HYPNOSIS BY NUMBERS
Discovering Statistics Using R – Andy Field, Jeremy Miles and Zoë Field
I know what you’re thinking. This was all a fun and interesting place, full of Amy’s reviews of the weird and wacky, and then I turned up to spoil it all and talk about dull things. But you’d be wrong!
Stats, as this is what this book is (no doubt) about, is absolutely amazing! Not as amazing as that other branch of mathematics that is the numerical equivalent of 60s erotica, modular arithmetic, but still worth a quickie in the dark alley of probability.
So naturally, a book on statistics has its place in a blog about hypnosis. Of course it does; if it didn’t then we would have no idea about what worked and what didn’t, and that way mayhem lies. Hang on, what you say? Most hypnotists don’t read academic studies? And those that do have no clue about the maths in them? I can already hear Amy wailing…
We know that it is quite possible (easy, some would say) to learn hypnosis without any recourse to science (check Reality Is Plastic and Deeper And Deeper by Jonathan Chase for examples); and I would suggest it is quite reasonable to read an academic paper (assuming you can acquire them – give Google Scholar a go and look for those PDF links) without understanding the maths involved and to just accept the claims made by the authors. But there is so much joy to be had in a) truly understanding what the numbers imply, and b) criticising and discounting them when they are wrong.
Much has been written about the failure of scientists, and particularly psychologists, when applying the scientific approach. Even more, I would think, has been written about their misapplication and failures when it comes to statistics. And the reasons are obvious – stats is hard. I mean, really hard. Luckily, fun-time-wannabe-rock-star, and statistics lecturer, Andy Field, is here to save the day. And he’s yet another hero from that wondrous place that is the Psychology Department of University of Sussex.
It appears that Andy wrote this book (and all his others, check Amazon) with the idea in mind that this shouldn’t be too difficult. However, even he admits that some of the more advanced topics he introduces he doesn’t fully understand himself. But from the outset, it’s just counting things, and then averaging things, and then looking at how spread out things are, and then standardising things, and then comparing standardised averages and spreads against others, before advancing. Even that itself is directly useful to psychology experiments and the understanding of psychology papers.
Here’s an example: Gorassini and Spanos wanted to measure the effectiveness of a hypnosis training programme on people who didn’t respond well to hypnotic suggestions. One way to do this would have been (I’m not getting the paper out and checking it right now for historical fact, so let’s just go with a hypothetical example that may or may not be close to the truth) to randomly assign people to one of two groups. Group one would receive standard information about hypnosis, and group two would receive the hypnosis training that might make them better at responding.
By being randomly assigned, we could assume that the standardised average and spread of measured hypnotic response (a fancy way of saying the scores the people in the groups achieved on a standardised hypnosis scale) would be roughly the same in the two groups, if the training was entirely ineffective. By comparing the averages and spread of the two groups we can see if one group is better at taking suggestions than the other. Further, we can understand how likely it was that the difference was down to luck rather than the intervention.
If we use the standard ways of comparing them, we end up with measures that other scientists, and potentially us, can understand. These are the things that get included in the analysis sections of papers. Andy explains these ideas clearly and then progresses to other types of tests and comparisons that academics use in psychology; and all the examples, convoluted or otherwise, are grounded in things that actual psychologists might research.
So what’s this ‘R’ thing, you ask? Well, where other books on stats are dry, mathematics books, where the closest you get to working something out for yourself is to bosh a load of numbers into a calculator and then look up the answer in a table; this book focuses on the free software package, ‘R’, instead to do most of the leg work. It doesn’t stop Andy from explaining (in detail) the maths behind the tests and the processes taking place, but it does mean that you can load the numbers into a thing on your computer and smash the buttons to get the appropriate answers. There is nothing quite like seeing it all happen, automatically, for yourself. Oh, and graphs; Andy and ‘R’ love graphs, and so will you when you get the software to make them for you.
I’ve since moved on to ‘R Studio’ with the ‘tidyverse’ library, but the same basic approaches still apply and are still relevant. In this internet-connected age, the world moves forwards and better things appear. The book was written for ‘R’ and ‘R Commander’ using libraries that existed at the time (2012), but now ‘R Studio’ makes everything easier to do and the ‘tidyverse’ library does everything the same way so you can think less and achieve more. I’ve heard on the grapevine that the second edition of the book will be fully committed to ‘R Studio’ and ‘tidyverse’, so it might be worth waiting for that to come out before rushing to spend your lunch money on this particular book on stats.
Honestly, I love maths but I didn’t have a great relationship with stats before reading this book. Now I’m wining and dining it at every opportunity and finding all sorts of ways to crunch numbers and check people’s conclusions. It’s like levelling up but for psychology paper understanding. Like you’ve finally played enough Mario to know the level layout and can get through without hiccups. Like you can see the matrix for what it is; that the world is just numbers and hypnosis is just a part of that world.
I no longer see the numbers; I just see the mistakes and the illogical conclusions.
