statistical claims unless they take a formal course that trains them in the mathematical techniques of statistical analysis.
This book won't turn you into a statistician--that would require a much longer and more technical discussion--but it will give you the tools to understand statistical claims and avoid common pitfalls associated with translating statistical information from the language of mathematics to plain English.
Well, no--people lie, and sometimes they use statistical language to do it. Knowing when you're being hoodwinked requires a degree of statistical literacy, but most people don't learn how to interpret statistical claims unless they take a formal course that trains them in the mathematical techniques of statistical analysis.
This book won't turn you into a statistician--that would require a much longer and more technical discussion--but it will give you the tools to understand statistical claims and avoid common pitfalls associated with translating statistical information from the language of mathematics to plain English.
- GenresEconomics
154 pages, Paperback
First published December 5, 2017
12 people are currently reading
29 people want to read
Ratings & Reviews
Friends & Following
Create a free account to discover what your friends think of this book!
Community Reviews
Displaying 1 - 4 of 4 reviews
August 31, 2019
Extremely easy starter for non-math, non-techy reades.
January 3, 2020
4 out of 5 stars for 4 out of 5 solid chapters.
TL;DR: The closing chapter felt almost shoehorned into these briefs pages, making remarks about GDP and Gross Output that seemed detached from the rest of the book. The conclusion came back for a strong, high-note finish. I would grab the free ebook just for the reading recommendations list at the end of the book.
Overall, I enjoyed the book and it's example-heavy delivery of statical concepts with a very light touch on the math. (Seriously, the most we get is a linear equation of 8 variables, written out more as words than strict algebraic notation.) From the introduction, Davies makes the distinction between statistics and statistical analysis plain: "statistics" are the output of statistical analysis and "statistical analysis" is the mathematical process, techniques, and methods used to compute some meaning from data.
Chapter 1 primes the reader for how statistics are often misused and misinterpreted, covering a lot of ground and several definitions.
Chapter 2 educates on Probability, associated biases, probabilistic fallacies (e.g. "conjunctive"- or "joint probability" believability), and p-value. Conditional probability and joint probability are juxtaposed. A sound quip about our empathetic hearts not being statistically tuned gave me pause and a smile. There's also an entertaining analysis of simulacra of gender discrimination that should encourage the thoughtful.
Chapter 3 digs more into p-value, and where it is needed, as well as how to properly measure things for comparison. Statistically sound arguments are covered, and income inequality is touched on. There is a strong subtext of the fact that different aggregations of the same data can tell a different story, though the math never lies; people do. Causality is touched on here, but we dig more into it in the next chapter.
Chapter 4 covers the third-variable effect (though doesn't mention it by name, as it did in Chapter 1) and the usage of regression analysis, getting a touch "mathy" as mentioned previously. Surprisingly, Davies covers multiple regression which helps us articulate what has - and has not - been accounted for numerically (via the error term), so we don't falsely place blame.
He then outlines several ways the reader can do a bit of mental math to see if a statistical report actually trumps misinformation (i.e. if they did B, when A is correct while B is arbitrary and incorrect, call "bullshit"). We dig more into causality (with the unspoken "third-variable effect"), trends vs snapshots, and drive home the argument that one variable isn't enough to explain any one problem in our world. We also get a remark about controlled experiments and how to compare studies on people.
Chapter 5 is where things fall down a bit. As mentioned, the comments about GDP and GO seem misplaced and it was the second time that I thought the book was pushing an agenda, rather than just talking about statistics with examples supporting Libertarian views. Davies paints an interesting picture here, almost pitching the same ideas he's trained us against: using a single variable to explain a particular problem/solution. It is here that it felt he was pushing "the Libertarian agenda" of economic freedom being an express lane to prosperity (read: ending poverty). Thankfully, he produces enough counter examples to this being the case, though he doesn't say outright that economic freedom isn't guaranteed to make things better. His final example doesn't sit well with me; after multiple reviews, it feels hollow compared to the thorough cases before.
The book concludes with a succinct send-off, encouraging us to look past our anecdotes and appreciate the large-scale statistical truth, even when it doesn't tell us what we already believe. There are reading recommendations after the endnotes/footnotes that are absolutely fantastic: you can dig more into statical analysis for non-statisticians (and go further than the [perfunctory] economic freedom conversation started at the end of chapter 5), learn the foundations of statistical analysis for yourself, or explore regression analysis and more advanced statistical analysis. It's very much a "I've shared the ideas with you, now arm yourself with the knowledge and tools of the trade".
Very cool book and gives me hope for the rest of Liberationism.org's books
TL;DR: The closing chapter felt almost shoehorned into these briefs pages, making remarks about GDP and Gross Output that seemed detached from the rest of the book. The conclusion came back for a strong, high-note finish. I would grab the free ebook just for the reading recommendations list at the end of the book.
Overall, I enjoyed the book and it's example-heavy delivery of statical concepts with a very light touch on the math. (Seriously, the most we get is a linear equation of 8 variables, written out more as words than strict algebraic notation.) From the introduction, Davies makes the distinction between statistics and statistical analysis plain: "statistics" are the output of statistical analysis and "statistical analysis" is the mathematical process, techniques, and methods used to compute some meaning from data.
Chapter 1 primes the reader for how statistics are often misused and misinterpreted, covering a lot of ground and several definitions.
Chapter 2 educates on Probability, associated biases, probabilistic fallacies (e.g. "conjunctive"- or "joint probability" believability), and p-value. Conditional probability and joint probability are juxtaposed. A sound quip about our empathetic hearts not being statistically tuned gave me pause and a smile. There's also an entertaining analysis of simulacra of gender discrimination that should encourage the thoughtful.
Chapter 3 digs more into p-value, and where it is needed, as well as how to properly measure things for comparison. Statistically sound arguments are covered, and income inequality is touched on. There is a strong subtext of the fact that different aggregations of the same data can tell a different story, though the math never lies; people do. Causality is touched on here, but we dig more into it in the next chapter.
Chapter 4 covers the third-variable effect (though doesn't mention it by name, as it did in Chapter 1) and the usage of regression analysis, getting a touch "mathy" as mentioned previously. Surprisingly, Davies covers multiple regression which helps us articulate what has - and has not - been accounted for numerically (via the error term), so we don't falsely place blame.
He then outlines several ways the reader can do a bit of mental math to see if a statistical report actually trumps misinformation (i.e. if they did B, when A is correct while B is arbitrary and incorrect, call "bullshit"). We dig more into causality (with the unspoken "third-variable effect"), trends vs snapshots, and drive home the argument that one variable isn't enough to explain any one problem in our world. We also get a remark about controlled experiments and how to compare studies on people.
Chapter 5 is where things fall down a bit. As mentioned, the comments about GDP and GO seem misplaced and it was the second time that I thought the book was pushing an agenda, rather than just talking about statistics with examples supporting Libertarian views. Davies paints an interesting picture here, almost pitching the same ideas he's trained us against: using a single variable to explain a particular problem/solution. It is here that it felt he was pushing "the Libertarian agenda" of economic freedom being an express lane to prosperity (read: ending poverty). Thankfully, he produces enough counter examples to this being the case, though he doesn't say outright that economic freedom isn't guaranteed to make things better. His final example doesn't sit well with me; after multiple reviews, it feels hollow compared to the thorough cases before.
The book concludes with a succinct send-off, encouraging us to look past our anecdotes and appreciate the large-scale statistical truth, even when it doesn't tell us what we already believe. There are reading recommendations after the endnotes/footnotes that are absolutely fantastic: you can dig more into statical analysis for non-statisticians (and go further than the [perfunctory] economic freedom conversation started at the end of chapter 5), learn the foundations of statistical analysis for yourself, or explore regression analysis and more advanced statistical analysis. It's very much a "I've shared the ideas with you, now arm yourself with the knowledge and tools of the trade".
Very cool book and gives me hope for the rest of Liberationism.org's books
Read
July 21, 2025Beautifully simple refresher
September 26, 2022
A simple summary of some key points in statistics that would help the general public be better able to understand data...sort of ruined by the author's clear attempt to reinforce key libertarian themes strategically placed within (best government is least government, rich pay their fair share of taxes, no regulation leads to good, market-based outcomes, etc.).
I get it and am not surprised as this is published by the Cato Institute (which I really respect), but the examples of this type placed in the book cause me concern on the intent of publishing this book: is it to educate the public on statistics or push libertarian themes? Probably both.
I get it and am not surprised as this is published by the Cato Institute (which I really respect), but the examples of this type placed in the book cause me concern on the intent of publishing this book: is it to educate the public on statistics or push libertarian themes? Probably both.
Displaying 1 - 4 of 4 reviews