Appendix.docx

Appendix – Signal Detection Theory A Signal Detection Experiment  In a signal detection experiment studying the detection of tones, we use only a single low-intensity tone that is difficult to hear, and we present this tone on some of the trials and present no tone at all on the rest of the trials o Thus, a signal detection experiment differs from a classical psychophysical experiment in two ways: in a signal detection experiment,  1) only one stimulus intensity is presented, and  2) on some of the trials, no stimulus is presented  Experiment, when the tone is presented o Says “yes” on 90 trials. This correct response – saying “yes” when a stimulus is present – is called a hit in signal detection terminology o Says “no” on 10 trials. This incorrect response – saying “no” when a stimulus is present – is called a miss  When no tone is presented o Says “yes” on 40 trials. This incorrect response – saying “yes” when there is no stimulus – is called a false alarm o Says “no” on 60 trials. This correct response – saying “no” when there is no stimulus – is called a correct rejection  Payoffs: a system of rewards and punishments used to influence a participants motivation in a signal detection experiment o Add some financial inducements to the experiment o Tell the participant that we are going to reward them for making correct responses and are going to penalize her for making incorrect responses by using payoffs  Receiver operating characteristic (ROC) curve: a graph in which the results of a signal detection experiment are plotted as the proportion of hits versus the proportion of false alarms for a number of different response criteria  Other than demonstrating that people will change how they respond to an unchanging stimulus, what does the ROC curve tell us? o The beauty of signal detection theory is that the person’s sensitivity is indicated by the shape of the ROC curve, so if the experiments on two people result in identical ROC curves, their sensitivities must be equal  What does signal detection theory tell us about functions such as the spectral sensitivity curve and the audibility function, which are usually determined using one of the classical psychophysical methods? o When the classical methods are used to determine these functions, it is usually assumed that the persons criterion remains constant throughout the experiment, so that the function measured is due not to changes in response criterion but to changes in the wavelength or some other physical property of the stimulus o This is a good assumption because changing the wavelength of the stimulus probably has little or no effect on factors such a motivation, which would shift the persons criterion Signal Detection Theory  Our purpose is to explain the theoretical bases underlying two ideas: o 1) the percentage of hits and false alarms depends on a persons criterion o 2) a person’s sensitivity to a stimulus is indicated by the shape of the persons ROC curve Signal and Noise  The signal is the stimulus presented to the person o Thus, in the signal detection experiment we just described, the signal is the tone o The noise is all the other stimuli in the environment, and because the signal is usually very faint, noise can sometimes be mistaken for the signal  Consider a typical signal detection experiment, in which a signal is presented on some trials and no signal is presented on the other trials o Signal detection theory describes this procedure not in terms of presenting a signal or no signal, but in terms of presenting signal plus noise (S+N) or noise (N) o That is, the noise is always present, and so on some trials, we add a signal Probability Distributions  Figure A.2 shows two probability distributions.  The one on the left represents the probability that a given perceptual effect will be caused by noise (N), and t