PSYC 380 Lecture Notes - Lecture 9: Detection Theory, Response Bias, Prior Probability

• School: Queen's University
• Department: Psychology
• Course: PSYC 380
• Professor: Niko Troje

Document Summary

Assessing threshold experiment: stimulus every time at different light intensities, threshold decreases as light intensity increases, response bias- responses vary depending on individual differences. Our way to measure sensitivity confounds true sensitivity and response bias! He (cid:396)espo(cid:374)ds (cid:862)i see it(cid:863) only when she is absolutely sure she saw it. Liberal criterion: laurie wants to appear supe(cid:396)se(cid:374)siti(cid:448)e. he (cid:396)epo(cid:396)ts (cid:862)i see it(cid:863) whenever there is only the slightest chance of seeing something. The smart guy: peter figures that every trial does contain a stimulus anyway a(cid:374)d al(cid:449)a(cid:455)s (cid:396)espo(cid:374)ds (cid:862)i see it(cid:863) How can we reduce response bias? (how can we come up with an unbiased sensitivity measure?: add trials in which no signal is present. 4 possible experimental outcomes: can measure in terms of frequency (#), rates (%) B = (hr + fa) / 2 (turns out these are not an accurate measure)