Attention, similarity, and the identification-categorization relationship

A unified quantitative approach to modeling subjects' identification and categorization of multidimensional perceptual stimuli is proposed and tested. Two subjects identified and categorized the same set of perceptually confusable stimuli varying on separable dimensions. The identification data were modeled using Shepard's (1957) multidimensional scaling-choice framework. This framework was then extended to model the subjects' categorization performance. The categorization model, which generalizes the context theory of classification developed by Medin and Schaffer (1978), assumes that subjects store category exemplars in memory. Classification decisions are based on the similarity of stimuli to the stored exemplars. It is assumed that the same multidimensional perceptual representation underlies performance in both the identification and categorization paradigms. However, because of the influence of selective attention, similarity relationships change systematically across the two paradigms. Some support was gained for the hypothesis that subjects distribute attention among component dimensions so as to optimize categorization performance. Evidence was also obtained that subjects may have augmented their category representations with inferred exemplars. Implications of the results for theories of multidimensional scaling and categorization are discussed.     

Improving Human‐Machine Cooperative Classification Via Cognitive Theories of Similarity

Acquiring perceptual expertise is slow and effortful. However, untrained novices can accurately make difficult classification decisions (e.g., skin‐lesion diagnosis) by reformulating the task as similarity judgment. Given a query image and a set of reference images, individuals are asked to select the best matching reference. When references are suitably chosen, the procedure yields an implicit classification of the query image. To optimize reference selection, we develop and evaluate a predictive model of similarity‐based choice. The model builds on existing psychological literature and accommodates stochastic, dynamic shifts of attention among visual feature dimensions. We perform a series of human experiments with two stimulus types (rectangles, faces) and nine classification tasks to validate the model and to demonstrate the model's potential to boost performance. Our system achieves high accuracy for participants who are naive as to the classification task, even when the classification task switches from trial to trial.     

The influence of monetary incentives on goal choice,goal commitment,and task performance

This study was designed to investigate the mechanism by which monetary incentives influence goal choice, goal commitment, and task performance. It is hypothesized that the explanation for incentive effects on goals and performance is in their influence on mediating events. A model attempting to explain the process of goal choice, work motivation, and performance in terms of cognitions is presented. Subjects were recruited for 5 days of part-time employment in a simulated organization to perform a clerical data transfer task. The 130 subjects were assigned randomly to 7 experimental conditions differing in terms of the magnitude of incentive offered for various levels of performance. The quantity and quality of performance was recorded daily and research questionnaires were administered at different times during their employment. Results provide support for the model presented and therefore suggest that the process of goal choice and commitment is central to understanding how incentives influence goals, motivation, and performance. The model provides a useful basis for investigating the relationships between organizational context and employee cognitions and for integrating goal setting with expectancy theory. The findings and future research issues are discussed.     

Effects of degree of interpersonal similarity on physical distance and self-reported attraction: A comparison of uniqueness and reinforcement theory predictions

The apparently divergent similarity-attraction findings of the reinforcement theory and uniqueness theory researchers were investigated by exposing male (N = 45) and female (N = 45) subjects to conditions of slight, moderate, or high similarity relative to one comparison other. Similarity was manipulated by means of a bogus personality and attitude inventory. For the self-report measures, results consistent with reinforcement theory were obtained in that greater self-reported attraction to the partner was linearly related to increased similarity. Additionally, interpersonal distance was curvilinearly related to similarity. This distancing pattern reflected the fact that moderate similarity subjects positioned themselves relatively closer to the other person's chair than the slight or high similarity subjects. The physical distance results support the uniqueness theory notion that high similarity may have some aversive properties. The discrepancy between the physical distance and self-report results is discussed in terms of an implicit causal theory that subjects may have about interpersonal similarity (i.e., greater similarity is good).     

More evidence for a relational shift in the development of analogy: Children's performance on a causal-mapping task

Gentner (1988) has proposed a relational shift whereby children interpret analogy and metaphor first in terms of object similarity and then in terms of relational similarity. Goswami (1996) argues against the relational shift hypothesis, citing as evidence a study performed by Goswami and Brown (1989) in which 3-, 4-, and 6-year-old children were able to correctly complete pictorial A:B::C:? analogies based on familiar causal relations, and, contrary to the predictions of the relational shift hypothesis, made very few object-similarity errors despite the presence of an object-similarity choice. In the present experiment we obtained similarity ratings of Goswami and Brown's stimuli and found that the materials did not contain a true object similarity choice and therefore that study was not an adequate test of the relational shift hypothesis. After appropriate modifications to their methodology, we found that 4- and 5-year-old children's performance was consistent with the relational shift hypothesis: First, object-similarity errors were highly frequent initially and decreased with age; second, the rate of relational (correct) responding increased with age; and third, performance on the analogues was positively related to children's knowledge about the participating causal relations. We conclude by proposing an explanation for the relational shift based on an alignment view of similarity comparison and, further, suggest a new role for object similarity in children's analogical development.     

Multidimensional vector model of stimulus-response compatibility

The present study proposes and examines the multidimensional vector (MDV) model framework as a modeling schema for choice response times. MDV extends the Thurstonian model, as well as signal detection theory, to classification tasks by taking into account the influence of response properties on stimulus discrimination. It is capable of accounting for stimulus-response compatibility, which is known to be an influential task variable determining choice-reaction performance but has not been considered in previous mathematical modeling efforts. Specific MDV models were developed for 5 experiments using the Simon task, for which stimulus location is task irrelevant, to examine the validity of model assumptions and illustrate characteristic behaviors of model parameters. The MDV models accounted for the experimental data to a remarkable degree, demonstrating the adequacy of the framework as a general schema for modeling the latency of choice performance. Some modeling issues involved in the MDV model framework are discussed.     

Theoretical analysis of an alphabetic confusion matrix

A study was undertaken to acquire a confusion matrix of the entire upper-case English alphabet with a simple nonserifed font under tachistoscopic conditions. This was accomplished with two experimental conditions, one with blank poststimulus field and one with noisy poststimulus field, for six Ss run 650 trials each. Three mathematical models of recognition, two based on the concept of a finite number of sensory states and one being the choice model, were compared in their ability to predict the confusion matrix after their parameters were estimated from functions of the data. In order to ascertain the facility with which estimates of similarity among the letters could lead to a psychological space containing the letters, η_ij, the similarity parameter of the choice model was input to an ordinally based multidimensional scaling program. Finally, correlation coefficients were computed among parameters of the models, the scaled space, and a crude measure of physical similarity. Briefly, the results were: (1) the finite-state model that assumed stimulus similarity (the overlap activation model) and the choice model predicted the confusion-matrix entries about equally well in terms of a sum-of-squared deviations criterion and better than the all-or-none activation model, which assumed only a perfect perception or random-guessing state following a stimulus presentation; (2) the parts of the confusion matrix that fit best varied with the particular model, and this finding was related to the models; (3) the best scaling result in terms of a goodness-of-fit measure was obtained with the blank poststimulus field condition, with a technique allowing different distances for tied similarity values, and with the Euclidean as opposed to the city-block metric; and (4) there was agreement among the models in terms of the way in which the models reflected sensory and response bias structure in the data, and in the way in which a single model measured these attributes across experimental conditions, as well as agreement among similarity ami distance measures with physical Similarity.     

Disruption of verbal STM by irrelevant speech, articulatory suppression, and manual tapping: Do they have a common source?

Under appropriate conditions, immediate serial verbal recall is impaired by irrelevant speech, articulatory suppression, and syncopated tapping. Interpretation of these variables in terms of the phonological loop component of working memory assumes separate phonological storage and articulatory rehearsal processes. In contrast, the Object-Oriented Episodic Record (O-OER) of Jones and the feature theory of Neath interpret these and other phenomena in terms of a unitary multimodal system. Three experiments investigate these disrupting tasks, with each experiment emphasizing one parameter. In each case, recall of phonologically similar and dissimilar letter sequences is compared as a marker of the presence or absence of phonological coding. In Experiment 1, subjects heard or articulated a single item, or tapped a single key at equal intervals. Only articulatory suppression impaired performance; it also abolished the effects of phonological similarity. Experiment 2 was identical, except that items were heard, or generated in a syncopated rhythm. Both suppression and tapping impaired performance to an equivalent extent and obliterated the effect of phonological similarity. Syncopated irrelevant speech caused a modest but significant impairment in performance. Experiment 3 was identical to Experiment 1, except that six tokens were used. Irrelevant speech and tapping had a clear impact on recall, but neither removed the phonological similarity effect. Again articulatory suppression had a major impact on performance and removed the effect of phonological similarity. We conclude that the pattern of results readily fits the phonological loop hypothesis, provided one accepts Saito's proposal that generating syncopated sequences uses common processes with speech production. It is not clear how the results can be explained by either the O-OER or the feature hypothesis.     

The length of the retention interval,forgetting, and subjective similarity

A long retention interval tends to result in the poor retention known as forgetting. A high subjective similarity between stimuli frequently produces their poor retention. Thus, a long retention interval may increase the subjective similarity between stimuli (the RIISS hypothesis), and this increase may produce forgetting. To examine this hypothesis, college students made speeded same-different discriminations between two lines or tones of different lengths or frequencies that were 400 ms or 3,300 ms apart, and they rated the similarity of these stimuli. The long interval produced poorer overall performance as expected, but also produced poorer performance on different than same stimuli, implying that it increased the subjective similarity between the initial and subsequent stimuli, and it also increased rated similarity, in support of the RIISS hypothesis. The position that stored stimuli lose less common information than distinctive information explains RIISS evidence better than does perturbation theory.     

Interpersonal congruency,attitude similarity,and interpersonal attraction

To examine relationships among interpersonal congruency, attitude similarity, and interpersonal attraction, the present study required pairs of subject to rate their attraction to each other in response to two proportions of similar attitudes orthogonally varied with two treatments of interpersonal congruency. Findings showed that attitude similarity determined attraction when congruency was uncontrolled, but that similarity and attraction were unrelated within the two congruency treatments. Additional correlational analyses suggest that different judgments of the accuracy of other's perceptions of oneself accompany previous findings on attitude similarity and attraction, and it is proposed that interpersonal attraction may be understood in terms of stabilizing processes postulated by the theory of interpersonal congruency.     

Factors affecting children's graphic symbol use in the third year: Language,similarity, and iconicity

The question addressed in this study is whether the claim that children understand the symbolic status of pictures by the middle of their third year is an overestimate of their ability. Specifically, we asked whether children use language if possible to facilitate their performance in graphic symbolic tasks. Language (availability of verbal labels) was manipulated along with iconicity (degree of resemblance between symbol and referent) and perceptual similarity (between choice items) in a series of four experiments. Children 2.5 and 3 years old were presented with a graphic symbol for 4 s and immediately asked to choose the object depicted (referent) from two choice objects. In Study 1, degree of iconicity between picture and referent was varied and both choice objects had the same verbal label. The 2.5-year-olds failed to use any pictures or replicas as symbols. The 3-year-olds performed well with all types of symbols and better with highly iconic symbols. In Study 2, verbal label availability was manipulated by presenting choice objects having the same or different labels and by varying familiarity of labels. The 2.5-year-olds performed at chance when verbal labels were unavailable but above chance when they were available. The 3-year-olds were above chance in all conditions but performed less well when verbal labels were unavailable. Study 3 confirmed that young children use language to mediate picture symbol use. When 2.5-year-olds were provided with subordinate verbal labels in the matching task, subsequent performance was good even when choice objects had the same basic level verbal label. In Study 4, verbal label availability was contrasted with perceptual similarity between choice objects. When verbal labels could be used and choice objects were dissimilar, performance was best, and when verbal labels could not be used and choice objects were similar, it was worst. The results suggest that children's developing understanding of the symbolic function of pictures is tenuous in the third year, and is supported by their use of verbal labels.     

Similarity, confusability, and the density hypothesis

Results from several letter- and digit-identification studies have been interpreted (Appelman & Mayzner, 1982; Krumhansl, 1982) as providing support for the hypothesis that psychological similarity is influenced by the local density of items in the stimulus space. This conclusion is questioned on the grounds that density was not directly manipulated in the studies, thus alternative explanations based on other stimulus characteristics cannot be excluded. In the present article six experiments are reported in which stimulus density was manipulated. In three experiments using similarity ratings and two using discrimination confusions, no effect of stimulus density was found. However, the identification counterpart of one of the discrimination studies did provide evidence of an effect of density on response probabilities. It is concluded that stimulus density can affect identification performance through its influence on the choice process implicit in any identification task, but is not an important determinant of psychological similarity.     

Concepts and conceptual structure

Research and theory on categorization and conceptual structure have recently undergone two major shifts. The first shift is from the assumption that concepts have defining properties (the classical view) to the idea that concept representations may be based on properties that are only characteristic or typical of category examples (the probabilistic view). Both the probabilistic view and the classical view assume that categorization is driven by similarity relations. A major problem with describing category structure in terms of similarity is that the notion of similarity is too unconstrained to give an account of conceptual coherence. The second major shift is from the idea that concepts are organized by similarity to the idea that concepts are organized around theories. In this article, the evidence and rationale associated with these shifts are described, and one means of integrating similarity-based and theory-driven categorization is outlined.     

Social comparison,attraction, and choice of a comparison other

The effects of attitudinal similarity and task-relevant affiliative status, together with three situational factors (presence or absence of a monetary reward, choice of information source or work partner, and choice made before or after Ss completed preliminary work on the task), on the choice of comparison others were examined. The most significant determinant of the comparison other chosen was attitude similarity (p < .01). There was also a significant interaction between attitude similarity, task-relevant affifliative status, and time of choice (p < .05). The nature of the interaction suggested that general impressions such as conveyed by general attitudinal information may be more important in the selection of referent others than comparability on situation-specific factors.     

Auditory detection and recognition in a two-alternative,directional uncertainty situation

The effects of directional uncertainty on the detection and recognition of auditory signals were investigated, with particular reference to predictions from the Luce-Shipley choice theory analysis. Sixteen Ss were used in a group detection and recognition task, involving auditory signals in a white noise background, from two alternative sources to the right and left of the Ss. Confidence ratings and false alarm rates were used to calculate detectability measures, and from confusion matrices signal recognition and response bias parameters were estimated. The results indicated that detectability is reduced by uncertainty of this kind, provided the initial detectability of the signal is fairly high. However, recognition was apparently unrelated to detection performance, except in very general terms. Certainly the data revealed no evidence for a precise, quantifiable relation.     

The coincidence effect in similarity and choice

Medin, Goldstone, and Markman (1995) recently described a series of parallel effects in similarity and choice. They suggested that similarity and choice are related in a nontrivial way such that choice may entail a similarity judgment to an explicit or constructed ideal. In this paper, the correspondences between similarity and choice were investigated with respect to a phenomenon in similarity known as thecoincidence effect. In coincidence (pronounced “coincide-ence”), two items that match on one dimension but have a large difference on another dimension receive a higher similarity rating than do two items that have only modest differences on both dimensions. We conducted five experiments in order to examine commonalities between similarity and choice processes with respect to coincidence. Four types of tasks were given: similarity ratings, desirability ratings, forced choice similarities (which of two items is most similar to a target), and forced choice preferences (which of two items one would prefer, given a target). We found a main effect for ratings as opposed to forced choices, with ratings showing greater coincidence effects than did choices. Similarity measures tended to produce more coincidence than did preference measures. The overall pattern of results suggests the presence of dimensional weighting processes sensitive to task characteristics and operating somewhat differently for similarity and decision making.     

K-means may perform as well as mixture model clustering but may also be much worse: comment on Steinley and Brusco (2011)

Steinley and Brusco (2011) presented the results of a huge simulation study aimed at evaluating cluster recovery of mixture model clustering (MMC) both for the situation where the number of clusters is known and is unknown. They derived rather strong conclusions on the basis of this study, especially with regard to the good performance of K-means (KM) compared with MMC. I agree with the authors' conclusion that the performance of KM may be equal to MMC in certain situations, which are primarily the situations investigated by Steinley and Brusco. However, a weakness of the paper is the failure to investigate many important real-world situations where theory suggests that MMC should outperform KM. This article elaborates on the KM-MMC comparison in terms of cluster recovery and provides some additional simulation results that show that KM may be much worse than MMC. Moreover, I show that KM is equivalent to a restricted mixture model estimated by maximizing the classification likelihood and comment on Steinley and Brusco's recommendation regarding the use of mixture models for clustering.     

Choice Environment, Market Complexity, and Consumer Behavior: A Theoretical and Empirical Approach for Incorporating Decision Complexity into Models of Consumer Choice

Most empirical models of choice in economics and consumer research assume that the decision maker assesses all alternatives and information in a perfect information-processing sense. The complexity of the choice environment, the ability of the individual to make complex decisions, and the effect of choice context on the decision strategy are generally not considered in statistical model development. One of the reasons for this omission is that theoretical literature on choice complexity and imperfect ability to choose that has developed in psychology and behavioral decision theory (BDT) literatures has not been translated into empirical econometric analysis. Second, the data used in economics and consumer research studies tend to be somewhat different from the data structures used in psychology and BDT literatures. In this paper we outline a theoretical model that simultaneously considers task complexity, effort applied by the consumer, ability to choose, and choice. We then construct a measure of task complexity and incorporate this in an analysis of a number of data series based on the random utility framework. We also examine the performance of our measure of task complexity in a composite data set that allows for increased variability in factors affecting decision context. Our approach provides a mechanism to link research in BDT and econometric modeling of consumer choice. Our findings suggest that task complexity does affect inferences about choice model parameters and that context effects, such as complexity, have a systematic impact on the parameters of econometric models of choice. The modeling approach provides a mechanism for inclusion of results from psychology and BDT in traditional economic models of consumer choice.     

Optimal choice

We present a classification and theoretical analysis of discrete-trial and free-operant choice procedures in which reinforcement is assigned to one alternative only, or independently to both, is either always available or conditionally available, and is either "held" or not from trial to trial. Momentary-maximizing and (globally) optimal choice sequences are defined in terms of initializing and marker events. Free-operant choice is analyzed in terms of a clock space whose axes are the times since the last A and B choices. The analysis shows that most molar matching data are derivable from momentary maximizing, and that the momentary-maximizing hypothesis has not been adequately tested in either discrete-trial or free-operant situations.