Interview With Robert E. McGrath About Quantitative Models in Psychology
In this video, author Robert E. McGrath talks about his book, Quantitative Models in Psychology. (4 minutes, 52 seconds)
Interviewer [female voice]: Before we begin to talk about your book, Quantitative Models in Psychology, perhaps you could provide us with your own definition of a quantitative model.
Robert McGrath: Sure. We actually use models all the time. You could think of a road map as a kind of model. Models are something that we use all the time to relay information to each other, to communicate, to figure out what's important in any situation. So, if you think about a road map as an example of a model — you don't have all the buildings there, you don't have the holes in the street, you don't have the fire hydrants. What it does is it reduces the information to what you absolutely need.
And that's exactly what a model is in science. It's a way of representing a situation or some phenomenon in a way that is useful for the purposes of the scientist. So, models are everywhere in science, and frankly, in everyday life as well. But one of the big differences about science is that they develop what's known as quantitative models, and that's a tradition that goes back all the way to Galileo and Newton — of taking situations and reducing them to quantitative information.
One of the things that was really exciting about the work of Newton, for example, was that he was able to take a quantitative model and apply it to the falling apple and to the motion of the moon around the earth. So that by simplifying it to essential numeric information, you're able to discover things that would surprise you and that you would never have expected to be true.
Interviewer: What would you say to those who express skepticism about psychology being or ever becoming a truly quantitative science?
Robert McGrath: Well, I think that the evidence is in that that's not a problem. I think for a long time, people really were concerned about trying to reduce human experience to numbers. And I think, let's face it, we're never going to reduce love to numbers. We're never going to reduce the relationships between people to nothing but a set of numbers.
But, to the extent that by creating quantitative models, by taking situations and trying to model them with the use of numbers, you can learn things that you didn't know about before. And recognizing that, of course, there's more to it than that. In the same way that there's more to a road than what a road map shows you, but it gives you something useful about the situation.
Interviewer: What makes for a good quantitative model in psychology?
Robert McGrath: Well, it's a good question because there are different standards. Of course, the most obvious sign that you have a good model is that it fits the data well. Because, let's face it, that's the number one criterion in science for something that's worth keeping. But also, sometimes you can have a model that's really exciting because it makes sense logically or just because it's useful.
Interviewer: Could you briefly describe one or two of the key controversies surrounding how psychologists currently use quantitative modeling?
Robert McGrath: Sure. One of the ways that we use models is in drawing inferences or conclusions about large groups of people, populations, from small groups of people, from samples. And there are a whole variety of ways of doing that, and there is something called Fisher inferential statistics or hypothesis testing. There is Pearson and Neyman, null hypothesis significance testing. There's something called Bayesian statistics. These are all different ways of drawing conclusions about larger groups from smaller groups, and there's been a tremendous amount of controversy about these things over the years.
Unfortunately, sometimes these get really confusing to students in psychology and in the social sciences because it seems like people are disagreeing with each other about what's the right thing to do. And in fact, there is a reason for that — because people are in fact disagreeing about the best way to draw conclusions based on data. And that's one of the things that I tried to do in my book — was to give you a really good sense of what these different approaches, these models to making inferences, involve and why they're different each other, and when each of them makes sense.