Chapter 1. Signal Detection Theory

Introduction

You are downstairs watching a movie. The volume is up on your surround sound system and you don’t have a phone in the room. At a moment of great tension and excitement, and perhaps volume, you think you hear the phone. You are not sure. What are you to do?

Most of the time, we are aware of our surroundings. You can easily identify the door in the room you are in. It is obvious, but there are times when something happens, like the phone ringing, and you have to decide if something has actually happened. Signal detection theory was developed to explain these ambiguous situations.

This activity will use a simulated experiment to guide you through the elements of signal detection theory and how it describes how you determine if you think the phone actually rang.

Experiment Setup

Instructions

In this experiment, you have to determine if something has changed on the screen. The situation is very simple. Dots will flash on the screen in front of you; sometimes there will be more dots than others. It is all random. Your job is to determine if the experimenter has added any dots to the screen.

Please follow the instructions on the experiment tab as you go through the activity to illustrate the elements of signal detection theory.

Experiment

Results

Debriefing

Signal detection theory describes how we go about determining if something happened in a situation where it is not clear that something happened. For example, you are downstairs without a phone, watching a movie, and at an intense moment, you thought you heard the phone ring. You have to decide if the phone rang. Signal detection theory separates this decision into two elements: how easy is it for you to detect the stimulus (your sensitivity) and how much you get by detecting the stimulus (your criterion). To illustrate criterion, think of waiting to hear about a possible date: You want to hear the phone. Or think of hoping to avoid a call from your parents: You don’t want to hear the phone.

For any stimulus, you have a fixed sensitivity but you can adjust your criterion to any level. You can be very lax where you say you see the stimulus all the time. You will have a lot of hits but also a lot of false alarms. You could be very strict and say you never see the stimulus unless you are absolutely sure. You would have few false alarms but also few hits. To summarize all of the possible outcomes for one stimulus, the ROC curve was developed. The ROC curve plots the false alarms on the x-axis and the hits on the y-axis. As your sensitivity increase, the curve bends away from the positive diagonal which represent what happens when you cannot detect the stimulus at all (hits = false alarms).

References:

Green, D. M., & Swets, J. A. (1966). Signal detection theory and psychophysics. Oxford, England: John Wiley.

Liu, S. S., Chen, J., Melara, R. D., & Massara, F. (2008). Consumer's product-locating behavior: Exploring the application of signal detection theory. Psychology & Marketing, 25(6), 506-520.

Quiz