Numerical Analog for a Psychophysical Task
Michael Schoeffler · The Journal of the Acoustical Society of America · 1965
In the theory of signal detectability, it is assumed that the perceived magnitude of a stimulus can be represented as number, and the biases resulting from response preferences or decision theoretic considerations affect only the criterion that is used. In order to examine these assumptions, an experiment was conducted in which observers were presented three digit numbers, each of which was drawn from one of two normal distributions. Observers were required to respond “high” or “low” to each number and feedback was provided. Three a priori probabilities were used. The results indicate that the variance of the criterion and response latencies decrease with practice; that correct responses for extreme values of the stimulus have shorter latencies than those for intermediate values; and that the response corresponding to the more frequently presented stimulus has a shorter latency than the other response. The relevance of the results to the interpretation of psychophysical data is discussed.