Intuitive Biostatistics
4th Edition, 2018
Harvey J. Motulsky
Oxford University Press
ISBN13: 978-0190643560 

amazon.com
amazon.co.uk

Overview

Intuitive Biostatistics is both an introduction and review of statistics. Compared to other books, it has:

  • Breadth rather than depth. It is a guidebook, not a cookbook.
  • Words rather than math. It has few equations.
  • Explanations rather than recipes. This book presents few details of statistical methods and only a few tables required to complete the calculations.

Who is it for?

I wrote Intuitive Biostatistics for three audiences:

  • Medical (and other) professionals who want to understand  the statistical portions of journals they read. These readers don’t need to analyze any data, but need to understand analyses published by others. I’ve tried to explain the big picture, without getting bogged down in too many details.
  • Undergraduate and graduate students, post-docs and researchers who will analyze data. This book explains general principles of data analysis, but it won’t teach you how to do statistical calculations or how to use any particular statistical program. It makes a great companion to the more traditional statistics texts and to the documentation of statistical software.
  • Scientists who consult with statisticians. Statistics often seems like a foreign language, and this text can serve as a phrase book to bridge the gap  between scientists and statisticians. Sprinkled throughout the book are “Lingo” sections that explain statistical terminology, and point out when statistics gives ordinary words very specialized meanings (the source of much confusion). 

What's new in the fourth edition?

 In this fourth edition, I edited every chapter for clarity, to introduce new material, and to improve the Q&A and Common Mistakes sections. I substantially rewrote two chapters, Chapter 26 on sample size calculations and Chapter 28 about case- control studies. I also added two new chapters. Chapter 47 discusses statistical concepts regarding the reproducibility of scientific data. Chapter 48 is a set of checklists to use when publishing or reviewing scientific papers. Other improvements:<

  • Chapter 1. Two new sections were added to the list of ways that statistics is not intuitive. One section points out that we don’t expect variability to depend on sample size. The other points out that we let our biases deter- mine how we interpret data.

  • Chapter 2.New sections on conditional probability and likelihood. Updated examples.

  • Chapter 4.Begins with a new section to explain different kinds of variables. New example (basketball) to replace a dated example about premature babies. Added section on Bayesian credible intervals. Improved discussion of “95% of what?” Took out rules of five and seven. Pie and stacked bar graphs to display a proportion.

  • Chapter 7. New Q&As. Violin plot.

  • Chapter 9. How to interpret a SD when data are not Gaussian. Differentways to report a mean and SD. How to handle data where you collect data from both eyes (or ears, elbows, etc.) in each person.

  • Chapter 11. Geometric SD factor. Mentions (in Q&As) that lognormal distributions are common (e.g., dB for sound, Richter scale for earthquakes). Transforming to logs turns lognormal into Gaussian.

  • Chapter 14.Error bars with lognormal data(geometric SD; CI of geometric mean). How to abbreviate the standard error of mean (SEM and SE are both used). Error bars with n = 2.

  • Chapter 15. Stopped using the term assume with null hypothesis and instead talk about “what if the null hypothesis were true?” Defines null versus nil hypothesis. Manhattan plot. Advantage of CI over P. Cites the 2016 report about P values from the American Statistical Association.

  • Chapter 16.Type S errors. What questions are answered by Pvalues and CIs?

  • Chapter 18. Added two examples and removed an outdated one (prednisone and hepatitis). Major rewrite.

  • Chapter 19. Rewrote section on very high P values. Points out that a study result can be consistent both with an effect existing and with it not existing.

  • Chapter 20. Distinguishing power from beta and the false discovery rate. When it makes sense to compute power.

  • Chapter 21. Fixed 90% versus 95% confidence intervals. Two one-sided t tests.

  • Chapter 22. Introduces the phrase (used in physics) look elsewhere effect.

  • Chapter 23. Two new ways to get trapped by multiple comparisons, the garden of forking paths, and dichotomizing in multiple ways.

  • Chapter 24. QQ plots. Corrected the explanation of kurtosis.

  • Chapter 25. Points out that outlier has two meanings.

  • Chapter 26. This chapter on sample size calculations has been entirely rewritten to clarify many topics.

  • Chapter 28. This chapter on case-control studies has been substantially rewritten to clarify core concepts.

  • Chapter 29. Improved definition of hazard ratio.

  • Chapter 31. Added discussion of pros and cons of adjusting for pairing or matching.

  • Chapter 32. New common mistake pointed out that if you correlate a variable A with another A-B, you expect r to be 0.7 even if data are totally random. Points out that r is not a percentage.

  • Chapter 33. Which variable is X, and which is Y? Misleading results if you do one regression from data collected from two groups.

  • Chapter 34. Defines the terms response variable and explanatory variable. Discusses three distinct goals of regression.

  • Chapter 39. Expanded discussion of two-way ANOVA with an example.

  • Chapter 42. Removed discussion of LOD score. Added example for HIV testing.

  • Chapter 43. Added a discussion of meta-analyses using individual participant data, enlarged the discussion of funnel plots, added more Q&As.

  • Chapter 45. New statistical traps: dichotomizing, confusing FDR with significance level, finding small differences with lots of noise, overfitting, pseudoreplication.

  • Chapter 47. New chapter on reproducibility.

  • Chapter 48. New chapter with checklists for reporting statistical methods. 

About the author 

After graduating from medical school and doing an internship in internal medicine, I switched to research in receptor pharmacology (and published over 50 peer reviewed articles). While I was on the faculty of the Department of Pharmacology at the University of California San Diego, I was given the job of teaching statistics to first year medical students and to graduate students. The syllabus for those courses grew into the first edition of this book. I hated creating graphs by hand, so I created some programs to do so. I also created some simple statistics programs after realizing that the existing statistical software, while great for statisticians, was overkill for most scientists. These efforts were the origins of GraphPad Software Inc., which has been a full-time endeavor for many years. In this role I email with students and scientists almost daily, making me acutely aware of the many ways that statistical concepts can be confusing or misunderstood.

Alternative shorter book: Essential Biostatistics

Essential Biostatistics (released in July 2015) is shorter (200 pages) and less expensive ($20) than Intuitive Biostatistics.