Category Adjustment Model Research Articles

Towards a Generalized Bayesian Model of Reconstructive Memory

AbstractPrior knowledge has long been known to shape reconstruction from memory. An individual stimulus from a category is often remembered to be closer to the center of that category than its true location. This effect, together with more complex memory effects that involve prior knowledge at multiple levels of abstraction, has been successfully explained by the Category Adjustment Model (CAM; Huttenlocher et al. Journal of Experimental Psychology: General, 129(2), 220, 2000) and its extensions. However, recent experimental results diverge from CAM’s predictions showing that reconstructive memory for atypical category examples is influenced by the category center less than that of typical category examples. To unify these findings, we propose a generalized Bayesian model of reconstructive memory, called the generalized CAM model (g-CAM). We demonstrate through simulations that g-CAM can account for previously known effects of reconstructive memory, while additionally capturing recent empirical findings involving atypical category examples.

On the category adjustment model: another look at Huttenlocher, Hedges, and Vevea (2000)

Huttenlocher, Hedges, and Vevea (2000) (Why do categories affect stimulus judgment? Journal of Experimental Psychology: General, 129, 220-241) introduce the category adjustment model (CAM). Given that participants imperfectly remember stimuli (which we refer to as “targets”), CAM holds that participants maximize accuracy by using information about the distribution of the targets to improve their judgments. CAM predicts that judgments will be a weighted average of the imperfect memory of the target and the mean of the distribution of targets. Huttenlocher, Hedges, and Vevea (2000) report on three experiments and conclude that CAM is “verified.” We attempt to replicate Experiment 3 from Huttenlocher et al. (2000). We analyze judgment-level data rather than averaged data. We find evidence of a bias toward a set of recent targets rather than a bias toward the running mean. We also do not find evidence of the joint hypothesis that participants learned the distribution of targets and employed this information in their judgments. The judgments in our dataset are not consistent with CAM. We discuss how the apparent defects in HHV – including dividing by zero – went unnoticed and how such mistakes can be avoided in future research. Finally, we hope that the techniques that we employ will be used to test other datasets that are currently regarded as consistent with CAM or any Bayesian model of judgment.

Context effects on reproduced magnitudes from short-term and long-term memory.

Extant research has demonstrated strong contextual dependencies in reproducing magnitudes of perceptual stimuli from short-term memory. Two experiments examined how context as defined by (a) the mean of the distribution, (b) stimulus ranks, (c) values of anchor stimuli used in the reproduction task, and (d) values from the most recent trial operate on estimates of square size. Experiment 1 demonstrated distributional contrast effects on ratings of squares and distributional assimilation effects on reproduction of squares from short-term memory for the same participants. The fit of a modified version of the category adjustment model demonstrated reliable effects of the running mean, start anchors, and previous stimulus on reproduction bias. In Experiment 2, participants first learned to associate labels with squares, then reproduced square sizes based on the label cues, a long-term memory task, followed by a reproduction from short-term memory task as in Experiment 1. Results for the short-term memory task were largely consistent with Experiment 1. Results for the long-term memory task showed a very different pattern of effects, with larger reproduced sizes when squares were drawn from positively skewed rather than negatively skewed distributions. This contrast effect was explained by a modified range-frequency model as the result of rank encoding of square values along with displacement away from the running mean and shifts towards the prior response and start anchors. The combined results identify multiple sources of context effects in estimation that depend critically on memory retrieval factors and show how they can be incorporated into existing models.

Omitted-variable bias and other matters in the defense of the category adjustment model: A comment on Crawford (2019)

The datasets from Duffy, Huttenlocher, Hedges, and Crawford (2010) [Psychonomic Bulletin and Review, 17(2), 224–230] were reanalyzed by Duffy and Smith (2018) [Psychonomic Bulletin and Review, 25(5), 1740–1750]. Duffy and Smith (2018) conclude that the datasets are not consistent with the category adjustment model (CAM). Crawford (2019) [Psychonomic Bulletin and Review, 26(2), 693–698] offered a reply to Duffy and Smith (2018) that is based on three main points. Crawford proposes regressions that are, in part, based on a “deviation” analysis. Crawford offers a different simulation of data and claims that the techniques employed by Duffy and Smith (2018) are not sufficiently sensitive to detect a specific relationship that is claimed to be consistent with CAM. Crawford also appeals to a figure showing that the responses appear to be biased toward the overall running mean, and presumably not toward recently viewed lines. We show that Crawford's analysis suffers from an omitted-variable bias. Once this bias is corrected, the evidence in support of CAM disappears. When we produce a simulated dataset that is consistent with the specification suggested by Crawford, the techniques of Duffy and Smith (2018) correctly detect the true relationship. Despite the assertion otherwise, the simulated dataset that was analyzed by Crawford is not publicly available. It remains our view that the datasets from Duffy, Huttenlocher, Hedges, and Crawford (2010) do not appear to be consistent with CAM or any Bayesian model of judgment.

What Do Biased Estimates Tell Us about Cognitive Processing? Spatial Judgments as Proportion Estimation

ABSTRACTSimilar estimation biases appear in a wide range of quantitative judgments, across many tasks and domains. Often, these biases (those that occur, for example, when adults or children indicate remembered locations of objects in bounded spaces) are believed to provide evidence of Bayesian or rational cognitive processing, and are explained in terms of relatively complex Bayesian models (e.g., the Category Adjustment Model). Here, we suggest that some of these phenomena may be accounted for instead within a simpler alternative theoretical framework that has previously been found to explain bias in common numerical estimation tasks across development. We report data from university undergraduate students and 7- through 10-year-olds completing a speeded linear position reproduction task. Bias in both adults’ and children’s responses was effectively explained in terms of a relatively simple psychophysical model of proportion estimation. These data clearly show that the proportion estimation framework is a viable alternative to theories that explain biases as the result of a Bayesian cognitive adjustment process. We also discuss our view that these data are not easily reconciled with the requirements of the more complex Category Adjustment Model that assumes estimates should exhibit a central tendency bias.

Reply to Duffy and Smith's (2018) reexamination.

Duffy, Huttenlocher, Hedges, and Crawford (2010, Psychonomic Bulletin & Review, 17[2], 224-230) examined whether the well-established central tendency bias in people's reproductions of stimuli reflects bias toward the mean of an entire presented distribution or bias toward only recently seen stimuli. They reported evidence that responses were biased toward the long-run mean and found no evidence that they were biased toward the most recent stimuli. Duffy and Smith (2018) reexamine the data using a different analytical strategy and argue that estimates are biased by recent stimuli rather than toward the long-run mean. I argue that this reanalysis misses a true effect of the running mean and that the data are (mostly) consistent with the claims in the original work. I suggest that these results, and many other null results presented by Duffy and Smith, do not have major theoretical significance for the category adjustment model and similar Bayesian models. (Code and data available: https://osf.io/tkqvn .).

Omitted-Variable Bias and Other Matters in the Defense of the Category Adjustment Model: A Reply to Crawford (2019)

The datasets from Duffy, Huttenlocher, Hedges, and Crawford (2010) [Psychonomic Bulletin & Review, 17(2), 224-230] were reanalyzed by Duffy and Smith (2018) [Psychonomic Bulletin & Review, 25(5), 1740-1750]. Duffy and Smith (2018) conclude that the datasets are not consistent with the category adjustment model (CAM). Crawford (2019) [Psychonomic Bulletin & Review, 26(2), 693-698] offered a reply to Duffy and Smith (2018) that is based on three main points. Crawford proposes regressions that are, in part, based on a “deviation” analysis. Crawford offers a different simulation of data and claims that the techniques employed by Duffy and Smith (2018) are not sufficiently sensitive to detect a specific relationship that is claimed to be consistent with CAM. Crawford also appeals to a figure showing that the responses appear to be biased toward the overall running mean, and presumably not toward recently viewed lines. We show that Crawford’s analysis suffers from an omitted-variable bias. Once this bias is corrected, the evidence in support of CAM disappears. When we produce a simulated dataset that is consistent with the specification suggested by Crawford, the techniques of Duffy and Smith (2018) correctly detect the true relationship. Despite the assertion otherwise, the simulated dataset that was analyzed by Crawford is not publicly available. Since the analysis of Crawford (2019) is incorrect, it remains our view that the datasets from Duffy, Huttenlocher, Hedges, and Crawford (2010) do not appear to be consistent with CAM or any Bayesian model of judgment.

Free the brand: How a logo frame influences the potentiality of brand extensions

Inspired by the logo redesign case of Starbucks, this research used four experiments to explore how marketers may broaden brand breadth and improve brand extension attitudes by removing the frame from a logo. A logo frame is speculated to create a mental boundary that dictates the scope to which products belong. Removing the logo frame thus frees the brand from such finite boundaries and encourages consumers’ relational associations. Study 1 investigated the mechanism underlying the impact of the logo frame on relational and item-specific elaborations. An open logo encouraged participants’ usage of relational elaborations, whereas a framed logo reinforced participants’ usage of item-specific elaborations. Study 2 demonstrated that participants associated broader product portfolios with a brand that had an open logo than a framed logo. Furthermore, Study 3 employed a brand extension scenario to support that removing the logo frame enhanced consumers’ attitudes toward a new extension. Study 4 illustrated that the perceived distance mediated the logo frame effect on extension attitudes as suggested by the Category Adjustment model. These findings demonstrate that marketers could leverage brand perception and favor certain brand extensions through an easily manipulated design feature, the logo frame.

A systematic spatial bias in remembering eye positions

ABSTRACTCategorical bias in location memory in geometric spaces is well established. The present study assessed the presence of the bias in locating eyes in images of human faces. Participants were presented with digital faces and indicated the position of one of the eyes in both upright and inverted orientations. Biases resulted from participants using multiple sources of accessible information. No differences were found for upright vs. inverted face images. Overall, the data were consistent with the retrieval category adjustment model. These distortions may be considered within the forensic context, when eyewitnesses work with a sketch artist or use a computer program to generate an image of the culprit.

Learning About Spatial and Temporal Scale: Current Research, Psychological Processes, and Classroom Implications

ABSTRACT Geoscientists analyze and integrate spatial and temporal information at a range of scales to understand Earth processes. Despite this, the concept of scale is ill defined and taught unevenly across the K–16 continuum. This literature review focuses on two meanings of scale: one as the magnitude of the extent of a dimension and the other as a relationship between objects or events. We review 42 papers from science education and discipline-based education research (DBER) literature on students' conceptions related to one or both meanings of scale. Analysis of this prior work reveals a broader (though still limited) research base on domain general concepts of scale as magnitude and scant research on scale as a relationship. Learners begin reasoning about spatial and temporal magnitudes categorically by working with scales based on standard units and nonmetric values, such as body length. Concepts of scale magnitudes outside human experience are nonlinear. Facility with fractions and proportional reasoning are positively associated with the ability to reason about scale as a relationship. Two constructs from the psychological literature, structure mapping and the category adjustment model, offer theoretical accounts for these findings. We borrow a typology from the psychological literature to frame common geoscience instructional models in the context of spatial and temporal scale and suggest how instructors might facilitate students' reasoning about scale models. We identify a number of avenues for possible future research, including a critical need to understand how conceptual understanding of scale develops across the K–16 continuum.

Category effects on stimulus estimation: Shifting and skewed frequency distributions-A reexamination.

Duffy, Huttenlocher, Hedges, and Crawford (Psychonomic Bulletin & Review, 17(2), 224-230, 2010) report on experiments where participants estimate the lengths of lines. These studies were designed to test the category adjustment model (CAM), a Bayesian model of judgments. The authors report that their analysis provides evidence consistent with CAM: that there is a bias toward the running mean and not recent stimuli. We reexamine their data. First, we attempt to replicate their analysis, and we obtain different results. Second, we conduct a different statistical analysis. We find significant recency effects, and we identify several specifications where the running mean is not significantly related to judgment. Third, we conduct tests of auxiliary predictions of CAM. We do not find evidence that the bias toward the mean increases with exposure to the distribution. We also do not find that responses longer than the maximum of the distribution or shorter than the minimum become less likely with greater exposure to the distribution. Fourth, we produce a simulated dataset that is consistent with key features of CAM, and our methods correctly identify it as consistent with CAM. We conclude that the Duffy et al. (2010) dataset is not consistent with CAM. We also discuss how conventions in psychology do not sufficiently reduce the likelihood of these mistakes in future research. We hope that the methods that we employ will be used to evaluate other datasets.

Spatial Category Bias in a Three-Dimensional Virtual Environment

One bias in spatial memory is that individuals remember a target toward the centroid of its region, the category bias or prototype effect. The standard approach to investigate the bias follows the category adjustment (CA) model by Huttenlocher, Hedges, and Duncan (1991) , with the use of a two-dimensional (2D) dot-in-circle task in which observers see a target in a blank circle and indicate its location after a brief interval. However, it is unclear whether the work has ecological validity. Investigations using real life spaces compromise control. The current paper investigated the effect in a 3D virtual environment, as it lends high levels of experimental control in a simulated naturalistic space. The pattern found with the 2D dot-in-circle task was replicated with its 3D virtual counterpart. We discuss how our method, in addition to increasing ecological validity under controlled conditions, eliminates concerns regarding studies that applied the CA model to more ecologically valid environments.

Category Effects on Stimulus Estimation: Shifting and Skewed Frequency Distributions - A Reexamination

Duffy, Huttenlocher, Hedges, and Crawford (2010) [Psychonomic Bulletin & Review, 17(2), 224-230] report on experiments where participants estimate the lengths of lines. These studies were designed to test the Category Adjustment Model (CAM), a Bayesian model of judgments. CAM predicts that there will exist a bias toward the running mean of the lines and that judgments will not be differentially affected by recent stimuli. The authors report that their analysis provides evidence consistent with CAM. We reexamine their data. First, we attempt to replicate their analysis and we obtain different results. Second, we conduct a different statistical analysis. We find significant recency effects and we identify several specifications where the running mean is not significantly related to judgment. Third, we conduct a test of an auxiliary prediction of CAM: that the bias towards the mean will increase with exposure to the distribution. We do not find such a relationship. Fourth, we produce a simulated dataset that is consistent with CAM and our methods correctly identify it as consistent with CAM. We conclude that the Duffy et al. (2010) dataset is not consistent with CAM. We also discuss how conventions in psychology do not sufficiently reduce the likelihood of these mistakes in future research.

On the Category Adjustment Model: Another Look at Huttenlocher, Hedges, and Vevea (2000)

Huttenlocher, Hedges, and Vevea (2000) (Why do categories affect stimulus judgment? Journal of Experimental Psychology: General, 129, 220-241) introduce the category adjustment model (CAM). Given that participants imperfectly remember stimuli (which we refer to as “targets”), CAM holds that participants maximize accuracy by using information about the distribution of the targets to improve their judgments. CAM predicts that judgments will be a weighted average of the imperfect memory of the target and the mean of the distribution of targets. Huttenlocher, Hedges, and Vevea (2000) report on three experiments and conclude that CAM is “verified.” We attempt to replicate Experiment 3 from Huttenlocher et al. (2000). We analyze judgment-level data rather than averaged data. We find evidence of a bias toward a set of recent targets rather than a bias toward the running mean. We also do not find evidence of the joint hypothesis that participants learned the distribution of targets and employed this information in their judgments. The judgments in our dataset are not consistent with CAM. We discuss how the apparent defects in HHV – including dividing by zero – went unnoticed and how such mistakes can be avoided in future research. Finally, we hope that the techniques that we employ will be used to test other datasets that are currently regarded as consistent with CAM or any Bayesian model of judgment.

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.

Categorical biases in spatial memory: the role of certainty.

Memories for spatial locations often show systematic errors toward the central value of the surrounding region. The Category Adjustment (CA) model suggests that this bias is due to a Bayesian combination of categorical and metric information, which offers an optimal solution under conditions of uncertainty (Huttenlocher, Hedges, & Duncan, 1991). A fundamental assumption of this model is that representations of locations are unbiased but uncertain; during combination, greater metric uncertainty results in relatively greater emphasis on categorical information, ultimately leading to increased bias (but also minimizing error across multiple estimates). Sampaio and Wang (2009) have demonstrated that metric information is not lost during this combination process, supporting the CA model's assumption that underlying spatial representations are undistorted. Here, we examine the 2nd half of the CA model's central assumption: that increasing metric uncertainty drives the combination process. Participants recognized point locations within visually complex images in a 4-choice task. Our results indicate that individuals recognized the correct location over other, biased alternatives, confirming that metric information is unbiased at the time of retrieval. In addition, we found that, when participants make errors, they are more likely to select locations that are biased toward the category prototype. In Experiment 2, we demonstrate that categorically biased locations are most likely to be chosen under conditions of uncertainty. Indeed, under these conditions, categorically biased locations were chosen more frequently than the correct location. These results suggest that systematic errors are the result of combination across multiple levels of spatial representations that are undistorted but somewhat uncertain.

Categorical Bias in Line Angle Judgments: Sex Differences and the Use of Multiple Categories

:According to the Category Adjustment (CA) model, spatial estimates (e.g., of location) involve Bayesian combination across multiple, hierarchical pieces information, each weighted by its relative certainty. Recent work, though, has shown that men and women differ in terms of their certainty regarding fine-grained and categorical information in location memory. Here we demonstrate that this reflects a more general sex difference in visuospatial processing by examining bias patterns in a line angle judgment task (JLAP-15). In addition, the data suggest that multiple, hierarchical levels of categorical information influenced spatial judgments. These results imply that the hierarchical combination outlined by the CA model extends beyond two levels, and that men and women apply differential weighting to these representations.

How many calories were in those hamburgers again? Distribution density biases recall of attribute values

AbstractDecisions that consumers make often rest on evaluations of attributes, such as how large, expensive, good, or fattening an option seems. Extant research has demonstrated that these evaluations in turn depend upon the recently experienced distribution of attribute values (e.g., positively or negatively skewed). In many situations decisions rely on recalling the attribute values of each option, a process that has been neglected in much of the previous literature. In two experiments, participants learned attribute information for labeled stimuli presented within either a positively or negatively skewed distribution and then they recalled values from labels after approximately one minute. The results demonstrated effects that are inconsistent with predictions of the category-adjustment model (Duffy, Huttenlocher, Hedges & Crawford, 2010) that recalled values would shift toward the mean of the distribution of values presented. Instead, results were consistent with predictions of the comparison-induced distortion model (Choplin & Hummel, 2002) that remembered values would depend on the density of stimuli within the attribute range. Reasons for these results, alternative models, and implications for decision-making are discussed.

Sex differences in the weighting of metric and categorical information in spatial location memory

According to the Category Adjustment model, remembering a spatial location involves the Bayesian combination of fine-grained and categorical information about that location, with each cue weighted by its relative certainty. However, individuals may differ in terms of their certainty about each cue, resulting in estimates that rely more or less on metric or categorical representations. To date, though, very little research has examined individual differences in the relative weighting of these cues in spatial location memory. Here, we address this gap in the literature. Participants were asked to recall point locations in uniform geometric shapes and in photographs of complex, natural scenes. Error patterns were analyzed for evidence of a sex difference in the relative use of metric and categorical information. As predicted, women placed relatively more emphasis on categorical cues, while men relied more heavily on metric information. Location reproduction tasks showed a similar effect, implying that the sex difference arises early in spatial processing, possibly during encoding.

Two-category place representations persist over body rotations

We explored a system that constructs environment-centered frames of reference and coordinates memory for the azimuth of an object in an enclosed space. For one group, we provided two environmental cues (doors): one in the front, and one in the rear. For a second group, we provided two object cues: a front and a rear cue. For a third group, we provided no external cues; we assumed that for this group, their reference frames would be determined by the orthogonal geometry of the floor-and-wall junction that divides a space in half or into multiple territories along the horizontal continuum. Using Huttenlocher, Hedges, and Duncan’s (Psychological Review 98: 352-376, 1991) category-adjustment model (cue-based fuzzy boundary version) to fit the data, we observed different reference frames than have been seen in prior studies involving two-dimensional domains. The geometry of the environment affected all three conditions and biased the remembered object locations within a two-category (left vs. right) environmental frame. The influence of the environmental geometry remained observable even after the participants’ heading within the environment changed due to a body rotation, attenuating the effect of the front but not of the rear cue. The door and object cues both appeared to define boundaries of spatial categories when they were used for reorientation. This supports the idea that both types of cues can assist in environment-centered memory formation.