RESUMEN

In a recently published article, Van de Calseyde and Efendic (2022) argue that inner-crowd wisdom (i.e., the reduction in error afforded by aggregating two estimates from a given person relative to a single initial estimate from that person) is enhanced when people are instructed to adopt the perspective of someone with whom they disagree prior to making a second estimate. Here, I present a reanalysis of Van de Calseyde and Efendic's data and argue that evidence supporting their primary claim spuriously arises from anticonservative multilevel models. Specifically, Van de Calseyde and Efendic assess their data via random-intercept models and fail to account for item-level effects of experimental condition. Such an approach generally allows analysts to reap the enhanced statistical power of multilevel models without implementing appropriate checks on that power; in this case, underestimation of item-level variance appears to have driven an illusory benefit of perspective taking.

Asunto(s)

RESUMEN

The fields of machine learning (ML) and cognitive science have developed complementary approaches to computationally modeling human behavior. ML's primary concern is maximizing prediction accuracy; cognitive science's primary concern is explaining the underlying mechanisms. Cross-talk between these disciplines is limited, likely because the tasks and goals usually differ. The domain of e-learning and knowledge acquisition constitutes a fruitful intersection for the two fields' methodologies to be integrated because accurately tracking learning and forgetting over time and predicting future performance based on learning histories are central to developing effective, personalized learning tools. Here, we show how a state-of-the-art ML model can be enhanced by incorporating insights from a cognitive model of human memory. This was done by exploiting the predictive performance equation's (PPE) narrow but highly specialized domain knowledge with regard to the temporal dynamics of learning and forgetting. Specifically, the PPE was used to engineer timing-related input features for a gradient-boosted decision trees (GBDT) model. The resulting PPE-enhanced GBDT outperformed the default GBDT, especially under conditions in which limited data were available for training. Results suggest that integrating cognitive and ML models could be particularly productive if the available data are too high-dimensional to be explained by a cognitive model but not sufficiently large to effectively train a modern ML algorithm. Here, the cognitive model's insights pertaining to only one aspect of the data were enough to jump-start the ML model's ability to make predictions-a finding that holds promise for future explorations.

Asunto(s)

RESUMEN

We investigated the effect of expertise on the wisdom of crowds. Participants completed 60 trials of a numerical estimation task, during which they saw 50-100 asterisks and were asked to estimate how many stars they had just seen. Experiment 1 established that both inner- and outer-crowd wisdom extended to our novel task: Single responses alone were less accurate than responses aggregated across a single participant (showing inner-crowd wisdom) and responses aggregated across different participants were even more accurate (showing outer-crowd wisdom). In Experiment 2, prior to beginning the critical trials, participants did 12 practice trials with feedback, which greatly increased their accuracy. There was a benefit of outer-crowd wisdom relative to a single estimate. There was no inner-crowd wisdom effect, however; with high accuracy came highly restricted variance, and aggregating insufficiently varying responses is not beneficial. Our data suggest that experts give almost the same answer every time they are asked and so they should consult the outer crowd rather than solicit multiple estimates from themselves.

Asunto(s)

RESUMEN

We investigated the cognitive processes that cause confidence to increase. Participants were asked 48 general-knowledge questions either once or three times, without feedback. After 2 min (Experiment 1) or 48 h (Experiment 2) they were asked the same questions again, and rated their confidence. Repeated questioning increased confidence but not accuracy. This increase, which replicated research on episodic memory in the eyewitness literature (e.g., Shaw, Journal of Experimental Psychology: Applied 2: 126-146, 1996), occurred even though accuracy was only around 25%. A mediation analysis identified response repetition, but not fluency, as a mechanism underlying growth in confidence. Thus, the basis for confidence judgments appears to be whether one's current response has been generated previously. In sum, answering a factual question increases confidence, but not accuracy, and this happens because learners use response repetition as a cue for confidence judgments.

Asunto(s)

RESUMEN

Testing is a powerful enhancer of memory. However, if initial encoding is poor, and subsequent retrieval practice is likely to fail, then the benefits of testing are diminished or even eliminated. Previous work has suggested that the benefits of testing may be preserved under difficult conditions with a scaffolded technique called diminishing-cues retrieval practice (DCRP; Fiechter & Benjamin, Psychonomic Bulletin & Review, 25(5), 1868-1876, 2018). DCRP provides increasing retrieval demands over practice, but does not adapt to individual learners or to materials of varying difficulty. Here, we evaluate a new technique called adaptive-cues retrieval practice (ACRP). ACRP adapts to an individual's moment-to-moment ability by providing within-trial accumulated cuing, generating more demanding retrieval practice for better learned items. Across six experiments, learners practiced English-Iñupiaq word pairs using ACRP, standard retrieval practice, restudy, and DCRP. ACRP is even more effective than DCRP in situations where standard retrieval practice is ineffective. When testing is most effective, ACRP, DCRP, and standard retrieval practice all enhance memory to approximately the same degree, but DCRP requires the least practice time. Our findings suggest that DCRP is a more efficient technique for learning, but that the benefits of ACRP extend to more learning scenarios than those of any other identified practice regimen.

Asunto(s)

RESUMEN

Video job interviews have become a common hiring practice, allowing employers to save money and recruit from a wider applicant pool. But differences in job candidates' internet connections mean that some interviews will have higher audiovisual (AV) quality than others. We hypothesized that interviewers would be impacted by AV quality when they rated job candidates. In two experiments, participants viewed two-minute long simulated Skype interviews that were either unedited (fluent videos) or edited to mimic the effects of a poor internet connection (disfluent videos). Participants in both experiments rated job candidates from fluent videos as more hirable, even after being explicitly told to disregard AV quality (experiment 2). Our findings suggest that video interviews may favor job candidates with better internet connections and that being aware of this bias does not make it go away.

RESUMEN

Retrieval practice has been shown to be a highly effective tool for enhancing memory, a fact that has led to major changes to educational practice and technology. However, when initial learning is poor, initial retrieval practice is unlikely to be successful and long-term benefits of retrieval practice are compromised or nonexistent. Here, we investigate the benefit of a scaffolded retrieval technique called diminishing-cues retrieval practice (Finley, Benjamin, Hays, Bjork, & Kornell, Journal of Memory and Language, 64, 289-298, 2011). Under learning conditions that favored a strong testing effect, diminishing cues and standard retrieval practice both enhanced memory performance relative to restudy. Critically, under learning conditions where standard retrieval practice was not helpful, diminishing cues enhanced memory performance substantially. These experiments demonstrate that diminishing-cues retrieval practice can widen the range of conditions under which testing can benefit memory, and so can serve as a model for the broader application of testing-based techniques for enhancing learning.

Asunto(s)

RESUMEN

Practice tests provide large mnemonic benefits over restudying, but learners judge practice tests as less effective than restudying. Consequently, learners infrequently utilize testing when controlling their study and often choose to be tested only on well-learned items. In 5 experiments, we examined whether learners' choices about testing and restudying are effective for improving subsequent memory performance. Learners studied a list of word pairs and chose which items to restudy and which to test. Some of learners' choices were honored (by assigning those items to the chosen activity) and some of learners' choices were dishonored (by assigning those items to the opposite study activity). Surprisingly, and in contrast with all work to date on the metacognitive monitoring of testing effects, honoring learners' testing choices consistently resulted in better memory performance than dishonoring choices. This effect occurred principally because learners often chose to restudy difficult items, and those items did not benefit from testing. The effectiveness of learners' choices about testing casts the metacognition of testing in a new light: learners may not appreciate the benefits of testing, but they do have an understanding of circumstances in which the benefits of testing are minimal. (PsycINFO Database Record

Asunto(s)

RESUMEN

In five experiments, we investigated whether expected retention intervals affect subjects' encoding strategies. In the first four experiments, our subjects studied paired associates consisting of words from the Graduate Record Exam and a synonym. They were told to expect a test on a word pair after either a short or a longer interval. Subjects were tested on most pairs after the expected retention interval. For some pairs, however, subjects were tested after the other retention interval, allowing for a comparison of performance at a given retention interval conditional upon the expected retention interval. No effect of the expected retention interval was found for 1 min versus 4 min (Exp. 1), 30 s versus 3 min (Exp. 2), and 30 s versus 10 min (Exps. 3 and 4), even when subjects were given complete control over the pacing of study items (Exp. 4). However, when the difference between the expected retention intervals was increased massively (10 min vs. 24 h; Exp. 5), subjects remembered more items that they expected to be tested sooner, an effect consistent with the idea that they traded off efforts to remember items for the later test versus items that were about to be tested. Overall, this set of results accords with much of the test-expectancy literature, revealing that subjects are often reluctant to adjust encoding strategies on an item-by-item basis, and when they do, they usually make quantitative, rather than qualitative, adjustments.

Asunto(s)