Journal of Experimental Psychology Applied

Irving, Dulcie; Clark, Robbie W. A.; Lewandowsky, Stephan; Allen, Peter J.

doi: 10.1037/xap0000408pmid: 35007097

Although retractions significantly reduce the number of references people make to misinformation, retracted information nevertheless persists in memory, continuing to influence reasoning. One hundred and twenty-nine lay participants completed an adaptation on the traditional continued influence paradigm, which set out to identify whether it is possible to debunk a piece of common statistical misinformation: inappropriate causal inference based on a correlation. We hypothesized that participants in the correction condition would make fewer causal inferences (misinformation) and more correlational inferences (correction) than those in the no-correction condition. Additional secondary hypotheses were that the number of references made to the misinformation and correction would be moderated by the level of trust in science and scientists, and the amount of television that participants watch. Although the secondary hypotheses were not supported, the data strongly supported the primary hypotheses. This study provides evidence for the efficacy of corrections about misinformation where correlational evidence has been inappropriately reported as causal.

Sanderson, Penelope; McLanders, Mia; Santomauro, Chiara; Tran, Jimmy; Fouhy, Sarah; Shapiro, Jesse; Brown, Dannielle; McNulty, Emma

doi: 10.1037/xap0000373pmid: 34435848

Healthcare workers often monitor patients while moving between different locations and tasks, and away from conventional monitoring displays. Vibrotactile displays can provide patient information in vibrotactile patterns that are felt regardless of the worker’s location. We examined how effectively participants could identify changes in vibrotactile representations of patient heart rate (HR) and oxygen saturation (SpO2). In Experiment 1, participants identified changes in HR and SpO2 with greater than 90% accuracy while using vibrotactile displays configured in either an integrated or a separated format. In Experiment 2, incidental auditory and visual cues were removed and performance was still greater than 90% for the integrated display. In Experiments 3 and 4, ongoing tasks with low or high task load were introduced; high load worsened participants’ response accuracy and speed at identifying vital signs. In Experiments 5 and 6, alternative designs were tested, including a design with a seemingly more natural mapping of HR to vibrotactile stimulation. However, no design supported more accurate performance than the integrated display. Results are interpreted with respect to multiple resource theory, and constraints on conforming to design guidelines are noted. Vibrotactile displays appear to be viable and therefore potentially suitable for use in healthcare and other contexts.

Hout, Michael C.; White, Bryan; Madrid, Jessica; Godwin, Hayward J.; Scarince, Collin

doi: 10.1037/xap0000295pmid: 34941350

In many important search tasks, observers must find what they are looking for using only visual information (e.g., X-ray baggage screening/medical screening). However, numerous other search tasks can only be effectively completed when the searcher uses their hands to find what they are looking for (e.g., “rummage” search). Unfortunately, it is not currently well understood how observers conduct such “interactive” searches nor what the best strategies might be for doing so. Here, we first review the limited literature on interactive search. We then present a novel methodology for the study of interactive search that involves having observers seek out LEGO® targets in a cluttered tray of assorted bricks. In our validation task, we confirm the validity of this approach by demonstrating that it produces sensible patterns of diminishing returns in response time as targets are removed from the set as well as hastened search times for larger targets. In our experiment, we modify the approach, refining its systematicity and experimental control. We also build on prior work exploring strategy use in visual search by investigating the extent to which active and passive strategy use impacts performance in interactive search. In contrast to our prior findings in hybrid visual search (Madrid & Hout, 2019), our current findings suggest that in interactive search, an active search strategy can be superior to a passive one. We close by offering a conceptual model (the Interactive Multiple Decision Model [i-MDM]) that explicates the steps involved in a search task of this nature, and we then provide suggestions for how to further refine the task to achieve higher internal validity and to delve deeper into questions of theoretical importance in the field of interactive search.

Jago, Arthur S.; Carroll, Glenn R.; Lin, Mariana

doi: 10.1037/xap0000365pmid: 35266793

In an increasing number of domains, people interact with automated agents (such as algorithms, robots, and computers) instead of humans. Across five studies, we explore the role of authenticity in shaping people’s reactions to automated agents’ work. In doing so, we examine two basic ways to generate authenticity in autonomous technological work: (a) highlighting the human origins of autonomous technologies and (b) anthropomorphizing autonomous technologies, or presenting them with human-like qualities. We find strong evidence that human origin stories generate authenticity, but much less evidence that simple anthropomorphic cues do so to the same degree (Studies 1–3). Simply prompting people to consider human origins can also generate attributions of authenticity (Study 4), which translates into intended and recommended support for automated work (Study 5). We discuss how managers of organizations can implement automated systems in ways that encourage attributions of authenticity.

Tullis, Jonathan G.; Qiu, Jiahui

doi: 10.1037/xap0000350pmid: 33939460

Students frequently generate mnemonic cues to help them remember difficult or abstract information (Tullis & Maddox, Metacognition and Learning, 2020, 15, 129). Self-generated mnemonics have the potential to be particularly effective means of remembering target information because they can transform abstract information into meaningful units, connect information to existing schema, and create distinct retrieval routes to the targets. Across five experiments, we compared the effectiveness of self-generated mnemonics to mnemonics generated by others for remembering chemistry information. Generating one’s own mnemonics consistently boosted recall for both the chemistry content and the mnemonic itself. However, experimentally boosting recall of mnemonics through retrieval practice did not affect recall of associated chemistry content. These results indicate that improved recall of chemistry content is not caused by better recall of the mnemonic itself; rather, generating a mnemonic involves deep and effortful processing of chemistry content that boosts recall more than reading someone else’s mnemonic.

Sarno, Dawn M.; McPherson, Rachel; Neider, Mark B.

doi: 10.1037/xap0000410pmid: 35157483

Most previous phishing interventions have employed discrete training approaches, such as brief instructions aimed at improving phishing detection. However, these discrete interventions have demonstrated limited success. The present studies focused on developing an alternative to discrete training by providing college-age adults with a persistent classification aid that guided them on what characteristics a phishing email might include. Experiment 1 determined if this classification aid improved email categorization performance relative to feedback and control. Experiment 2 continued the evaluation of the classification aid to determine whether performance improvements were due to increased systematic processing of emails. Experiment 3 explored whether the classification aid would be more effective when embedded directly into the email interface. The results suggested three major findings. (a) Persistent interventions can improve phishing email detection. (b) Performance improvements were largest when the classification aid was embedded into the task. (c) These benefits were likely driven by an improved systematic processing of the emails. This novel phishing classification aid serves as a promising persistent intervention that can be adaptable to specific email environments and individuals.

Hartwig, Marissa K.; Rohrer, Doug; Dedrick, Robert F.

doi: 10.1037/xap0000391pmid: 34990156

Many randomized controlled experiments in the classroom have found that mathematics learning is improved dramatically when practice problems of one kind are distributed across multiple assignments (spaced) and mixed with other kinds of problems (interleaved). In two studies, we investigated students’ knowledge of spacing and interleaving. In Study 1, 193 undergraduates designed learning schedules for a hypothetical math class. In Study 2, 175 undergraduates selected from among five hypothetical schedules in response to a variety of questions, provided reasons for their selections, and rated the utility of spacing and interleaving. In both studies, most participants incorrectly judged schedules with minimal degrees of spacing and interleaving to be most effective. Also, schedules with more spacing and interleaving were perceived as more difficult, less enjoyable, and less common. Participants’ ratings of utility revealed mixed perspectives on spacing and an underappreciation of interleaving. Altogether, these findings demonstrate that most students fail to recognize the benefits of spaced and interleaved practice. Further, by identifying specific ways in which their beliefs about spacing and interleaving fall short, we reveal opportunities to reshape students’ beliefs to foster these effective learning techniques.

Uner, Oyku; Tekin, Eylul; Roediger, Henry L.

doi: 10.1037/xap0000363pmid: 34110858

Testing with various formats enhances long-term retention of studied information; however, little is known whether true–false tests produce this benefit despite their frequent use in the classroom. We conducted four experiments to explore the retention benefits of true–false tests. College students read passages and reviewed them by answering true–false questions or by restudying correct information from the passages. They then took a criterial test 2 days later that consisted of short-answer questions (Experiments 1 and 2) or short-answer and true–false questions (Experiments 3 and 4). True–false tests enhanced retention compared to rereading correct statements and compared to typing those statements while rereading (the latter in a mini meta-analysis). Evaluating both true and false statements yielded a testing effect on short-answer criterial tests, whereas evaluating only true statements produced a testing effect on true–false criterial tests. Finally, a simple modification that asked students to correct statements they marked as false on true–false tests improved retention of those items when feedback was provided. True–false tests can be an effective and practical learning tool to improve students’ retention of text material.

Erdmann, Julia; Rummel, Nikol

doi: 10.1037/xap0000377pmid: 34323546

Extinction learning and recovery effects are known to be essential learning mechanisms. We investigated whether spontaneous recovery effects transfer to skill acquisition and play a role in the context of widely used but also criticized German orthography instruction methods that implement a prolonged phase of phonological spelling without correction. To test our hypothesis in a controlled manner, in the current experimental study, adult participants learned how to spell an artificial vocabulary with the help of a phonological spelling instruction method that emulates the criticized German orthography instruction methods. We compared participants (148 University students) who did not receive correction in the initial learning phase (delayed correction) with participants who received correction from the beginning (immediate correction). Although the incorrect spellings which were practiced in the initial learning phase (acquisition phase) were extinguished in the subsequent learning phase (extinction phase), the erroneous spellings recovered after a break of 24 hr in the delayed correction condition. The results indicate that learning with methods that emphasize phonological spelling, without correction in the initial learning phase, can lead to spontaneous recovery of initially learned incorrect spellings after the incorrect spellings have seemingly been extinguished.

Pena-Marin, Jorge; Bhargave, Rajesh

doi: 10.1037/xap0000401pmid: 34735184

Decision-makers frequently use numerical estimates, such as distances, future prices, and the expected timing of events. How do they evaluate these estimates, once they observe the “correct” values? This research finds that when people evaluate estimates for their accuracy, they judge underestimates to be better than overestimates in these comparisons. For instance, if an obtained stock price is $25, an underestimate interval ($21–$23) is evaluated better than an overestimate interval ($27–$29), and a point value underestimate of $22 is evaluated better than an overestimate ($28). Perceptions of under (vs. over) estimates in turn influence important decisions, such as which forecasters to hire, based on their past estimates. Six experiments demonstrate this effect across multiple decision domains and support a numerical processing framework for the finding. The results suggest that people find it easier (more difficult) to calculate the difference between an underestimate (overestimate) and the correct value, which in turn influences the perceived accuracy of the estimate.