Purchase this article with an account.
Kazuya Ishibashi, Shinichi Kita; Effect of subjective probability on search termination. Journal of Vision 2008;8(6):1083. doi: https://doi.org/10.1167/8.6.1083.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Observers make a perceptual decision based on subjective probability of visual stimuli, and thereby achieve a solution against tasks such as luminance judgment, motion detection, temporal order judgment and so forth. The present study offers arguments for the usage of subjective probability, especially for the adherence of prior probability, which is likely to be neglected in cognitive decision. To examine the hypothesis that observers search visual targets effectively relying on subjective probability, we conducted a conjunctive visual search experiment. For the analysis of subjective probability, we manipulated posterior probability which was defined with the combination of prior and conditional probability. We defined prior probability by setting present-trial ratios to 10%, 50% and 90%, and conditional probability by setting a high and a low probability cue appearing in 80% and 20% of the present-trial, respectively. For the evaluation of the effect of subjective probability, two types of measurement were adopted: values of criterion in the sense of signal detection theory and search termination times, i.e., “no” response times, which have been shown to increase in accordance with target frequency. The results of the experiment showed that increasing prior and conditional probability brought about slower response times for “no” responses and a remarkable shift of the criterion value towards the peak of “yes” responses. Inquiries on target frequency of each cue in each condition showed that the observers were aware of subjective probability almost identical with posterior probability calculated on the basis of prior and conditional probability. These results suggest that subjective probability in visual search can be approximated with great precision by taking into account both prior and conditional probability and that search termination times rely on the probability information.
This PDF is available to Subscribers Only