Psychometric curves of time judgments across oddball durations for experiments 3 & 4.
a: Percent “longer” responses across stimulus durations around 1 s with an asymmetric oddball placement (experiment 3) around the standard duration (1050 ms, black vertical line). The probability of responding “oddball > standard” increased with longer oddball durations (triangles) and the point of subjective equality (PSE, circles; 50% threshold of a Weibull fitted curve, solid lines) was for both correct (red) and incorrect trials (blue) significantly shorter than the standard duration. b: Absolute counts of PSEs resulting from 1000 bootstrap simulations for correct (red) and incorrect trials (blue). Threshold estimates from both correct and incorrect trials were both significantly shorter than the standard duration (black vertical line). In addition, PSEs from correct trials were significantly lower than from incorrect trials. c: Percent “longer” responses across stimulus durations around 100 ms with an asymmetric oddball placement (experiment 4) around the standard duration (70 ms, black vertical line). The probability of responding “oddball > standard” increased with longer oddball durations (triangles) and the point of subjective equality (PSE, circles; 50% threshold of a Weibull fitted curve, solid lines) exceeded the standard duration significantly for both correct (red) and incorrect trials (blue). d: Absolute counts of PSEs resulting from 1000 bootstrap simulations for correct (red) and incorrect trials (blue). Threshold estimates from both correct and incorrect trials significantly exceeded the standard duration (black vertical line). PSEs from correct trials were significantly lower than from incorrect trials. (a & c) Error bars for real data (triangles) indicate one standard error of the mean (SEM) for within-subject designs (similar to the description in [26]). Individual proportions of longer responses of each subject have been centered on their average proportions of longer responses across conditions before calculating the SE. Error bars for estimated thresholds (circles) indicate 95% confidence intervals of the bootstrapped threshold distributions. (b & d) Dashed vertical lines indicate 95% confidence intervals of the bootstrapped threshold distributions [28].
