Interval bias in 2AFC detection tasks: sorting out the artifacts

Proportion correct in two-alternative forced choice (2AFC) detection tasks often varies when the stimulus is presented in the first or in the second interval. Reanalysis of published data reveals that these order effects (or interval bias) are strong and prevalent, refuting the standard difference model of signal detection theory. Order effects are commonly regarded as evidence that observers use an off-center criterion under the difference model with bias. We consider an alternative difference model with indecision whereby observers are occasionally undecided and guess with some bias toward one of the response options. Whether or not the data show order effects, the two models fit 2AFC data indistinguishably, but they yield meaningfully different estimates of sensory parameters. Under indeterminacy as to which model governs 2AFC performance, parameter estimates are suspect and potentially misleading. The indeterminacy can be circumvented by modifying the response format so that observers can express indecision when needed. Reanalysis of published data collected in this way lends support to the indecision model. We illustrate alternative approaches to fitting psychometric functions under the indecision model and discuss designs for 2AFC experiments that improve the accuracy of parameter estimates, whether or not order effects are apparent in the data.     

Decision strategies for the two-alternative forced choice reminder paradigm

The decision strategy adopted by an observer in a psychophysical procedure may be established by comparing estimates of sensitivity obtained by using detection-theoretic models based on each of the available decision strategies. Estimates of sensitivity obtained from 10 observers on auditory-level discrimination tasks using the yes/no, two-alternative forced choice (2AFC), and 2AFC with reminder (2AFCR; sometimes called duo-trio) procedures were compared in order to determine the decision strategy used by the judges in the 2AFCR procedure. 2AFCR permits at least three decision strategies: differencing, likelihood ratio, and comparison of distances. The latter strategy has frequently been assumed in the analysis of duo-trio data in the sensory evaluation literature. The present study demonstrates that the comparison-of-distances strategy was not used by 9 of the 10 observers in this auditory experiment. The nature of the instructions given to the observers had no measurable impact on the decision strategy that they adopted.     

Measuring, estimating, and understanding the psychometric function: A commentary

The psychometric function, relating the subject’s response to the physical stimulus, is fundamental to psychophysics. This paper examines various psychometric function topics, many inspired by this special symposium issue ofPerception & Psychophysics: What are the relative merits of objective yes/no versus forced choice tasks (including threshold variance)? What are the relative merits of adaptive versus constant stimuli methods? What are the relative merits of likelihood versus up-down staircase adaptive methods? Is 2AFC free of substantial bias? Is there no efficient adaptive method for objective yes/no tasks? Should adaptive methods aim for 90% correct? Can adding more responses to forced choice and objective yes/no tasks reduce the threshold variance? What is the best way to deal with lapses? How is the Weibull function intimately related to thed’ function? What causes bias in the likelihood goodness-of-fit? What causes bias in slope estimates from adaptive methods? How good are nonparametric methods for estimating psychometric function parameters? Of what value is the psychometric function slope? How are various psychometric functions related to each other? The resolution of many of these issues is surprising.     

A measure of within-participant response consistency

In this article, we introduce a measure of within-participant response consistency for use in the analysis of performance in decision-making tasks. The measure is an estimate of the correlation between the responses associated with two identical blocks of trials, the second of which has yet to be conducted. We derive a formula for the measure that can be applied to data from any two-choice decision task, including yes/no detection and two-alternative forced choice (2AFC). The estimate is easily calculated from the observed frequencies of hits, misses, false alarms, and correct rejections. We utilized data from an actual 2AFC experiment to compare estimated and observed consistency values; the estimates accounted for more than 90 % of the variability in observed consistency scores. We also discuss potential applications of the measure.     

Indicating when you do not know the answer: The effect of question format and interviewer knowledge on children's ‘don't know’ responses

Children are interviewed in a variety of contexts, for example, in the legal setting and in experimental research. In these situations, it is often very important that children indicate when they do not know the answer to a question, rather than guess. In the present experiment, one hundred and forty‐nine 5‐ to 9‐year‐olds witnessed a staged event in one of two conditions. The interviewer was either present at the event (knowledgeable interviewer) or absent from the event (uninformed interviewer). Children were then interviewed using yes/no questions and wh‐questions. Within each type of question, half were answerable based on the information provided; the other half were not answerable (i. e. the correct answer was ‘don't know’). The children performed consistently well with the answerable questions. With the unanswerable questions, there was an effect of format and interviewer knowledge. Children were more likely correctly to indicate that they did not know the answer to an unanswerable wh‐question than an unanswerable yes/no question. Also, children were more likely correctly to say ‘don't know’ to unanswerable questions when the interviewer had been absent from the event.     

The stories we tell,the reputations we form: Narrative identity and person perception

We explored the degree of consensus and accuracy in observers’ ratings of targets’ traits and autobiographical narratives. Targets narrated life high, low, and turning points and reported their personality traits. The conceptual content of these stories had previously been quantified. Here, groups of observers provided ratings of targets’ traits, knowability (how well the observer felt they ‘knew’ the target), and story conventionality, after reading each narrative. Observers exhibited consensus in perceptions of traits, knowability, and story conventionality. These ratings corresponded with targets’ self-ratings of traits, to a modest degree. The knowability of narrators did not moderate this relation, nor did key scene type. The conceptual content of targets’ narratives partially accounted for the relation between observer-reported and target-reported traits.     

Measuring, estimating, and understanding the psychometric function: a commentary.

The psychometric function, relating the subject's response to the physical stimulus, is fundamental to psychophysics. This paper examines various psychometric function topics, many inspired by this special symposium issue of Perception & Psychophysics: What are the relative merits of objective yes/no versus forced choice tasks (including threshold variance)? What are the relative merits of adaptive versus constant stimuli methods? What are the relative merits of likelihood versus up-down staircase adaptive methods? Is 2AFC free of substantial bias? Is there no efficient adaptive method for objective yes/no tasks? Should adaptive methods aim for 90% correct? Can adding more responses to forced choice and objective yes/no tasks reduce the threshold variance? What is the best way to deal with lapses? How is the Weibull function intimately related to the d' function? What causes bias in the likelihood goodness-of-fit? What causes bias in slope estimates from adaptive methods? How good are nonparametric methods for estimating psychometric function parameters? Of what value is the psychometric function slope? How are various psychometric functions related to each other? The resolution of many of these issues is surprising.     

Contrast discrimination by the methods of adjustment and two-alternative forced choice

Fifty years after the advent of signal-detection theory, some visual scientists still adhere to the notion of a high, fixed threshold. Indeed, experienced observers are able to deliver consistent settings of thresholds using the method of adjustment, but those settings are different from the thresholds obtained with two-alternative forced choice (2AFC) procedures. This article illustrates the difference by reference to four existing studies of contrast discrimination. All four sets of data show clear subthreshold facilitation, but the method of adjustment produces a rectilinear characteristic that corresponds to the use of different criteria over different ranges of contrast, whereas 2AFC procedures produce a curvilinear characteristic with generally lower thresholds. This article proposes a theoretical explanation for the difference.     

The hubris hypothesis: The downside of comparative optimism displays

According to the hubris hypothesis, observers respond more unfavorably to individuals who express their positive self-views comparatively than to those who express their positive self-views non-comparatively, because observers infer that the former hold a more disparaging view of others and particularly of observers. Two experiments extended the hubris hypothesis in the domain of optimism. Observers attributed less warmth (but not less competence) to, and showed less interest in affiliating with, an individual displaying comparative optimism (the belief that one’s future will be better than others’ future) than with an individual displaying absolute optimism (the belief that one’s future will be good). Observers responded differently to individuals displaying comparative versus absolute optimism, because they inferred that the former held a gloomier view of the observers’ future. Consistent with previous research, observers still attributed more positive traits to a comparative or absolute optimist than to a comparative or absolute pessimist.     

Seeing mountains in mole hills: geographical-slant perception

When observers face directly toward the incline of a hill, their awareness of the slant of the hill is greatly overestimated, but motoric estimates are much more accurate. The present study examined whether similar results would be found when observers were allowed to view the side of a hill. Observers viewed the cross-sections of hills in real (Experiment 1) and virtual (Experiment 2) environments and estimated the inclines with verbal estimates, by adjusting the cross-section of a disk, and by adjusting a board with their unseen hand to match the inclines. We found that the results for cross-section viewing replicated those found when observers directly face the incline. Even though the angles of hills are directly evident when viewed from the side, slant perceptions are still grossly overestimated.     

Testing theories of recognition memory by predicting performance across paradigms

Signal-detection theory (SDT) accounts of recognition judgments depend on the assumption that recognition decisions result from a single familiarity-based process. However, fits of a hybrid SDT model, called dual-process theory (DPT), have provided evidence for the existence of a second, recollection-based process. In 2 experiments, the authors tested predictions of DPT and SDT by comparing the invariance of parameter estimates between yes/no (Y/N) and 2-alternative forced-choice (2AFC) testing paradigms. Both experiments showed DPT recollection estimates in Y/N to be poorly correlated with recollection estimates in 2AFC. In Experiment 2, SDT predictions explained more variance than DPT predictions. The authors evaluate and discuss the extent to which each model possesses theoretical validity versus computational flexibility in curve fitting.     

Social learning in juvenile lemon sharks, Negaprion brevirostris

Social learning is taxonomically widespread and can provide distinct behavioural advantages, such as in finding food or avoiding predators more efficiently. Although extensively studied in bony fishes, no such empirical evidence exists for cartilaginous fishes. Our aim in this study was to experimentally investigate the social learning capabilities of juvenile lemon sharks, Negaprion brevirostris. We designed a novel food task, where sharks were required to enter a start zone and subsequently make physical contact with a target in order to receive a food reward. Naive sharks were then able to interact with and observe (a) pre-trained sharks, that is, ‘demonstrators’, or (b) sharks with no previous experience, that is, ‘sham demonstrators’. On completion, observer sharks were then isolated and tested individually in a similar task. During the exposure phase observers paired with ‘demonstrator’ sharks performed a greater number of task-related behaviours and made significantly more transitions from the start zone to the target, than observers paired with ‘sham demonstrators’. When tested in isolation, observers previously paired with ‘demonstrator’ sharks completed a greater number of trials and made contact with the target significantly more often than observers previously paired with ‘sham demonstrators’. Such experience also tended to result in faster overall task performance. These results indicate that juvenile lemon sharks, like numerous other animals, are capable of using socially derived information to learn about novel features in their environment. The results likely have important implications for behavioural processes, ecotourism and fisheries.     

Bayesian adaptive estimation of arbitrary points on a psychometric function

Bayesian adaptive methods have been extensively used in psychophysics to estimate the point at which performance on a task attains arbitrary percentage levels, although the statistical properties of these estimators have never been assessed. We used simulation techniques to determine the small‐sample properties of Bayesian estimators of arbitrary performance points, specifically addressing the issues of bias and precision as a function of the target percentage level. The study covered three major types of psychophysical task (yes‐no detection, 2AFC discrimination and 2AFC detection) and explored the entire range of target performance levels allowed for by each task. Other factors included in the study were the form and parameters of the actual psychometric function Ψ, the form and parameters of the model function M assumed in the Bayesian method, and the location of Ψ within the parameter space. Our results indicate that Bayesian adaptive methods render unbiased estimators of any arbitrary point on ψ only when M=Ψ, and otherwise they yield bias whose magnitude can be considerable as the target level moves away from the midpoint of the range of Ψ. The standard error of the estimator also increases as the target level approaches extreme values whether or not M=Ψ. Contrary to widespread belief, neither the performance level at which bias is null nor that at which standard error is minimal can be predicted by the sweat factor. A closed‐form expression nevertheless gives a reasonable fit to data describing the dependence of standard error on number of trials and target level, which allows determination of the number of trials that must be administered to obtain estimates with prescribed precision.     

Psychophysics with children: Investigating the effects of attentional lapses on threshold estimates

When assessing the perceptual abilities of children, researchers tend to use psychophysical techniques designed for use with adults. However, children’s poorer attentiveness might bias the threshold estimates obtained by these methods. Here, we obtained speed discrimination threshold estimates in 6- to 7-year-old children in UK Key Stage 1 (KS1), 7- to 9-year-old children in Key Stage 2 (KS2), and adults using three psychophysical procedures: QUEST, a 1-up 2-down Levitt staircase, and Method of Constant Stimuli (MCS). We estimated inattentiveness using responses to “easy” catch trials. As expected, children had higher threshold estimates and made more errors on catch trials than adults. Lower threshold estimates were obtained from psychometric functions fit to the data in the QUEST condition than the MCS and Levitt staircases, and the threshold estimates obtained when fitting a psychometric function to the QUEST data were also lower than when using the QUEST mode. This suggests that threshold estimates cannot be compared directly across methods. Differences between the procedures did not vary significantly with age group. Simulations indicated that inattentiveness biased threshold estimates particularly when threshold estimates were computed as the QUEST mode or the average of staircase reversals. In contrast, thresholds estimated by post-hoc psychometric function fitting were less biased by attentional lapses. Our results suggest that some psychophysical methods are more robust to attentiveness, which has important implications for assessing the perception of children and clinical groups.

     

Touch dominates haptic estimates of discordant visual-haptic size

Observers (72 college students) estimated the size of plastic squares that they held in their fingers and simultaneously viewed through a reducing lens that halved the squares’ visual size. The squares were grasped from below through a cloth that prevented direct sight of the hand. Each estimate was a match selected later, either haptically or visually, from a set of comparison squares. Vision dominated the visual estimates and touch dominated the haptic estimates, whether or not the observers knew in advance which type of estimate they would be asked to make. Neither modality inherently dominates perceived size.     

Effects of alternating set for speed or accuracy on response time,accuracy and confidence in a unidimensional discrimination task

Three approaches to a theoretical analysis of confidence judgments are considered: one linking confidence to the number of sensory observations, one based on a distinction between ‘state’ and ‘process’ factors, and a ‘balance of evidence’ hypothesis developed from an accumulator model of discrimination. An experiment is described in which observers were asked to decide which of two parallel lines was the longer, and to rate their confidence in each judgment. Each observer's set for speed or accuracy was manipulated over successive blocks of trials, and effects on response time, accuracy, and confidence were examined. Under an accuracy set, observers produced a higher percentage of correct responses, had longer response times, and made more confident judgments than under a set for speed. Within both speed and accuracy blocks, however, confidence ratings were inversely related to response time. The data on response accuracy, time, and confidence indicate certain deficiencies in either of the first two approaches, but were well accounted for by the third.     

Training Observers to Detect Deception Effects of Self-Monitoring and Rehearsal

It was predicted that trained observers would detect deception more accurately than untrained observers. More specifically, it was predicted that the highest deception detection accuracy would be found among trained observers judging the veracity of low self-monitors and unrehearsed liars, whereas the lowest detection accuracy would be found among untrained observers judging the veracity of high self-monitors and rehearsed deceivers. It also was hypothesized that the discrepancy between observers‘actual ability to detect deception and their certainty in the accuracy of their judgments would be smaller for trained observers than for untrained observers. Observers trained to detect deception used six behavioral cues based on research by deTurck and Miller (1985): (a) message duration, (b) response latency, (c) adaptors, (d) pauses, (e) nonfluencies, and (f) hand gestures. Results confirmed both hypotheses.     

How the Eyes Tell Lies: Social Gaze During a Preference Task

Social attention is thought to require detecting the eyes of others and following their gaze. To be effective, observers must also be able to infer the person's thoughts and feelings about what he or she is looking at, but this has only rarely been investigated in laboratory studies. In this study, participants' eye movements were recorded while they chose which of four patterns they preferred. New observers were subsequently able to reliably guess the preference response by watching a replay of the fixations. Moreover, when asked to mislead the person guessing, participants changed their looking behavior and guessing success was reduced. In a second experiment, naïve participants could also guess the preference of the original observers but were unable to identify trials which were lies. These results confirm that people can spontaneously use the gaze of others to infer their judgments, but also that these inferences are open to deception.     

Estimation of lifted weight and produced effort through perception of point-light display

It has been shown that human observers can estimate the weight of a box from the observation of a point-light display of a lifting motion. We asked observers to report the weight of the box and the effort produced by five lifters ranging in size and strength to determine if observers can perceive lifter size. In experiment 1, five or six weights from each of five lifters were shown to fourteen observers in a random order. Observers showed less error in estimating the amount of effort each lifter produced than in estimating the actual weight of the box. In experiment 2, the lifters were presented individually to forty observers to remove any effect observing a previous lift might have had on estimating the subsequent lift by a different lifter. The results showed an improvement in estimated weight but not in estimated effort. In experiment 3, the actual size of the lifters was given to thirty-four observers, and the estimations of both weight and effort improved. In experiment 4, observers did not improve when observing practice trials and estimating either only weight or only effort. The results from the four experiments suggest that observers are more sensitive to lifter's effort than to the weight lifted, and that observers tend to use changes in the velocity profile of the lift when making their estimates.     

Interviewing children and adults: the effect of question format on the tendency to speculate

In formal interviews it is important that interviewees indicate when they do not know the answer, rather than speculate. In this study we investigated whether question format affected the tendency to speculate. One hundred and twenty‐eight 5‐ to 9‐year‐olds, and 23 adults, were told two short stories, and were then asked questions about the stories. Half of the questions were answerable based on the information provided; the other half were not answerable. Within these categories, half of the questions were closed questions (i.e. only required a yes/no response), and half were wh‐questions (i.e. requested particular details to be provided). All participants performed at ceiling with the answerable questions. With the unanswerable questions, there was an effect of format. The majority of children and adults correctly indicated that they did not know the answer when asked unanswerable wh‐questions. However, the majority of children, and just over one‐fifth of adults, provided a response (i.e. ‘yes’ or ‘no’) to the closed unanswerable questions. The implications for interviews, particularly within a forensic context, are discussed. Copyright © 2001 John Wiley & Sons, Ltd.