Ten Books to understand randomised controlled trials

Randomised controlled trials are a necessary, but not sufficient part of evidence-informed practice. I have included books that range from ones that I found very accessible and enjoyable reads to those that have genuinely upset me with how difficult they are.

1. Testing Treatments by Imogen Evans and colleagues

This is a book I wish I had read sooner. It’s written about healthcare, but it’s incredibly accessible. There is also a free version online and lots of accompanying materials.

My favourite section is the first few chapters which challenge some intuitive notions and help you dial up your scepticism. The chapter titles are instructive:

  1. New, but is it better?
  2. Hoped-for effects that don’t materialise
  3. More is not necessarily better
  4. Earlier is not necessarily better

I cannot recommend this book highly enough – get it, read it, think about it.

2. Designing randomised trials in health, education and the social sciences by David Torgerson and Carole Torgerson

This is one of the first books I read about trials, so perhaps I’m biased, but it is still one of my favourites. Some sections are now a little dated, given the rapid expansion of trials since 2008, but I would still recommend it as a great starting point for anyone wanting to understand more about the principles of RCTs.

My favourite part is chapter 2, which concisely explains the limitations of before and after designs by introducing selection bias and regression to mean very clearly.

3. Research Design by Stephen Gorard

This masterful book covers much more than trials, and it completely changed how I think about research. Sections of it are hard going, but Stephen explains tricky ideas clearly.

My favourite section is chapter 1 on design notation. It takes a bit of effort to make sense of, but once you get design notation, you have a great way of making sense of research.

Another favourite section is about ‘research warrants’, which is essentially the logic that connects the data with the claims made. Again, it’s not a light read, but it’s profoundly helpful.

4. How to Read a Paper: the Basics of Evidence-Based Medicine and Healthcare by Trisha Greenhalgh

Sorry for giving another example from healthcare, but education must shamelessly and critically learn from other fields.

This book is now in its sixth edition for a reason. The sections relating to trials are obviously relevant. More broadly, How to Read a Paper provides a brilliant introduction to critical methodological ideas and how to think about evidence use.

I particularly like chapter 3, Getting Your Bearings: What is this Paper About? This chapter deconstructs the thought process of an expert appraising a research paper.

5. Teacher-Led Research by Richard Churches and Eleanor Dommett

This book is written for people who want to produce rather than just use research. If that’s your aim, then this is the book for you. However, if, like me, your primary interest is in consuming research, then this book goes into more of the craft of trials and will build very nicely on the foundational texts already mentioned.

My favourite part is the third chapter about designing experimental research. This chapter also introduces alternative designs like within-participant designs instead of between-participant designs. However, in truth, these designs are rare in educational research.

Teacher-Led Research is a great book, but if you have already read more than one of the earlier books on this list, you could give it a miss.

6. Bad Pharma by Ben Goldacre

Randomised controlled trials are great in principle, but in practice, many are useless. Ben Goldacre offers a funny and thought-provoking tour of the issues that have blighted medical research linked to the pharmaceutical industry. Many of the ills that Goldacre identifies are also present in education trials. Some of the ills that healthcare has suffered may come to education in the future unless we manage to heed the warnings from other fields.

My favourite part was the recurring theme about the need for all trials to be published, which Goldacre subsequently expanded into the All Trials campaign.

7. Using randomised controlled trials in education by Paul Connolly and colleagues

Using Randomised Controlled Trials in Education offers one of the gentlest introductions that you could hope for to some of the finer points of trials, including how to analyse trials. Like Teacher-Led Research, this book is ideal if you are considering running trials.

My favourite section of this book is the step-by-step guidance on analysing trials, including more complex designs. There is also a good section on logic models and rich examples.

8. Experimental and quasi-experimental designs for generalised causal inference by William Shadish, Thomas Cook and Donald Campbell

This book was published in 2002 and built on two earlier books about ‘generalised causal inference’ published in 1963 and 1979. Campbell is the Donald Campbell after whom the Campbell Collaboration is named. Let’s be clear: this is not some light reading; this is a textbook. I’ve heard multiple triallists refer to this book as their bible.

My favourite section covers the different threats to internal validity. For a book that is 20 years old, it’s still worth reading, but I would skip over many of the statistics sections and turn to more recent texts instead.

Suppose you don’t want to read a textbook, then I strongly recommend the Research Methods Knowledge Base, developed by a collaborator of the authors, Bill Trochim. It’s an excellent reference resource – particularly the section on research design.

9. The Effect by Nick Huntington-Klein

Okay, at this point, you’ve reached an unhealthy relationship with trials. Fortunately, you get to meet Nick Huntington Klein, who is witty, writes elegantly and makes your obsession seem mild by comparison. Nick offers a superb overview of causal inference covering trials and a heap of other methods. Crucially, Nick helps you to see the relationships between these different methods.

My favourite section is the first half, which explains causal inference very clearly and differently from the texts mentioned so far before the second half gives more of the technical minutiae. Beware, the second section includes some complicated maths – it’s the kind of maths with symbols you struggle to find on your keyboard.

Finally, the book is freely available online, and Nick has recorded an engaging video series to accompany the text.

10. Cluster randomised trials by Richard Hayes and Lawrence Moulton

This book genuinely upset me with how difficult it was: it’s not for the casual reader! It is an excellent book if you want to know more about a special type of trial. Cluster randomised controlled trials involve allocating groups to treatment conditions rather than individuals, like classes or whole schools. Unfortunately, the individuals within a group are more similar to each other than they are to people outside of the group; this means that fancy designs and analyses are needed.

My favourite part of this book was the first few chapters – before it got too hard, obviously. The early chapters explain the overarching principles of clustering, why it matters, and how it can be handled. The remainder of the book is rather technical.

Finally

Books are great, but they are also expensive. Here are some of my favourite free things I have not mentioned so far.

  1. A Guide to Running Randomised Controlled Trials for Educational Researchers (2010). This is a brilliant resource – it also teaches you what equipoise is, which you can then drop into conversations about ethics.
  2. Test Learn Adapt: Developing Public Policy with Randomised Controlled Trials (2012). The font size used in this resource is reassuringly large. The diagrams are also excellent.
  3. Randomised trials in education: an introductory handbook (2013). This is an updated and abridged version of the book mentioned above. The FAQ section is useful.