A Bayesian prediction of four-look recognition performance from one-look data

The hypothesis that a human O’s (S’s) performance in a visual recognition task can be modelled by Bayes’ theorem was investigated. Two Ss were run for 40 experimental sessions each. Their task was to specify the direction of the gap of tachistoscopically presented Landolt rings (Cs). There were four possible gap directions, and two experimental conditions. In one condition, S responded after each stimulus presentation. In the other, a fixed-observation condition, Ss responded after four consecutive presentations of a C. Exposure durations were such that performance under both conditions was greater than chance, but less than unity. Predictions of four-look performance from one-look data were made. Overall hit rates were predicted closely. The entire pattern of each S’s four-look data was also predicted reasonably well. Further tests of the model are currently under way.     

A Bayesian prediction of multiple-look identification performance from one-look data: The effect of unequal prior probabilities

Bayes’s theorem had accurately predicted multiple-look visual identification performance from one-look data in two previous studies when the prior probabilities of the visual stimuli were equal. In this study, prior probabilities were manipulated to determine whether these values were important in determining agreement between man and statistical man. The prior probabilities of the stimuli, four gap orientations in Landolt rings, were .7, .1, .1, and .1 for four Ss and .3, .3, .3, and .1 for three Ss. One-look data were used to predict two-look and four-look performance. In the 7-1-1-1 condition, the empirical percent correct increased as a function of repeated observations for the rare stimuli, but not for the high-probability stimuli. In the 3-3-3-1 condition, Ss improved over repeated observations for both types of stimuli. Predictions of multiple-look performance under unequal prior probability conditions in this study were less accurate than those under equal prior probability conditions in previous studies. Three possible determinants of this difference were discussed.     

Counselor Confidence and Accuracy of Prognoses of Success or Failure

This study assessed the accuracy of predictions of freshman and overall college scholastic performance made by groups of high school counselors, college advisors, and counseling psychologists from a university counseling center in relation to the confidence of these judges that their prognoses were accurate. Predictions were made from three sets of case information. The results revealed that: (1) the degree of confidence counselors indicated in their freshman and overall college “pass” predictions was appropriately related to accuracy; (2) counselor confidence in freshman “fail” predictions was not related to accuracy although the “fail” judgments tended to be more accurate than the “pass” prognoses; (3) counselor confidence in their overall “fail” predictions was not significantly related to accuracy and, unlike the results for the freshman judgments, the overall “fail” predictions were not more accurate than the “pass” predictions; (4) the amount of case data available was not related to counselor predictive accuracy.     

Visual and auditory processing of common pattern class structure

Visual and auditory classification of equivalent class structured patterns were examined. Underlying patterns from two classes were translated into auditory tone sequences and visual polygons. All Ss classified 50 visual patterns and their direct auditory analogs. Visual classification accuracy exceeded auditory accuracy (p < .01); however, auditory accuracy improved when auditory classification was preceded by the visual task (p < .01). Based on group data, classification strategies appeared similar across modalities, with accuracy of classification of individual patterns predicted to the same degree by common measures of physical class structure across modalities. Ss’ drawings of the prototypes also suggested a common strategy across modalities. While group data suggest some consistency of classification strategy across modalities, individual Ss were not at all consistent in their visual and auditory classifications.     

Speech identification and “knowing that you know”

Twenty-five Ss heard a random list and a context (food-elated) list of 50 words masked in white noise for a total of three trials each. The Ss were required to write the word and indicate their certainty that it was correct. Confidence rating accuracy (Type 2 d’) was greater for context than for random lists (p < .001). Probability correct also improved with context (p < .001). Unlike the context factor, repetitions led to greater probability correct (p < .001) but did not lead to changes in Type 2 d’. The results were interpreted as support for a feedback model of context and confidence rating accuracy.     

Binocular summation of equal-energy flashes of unequal duration

To determine if binocular summation occurs when increment flashes are of equal energy (Bloch’s law) but unequal in luminance-duration parameters, three Ss made temporal forced-choice judgments: (1) monocularly, (2) binocularly when the flashes to each eye were identical, (3) binocularly when the flashes to each eye were of equal energy but different in terms of their luminance and duration parameters, and (4) binocularly when flashes to each eye were separated by 100 msec. Binocular detection rates were consistently superior to monocular detection rates. Similarity in performance between Conditions 2 and 3 indicates that the binocular visual system responds only to the total energy of each monocular flash. The data from two Ss reveal that binocular performance was greater than that predicted on the basis of probability summation.     

Confidence ill stimulus predictions and choice reaction time

In two choice reaction time (RT) experiments, a stimulus prediction and a confidence judgment in the prediction preceded each occurrence of one of two stimulus alternatives. Ss identified each stimulus presentation by pressing a left-hand or right-hand telegraph key. In Experiment I the source of the stimulus predictions and confidence estimates was varied between groups of 20 Ss. For each condition, RT to correctly predicted stimuli was an inverse function of prediction confidence. Following incorrectly predicted stimuli, RT was not reliably influenced by confidence when S gave both predictions and confidence judgments; but RT to incorrectly predicted stimuli was an increasing function of confidence when E verbalized the predictions and confidence estimates or when S predicted and E indicated confidence. In Experiment II Ss made predictions and the validity of Es confidence estimate was manipulated between Ss. When Es confidence was perfectly related to the probability of a correct prediction, choice RT to nonpredicted stimuli was inversely related to confidence. However, choice RT to nonpredicted stimuli was not affected by prediction confidence when Es judgments were random.     

The probability-matching decision rule in the visual discrimination of order

In three experiments, human Ss were required to decide, on each trial, which of two spatially proximate, nearly simultaneous, lights was illuminated first. The right light preceded the left on either .33 or .83 of the trials. When Ss were frequently informed of these values, then, regardless of the accuracy of performance, Ss adjusted the rate at which they made “right” responses to approximate the rate at which the right-first pattern was presented. These results are in clear agreement with an adaptation of Atkinson’s three-state detection model.     

Response Speed and Accuracy During Anticipatory Stress

Response speed and accuracy of 48 general aviation pilots were determined under condition of anticipatory physical threat stress (APTS). Variables included unpleasantness of the event (electrical shock), proximity of the event, and relationship of Ss’ performance to the occurrence of the event. The data support inclusion of these variables in Wherry’s model of APTS. Ss who believed they could avoid the shock by an adequate performance were able to maintain or improve that performance while the performance of Ss who perceived the shock as inevitable was deteriorating. Two possible modifications of the APTS model were discussed: (1) weighting APTS determiners according to their effectiveness in generating anticipatory stress, (2) recognizing the interaction of the APTS variables with S’s personality structure.     

Signal detection and localization by real observers

A model for detection and localization (DAL) was applied to the processing of a noise-masked tonal signal by human Ss. The model was supported in that consistent values of d’ were obtained from rating. forced-choice (FC). and DAL procedures. Also. the rating d’ accurately predicted per cent correct on the FC and DAL trials. A shortcoming of the theory was its failure to account satisfactorily for the details of FC and DAL stimulus-response matrices.     

Eyewitness memory and aging: predictors of accuracy in recall and person recognition.

This study examined adult age differences in identification accuracy, confidence, and verbal recall of crime films. A total of 120 Ss in 3 adult age groups watched 2 crime films that contained characters varying in role and visibility. Results suggested a modest negative relationship between age and identification accuracy but no age differences in recall of main points or details. Education and self-reported health positively predicted verbal recall, but higher education was associated with lower identification accuracy. No age differences were found in confidence or in the relationship between confidence and accuracy.     

Reevaluation of two models for judgments of perceptual intervals

In Marks and Cain’s (1972) experiment, 12 Ss estimated magnitudes of stimuli presented one at a time and differences in magnitudes of stimuli presented in pairs for heaviness of weight, roughness of emery papers, and area of circles. Their results appeared inconsistent with the two-stage model of magnitude estimation. But the expressions fitted to the data were different from those previously used in tests of the model. An additive constant was omitted. A reanalysis of their data with the constant included yielded support for the model for all three continua. An alternative model proposed by Marks and Cain was reasonably consistent with the da~a for weight, while inconsistent with the data for roughness and area.     

A model of memory in concept learning

A concept learning model was developed and tested in two conjunctive attribute identification tasks. The model includes assumptions about the focus of attention, decision making, and memory for stimulus information and prior decisions. Predictions are made about how S changes his hypothesis following an error. Procedures in both tasks allowed inference of the subject's current hypothesis. The hypothesis selections and error statistics were in the majority of cases accurately predicted by the model. Deviations from predictions on hypothesis sampling occurred for naive Ss but not for trained Ss who were required to state a hypothesis on each trial.     

Differences in heaviness in relation to density and weight

Eight Ss (Group L) judged differences in heaviness between all pairs of stimuli in three series of five weights each, in the range of 100–300 g, presented randomly. Series L varied in weight and density, Series LVW varied only in weight, and Series LVD varied only in density. Another eight Ss (Group H) judged heaviness differences between all pairs of stimuli in three similar series (H, HVW, and HVD) in the range 700–900 g. Differences in density alone yielded relatively greater impressions of difference in heaviness in Group H than in Group L. Predictions from Ross and Di Lollo’s vector theory were confirmed.     

On the relationship between detection and recognition

Predictions derived from a “unitary-process” theory on the relationship between detection and recognition were examined in two experiments on visual detection and recognition. As predicted on the basis of this theory, recognition accuracy was superior to detection accuracy. An analysis of the conditional probabilities from a joint detection-recognition task revealed the pattern of relationships predicted by the theory. Increasing the number of alternative stimuli while maintaining only two response categories was found not to significantly affect detection or recognition performance, contrary to expectations based on the idea of multiple independent detectors.     

The detectability of the odor of butanol

Three experiments were conducted to measure the sensitivity of two Ss to the odor of butanol. In the first two experiments, the method of double random, yes-no staircases was used. A practice effect of over a log₁₀ unit in millimoles decrease in apparent threshold was observed in both Ss. Consistent shifts in the response criterion were induced when Ss were paid to meet an arbitrarily determined physical criterion. In Experiment S, the confidence rating procedure was used. Results at eight different signal intensities are of the form predicted by signal detection theory. d’ is shown to be related to signal strength by a power function with a slope of about .30 which suggests that the olfactory transducer compresses sensory input produced by weak concentrations of butanol.     

The role of construal processes in overconfident predictions about the self and others

Overconfident behavioral predictions and trait inferences may occur because people make inadequate allowance for the uncertainties of situational construal. In Studies 1-3, Ss estimated how much time or money they would spend in various hypothetical, incompletely specified situations. Ss then offered associated "confidence limits" under different "construal conditions". In Study 4, Ss made trait inferences about someone they believed had responded "deviantly"--again with situational details unspecified and construal conditions manipulated. In all 4 studies, Ss who made predictions or trait inferences without being able to assume the accuracy of their situational construals offered confidence limits no broader than those of Ss who made their responses contingent on such accuracy. Only in conditions where Ss were obliged to offer alternative construals did they appropriately broaden their confidence limits or weaken their trait inferences.     

Compensating for distortion in viewing pictures obliquely

Gombrich (1972) suggested that interpretation of obliquely viewed pictures depends largely on the foreshortened image projected to the eye, the attendant distortion rarely being important. Sixteen university students judged whether 128 drawings viewed at 41 and 26 deg to the page pictured rectangular or nonrectangular boxes. Projective geometry predicted Ss’ classifications; for many pictures, predictions for normal and oblique viewing differed. Nevertheless, at 41 deg, Ss classified 77% of the pictures where predictions differed according to the normal prediction. Furthermore. performance on cases where the predictions agreed was no better. Even at 26 deg, normal classifications still dominated. Ss demonstrated substantial capacity to compensate for oblique viewing and small tendency to judge according to the foreshortened image.     

Second choices in a visual span of apprehension task

Four Ss were run in a visual span of apprehension experiment to determine whether second choices made following incorrect first responses are at the chance level, as implied by various high threshold models proposed for this situation. The relationships between response biases on first and second choices, and between first choice biases on trials with two or three possible responses, were also examined in terms of Luce’s (1959) choice theory. The results were: (a) second choice performance in this task appears to be determined by response bias alone, i.e., second choices were at the chance level; (b)first and second choice response biases were not related according to Luce’s choice axiom; and (c) the choice axiom predicted with reasonable accuracy the relationships between first choice response biases corresponding to trials with different numbers of possible response alternatives.