I like how you present various ways that you can make mistakes with statistics. One thing that is missing is like a Summary/Checklist chapter that tells you wh

I like how you present various ways that you can make mistakes with statistics. One thing that is missing is like a Summary/Checklist chapter that tells you what you SHOULD do, in a few common scenarios to avoid all the mistakes presented in the previous chapter. I know its not that simple, and it depends a lot on how you are testing and what you're actually trying to achieve, but a few examples wouldn't hurt. For example: I have two sets of measurements, and I want to know: * is there a statistically significant difference between them * if yes, how much is the difference A somewhat simplistic way I'd do that is to do a two-sample t-test for question #1, and to compute statistics for the difference between the samples (mean, median, confidence intervals), but doing just that I might've already committed some of the mistakes that your site warns about, for example I completely disregarded the power of the test. FWIW I like parts of this book on statistics which focuses on statistics in the domain of computer systems / network, although it is rather too long: http://perfeval.epfl.ch/ http://perfeval.epfl.ch/