Statistics As Principled Argument

This book articulates critical use of statistics. It is extremely informative and evidently written by somebody who is well-experienced in this area. The narrative is devoid of formulas for easy reading, yet still requires the reader have a background in formal understanding of statistics, with focus on hypothesis testing and experimental design. It is well-framed around the title of the book. I thank the writer for sharing his expertise.

Abelson's Statistics as Principled Argument is a must read for advanced undergraduates or graduate students learning statistics.Abelson really explains why we do statistics the way we do, and how to construct and deconstruct a statistical argument. Abelson has a very lucid writing style which makes the book a very quick read.I recommend assigning this book in your statistics courses. It will really help new researchers understand how and why statistics is the language of science.

The most fundamental mistake that most students make in misconstruing statistics is thinking the subject laborious and too abstract for an analysis that most of us have "gut feelings" for in any case. Everyone knows, for example, that basketball players have hot streaks, that betting on black is safer after several rounds of a roulette ball landing on red, or that the mind possesses a mystical way of knowing the world (ESP). In Statistics as Principled Argument, Abelson does the reader a great favor by making the stakes of statistical literacy relevant by turning these intuitions on their head. Null hypothesis testing, in his words, are an extremely ritualized exercise of "playing the devil's advocate" against our intuitions, against the fundamental attribution error. The fact that he is able to make this exercise fun and meaningful for nearly 200 pages is a testimony to a rare author who has both a clear mastery of the subject and the ability to communicate that mastery to naïve and expert audiences. Upon completion of the book, it is hard not to like, because it is authoritative and comprehensive enough, and at the same time, approachable, entertaining and even light-hearted. In this reviewer's opinion, such a combination makes this the perfect text (or at least a well-placed additional text) for teaching courses in statistical analysis, particularly in psychology departments, a topic from which Abelson concentrates on throughout the text. The greatest strength of the text is Abelson's talent to simplify the process of making statistical arguments. He prescribes a system that bins results into ticks, significant results supporting a theoretical argument, and buts, statistically significant exceptions to these arguments. He uses this system to paint a thought-provoking picture of how fields are advance knowledge systematically. He devotes considerable attention to importance of parsimony in this framework as well, a term he systematized by suggesting that claims with the fewest numbers of ticks and buts are, by definition, the most parsimonious claims (and the best kinds of claims). Abelson, echoing long held sentiments in the field of null-hypothesis testing, also attacks the largely derided .05 level of significance as beautified with false saintliness. His suggestions, taken from his mentor John Tukey, prescribe using less categorical or more flexible conclusions from null-hypothesis testing. For example findings that may "lean", or "hint" at p =.06 may be useful when subject to cumulative replication, a process which is the ultimate rectifier of misleading analyses in any case, according to Abelson. There are only two areas where Ableson's analysis may leave some readers slightly disconcerted: MAGIC and mathematical formalism. MAGIC (magnitude, articulation, generalization, interestingness, and credibility) is Abelson's way of systematizing our folk-intuitions about the importance or worthwhileness of a particular claim. It is unclear what version of the universe requires that we systematize principles of "interestingness" precisely because such judgments are often made post-hoc. Here, Abelson may be falling victim to the very fundamental attribution error he so vigilantly crusades against at the onset of the book by expounding on and formalizing what has historically made findings interesting to him. There are several cases where findings, which at first seemed marginal (and which would certainly fail MAGIC standards if the authors were choosing their projects by this system) came to re-define fields in unexpected ways. In no way did an environmental biologist studying algal proteins expect to revolutionize the field of neuroscience by discovering optical tools for neuroscientists, as an example. As the ecosystem of scientific findings expands exponentially, increasingly, this complexity will likely defy the author's intuitions about what makes a finding "MAGICal". Such systems, by design, exclude the sorts of intellectual peregrinations that are yielding increasing and unexpected rewards in an era of high-throughput science. The last failing of the text, if it can be called that, is that it is very sparse on both graphs and mathematical and theoretical formalism. This is not to say the author doesn't deal with these subjects - Abelson's review of multiple comparisons, for example, is thorough and helpful. As a stand-alone text however, this makes the prescription for this book as a standard textbook extremely challenging without a more formal and dense theoretical companion volume for those more mathematically inclined. In total, Abelson's book accomplishes the aim of making statistics relevant for researchers and makes this reviewer wish that Statistics as Principled Argument had been recommended before a quantitative analysis class or at least at the beginning of one. Since making stats relevant and interesting is what commends the book so strongly, its best deployment would either be in concert with or preceding more formal analysis. Very few books succeed in conveying a sense of fun and excitement about the topic of statistics, and fewer succeed in in planting conviction for the importance of principled argument and critical thinking, but hats off to Ableson for succeeding with both at the same time. Overall, I would strongly recommend this book to lay person and expert alike.

This book finally made me understand the core of the statistics, why it appears to be so confusing for people coming from science and how it can be dealt with.

Few works emphasize argumentation on top of calculation. Numbers are nice, but they are truly meaningful when a context is explained, and this book has achieved the lot as concerns for statistics. From the beginning is explained that the word "statistics" from data banking about real state, but having data is the beginning and not the goal. The goal is achieved when such data is analyzed, criticized, and explored within the context it was taken from.

Abelson is experienced, thoughtful, and practical, with an entertaining writing style. A great deal of wisdom is distilled into this little book.

This book may be all the positive reviewers say it is, but its design is atrocious. The type is far too small, the lines way too long for the point size, and the lines packed too tightly for the line width.I'm a book designer, so I know whereof I speak. This book violates every principle of good layout and design of which I'm aware.I started reading, and it's going to be a long, hard slog. It's a challenging topic for me: I don't need to have the process of reading be so painful on top of that.

Item arrived as described, in a timely manner, no issues.

Brilliant book on the use of stats. For non quant researchers it gets a bit heavy in places, but if you are a qual researcher like me, it is still an excellent tool for helping you critique quant papers and any statistical analyses or evaluations that you may come across. I read it in conjunction with a few other books on the nuts and bolts of statistical analyses like Naked Statistics, Thinking Statistically, and Tiger That Isn't, so I do now feel fully confident about unpicking statistical assertions in policy evaluations.

It's not a new book so it's not surprising that here and there it seems a bit dated. But I'm glad I bought this book simply for Abelson's very useful and insightful discussion of how to interpret anova results.

As a doctoral student who wanted to start learning statistical methods this book is very helpful. I have an education in design, and last time I studied something math-related was in high-school!The book is about concepts, reasoning and argumentations, which are the basis of good research, either you use quantitative or qualitative data. The book is not dogmatic and pedantic as one might fear approaching such subject, and it is pleasantly and clearly written!

An engaging book about to statistics can be used to support an argument. There are explanations how statistics can be misused to help the reader recognize such misuse and hopefully avoid it in their own data analysis projects. Too bad this book is not required as a complement to traditional first year statistics courses; it would give context to the formulaic focused introduction to statistics.