Why do categories affect stimulus judgment?

The authors tested a model of category effects on stimulus judgment. The model holds that the goal of stimulus judgment is to achieve high accuracy. For this reason, people place inexactly represented stimuli in the context of prior information, captured in categories, combining inexact fine-grain stimulus values with prior (category) information. This process can be likened to a Bayesian statistical procedure designed to maximize the average accuracy of estimation. If people follow the proposed procedure to maximize accuracy, their estimates should be affected by the distribution of instances in a category. In the present experiments, participants reproduced one-dimensional stimuli. Different prior distributions were presented. The experiments verified that people's stimulus estimates are affected by variations in a prior distribution in such a manner as to increase the accuracy of their stimulus reproductions.     

Children use categories to maximize accuracy in estimation

The present study tests a model of category effects upon stimulus estimation in children. Prior work with adults suggests that people inductively generalize distributional information about a category of stimuli and use this information to adjust their estimates of individual stimuli in a way that maximizes average accuracy in estimation (see Huttenlocher, Hedges & Vevea, 2000). However, little is known about the developmental origin of this cognitive process. In the present study, 5- and 7-year-old children viewed stimuli that varied in size and reproduced each from memory. Consistent with the predictions of a Bayesian model of category effects on estimation, responses were adjusted toward the central value of the stimulus distribution. Additionally, the dispersion of the stimulus distribution affected the pattern of bias and variability of responses in a way that is predicted by the model. The results suggest that, like adults, children use categories for increasing average accuracy in estimating inexact stimuli.     

不同学习方式下归类不确定时的特征推理

采用学习-迁移模式, 探讨了同时学习和继时学习两种方式下归类不确定时的特征推理。共包括2个实验, 其中实验1探讨了固定学习轮次的情况, 实验2探讨了固定学习正确率的情况。实验结果表明：同时呈现类别要素的同时学习方式下, 被试习得序列式的单类别表征(原型表征), 在归类不确定时的特征推理中按照“单类的Bayesian规则”进行特征推理, 即P(j\F) =P(k\F)·P(j\k); 相继呈现类别要素的继时学习方式下, 被试习得并列式的多类别表征, 在归类不确定时的特征推理中按照“理性模型”进行推理, 即 P(j\F) =Σk P(k\F)·P(j\k)。     

Estimating stimuli from contrasting categories: truncation due to boundaries

In this article, the authors present and test a formal model that holds that people use information about category boundaries in estimating inexactly represented stimuli. Boundaries restrict stimuli that are category members to fall within a particular range. This model posits that people increase the average accuracy of stimulus estimates by integrating fine-grain values with boundary information, eliminating extreme responses. The authors present 4 experiments in which people estimated sizes of squares from 2 adjacent or partially overlapping stimulus sets. When stimuli from the 2 sets were paired in presentation, people formed relative size categories, truncating their estimates at the boundaries of these categories. Truncation at the boundary of separation between the categories led to exaggeration of differences between stimuli that cross categories. Yet truncated values are shown to be more accurate on average than unadjusted values.     

归类不确定情境下特征预测的理性模型研究

探讨了Bayesian规则计算公式的具体含义,进而考察了归类不确定情境下的特征预测是否符合Bayesian规则.包括2个实验:实验1探讨保持类别内目标成员中预测特征的比例不变,而改变类别内所有成员中预测特征的比例是否影响特征预测.实验2探讨保持类别内目标成员中预测特征的比例不变,进一步扩大类别内所有成员中预测特征比例的差异是否会影响特征预测.结果表明: 类别内目标成员中预测特征的比例影响特征预测,而类别内所有成员中预测特征的比例不影响特征预测.据此,本研究认为,Murphy对Bayesian规则计算公式的含义理解有偏差,其由此提出的"单类说"值得商榷.     

不同维度特征的共存对归类不确定性特征推理的影响

探讨在归类不确定的情境下,目标特征和预测特征的共存对特征推理的影响。共包括了三个实验,其中实验1、2考察了非靶类别中目标特征和预测特征的共存性对特征推理的影响,实验3考察了靶类别中目标特征和预测特征的共存性对特征推理的影响。三个实验五个分实验的结果一致支持了修正后的Bayesian规则,排除了“单类说”、“综合条件概率模型”的假设,并进一步修正了Bayesian规则为：     

Spatial categories and the estimation of location

Four experiments are reported in which people organize a space hierarchically when they estimate particular locations in that space. Earlier work showed that people subdivide circles into quadrants bounded at the vertical and horizontal axes, biasing their estimates towards prototypical diagonal locations within those spatial categories (Psychological Review 98 (1991) 352). In this work Huttenlocher, Hedges, and Duncan showed that the use of such spatial categories can increase the accuracy of estimation of inexactly represented locations. The stimulus locations we examined were uniformly distributed across the circle. In the present study we explore whether variation in the distribution of locations affects how the circle is categorized. Other things being equal, categories that capture high density regions in a stimulus space should contribute most to accuracy of estimation. However, precision of boundaries is also important to accuracy; with imprecise boundaries stimuli may be misclassified, leading to large errors in estimation. We found that people use the same spatial categories regardless of the distribution of the locations. We argue that this spatial organization nevertheless can maximize the accuracy of estimates because vertical and horizontal category boundaries are the most exact, minimizing misclassification of stimuli.     

Learning overhypotheses with hierarchical Bayesian models

Inductive learning is impossible without overhypotheses, or constraints on the hypotheses considered by the learner. Some of these overhypotheses must be innate, but we suggest that hierarchical Bayesian models can help to explain how the rest are acquired. To illustrate this claim, we develop models that acquire two kinds of overhypotheses--overhypotheses about feature variability (e.g. the shape bias in word learning) and overhypotheses about the grouping of categories into ontological kinds like objects and substances.     

The relationship between estimates of foreperiod duration and simple time reaction

The relationship between estimates of foreperiod length and simple reaction time (RT) was examined using constant foreperiods. It was learned that neither the average accuracy nor the variability of the foreperiod estimates necessarily manifested themselves in appropriate simple RT change. Of secondary interest was the discovery that the average accuracy of foreperiod duration estimations differed greatly from that obtained for similar time lengths in a time-estimation task. Finally, an attempt was made to determine whether subjects actually utilized the directional knowledge of results inherent in the imperative stimulus to influence their foreperiod duration judgments. Unfortunately, the evidence obtained did not permit a definite answer regarding this possibility.     

维度的结合与分离对归类不确定性预测的影响

探讨在归类不确定的情境下目标与预测特征两个维度的结合或分离对被试特征预测的影响。共包括 3个实验 :实验 1在Murphy和Ross的研究的基础上进一步提高非靶类型中目标及预测特征的基本概率 ,考察被试的特征预测是否会受非靶类型信息的影响。实验 2探讨非靶类型的目标与预测特征结合与否是否会影响被试预测特征时对非靶类型信息的使用。实验 3探讨提高靶类型中目标与预测特征结合的比例是否影响被试对特征的预测。结果表明 :当非靶类型中目标与关键特征处于分离的状态时 ,被试在进行特征预测时没有利用非靶类型的信息 ,符合单类说的假设 ;而当非靶类型中目标与关键特征结合时 ,被试在进行特征预测时则会利用非靶类型的信息 ,符合Bayesian规则 ;靶类型中的目标与关键特征结合的比例提高 ,被试对特征预测的概率也随之提高。据此 ,本研究将目标与预测特征结合比例这个变量加入Bayesian规则的计算公式 ,对该预测模型进行了修正     

Information measures in perceptual experiments: Some pitfalls encountered by Kintz,Parker & Boynton

The use of information transmission to calculate optimum numbers of response or stimulus categories has the severe disadvantage that empirical information measures are biased estimates of their true values, and the extent of the bias usually depends on the number of stimulus and response categories in use. By neglecting bias, Kintz, Parker, and Boynton have overestimated maximum transmission in their experiments, have wrongly identified the optimal number of response categories, and, by implication, have mistaken the stimulus conditions likely to give greatest transmission.     

Dynamic instability of visuospatial images

Five experiments using a visuospatial task were conducted to study memory accuracy and variability and to identify the origin of variations in steady states. This research was conducted from a dynamical perspective, that is, by analyzing the temporal course of discrepancies between the perceptual configuration and its memory (accuracy) and the temporal course of discrepancies between 2 successive memories (variability). In Experiment 1 the stimulus (12 black dots randomly disposed) was presented repeatedly to assess the general evolution of accuracy and variability. In Experiments 2 and 3 memory accuracy and memory variability were separated to identify their relationship. In Experiments 4 and 5 memory variability was studied to determinate the origin of steady state variations. Results show that memory accuracy and memory variability evolved independently and that memory variability reached a threshold that was subject-dependent. The dynamic properties of image construction and stability are discussed.     

Seeing Like a Geologist: Bayesian Use of Expert Categories in Location Memory

Memory for spatial location is typically biased, with errors trending toward the center of a surrounding region. According to the category adjustment model (CAM), this bias reflects the optimal, Bayesian combination of fine‐grained and categorical representations of a location. However, there is disagreement about whether categories are malleable. For instance, can categories be redefined based on expert‐level conceptual knowledge? Furthermore, if expert knowledge is used, does it dominate other information sources, or is it used adaptively so as to minimize overall error, as predicted by a Bayesian framework? We address these questions using images of geological interest. The participants were experts in structural geology, organic chemistry, or English literature. Our data indicate that expertise‐based categories influence estimates of location memory—particularly when these categories better constrain errors than alternative (“novice”) categories. Results are discussed with respect to the CAM.     

Feature saliency affects delayed matching of an attended feature

We examined how top-down attentional modulation and bottom-up stimulus saliency interact with feature memory. Experiment 1 used a delayed-matching-to-sample (DMS) task to examine the relative saliency between features by observing the relative accuracy of recognition at different stimulus durations. Feature salience decreased according to the following order: colour, form, and texture. In a modified DMS task (Experiments 2 and 3), participants were required to attend to one of three features and ignore the others. After a delay, they were required to choose which of the two test stimuli matched the reference stimulus on the attended feature, disregarding other task-irrelevant features. The target was either identical to the reference stimulus or mismatched the reference stimulus on one of the irrelevant features. The results showed that colour matching was affected neither by a form change nor by a texture change. Form matching was affected by a colour change, and texture matching was affected by a colour or form change. These results are consistent with the relative saliency hypothesis. Even when all features of an attended object are maintained, a relatively more salient task-irrelevant feature can interfere with the delayed recognition of a less salient feature.     

Temporal stimulus-response compatibility

The study of element-level stimulus-response compatibility (SRC) has predominantly focused on spatial and symbolic relationships and has involved measures of response time and (dichotomous) error rate. This article explores a new form of SRC that is observed when duration is the relevant feature of both the stimulus and the response, using a more extensive analysis of performance accuracy and variability. The results indicate that element-level SRC generalizes to situations involving time as the relevant dimension of stimuli and responses. Evidence of this was found in all of the extracted measures of performance; however, temporal SRC was shown to have independent effects on when and how accurately a response was made. Implications for SRC research are discussed.     

Is memory for stimulus magnitude Bayesian?

This study was designed to determine whether memory for stimulus values is a Bayesian weighting of the magnitude of a stimulus and the central tendency of an exemplar's category (Huttenlocher, Hedges, & Vevea, 2000). In five experiments, participants reproduced the remembered size of a geometric figure drawn from one of two categories whose means for size differed. Reproductions were biased toward the mean of the combined distribution rather than the mean of either category. Reproductions were also influenced by the size of the stimulus on the preceding trial. Neither of these results is entirely consistent with the view that recollections are partially constructed from a consideration of the long-run probabilities established by category membership.     

A comparison of stimulus preference assessment methods in adults with schizophrenia

Three methods of assessing preference for stimuli were compared in four adults with a diagnosis of schizophrenia. During phase 1, a survey method, a verbal stimulus choice method, and a pictorial stimulus choice method of assessing preference for four categories of stimuli were administered. During phase 2, a coupon system was used to determine which categories of stimuli actually functioned as reinforcers for each participant. Comparisons between the three preference assessment methods were then conducted based on the results of the reinforcer assessment. Results showed that, overall, there were few differences in total accuracy among the preference assessment procedures. Copyright © 2003 John Wiley & Sons, Ltd.     

Further attempts to clarify the importance of category variability for categorisation

The issue of how category variability affects classification of novel instances is an important one for assessing theories of categorisation, yet previous research cannot provide a compelling conclusion. In five experiments we reexamine some of the factors thought to affect participant performance. In Experiments 1 and 2, participants almost always classified the test item as belonging to the high variability category. By contrast, in Experiment 3 we employed an alternative experimental paradigm, where the difference in variability of the two categories was less salient. In that case, participants tended to classify a test item as belonging to the low variability category. Two additional experiments (4 and 5) explored in detail the differences between Experiments 1, 2 on the one hand, and 3 on the other. Some insight into the underlying psychological processes can be provided by computational models of categorisation, and we focus on the continuous version of Anderson's (1991) Rational Model, which has not been explored before in this context. The model predicts that test instances exactly halfway between the prototypes of two categories should be classified into the more variable category, consistent with the bulk of empirical findings. We also provided a comparison with a slightly reduced version of the Generalised Context Model (GCM) to show that its predictions are consistent with those from the Rational Model, for our stimulus sets.     

类别不确定下的特征推理是基于类别还是基于特征联结

共有3个实验探讨归类不确定情况下的特征推理是基于类别进行还是基于特征联结进行。实验1在中文条件下重复了Verde等人2005的实验,得出了与之相符的结果,这个结果用基于类别的理性模型的设想或者是用基于特征联结的设想都可以解释。实验2考察被试在靶类别的类别特征频次并且特征结合出现频次高低不同的条件下特征推理的情况,实验2的结果表明,高集中与低集中两种条件下特征推理没有显著差异,不符合特征推理是基于类别进行的设想,而与特征推理是基于特征联结进行的设想吻合。实验3进一步考察被试在特征结合出现的总频次并且靶类别中特征结合出现的总频次高低不同的条件下特征推理的情况,结果表明,在高结合条件下进行特征推理要优于在低结合条件,支持了在归类不确定情况下的特征推理是基于特征联结进行的设想。据此可以认为,人们的特征推理是基于特征之间联结的频次进行,而不是基于类别进行     

Stimulus congruence and the Simon effect

Summary A standard experimental procedure was implemented with novel response requirements to assess the hypothesis that the Simon effect is attributable not to the irrelevant stimulus-response relationship, but to the congruence between stimulus attributes. The stimulus ensemble consisted of the words LEFT and RIGHT, one of which was presented on each trial to the left or right of a central fixation point. The distinctive feature of the task is that subjects were asked to respond, by laterally placed keys, whether or not the stimulus word was in accord (i. e., congruent) with its location on the display. Asking subjects to judge stimulus congruence directly enables the effect of congruence to be assessed, as well as independent estimates of the two irrelevant S-R relationships that apply in the task — that is, between the response location and (1) the stimulus location (the Simon effect) and (2) the stimulus word (the reverse Simon effect). Marked effects were obtained in all three cases. Stimulus congruence remains in contention as a factor in the explanation of the Simon effect, but the strong effects of irrelevant S-R factors suggest that a more comprehensive account of the Simon effect is needed. An explanation in terms of stimulus salience, based on an interactive activational model, is briefly discussed.