Base rates versus sample accuracy: competition for control in human matching to sample

People often place undue weight on specific sources of information (case cues) and insufficient weight on more global sources (base rates) even when the latter are highly predictive, a phenomenon termed base-rate neglect. This phenomenon was first demonstrated with paper-and-pencil tasks, and also occurs in a matching-to-sample procedure in which subjects directly experience case sample (cue) accuracy and base rates, and in which discrete, nonverbal choices are made. In two nonverbal experiments, subjects were exposed to hundreds of trials in which they chose between two response options that were both probabilistically reinforced. In Experiment 1, following one of two possible samples (the unpredictive sample), either response was reinforced with a .5 probability. The other sample (predictive) provided reinforcement for matching on 80% of the trials in one condition but in only 20% of the trials in another condition. Subjects' choices following the unpredictive sample were determined primarily by the contingencies in effect for the predictive sample: If matching was reinforced following the predictive sample, subjects tended to match the unpredictive sample as well; if countermatching the predictive sample was generally reinforced, subjects tended to countermatch the unpredictive sample. These results demonstrate only weak control by base rates. In Experiment 2, base rates and sample accuracy were simultaneously varied in opposite directions to keep one set of conditional probabilities constant. Subjects' choices were determined primarily by the overall accuracy of the sample, again demonstrating only weak control by base rates. The same pattern of choice occurred whether this pattern increased or decreased rate of reinforcement. Together, the results of the two experiments provide a clear empirical demonstration of base-rate neglect.     

未意识到错误影响错误后调整的电生理证据

现有研究一致认为意识到错误可引起错误后调整, 但是未意识到错误能否促使个体进行错误后调整尚存争议。本实验采用基于go/no-go范式的错误意识任务考察上述问题, 并根据被试对自己按键反应正误主观报告将no-go错误反应分为意识到错误和未意识到错误。行为结果发现, 意识到错误和未意识到错误后正确率均显著高于正确击中试次(正确go试次)后正确率; 但是, 意识到错误后试次反应时显著快于正确击中后反应时, 未意识到错误后反应时显著慢于正确击中后反应时。该结果表明两类错误均优化了错误后行为表现, 但是意识到错误后被试调整速度加快, 未意识到错误后被试调整速度减慢。进而, 时频分析发现意识到错误相较于未意识到错误诱发显著更强的alpha波能量。并且, 前者在错误意识主观报告前已诱发alpha波, 后者在错误意识主观报告反应后诱发alpha波。该结果表明意识到错误一直处于持续的注意监控中, 而未意识到错误是任务引起的暂时注意控制。因此, 本实验说明错误意识影响错误后调整, 意识到错误可能采用类似主动性控制的策略调节错误后行为, 而未意识到错误可能采用类似反应性控制的策略调节错误后行为。     

Why Do Eyewitnesses Make So Many Mistakes?

Previous research on eyewitness identification has demonstrated high rates of error. Subjects have frequently identified innocent targets as the "criminal" they had seen earlier (false identifications) or had falsely claimed that the criminal was not in the line-up (misses). The present study examines whether identification error rates are inflated by pressures in the typical experimental situation to "make a guess" regardless of one's confidence in the accuracy of the response. It was found that providing an explicit option for subjects to respond "don't know" significantly decreased false identifications and misses with no cost to the proportion of correct identifications. The addition of written and verbal instructions emphasizing the acceptability of the "don't know" option produced a marginally significant further decline in identification errors, again without cost to correct identifications. The discussion considered implications of the present results for experimental and actual police line-up procedures.     

Evidence for transliminality from a subliminal card-guessing task

In this experiment we sought to provide evidence for transliminality from a test of subliminal perception that was disguised as a computerised ESP card-guessing task. It was predicted that highly transliminal individuals would outperform those with low levels of transliminality when given subliminal primes or 'clues' to the correct choice of card, but not when no primes were given. In line with the prediction, higher levels of transliminality were found to be associated with a greater number of correct selections of the target card on the primed trials, but not on the unprimed trials. In addition, a positive correlation was obtained between transliminality and detection accuracy, suggesting that higher levels of transliminality are associated with a greater sensitivity to visual stimulation. The results are discussed with reference to the possibility that transliminality might offer an alternative explanation for some ostensibly psychic perceptual experiences if subliminally acquired material is wrongly attributed to psychic sources.     

Toward a unified theory of decision criterion learning in perceptual categorization

Optimal decision criterion placement maximizes expected reward and requires sensitivity to the category base rates (prior probabilities) and payoffs (costs and benefits of incorrect and correct responding). When base rates are unequal, human decision criterion is nearly optimal, but when payoffs are unequal, suboptimal decision criterion placement is observed, even when the optimal decision criterion is identical in both cases. A series of studies are reviewed that examine the generality of this finding, and a unified theory of decision criterion learning is described (Maddox & Dodd, 2001). The theory assumes that two critical mechanisms operate in decision criterion learning. One mechanism involves competition between reward and accuracy maximization: The observer attempts to maximize reward, as instructed, but also places some importance on accuracy maximization. The second mechanism involves a flat-maxima hypothesis that assumes that the observer's estimate of the reward-maximizing decision criterion is determined from the steepness of the objective reward function that relates expected reward to decision criterion placement. Experiments used to develop and test the theory require each observer to complete a large number of trials and to participate in all conditions of the experiment. This provides maximal control over the reinforcement history of the observer and allows a focus on individual behavioral profiles. The theory is applied to decision criterion learning problems that examine category discriminability, payoff matrix multiplication and addition effects, the optimal classifier's independence assumption, and different types of trial-by-trial feedback. In every case the theory provides a good account of the data, and, most important, provides useful insights into the psychological processes involved in decision criterion learning.     

Conditioned reinforcement and discrimination performance

Pigeons were trained in a three-key chamber to peck one side key in the presence of a vertical line on the center key and to peck the other side key in the presence of a horizontal line. Correct choice responses were reinforced with food according to fixed- and variable-ratio, fixed-interval, and differential-reinforcement-of-long-latency schedules of reinforcement. For each schedule, the birds performed under each of two conditions: (1) each correct choice response produced a brief presentation of stimuli intermittently paired with food, then the next trial; (2) each correct choice response produced an intertrial interval only. For all schedules except one long latency schedule, response rates were higher under the condition of brief stimulus presentation than under the comparable control condition. Presentation of brief magazine stimuli increased choice accuracy. The amount of change in accuracy was correlated with the rate of food presentation. Performance under the schedules with highest food reinforcement rates showed no enhancement; performance under the schedules with the lowest reinforcement rates showed the greatest enhancement.     

Covert verbal mediation in arbitrary matching to sample

Covert verbal mediation was examined in an arbitrary matching‐to‐sample (MTS) preparation with a high‐verbal group (college students) and a low‐verbal group (adults with intellectual disabilities). Arbitrary relations were established between nonsense words, visual symbols, objects, and hand signs. Task difficulty was balanced for the groups based on errors during acquisition. All participants experienced a hand sign condition, and three MTS conditions each with a unique configuration of the comparison array: fixed location, random location, and all symbols the same. The same symbol condition was designed to impede a participant's ability to label individual symbols. The results showed that disrupting labeling adversely affected MTS performance for high‐verbal participants, but not for low‐verbal participants. The data suggest that high‐verbal participants depended on mediating verbal behavior and joint control to assist them in finding the correct comparison stimulus. Low‐verbal participants could not benefit from verbal mediating variables and likely relied on unmediated contingencies, or some form of nonverbal mediation. For the high‐verbal group, 19 different putative emergent relations were identified as occurring at various stages of acquisition between the sample stimulus and the selection response. These emergent relations likely provided supplementary sources of stimulus control that participated in evoking MTS selection behavior.     

Asymmetric attributions for compliance: Reward vs punishment

It was proposed that people attribute an individual's behavior more to internal factors when that individual's actions are influenced by reward than when those actions are influenced by punishment. Previous research has failed to control for the power of reward versus punishment which, in effect, creates a confounding of behavioral base rates (consensus) with the reward-punishment manipulation. The current research created reward and punishment contingencies that were equal in their base rates for producing a compliant response. In Experiment 1, subjects (n = 63) who produced the base-rate data also made attributions regarding a compliant target person. The results supported the reward-punishment attributional asymmetry hypothesis in that the target person was held more responsible for his actions in the reward than in the punishment conditions. A second experiment (n = 72) provided some attributors with information regarding base rates for compliance and measured perceived base rates for compliance. Knowledge of the base rates for compliance eliminated the reward-punishment attributional asymmetry phenomenon. Subjects not provided with such knowledge erroneously assumed different base rates for reward and punishment and maintained the perception of reward-punishment attributional asymmetry. Using subjects' estimates of base rate for compliance as a covariate eliminated the attributional asymmetry effect. It is suggested that erroneous base-rate assumptions mediate the attributional asymmetry phenomenon.     

Optimal classifier feedback improves cost-benefit but not base-rate decision criterion learning in perceptual categorization

Unequal payoffs engender separate reward- and accuracy-maximizing decision criteria; unequal base rates do not. When payoffs are unequal, observers place greater emphasis on accuracy than is optimal. This study compares objective classifier (the objectively correct response) with optimal classifier feedback (the optimal classifier's response) when payoffs or base rates are unequal. It provides a critical test of Maddox and Bohil's (1998) competition between reward and accuracy maximization (COBRA) hypothesis, comparing it with a competition between reward and probability matching (COBRM) and a competition between reward and equal response frequencies (COBRE) hypothesis. The COBRA prediction that optimal classifier feedback leads to better decision criterion leaning relative to objective classifier feedback when payoffs are unequal, but not when base rates are unequal, was supported. Model-based analyses suggested that the weight placed on accuracy was reduced for optimal classifier feedback relative to objective classifier feedback. In addition, delayed feedback affected learning of the reward-maximizing decision criterion.     

The effect of presentation format and encoding strategy on associative memory distortion

IntroductionPrevious studies using semantically related words revealed more accurate memory when the items were encoded visually rather than auditorily and when mental images were created during encoding. However, how the level of memory distortion is affected by the creation of different mental imagery formats or by techniques that should suppress generation of mental images has rarely been investigated.ObjectiveThe aim of the present studies was to investigate the ways in which the encoding strategy affects the accuracy of memory reports for two presentation formats of semantically related words: verbal and pictorial.MethodIn experiment 1, the participants were asked to memorize either pictures or their verbal equivalents (words) from the same category, using one of two encoding strategies: uttering the words or counting backwards. In experiment 2, pictorially or auditory presented material was encoded together with the creation of either visual or auditory mental images of the items. The results of the experimental groups were compared to control groups that received no specific instruction.ResultsHigher levels of false recognition, together with lower rates of correct recognition, were observed for words, presented either visually or auditory, relative to pictures. Moreover, self-generation of additional code during the processing of information favored the reduction of false recognitions.ConclusionEncoding strategies that engaged dual coding reduced false recognition. The results are discussed within the distinctiveness heuristic phenomenon.     

Children’s eyewitness memory: A comparison of two interviewing strategies as realized by forensic professionals

A critical issue for developmental psychology is how to obtain accurate and complete eyewitness memory reports from preschoolers without offering suggestions that might result in false allegations. We examined effects of two interviewing strategies (police/verbal interviews and clinician/prop-assisted interviews) on young children’s reports about a medical examination. A total of 58 4-year-olds participated in the study, which conformed to a 2 (Interview Type) × 2 (Number of Interviews) factorial design. Analyses revealed that interviewers spent less time off topic and asked more free recall questions in the police/verbal interviews than in the clinician/prop-assisted interviews. Compared with police/verbal interviews, clinician/prop-assisted interviews resulted in significantly more correct rejections and commission errors in children’s memory reports. However, on a final free recall test, error rates were comparable across conditions. Higher child verbal intelligence predicted memory accuracy in police/verbal interviews, and greater parental attachment anxiety predicted children being asked a higher number of misleading questions. The study provides new insights into interview techniques that promote preschoolers’ accurate eyewitness reports.     

Multivariate repeated measures designs

In the social, behavioral, and health researches it is a common strategy to collect data along time on more than one group of participants on multiple dependent variables. To analyse this kind of data is very complicated due to the correlations between the measures taken in different points of the time, and between the answers. Usually to analyse these data the multivariate mixed model, or the doubly multivariate model, are the most frequent approaches. Both of them require combined multivariate normality, equal covariance matrices, independence between the observations of different participants, complete measurements on all subjects, and time-independent covariates. When one ore more of these assumptions are not accomplished these approaches do not control in the correct way the Type I error, and this affects the validity and the accuracy of the inferences. In this paper some solutions that solve the problems with the error Type I will be shown. Several programs for a correct realization of the analyses through the SAS Proc Mixed procedure are presented.     

“I think you did it!”: Examining the effect of presuming guilt on the verbal output of innocent suspects during brief interviews

Innocent suspects interviewed by a guilt‐presumptive versus innocence‐presumptive or neutral interviewer may tend more to display non‐verbal behaviours which neutral judges consider indicative of guilt. We examined the effects of interviewer's presumption of guilt on innocent mock suspects' alibis. Participants (N = 90) provided an alibi to convince an interviewer of their innocence of a theft after she implied that she believed that they were guilty or innocent or that she had no belief about their veracity. On the basis of existing conflicting findings for suspects' verbal behaviour during accusatory interviews, we predicted that alibis in the guilt‐belief condition would contain the highest or lowest number of correct details with overall higher or poorer accuracy rates, respectively. Although participants perceived the interviewer's presumptive approach, the number of correct details provided and accuracy rates of alibis did not differ significantly between conditions. We propose explanations to these findings and future research paths.     

Distinguishing choice and subjective probability estimation processes: Implications for theories of judgment and for cross-cultural comparisons

Research on subjective probability (SP) often focuses on the difference between the fraction of correct binary choices and the SP that they are correct or on the deviation of the calibration curve from the diagonal. On the assumption that choice and SP judgments reflect different, albeit closely related, processes, it is important to analyze the two sources of data separately. This article uses data collected by Yates, Lee, Shinotsuka, Patalano, and Sieck (1998) from Chinese, Japanese, and Americans to demonstrate the advantages of doing so and to yield new conclusions regarding cross-cultural comparisons of judgment. One new result is that a discrete threshold model with an error term best described the choice data of all groups. From their analyses, Yates et al. (1998) had concluded that the Chinese were more overconfident than the American or Japanese. According to the present results, the Chinese appeared either more overconfident or more underconfident, depending on how the data are analyzed. The dual outcome is a consequence of the greater variability both in their SP judgments and in their binary choices. Additionally, the Chinese participants’ choices and SPs were more responsive to the base rate than were those of the other groups.     

Effects of type and amount of feedback upon hypothesis sampling systems among 7- and 8-year-old children

Second- and third-grade children (age 7 to 9 years) received a series of discrimination-learning problems containing verbal probes for hypotheses. The children received either (a) verbal-directional feedback in which the correct stimulus complex was indicated, (b) verbal feedback in which they were told whether they were correct or wrong, or (c) material feedback in which a marble was moved by the experimenter following choice responses. The three types of feedback were combined with three amounts: right-wrong, wrong-blank (blank = no feedback), and right-blank, to yield a 3 × 3 factorial design. Children in the verbal-directional right-wrong and wrong-blank conditions showed systematic problem-solving strategies in about 80% of their problems, while those in the remaining cells showed strategies in only 32 to 43% of theirs. The most prevalent processing error among children in the latter conditions was failure to resample according to a local consistency criterion following errors.     

The acquisition of behavioral chains

The present investigation was designed to study some conditions which facilitate the acquisition of behavioral chains by young children. Three experiments were performed. In Expt I subjects were required to respond overtly to the internal components of the chain. In Expt II an aversive contingency was imposed following errors. In Expt III instructions were added at the beginning of training. The results supported the following conclusions: (a) Requiring subjects to mark and name the correct stimuli on each trial increased the accuracy of performance; (b) introducing time-out contingent on incorrect responses depressed the error rate; and (c) adding instructions concerning a pending change in the experimental task resulted in a lower error rate following withdrawal of the instructional stimuli.     

Speed-accuracy tradeoff in double stimulation: Effects on the first response

A single-stimulation and two double-stimulation response conditions were compared using explicit payoff matrices to vary speed-accuracy tradeoff. Under accuracy payoff, response latency (RT(1)) to the first stimulus increased as ISI dropped but accuracy remained high and relatively constant. Under speed payoff, RT(1) was only slightly affected by ISI but accuracy dropped as ISI decreased. Transmitted information rates consistently reflected detrimental effects of short ISI. In double stimulation, but not in single stimulation, error response latency exceeded correct response latency. Furthermore, error response latencies were found to be far more variable and more sensitive to changes in speed-accuracy condition than were correct response latencies. Finally, under both speed and accuracy conditions, response latency to the first of two successive stimuli was faster if a response was also required to the second stimulus. Implications of the data for possible models of double-stimulation speed-accuracy tradeoff are considered.