Customer Reviews

Stat-Spotting: A Field Guide to Identifying Dubious Data

5 star:		(3)
4 star:		(3)
3 star:		(0)
2 star:		(0)
1 star:		(0)

 

 

There are many books written on the interpretation of statistical information for the non-specialist. Some are quite engaging and difficult to put down; others are quite boring. I would place this little book roughly in the middle of that range - it's not the most gripping but it is certainly an eye-opener for those seeking guidance in this field. The author, a sociology professor, has divided the book such that each chapter addresses a type of question that a reader should ask when reviewing statistical information that may seem dubious or misleading. Each chapter opens with a brief digression on the issue at hand. This is followed by a practical example, i.e., a statistical "fact" that has been published in some medium and which the author analyzes in order to put this "fact" into perspective and in a manner that the reader can more easily interpret. The writing style is clear, authoritative and accessible to a wide audience. People seeking to make sense of the statistical data that floods our lives every day would benefit greatly from reading this concise little book.

One of the most important things one can do in this time of too much information is to be able to evaluate the information being presented to you. Stat-Spotting delivers by giving you tools and yardsticks to use in evaluating statistical and numerical information presented in the plethora of sources available to someone researching a topic.

This is a book that I suggest everyone read...

...but it's too slight. With some judicious editing, it's a journal article. But this is a case where the complaint is not that the food was bad and the portions too small: In fact, the food was so great, one wishes it were all-you-can-eat.

And then I'm left wondering if perhaps the book's very brevity might not be part of what makes it so splendind. I have often complained that too many books are mere pamphlets exploded out to book size to make them marketable. I guess if one does not tie value to mass, then even though this book might seem too little of a good thing, it is a remarkable value for what it does contain.

What it contains, chiefly, is a compact description of several particular species of stat problem (and ultimately, sign of a healthy stat) to watch out for, followed by an example of such a stat's use from real life. In this regard, it follows pretty literally its description as a "field guide," and the samples selected are key to what makes this book excellent.

I am not confident that after reading this book I will be as good at teasing out the issues as Best and other statisticians are, but he has tried to give me some tools to recognize some of the more egregious examples of number abuse. As someone who has long struggled with numeracy and come to see just how crucial it is for a rounded understanding of science and politics and other key subjects, I am grateful for the assistance such a guide provides. I think "Spotting" is a terrific book for people who find themselves in a similar position.

Statistics are used to support a variety of claims. It takes a trained eye to interpret the validity of a massive amount of statistical data particularly surrounding health care claims. Statistics are not synonymous with facts. Although statistics play a major role in data integrity one must look closer with a dubious eye. Especially since statistics are often used to support a claim or sell a concept. Joel Best takes the most common data marketing tricks and explains them to the lay person in a creative way. The dubious data is laid out before the reader as one would expect to see in an ecology field guide for nature enthusiasts. Therefore, Stat-Spotting fulfills its mission delightfully as promised by presenting a guide for spotting dubious data, "questionable stats". Along with the creative format the author also appeals to the reader's sense of reporting order by covering the "how", the "who" and the "what".
Be on the look out for "fictoids" which the author defines as colorful or erroneous stats used as a hyperbole. Where upon discovery a loud bell should ring in one's head as a reminder to not take literally but instead look deeper. In order to master stat spotting, here are a few simple rules to keep in mind:
1. Having a sense of scale allows you to understand the magnitude and validity of data.
2. The more severe or dramatic the case the more likely it is to be extremely rare.
3. Most people are innumerate or mathematically illiterate and subsequently easily fooled.
4. Keep an eye out for numbers that are surprising large or small.
5. The unit of measure is deceiving, for example using minutes to report a crime rate (% of total population)
6. All stats should be reported in simple language. The language used may change the implication of the data.
7. The graph is a visual representation of the statistic and may be misleading.
8. If the number is too high or too low, it most likely errs on the side of exaggeration and is therefore a guestimate. Think about how it was calculated.
9. Watch for superlatives by their nature they imply comparison.
10. Since every stat implies a definition realize the ever changing nature of definitions, for example "overweight"
11. Methodologies of data collection affect the outcome.
12. The "who" of data collection, that is, who supported the research will most likely affect the outcome.
13. Be aware of "meaningless milestones" which are the underlying trend and its cause.
14. The" law of averages" is very tricky. Therefore the median is a more useful measure particularly in a wide variation of numbers.
The underlying mission of Stat-Spotting is for the reader to become more mathematically astute. The frame of reference for many of the examples in the book is the health care field. As health care professionals are we rapidly accepting statistics as facts? Just a few of the book's statistical concepts that surround the health care profession are:
Epidemics compare old data and new data to draw conclusions but what if the definition of the disease changes?
Correlation does not necessarily prove causality therefore consider what other factors might explain the relationship.
Dramatic discoveries are a cautionary tale since a later study can prove them wrong.
The book has several pages of references to support Joel Best's concrete examples used throughout. There is also a list of helpful websites, blog sites, and suggested supplemental reading. If after reading Stat-Spotting you are still suspicious, then the guide has successfully served its purpose.

Ever wonder where they get those numbers everybody is quoting? This book will help you decide which are real and which are bogus. The author has two other books that go into the how and why of twisting the numbers. This one show you how to spot data that is off, or someone who is trying to make a conclusion not supported by the data.

This is a well-written, very useful book that fills a need: How can we know whether the statistics we see in the news are genuine?

This book is a field guide; it serves the same purpose as a field guide to birds, for example. It is a basic guide to recognizing questionable statistics based on the characteristics of those statistics, not any particular expertise about the subject at hand. The book first offers a handful of basic statistics about the US, such as the population, the numbers of births and deaths per year, the leading causes of deaths, etc. Then, each chapter puts those statistics-- and some basic critical thinking skills-- into various contexts.

The result is a short, very readable guide to the quality of the numbers we see in the news every day, one that doesn't require much math skill.

Why is this important? Because those numbers we see in the news are very often misleading, and some times they're no more than errors, blind guesses, and gross over- or understatements. If we don't have accurate data, we won't make good decisions: Garbage in, garbage out.

This is a very interesting and useful book, and I recommend it highly.