[MUSIC] So here we are with Zoltan Dienes, the author of this book, Understanding Psychology as a Science. Zoltan, it's really nice that you want to be here. >> Nice to be here. >> Take some time from this conference that we're at. Taking some time for this interview. So first of all I was wondering, how did you get into this topic? What decided that you would write about things like this or find them interesting? >> My interest goes back quite a way, actually. My mother was a Scientologist. And the headmaster of the sixth form I was in gave me a book on Popper. So I read about Karl Popper and he discusses what makes something a science or not a science. And that had a pretty big impact on me and how I viewed various movements and what they might call themselves a science or not. The fact that also it's not just a label, but Popper wanted to point out that how you approach that really matters to the quality of the knowledge that you get. >> I see that you have some interests in one of the examples comes from precognition research and psi research. And you did some studies about it yourself, right? >> Yes, now there's a topic where the philosophy of science, the philosophy of statistics really comes up. So if you investigate a topic where a lot of people think the hypothesis theory has a very low or high probability. Then just using p-values and p less than 0.05, somehow doesn't seem right. And then to make a binding decision at the end of that doesn't seem right. Whereas saying there's a certain amount of evidence here and that will change your opinion by a certain amount. Just seeing the better way of dealing with theories where people might have radically views about them in the first place. So that was in fact my first application of Bayesian statistics. >> All right, so mainly you were doing that research, that's when you ventured into this direction of Bayesian statistics. >> Yeah, initially, 1980s I was doing my PhD at Oxford and Oxford invited this statistician from UCL, John Kier, who had been Piaget's statistician and Hans Eysenck's statistician. And he came to teach the PhD students at Oxford, every Wednesday afternoon, statistics. I found him quite captivating, charismatic. But he would teach just whatever he wanted to teach. And one term he said I'm going to teach you Bayes. And by the end of the term, there's just me and one other person sitting there. >> [LAUGH] >> And I found it absolutely fascinating, but the other people just thought this is not relevant to my thesis, I'll disappear. >> All right. >> So my little Bayesian base factor calculator actually programmed at that point. >> Okay. >> Way back then. Like 1980s. >> So that's actually an interesting point, right? So you say that most of these students slowly left. But now, of course, if you write this book, there's a lot of topics in here that has to do with Bayesian statistics and likelihood. So apparently you thought it was worthwhile that the students shouldn't leave. Actually you thought it was worthwhile to explain more to students about this. So why do you think it's important that students learn about these things? >> It's important because there's a debate about how one should do statistics that isn't settled, or it seemed to be settled but it wasn't. And the way we had settled on doing things caused problems, that is why we had the credibility crisis, that was a big part of it. And now people are really thinking about how should we do statistics, how should we do science and I think in ten years' time results sections are going to look rather different than they do now. And you can either be ahead of the curve on that or you can be desperately trying to catch up later. And the people we need to reach are the current students coming up. They need to understand these issues, and where researchers have gone wrong in the past that led to science that wasn't really that credible in the end. And if you don't understand statistics, and the philosophy of science more generally, then you'll make the same mistakes again. So, I think this is the time to change. Change is happening. Ten years' time sounds not very different. I actually think in a hundred years time, people will look back on how we do science now and it will be a bit like how we look back hundred years ago. It's like really, they did it like that? That's pretty primitive. >> Yeah. >> And a lot of the changes that will be coming about and affecting how we do science in the future, they're happening right now. >> All right. >> So that's quite exciting, in a way. >> Yeah, yeah I agree. Okay, very good. So when you were you exhibiting all these things for yourself, these statistical techniques and figuring out, mainly this was because you needed to analyze your own data, right? I mean, you're not In principle statistician yourself. >> No trained as a psychologist, exactly. >> So you're trying to figure out how to use this, so what things were most helpful to you or most insightful or eye opening when you were learning about these things? >> I think when I wrote the book and I don't know if it comes across there but probably as leaning towards a likelihood approach. That sort of of. >> It comes across a little bit [INAUDIBLE]. >> Yeah. >> I noticed a little bit. >> Yes and either that obeys for me a big issue was how you deal with non-significant results. because in my field, non-significant results were often used [INAUDIBLE]. My field being unconscious knowledge, unconscious perception, and perceived learning and how you establish perception is unconscious, subliminal. Or learning is implicit is often by getting a non-significant result. And but just getting a non-significant result in itself does not have evidential value, so what do you do? So I was looking looking at light and bayes. And then just through having to really apply them in real situations. Not is about statistics but do psychology and apply those techniques. I found bayes was just most useful. Because typically when you have a theory, you're saying, well there's a range of possibilities for what the population hypothesis could be, some are more plausible than others. And as soon as you accept that and want to get evidence from the hypothesis you'll have to base factors. And I've just found that enormously helpful. If you look at how I write papers now even compared to three years ago, it's just radically different. I look back to how I read papers, really just bear in mind, just following through the logic that's contained in that book, but just following it through and actually applying it every step of the way. It's turned around how I do science. >> So you'd say you're still learning about this book. Looked into yourself. >> Still learning, exactly. So since 2008, when the book was published, I've been thinking about how to actually apply it in real situations. So exactly the same logic that's in there, but putting it into practice is what I've been working on. >> All right, great. Thanks so much. [MUSIC]