ArticlePublisher preview available

Not Just Stimuli Structure: Sequencing Effects in Category Learning Vary by Task Demands

Authors:
To read the full-text of this research, you can request a copy directly from the authors.

Abstract

General Audience Summary Research has shown that generally when learning new categories, it is best to intersperse examples from different categories between one another (e.g., ABCABCABC) rather than show examples from one category at a time (e.g., AAABBBCCC). This finding is called the interleaving effect. In this study, we show that interleaved study may not always confer an advantage over blocked study when learners do not have much information about the categories being learned (i.e., unsupervised learning tasks). Participants were presented with a list of Chinese characters and their meanings and then were shown new characters and asked to guess at their meanings. What learners needed to discover for themselves in Experiments 1 and 2, is that these words fit into “categories” (e.g., water-related: rapids, damp, boil, harbor) and that characters within a category share a common component. We show blocked study (one category at a time) can help learners notice the shared component; interleaving makes it harder. But when learners were alerted to the existence of the rules (Experiment 3), interleaving led to much better performance on the new character identification test. These results highlight the role that sequencing plays not just in driving learners’ attention but also facilitating memory and suggests that instructors’ and learners’ sequencing decisions should depend not only on the category stimuli being learned but also on the specific nature of the learning task being attempted.
EMPIRICAL ARTICLE
Not Just Stimuli Structure: Sequencing Effects in Category
Learning Vary by Task Demands
Veronica X. Yan and Brendan A. Schuetze
Department of Educational Psychology, The University of Texas at Austin
Attention- and memory-based accounts of sequencing effects in category learning are often pitted against
one another, but we propose that both are important. We created an unsupervised learning task in which the
rules governing categories would be difcult to notice under interleaved sequences. Specically, parti-
cipants were presented with Chinese characters and their meanings. Category-related characters all shared a
subcomponent (radical), but participants had to abstract this rule. No character was repeated. On the day-
delayed test, participants were shown new Chinese characters and asked to select a possible meaning to test
category induction. Under both passive (Experiment 1) and active (Experiment 2) study, we found no
interleaving benet. However, when we eliminated the demand on attentional processes by directing
attention to the rules (Experiment 3), we obtained an interleaving benet. We discuss implications for how
sequencing decisions should not only depend on the stimuli but also the learning task.
General Audience Summary
Research has shown that generally when learning new categories, it is best to intersperse examples from
different categories between one another (e.g., ABCABCABC) rather than show examples from one
category at a time (e.g., AAABBBCCC). This nding is called the interleaving effect. In this study, we
show that interleaved study may not always confer an advantage over blocked study when learners do
not have much information about the categories being learned (i.e., unsupervised learning tasks).
Participants were presented with a list of Chinese characters and their meanings and then were shown
new characters and asked to guess at their meanings. What learners needed to discover for themselves in
Experiments 1 and 2, is that these words t into categories(e.g., water-related: rapids, damp, boil,
harbor) and that characters within a category share a common component. We show blocked study (one
category at a time) can help learners notice the shared component; interleaving makes it harder. But
when learners were alerted to the existence of the rules (Experiment 3), interleaving led to much better
performance on the new character identication test. These results highlight the role that sequencing
plays not just in driving learnersattention but also facilitating memory and suggests that instructorsand
learnerssequencing decisions should depend not only on the category stimuli being learned but also on
the specic nature of the learning task being attempted.
Keywords: sequencing, interleaving, category learning, Attentional bias, spacing, Chinese characters
To do two things at once is to do neither(Syrus, 1856, p. 13, as
translated from Latin by Lyman). This saying may hold true at many
levels and make strong intuitive sense, but when it comes to learning
categories and concepts, the opposite has often been demonstrated to
be true. Imagine that you are a student enrolled in a statistics class,
and you have four different types of statistics problems (A, B, C, and
D) to learn. Do you focus on learning one problem type at a time
(e.g., AAABBBCCC)? This one at a timesequence is referred to
as a blocked sequence and is everywhere in the real world. It is
common, for example, to see course syllabi and textbooks organized
into modules, with students practicing just one type of problem at a
time (Rohrer et al., 2020). Alternatively, do you switch back and
forth randomly between the four different types of problems (e.g.,
ABCBCACAB) in an interleaved sequence?
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This article was published Online First November 15, 2021.
The studies were conceived and designed by Veronica X. Yan Material
preparation, data collection, and analyses were performed by Brendan A.
Schuetze Both authors interpreted the data and each drafted different sections
of the manuscript. Both have read and approved the nal manuscript.
All pre-registrations, materials, raw data, R output and supplemental
materials are available on OSF, https://doi.org/10.17605/OSF.IO/3UV7T.
The authors declare that they have no conict of interest.
Brendan A. Schuetze was supported by a Donald D. Harrington doctoral
fellowship. We also thank the members of the SLAM lab for their construc-
tive feedback.
Correspondence concerning this article should be addressed to Veronica
X. Yan, Department of Educational Psychology, The University of
Texas at Austin, Austin, United States. Email: veronicayan@austin.ute
xas.edu
Journal of Applied Research in Memory and Cognition
© 2021 American Psychological Association 2022, Vol. 11, No. 2, 218228
ISSN: 2211-3681 https://doi.org/10.1016/j.jarmac.2021.09.004
218
... Specifically, under an interleaving schedule, learners are more likely to notice the unique features of a category as learners pay more attention to between-categories differences. Those features, once identified, can be remembered better due to the memory benefits of an interleaving schedule (i.e., the two-stage framework of sequencing effects; Yan et al. 2020;Yan and Schuetze 2022). ...
... For example, in a learning task of science categories (i.e., organic chemical compounds), Eglington and Kang (2017) found the robust benefits of interleaving schedule over blocking both when the rules that define category membership were and were not visually highlighted on the exemplars. Using a learning task of Chinese characters, Yan and Schuetze (2022) also found the positive effect of interleaving schedule, but this only emerged when the category-level rules were visually highlighted to the participants. No benefits of interleaving were found when the rules were not provided. ...
... Meagher et al. (2022) found that some rock pairs contain discrete features that are relatively easier to verbalize, while others do not. If feature descriptions are beneficial for rock category learning, providing them may boost the interleaving effect as they draw learners' attention to the relevant dimensions when learners compare exemplars from different categories (Yan and Schuetze 2022). Further, interleaving schedule can also enhance learning as it improves learners' memory of the provided features (Yan et al. 2020; also see Carpenter 2014). ...
Article
Full-text available
The present study examined the effects of study schedule (interleaving vs. blocking) and feature descriptions on category learning and metacognitive predictions of learning. Across three experiments, participants studied exemplars from different rock categories and later had to classify novel exemplars. Rule-based and information-based categorization was also manipulated by selecting rock sub-categories for which the optimal strategy was the one that aligned with the extraction of a simple rule, or the one that required integration of information that may be difficult to describe verbally. We observed consistent benefits of interleaving over blocking on rock classification, which generalized to both rule-based (Experiment 1) and information-integration learning (Experiments 1–3). However, providing feature descriptions enhanced classification accuracy only when the stated features were diagnostic of category membership, indicating that their benefits were limited to rule-based learning (Experiment 1) and did not generalize to information-integration learning (Experiments 1–3). Furthermore, our examination of participants’ metacognitive predictions demonstrated that participants were not aware of the benefits of interleaving on category learning. Additionally, providing feature descriptions led to higher predictions of categorization even when no significant benefits on actual performance were exhibited.
... One possible explanation for the lack of interleaving effect in self-regulated learning might be that the discriminative contrast between subsequent images is impaired by a study sequence choice, which probably imposes additional cognitive load due to metacognitive processes (van Gog et al., 2020) and diminishes the visual representation of the preceding image due to temporal spacing (cf. Birnbaum et al., 2013;Yan & Schuetze, 2021). One further explanation could be that learners did not sufficiently understand how the mushroom doubles were distributed on the selection page. ...
... Discovering commonalities via blocking, in turn, is beneficial for category induction if the category exemplars are too diverse to be able to recognize their relevant differences among the irrelevant ones (Abel et al., 2021;Carvalho & Goldstone, 2015). Accordingly, studies that used very distinctive categories sharing only predictive features but providing no explicit information on critical features typically showed a benefit of blocking over interleaving (Carpenter & Mueller, 2013;Sorensen & Woltz, 2016;Yan & Schuetze, 2021). However, in general, learners are unaware of the implicit requirements of a category learning task with confusable stimuli that are distinct only regarding their predictive features -that is, the awareness-that-differences-matter. ...
Article
Full-text available
Distinguishing easily confusable categories requires learners to detect their predictive differences. Interleaved sequences — switching between categories — help learners to detect such differences. Nonetheless, learners prefer to block — switching within a category — to detect commonalities. Across two 2 × 2-factorial experiments, we investigated why learners scarcely engage in interleaving when learning confusable categories. In Experiment 1 (N = 190), we investigated the role of the utility value of being able to distinguish confusable mushroom doubles on their spontaneous study sequence choices and of the conditional knowledge component that for distinguishing, the detection of differences (between the doubles) matters. In Experiment 2 (N = 134), we again investigated the role of the latter and additionally of the conditional knowledge component that interleaving highlights differences. Results showed that combining two factors — increasing the utility value of distinguishing and informing learners that for distinguishing, the detection of differences matters — fostered learners’ use of interleaving. In conclusion, learners are more aware that interleaving highlights differences than previously thought. Nonetheless, learners prefer blocking because they do not recognize the utility value of distinguishing, and they lack the conditional knowledge that distinguishing requires finding predictive differences. Their blocked study sequence choices reflect a deliberate investment of effort to find commonalities rather than just avoiding effort. To make learners shift their effort allocation from finding commonalities to finding differences and engage them in spontaneous interleaving, we recommend highlighting the utility value of distinguishing and informing learners about the importance of finding differences for distinguishing.
... In many instances, the interleaving effect is eliminated or even reversed. For example, blocking advantages occur when learning concepts are easy to tell apart but hard to integrate (e.g., classic physics versus quantum physics; Carvalho & Goldstone, 2014;Yan & Sana, 2021b), when verbalizable rules of a category are difficult to discover (Carvalho & Goldstone, 2015;Noh et al., 2016;Yan & Schuetze, 2022), when the test does not require knowing small differences , and when skills are complex and require attention to multiple components (e.g., learning different tennis or golf strokes; Guadagnoli et al., 1999;Hebert et al., 1996). ...
... That is, what a learner will learn from an interleaved sequence differs from what a learner would learn from a blocked sequence of practice. When examples from different concepts are juxtaposed (i.e., interleaving), attention is drawn to the critical features that distinguish between them; when examples from the same concept are juxtaposed, attention is drawn to the critical features (i.e., blocking) that tie a concept together (Brunmair & Richter, 2019;Carvalho & Goldstone, 2017;Yan & Schuetze, 2022). This general proposal is consistent with the findings described before that interleaved practice improves learning of hard-to-discriminate concepts, but blocked practice improves learning of easy-to-discriminate concepts. ...
Article
Full-text available
Cognitive science of learning points to solutions for making use of existing study and instruction time more effectively and efficiently. However, solutions are not and cannot be one-size-fits-all. This paper outlines the danger of overreliance on specific strategies as one-size-fits-all recommendations and highlights instead the cognitive learning processes that facilitate meaningful and long-lasting learning. Three of the most commonly recommended strategies from cognitive science provide a starting point; understanding the underlying processes allows us to tailor these recommendations to implement at the right time, in the right way, for the right content, and for the right students. Recommendations regard teacher training, the funding and incentivizing of educational interventions, guidelines for the development of educational technologies, and policies that focus on using existing instructional time more wisely.
... The benefit of interleaving over blocking has been found many times in empirical research on learning visual materials like paintings, naturalistic photos, and artificial images [12,15,32]. However, under specific circumstances, blocking can be more beneficial compared to interleaving [2,3,4,5,10,11,30,32]. For example, Carvalho and Goldstone [3] have found that lowsimilarity stimuli were more effectively learned when presented in blocked sequences. ...
Preprint
In category learning, a growing body of literature has increasingly focused on exploring the impacts of interleaving in contrast to blocking. The sequential attention hypothesis posits that interleaving draws attention to the differences between categories while blocking directs attention toward similarities within categories. Although a recent study underscores the joint influence of memory and attentional factors on sequencing effects, there remains a scarcity of effective computational models integrating both attentional and memory considerations to comprehensively understand the effect of training sequences on students' performance. This study introduces a novel integration of attentional factors and spacing into the logistic knowledge tracing (LKT) models to monitor students' performance across different training sequences (interleaving and blocking). Attentional factors were incorporated by recording the counts of comparisons between adjacent trials, considering whether they belong to the same or different category. Several features were employed to account for temporal spacing. We used cross-validations to test the model fit and predictions on the learning session and posttest. Our findings reveal that incorporating both attentional factors and spacing features in the Additive Factors Model (AFM) significantly enhances its capacity to capture the effects of interleaving and blocking and demonstrates superior predictive accuracy for students' learning outcomes. By bridging the gap between attentional factors and memory processes, our computational approach offers a more comprehensive framework for understanding and predicting category learning outcomes in educational settings.
... The benefit of interleaving over blocking has been found many times in empirical research on learning visual materials like paintings, naturalistic photos, and artificial images [12,15,32]. However, under specific circumstances, blocking can be more beneficial compared to interleaving [2,3,4,5,10,11,30,32]. For example, Carvalho and Goldstone [3] have found that lowsimilarity stimuli were more effectively learned when presented in blocked sequences. ...
Conference Paper
Full-text available
In category learning, a growing body of literature has increasingly focused on exploring the impacts of interleaving in contrast to blocking. The sequential attention hypothesis posits that interleav-ing draws attention to the differences between categories while blocking directs attention toward similarities within categories [4, 5]. Although a recent study underscores the joint influence of memory and attentional factors on sequencing effects [31], there remains a scarcity of effective computational models integrating both attentional and memory considerations to comprehensively understand the effect of training sequence on students' performance. This study introduces a novel integration of attentional factors and spacing into the logistic knowledge tracing (LKT) models [22] to monitor students' performance across different training sequences (interleaving and blocking). Attentional factors were incorporated by recording the counts of comparisons between adjacent trials, considering whether they belong to the same or different category. Several features were employed to account for temporal spacing. We used crossvalidations to test the model fit and predictions on the learning session and posttest. Our findings reveal that incorporating both attentional factors and spacing features in the Additive Factors Model (AFM) significantly enhances its capacity to capture the effects of interleaving and blocking and demonstrates superior predictive accuracy for students' learning outcomes. By bridging the gap between attentional factors and memory processes, our computational approach offers a more comprehensive framework for understanding and predicting category learning outcomes in educational settings.
... However, whether a strategy is desirable depends on the task demands. We should, therefore, acknowledge that students may benefit from blocked practice in certain conditions, depending on the complexity of learning tasks 50 , (dis)similarity of to-be-learned materials 51,52 , and test type 53 . Second, in the interleaved practice literature, it is common to rely on learning tasks with relatively low authenticity, such as learning painting styles and bird species. ...
Article
Full-text available
During category learning, students struggle to create an optimal study order: They often study one category at a time (i.e., blocked practice) instead of alternating between different categories (i.e., interleaved practice). Several interventions to improve self-study of categorical learning have been proposed, but these interventions have only been tested in learning tasks where students did not create the study order themselves. Instead, they decided which type of study order to follow. This pre-registered experiment examined whether an intervention that combines refutations and metacognitive prompts can enhance students’ engagement in interleaved practice, specifically when they organize the learning materials themselves. Ninety-one undergraduate students were randomized into the intervention and control condition and learned visual categories. Prior to the intervention, students used more blocked practice. After the intervention, the use of interleaved practice significantly increased in both immediate and delayed-transfer tasks. More interleaved practice was associated with better classification performance. Our findings indicate that refutations and metacognitive prompts form a strong intervention that corrects students’ erroneous beliefs and increases their engagement in interleaved practice.
... Indeed, shortterm performance can be inflated in ways that circumvent the deeper cognitive processing that supports true learning. In each section of Figure 1, the strategy that creates the appearance of learning in the short-term leads to worse long-term retention (Rawson & Kintsch, 2005;Yan & Schuetze, 2022;Yeo & Fazio, 2019). LA research that focuses on optimizing in-the-moment performance might ultimately lead to recommendations that hinder learning. ...
Article
Full-text available
To promote cross-community dialogue on matters of significance within the field of learning analytics (LA), we as editors-in-chief of the Journal of Learning Analytics (JLA) have introduced a section for papers that are open to peer commentary. An invitation to submit proposals for commentaries on the paper was released, and 12 of these proposals were accepted. The 26 authors of the accepted commentaries are based in Europe, North America, and Australia. They range in experience from PhD students and early-career researchers to some of the longest-standing, most senior members of the learning analytics community.
Preprint
Full-text available
Numerous studies have shown that an interleaved study sequence of examples (e.g., ABCBACACB) from different categories, relative to a blocked sequence (e.g., AAABBBCCC), often yields superior category learning. Some explanations for sequencing effects centers on attentional processes, others focus on memory processes, and the two are often pitted against each other. We propose a new integrative two-stage framework for sequencing effects incategory-learning and support this framework using a meta-analytic approach. We show, using a meta-analytic approach, that a combination of memory and attentional predictors explains significantly more variance in sequencing effects than attentional factors alone. This approach also allowed us to examine the nature of the existing evidence, which revealed inferential limitations due to how researchers typically design experiments. We provide suggestions for future research on sequencing effects in category learning that would both test the two-stage framework and increase the impact of future research.
Article
Full-text available
A typical mathematics assignment consists of a block of problems devoted to the same topic, yet several classroom-based randomized controlled trials have found that students obtain higher test scores when most practice problems are mixed with different kinds of problems—a format known as interleaved practice. Interleaving prevents students from safely assuming that each practice problem relates to the same skill or concept as the previous problem, thus forcing them to choose an appropriate strategy on the basis of the problem itself. Yet despite the efficacy of interleaved practice, blocked practice predominates most mathematics textbooks. As an illustration, we examined 13,505 practice problems in six representative mathematics texts and found that only 9.7% of the problems were interleaved. This translates to only one or two interleaved problems per school day. In brief, strong evidence suggests that students benefit from heavy doses of interleaved practice, yet most mathematics texts provide scarcely any.
Article
Full-text available
Researchers commonly interpret effect sizes by applying benchmarks proposed by Cohen over a half century ago. However, effects that are small by Cohen's standards are large relative to the impacts of most field-based interventions. These benchmarks also fail to consider important differences in study features, program costs, and scalability. In this paper, I present five broad guidelines for interpreting effect sizes that are applicable across the social sciences. I then propose a more structured schema with new empirical benchmarks for interpreting a specific class of studies: causal research on education interventions with standardized achievement outcomes. Together, these tools provide a practical approach for incorporating study features, cost, and scalability into the process of interpreting the policy importance of effect sizes.
Article
Full-text available
An interleaved presentation of items (as opposed to a blocked presentation) has been proposed to foster inductive learning (interleaving effect). A meta-analysis of the interleaving effect (based on 59 studies with 238 effect sizes nested in 158 samples) was conducted to quantify the magnitude of the interleaving effect, to test its generalizability across different settings and learning materials, and to examine moderators that could augment the theoretical models of interleaved learning. A multilevel meta-analysis revealed a moderate overall interleaving effect (Hedges’ g = 0.42). Interleaved practice was best for studies using paintings (g = 0.67) and other visual materials. Results for studies using mathematical tasks revealed a small interleaving effect (g = 0.34), whereas results for expository texts and tastes were ambiguous with nonsignificant overall effects. An advantage of blocking compared to interleaving was found for studies based on words (g = -0.39). A multiple meta-regression analysis revealed stronger interleaving effects for learning material more similar between categories, for learning material less similar within categories, and for more complex learning material. These results are consistent with the theoretical account of interleaved learning, most notably with the sequential theory of attention (attentional bias framework). We conclude that interleaving can effectively foster inductive learning but that the setting and the type of learning material must be considered. The interleaved learning, however, should be used with caution in certain conditions, especially for expository texts and words.
Article
Full-text available
The use of linear mixed effects models (LMMs) is increasingly common in the analysis of biological data. Whilst LMMs offer a flexible approach to modelling a broad range of data types, ecological data are often complex and require complex model structures, and the fitting and interpretation of such models is not always straightforward. The ability to achieve robust biological inference requires that practitioners know how and when to apply these tools. Here, we provide a general overview of current methods for the application of LMMs to biological data, and highlight the typical pitfalls that can be encountered in the statistical modelling process. We tackle several issues regarding methods of model selection, with particular reference to the use of information theory and multi-model inference in ecology. We offer practical solutions and direct the reader to key references that provide further technical detail for those seeking a deeper understanding. This overview should serve as a widely accessible code of best practice for applying LMMs to complex biological problems and model structures, and in doing so improve the robustness of conclusions drawn from studies investigating ecological and evolutionary questions.
Article
Full-text available
General Audience Summary When learning multiple categories by viewing members of the categories, the order in which these examples are presented has been shown to influence learning. For instance, if you were trying to learn to identify and differentiate hawks and falcons, presenting examples of each category in an interleaved sequence during training (HFHFHF) would likely improve your ability to correctly categorize hawks and falcons, relative to a blocked training sequence (HHHFFF). We investigated whether this apparent benefit of interleaved presentation extended to the learning of educationally relevant scientific categories (i.e., organic chemicals), and also whether the following factors played a role: (a) the level of similarity among the examples and (b) highlighting of the diagnostic features of each category during training. Across four experiments, we found that interleaved presentation benefitted learning of the chemical categories, with simple as well as more complex categories. Neither the level of similarity among the examples nor highlighting the diagnostic features of a category influenced the interleaving advantage, although highlighting the diagnostic features did improve learning overall. These findings suggest that students may benefit from the interleaving of different categories during science instruction, even when the teacher (or the textbook) explicitly points out the diagnostic features of each category to help learning.
Article
Interleaving examples of to-be-learned categories, rather than blocking examples by category, can enhance learning. We examine the reliability of the interleaving effect between- (Experiments 1 and 2) and within-participants (Experiment 3). As a between-participant effect, we examined a broad spectrum of working memory by both measuring individual capacity and manipulating the task demand. Findings reveal a robust interleaving effect across the spectrum, eliminated only at the lowest and highest ends, but never reversed. In Experiment 3, we used an empirically defined source of potential heterogeneity by examining whether the size of the interleaving benefit a participant experiences on one set of stimuli predicts the size of the interleaving benefit that same participant experiences on two other sets of stimuli. It did not, with only a very small correlation between the two more similar stimuli sets. Taken together, these results add to the burgeoning literature on the robustness of the interleaving benefit.
Article
Visual categorization is fundamental to expertise in a wide variety of disparate domains, such as radiology, art history, and quality control. The pervasive need to master visual categories has served as the impetus for a vast body of research dedicated to exploring how to enhance the learning process. The literature is clear on one point: no category learning technique is always superior to another. In the present review, we discuss how two factors moderate the efficacy of learning techniques. The first, category similarity, refers to the degree of featural overlap of exemplars. The second moderator, category type, concerns whether the features that define category membership can be mastered through learning processes that are implicit/non-verbal (information-integration categories) or explicit/verbal (rule-based categories). The literature on each moderator has been conducted almost entirely in isolation, such that their potential interaction remains underexplored. We address this gap in the literature by reviewing empirical and theoretical evidence that these two moderators jointly influence the efficacy of learning techniques.
Article
Interleaved practice involves studying exemplars from different categories in a non-systematic, pseudorandom order under the constraint that no two exemplars from the same category are presented consecutively. Interleaved practice of materials has been shown to enhance test performance compared to blocked practice in which exemplars from the same category are studied together. Why does interleaved practice produce this benefit? We evaluated two non-mutually exclusive hypotheses, the discriminative-contrast hypothesis and the distributed-practice hypothesis, by testing participants’ performance on calculating the volume of three-dimensional geometric shapes. In Experiment 1, participants repeatedly practiced calculating the volume of four different-sized shapes according to blocked practice, interleaved practice, or remote-interleaved practice (which involved alternating the practice of volume calculation with non-volume problems, like permutations and fraction addition). Standard interleaving enhanced performance compared to blocked practice but did not produce enhanced performance compared to remote interleaving. In Experiment 2, we replicated this pattern and extended the results to include a remote-blocked group, which involved blocking volume calculation with non-volume problems. Performance on key measures was better for remote-interleaved groups compared to remote-blocked groups, a finding that supports the distributed-practice hypothesis.