Study of Choice

Study of Choice

Facilitator: Stephen Bush is a lecturer in the School of Mathematical Sciences at UTS. He's recently submitted his PhD which looks at the statistical study of choice.

So Stephen Bush, you're a lecturer in the School of Mathematical Sciences but you're studying the statistical analysis of choice. What does that mean?

Stephen: So when we talk about the statistical analysis of choice, what we want to do is we want to analyse the way people make choices so we want to get people to make choices. Then we want to decompose the choices that people make into the reasons why people are actually making the choices.

Facilitator: Can you explain that in a little bit more detail?

Stephen: So suppose you're buying a mobile phone. There are lots of different reasons why you might select a particular mobile phone. Maybe it's the way it looks; maybe it's the memory that's in the mobile phone; maybe it's the price - it's a good price.

So what we want to do there is actually get people to make real life choices and then take that information and actually build a statistical model figuring out exactly what is driving that choice.

Facilitator: So how do you go about your research?

Stephen: So my research is looking at exactly how to lay out one of these experiments. So when a respondent sits at a computer screen and does one of the experiments, they'll be presented with a series of what we call choice sets. In each choice set is a single decision.

So we present different competing products or different competing policy directions and we ask people to make a choice based on the information that we've provided. We take that information and we use statistical models to break down that series of choices into the different attributes or features.

Facilitator: How would this play out in our everyday lives?

Stephen: So people make choices all the time. As soon as someone gets up in the morning, they're thinking about what they're going to wear for the day and they need to trade-off between different alternatives. We might want to choose between the T-shirt or a business shirt. If you're going to a business meeting, you probably don't want to wear the T-shirt. So people think about why they're making the choices and the features of, say, that shirt that make it that more desirable - make that option more desirable.

Facilitator: So what about in retail? Have retailers cottoned on to what you're doing?

Stephen: Some of them have. A lot of companies that actually produce consumer goods are actually starting to contact places like CenSoC and doing projects with us. So if you think about a supermarket shelf, people are being faced with a choice every time they pick up an item.

There are usually different brands. There may be lights; there may be full strength; there may be cheaper alternatives; more expensive alternatives. Every time someone, say, buys a can of soup, they need to choose between all of these different options.

Facilitator: Can you look at what happens if somebody can't decide?

Stephen: Yes. One of the areas of my research was actually developing a model where I present two or three options. But I allow people to be indecisive. I allow people to say well I like those two or those three things just as much. So I want to have the option to tell the researcher that because that's actually really useful information for us as the statistician. Because that's telling us these two things are just as good.

Facilitator: Give me an example of a survey that you've done recently?

Stephen: So one survey that I worked on recently was actually looking into Australians' views towards different countries; how much people trust those different countries and also different threats to Australia's future.

Facilitator: So what's next then?

Stephen: So the next thing we're looking at is what sort of light bulbs people would like to buy. Now this project's actually being funded by the Australian Research Council. It's looking into different sorts of energy efficient lighting. Again, we're looking towards Australia's future.

What we want to look at is, let's say, different types of light; whether it be a bright yellow light or a blue light or a white light; how long different globes last; how much these globes actually cost. So these sorts of things we think might be important to light globe consumers but we want to make sure.

Facilitator: It's fascinating to think that what might seem like a very complex process going on in my mind when I'm going to make a choice about something, can actually be broken down into very simple mathematical models.

Stephen: Let's be honest. All of these models are just an approximation to the way people behave. The way people behave is so complex. But if we can simplify the process, we may get some really useful information out of that and that's what I really aim to do here.

Facilitator: Stephen Bush, thank you very much and good luck with your statistical analysis of choice.

Stephen: Thank you.