Reviews

Statistics for people who (think they) hate statistics

Strong quantitative research methods are one of the patient’s strongest safeguards against dangerous bias and prejudice on the part of researchers. In addition, the evidence that algorithmic decision making outperforms ‘clinical’ or other human decision making by a fair margin is very strong. It is therefore sad that statistics are so feared by students, and in general are not well taught. Most books are sound on the elementary parts of statistics, but start to become difficult when more abstract concepts are covered. This book is in that tradition.

It is a standard elementary introduction to statistics, with worked examples, a brief guide to the use of SPSS, and reference to a sample of statistical web sites. It has some other good features: very simple concepts like mean, median and mode are clearly covered; for graphs it suggests that students produce the graphs with pencil and paper before using the computer; it tries to explain words like inference and data (noting correctly that the latter word is in the plural); it is better than many on the general idea of hypothesis testing; it discusses reliability and validity sensibly; and it gives citations to further reading on some of its major examples.

Once it starts to get into the intellectually demanding areas, however, it starts to become less clear. Although it deals with topics like statistical significance, correlation and regression, analysis of variance and factor analysis, it covers them in

...

It is a standard elementary introduction to statistics, with worked examples, a brief guide to the use of SPSS, and reference to a sample of statistical web sites. It has some other good features: very simple concepts like mean, median and mode are clearly covered; for graphs it suggests that students produce the graphs with pencil and paper before using the computer; it tries to explain words like inference and data (noting correctly that the latter word is in the plural); it is better than many on the general idea of hypothesis testing; it discusses reliability and validity sensibly; and it gives citations to further reading on some of its major examples.

Once it starts to get into the intellectually demanding areas, however, it starts to become less clear. Although it deals with topics like statistical significance, correlation and regression, analysis of variance and factor analysis, it covers them in insufficient depth to convey to the reputed beginner the reason for using such methods.

This is the perennial pattern in statistics books, largely because the task is impossible. I suspect that the problem will not be overcome until syllabuses on nursing and social science courses set aside sufficient time to cover such important topics in adequate depth, and give students (and some teachers) a chance to get to know them as apprentices, rather than merely as readers.

Overall, its US origins show (for example, many techniques are described as ‘cool’). There are gaps as well. Although the list of statistical websites looks comprehensive, there were some sites that I could not find. If you like the north American approach to teaching, this book will be a useful addition to your armoury. If not, it has nothing special to recommend it.

Want to read more?

Subscribe for unlimited access

Enjoy 1 month's access for £1 and get:

  • Full access to nursing standard.com and the Nursing Standard app
  • Monthly digital edition
  • RCNi Portfolio and interactive CPD quizzes
  • RCNi Learning with 200+ evidence-based modules
  • 10 articles a month from any other RCNi journal

This article is not available as part of an institutional subscription. Why is this?