Effect of intermixed foveal and parafoveal presentation onsame-different judgments: Evidence for a criterion-inertia model

Abstract

Subjects judged whether two adjacent letters, which were presented either 0.5 0 (foveal or near condition) or 2.0 0 (parafoveal or far condition) from fixation, were identical or different. The preponderance of false-dijferent responses (i.e., errors on same trials) over false-same ones increased, whereas the fast-same effect was eliminated, on the far pairs, but only when they were intermixed with near pairs rather than presented in separate (pure) blocks of trials. Intermixing the near and far pairs produced the opposite trends on near trials (i.e., smaller preponderance offalse-dijferent errors, larger fast-same effect). Carr et al., who presented trigrams, found a similar criterion misadjustment, which likewise depended on intermixed presentation. They proposed that their easy (familiar or orthographically regular) pairs produced a bias or criterion shift towards "same." No obvious biasing features were present in the near or far pairs, however, so the criterion misadjustment found here was attributed to the combined effect of internal noise and criterion inertia, not to criterion shifting. Increasing the level of internal noise on far trials produced more spurious perceived mismatches, but on mixed blocks of near and far trials, subjects relied on a common, compromise criterion for responding versus rechecking (recheck moderate perceived differences), instead of the separate, more appropriate criteria for near trials (recheck small differences) and far trials (recheck large differences) used on pure blocks.When people judge whether two letters are identical or not, same judgments typically are faster than different judgments. According to the noisy-operator theory (Krueger, 1978), internal noise more often changes physical matches into spurious perceived mismatches than vice versa. As a result, different judgments are slower, owing to more rechecking of perceived mismatches, and there typically is a tendency to make more false-different responses (i.e., errors on same trials) than false-same ones.If internal noise is so important a factor, then the fast-

Limited translation invariance of human visual pattern recognition

“…It is worth noting that despite many investigations on the influence ofstimulus position in the visual field (e.g., 1.R. Irwin & Francis, 1995;Krueger, 1985;Schweitzer, 1991) on the fast-same effect, sensitivity to retinal translations (i.e., location changes) has not been studied in any detail (but see Chignell & Krueger, 1984;Kwak, Dagenbach, & Egeth, 1991). Our experiments were not designed to enter this field of research, yet they indicate some positional specificity of the fast-same effect.…”

“…As our pilot studies did not involve additional rotations, we wondered whether the intermixing of two kinds of experimental conditions could have interfered with different performance in Foster and Kahn's (1985) study, leading to an overestimation of the displacement effect on d'. There are several reports in the literature showing differential influences of mixed conditions in same/different experiments (e.g., Carr, Posner, Pollatsek, & Snyder, 1979;Krueger, 1985Krueger, , 1986Krueger & Allen, 1987).…”

Positive effect of heterogeneity of difference on thesame-different disparity in letter matching

“…No RT shift was found, however. The criterion for rechecking is not always properly adjusted when the internal noise level changes (Krueger, 1985) or, apparently, whee. the external noise level changes, as the present data indicate, so a given effect might be expressed solely on RT in some cases and solely on errors in other cases.…”

“…In a line-matching task, Krueger and Shapiro (1985) likewise found a significant effect of external noise on the same-different disparity for errors but not for RT. Perhaps criterion inertia (Krueger, 1985) prevented subjects from properly adjusting both RT and errors to the changed conditions on mixed blocks. The error data suggest that subjects used the same criterion on mixed blocks as on pure dissimilar blocks.…”

“…By intermixing the two types of different pairs, Chignell and Krueger forced their subjects to use a common decision criterion (Krueger, 1985), irrespective of different type. The evidence supporting the responsecompetition model (equally fast RT on same and dissimilar different pairs; smaller preponderance of false-different errors with similar than with dissimilar different pairs) may have resulted because the criterion for responding versus rechecking was set too low on the perceived difference count relative to the dissimilar different pairs and too high relative to the similar different pairs.…”