Test-potentiated learning: three independent replications,a disconfirmed hypothesis,and an unexpected boundary condition

Arnold and McDermott [(2013). Test-potentiated learning: Distinguishing between direct and indirect effects of testing. Journal of Experimental Psychology: Learning, Memory, and Cognition, 39, 940–945] isolated the indirect effects of testing and concluded that encoding is enhanced to a greater extent following more versus fewer practice tests, referred to as test-potentiated learning. The current research provided further evidence for test-potentiated learning and evaluated the covert retrieval hypothesis as an alternative explanation for the observed effect. Learners initially studied foreign language word pairs and then completed either one or five practice tests before restudy occurred. Results of greatest interest concern performance on test trials following restudy for items that were not correctly recalled on the test trials that preceded restudy. Results replicate Arnold and McDermott (2013) by demonstrating that more versus fewer tests potentiate learning when trial time is limited. Results also provide strong evidence against the covert retrieval hypothesis concerning why the effect occurs (i.e., it does not reflect differential covert retrieval during pre-restudy trials). In addition, outcomes indicate that the magnitude of the test-potentiated learning effect decreases as trial length increases, revealing an unexpected boundary condition to test-potentiated learning.     

Discrepancy detection in the retrieval-enhanced suggestibility paradigm

Retrieval-enhanced suggestibility (RES) refers to the finding that immediately recalling the details of a witnessed event can increase susceptibility to later misinformation. In three experiments, we sought to gain a deeper understanding of the role that retrieval plays in the RES paradigm. Consistent with past research, initial testing did increase susceptibility to misinformation – but only for those who failed to detect discrepancies between the original event and the post-event misinformation. In all three experiments, subjects who retrospectively detected discrepancies in the post-event narratives were more resistant to misinformation than those who did not. In Experiments 2 and 3, having subjects concurrently assess the consistency of the misinformation narratives negated the RES effect. Interestingly, in Experiments 2 and 3, subjects who had retrieval practice and detected discrepancies were more likely to endorse misinformation than control subjects who detected discrepancies. These results call attention to limiting conditions of the RES effect and highlight the complex relationship between retrieval practice, discrepancy detection, and misinformation endorsement.     

Retrieval activates related words more than presentation

Retrieving information enhances learning more than restudying. One explanation of this effect is based on the role of mediators (e.g., sand-castle can be mediated by beach). Retrieval is hypothesised to activate mediators more than restudying, but existing tests of this hypothesis have had mixed results [Carpenter, S. K. (2011). Semantic information activated during retrieval contributes to later retention: Support for the mediator effectiveness hypothesis of the testing effect. Journal of Experimental Psychology: Learning, Memory, and Cognition, 37(6), 1547–1552. doi:10.1037/a0024140; Lehman, M., &; Karpicke, J. D. (2016). Elaborative retrieval: Do semantic mediators improve memory? Journal of Experimental Psychology: Learning, Memory, and Cognition, 42(10), 1573–1591. doi:10.1037/xlm0000267]. The present experiments explored an explanation of the conflicting results, testing whether mediator activation during a retrieval attempt depends on the accessibility of the target information. A target was considered less versus more accessible when fewer versus more cues were given during retrieval practice (Experiments 1 and 2), when the target had been studied once versus three times initially (Experiment 3), or when the target could not be recalled versus could be recalled during retrieval practice (Experiments 1–3). A mini meta-analysis of all three experiments revealed a small effect such that retrieval activated mediators more than presentation, but mediator activation was not reliably related to target accessibility. Thus, retrieval may enhance learning by activating mediators, in part, but these results suggest the role of other processes, too.     

How is the visual working memory load of driver influenced by information density of traffic signs?

The rapid development of urban roads has resulted in an increase in the amount of information that needs to be displayed on traffic signs. The overloading of information on traffic signs may increase drivers’ visual cognition burden and cause driving errors. It could be an important factor affecting driving safety and result in accidents. To enable drivers to cognize traffic signs, the information density that they could endure under reasonable workload levels should be determined. In this study, we designed and implemented trials to analyze drivers’ visual working memory load (VWML) caused by different information densities (including stimulus intensity (SI) and interval time of stimulus (IS)) of traffic signs. Twenty-four paid participants (12 females) with a mean age of 23.38 years (SD = 1.38) participated in laboratory simulation trials based on the memory retrieval approach. Subjective ratings (rating scale mental effort (RSME)), and performance measurement based on reaction time and accuracy of information-matching task were measured to reflect drivers’ VWML at different information densities. The results show that information density is an important factor affecting drivers’ VWML. The following are the significant observations of the study: (1) The subjective ratings on mental effort increased with increase in the SI. Meanwhile, it first decreased and then increased with increase in the IS. (2) The response accuracy results show that a high accuracy (≥0.7) was achieved by drivers in an IS of 15 s when the SI was at most 14.16 bit/m². Correspondingly, the drivers underwent less mental workload as a result of their better short-term memory characteristic. (3) The mean reaction time was over 1800 ms when the SI was over 14.16 bit/m². This indicated a reduction in the information-processing capability of the drivers. The results indicate that cognition of information of unreasonable density on traffic signs would increase drivers’ VWML. This would, in turn, result in an increase in their reaction time and decrease in accuracy. The threshold of information density was also determined through the experiment. It would satisfy the information demand of drivers and prevent information redundancy. The results contribute to an understanding of drivers’ VWML while cognizing traffic signs with different information densities and provide safety, effectiveness, and information continuity considerations for the design and setting of traffic signs.     

不同情境转换对重复学习条件下类别-例词回忆的影响

为了考察重复学习条件下内部心境转换和外部情境转换对类别-例词回忆的影响,本研究参考提取练习范式,进行了两个实验。实验1选取96名被试,在初始学习与重复学习阶段之间通过指导语操纵被试维持初始学习时的心境,或转换为合作心境,或转换为竞争心境;实验2选取64名被试,被试在不同外部情境中进行初始学习和重复学习,测试时或处于初始学习的情境,或处于重复学习的情境。回忆测验的结果显示:与未经历重复学习的基线词相比,同类别重学的例词回忆率高于基线词,表现出促进现象;当发生内部心境转换,或发生外部情境转换且在重复学习的情境测试时,同类别中未重学例词的回忆率低于基线词,表现出削弱现象;当发生外部情境转换但在初始学习情境测试时,同类别未重学例词的回忆却高于基线词。可采用联结凸显-曝光-线索扩展性的三维度观点解释这一结果。     

Affective matching moderates S–R binding

We investigated the moderating influence of affective matching on S–R binding processes in a sequential priming study in which positive and negative nouns had to be categorised as referring to a person or to an object. Irrelevant positive and negative distractor words (adjectives) were integrated with responses into S–R episodes if they had the same valence as the target (affective match condition). In this case, repeating the prime distractor in the probe led to a retrieval of the prime response, which facilitated performance for response repetition sequences but had no effect on performance when responses changed between prime and probe. However, if target and distractor had different valences (affective mismatch condition), no interaction of distractor relation and response relation occurred, indicating that distractors were less likely to be associated with responses into event files during the prime trial episode. Findings reveal that affective mismatches are detected automatically and modulate a binding of irrelevant information with responses.     

Increasing the number of contacts generated during contact tracing interviews

ABSTRACT

To stem the spread of infectious diseases (e.g., Ebola), epidemiologists conduct contact tracing interviews (CTIs) with infected individuals regarding their contacts who may also be infected. These contact tracing interviews, however, may be vulnerable to deadly errors of omission. A promising technique to maximise recall is the Cognitive Interview (CI), which is grounded in psychological theory. In the present study, participants imagined they were infected with meningococcal meningitis and reported their contacts over the previous three days during either a control interview or a CI. To model the cognitive impairment associated with being sick, half of the participants were interviewed while simultaneously completing a cognitive impairment task. The CI generated more contacts than the control interview. However, when the cognitive impairment task was completed, the CI and the control interview performed similarly. We recommend the CI be considered as an alternative to the control interview, particularly if the interviewee is not temporarily impaired due to distraction or illness.     

Reminders can enhance or impair episodic memory updating: a memory-for-change perspective

ABSTRACT

The Memory-for-Change framework proposes that retrieving episodic memories can facilitate new learning when changes between existing memories and new information are integrated during encoding and later recollected. Four experiments examined whether reminders could improve memory updating and enhance new learning. Participants studied two study lists of word pairs and were given a cued recall test on responses from both lists. Reminders of List 1 words pairs (A-B) appeared immediately before List 2 words pairs that included repeated cues and changed responses (A-D). Across experiments, we varied the types of reminders to determine whether differences in their effectiveness as retrieval cues would influence memory for the list membership of responses. We found that presenting intact reminders (cue-response) enhanced the memory benefits associated with recollection-based retrieval of changes relative to when no reminders appeared and when partial reminders (cue-only) appeared with and without feedback. Importantly, cue-response reminders benefitted memory when they were recognised in List 2 and when changes were later recollected. This suggests that integrative encoding can be facilitated when substantial environmental support is available to cue retrieval of existing memories. These findings have practical implications for understanding which reminders best aid the correction of memories for inaccurate information.     

Using prequestions to enhance learning from reading passages: the roles of question type and structure building ability

ABSTRACT

Answering questions before learning something (“prequestions”) enhances learning. However, these benefits usually occur for information that was asked in the prequestions (i.e. prequestioned material), and not for non-prequestioned material. We reasoned that this narrow benefit may be due to the fact that studies typically use fairly simple prequestions that have a clear answer within one part of the learning material – isolative prequestions. We explored the effects of integrative prequestions that required participants to make connections across different parts of a reading passage. Experiment 1 showed the usual benefit of isolative prequestions on prequestioned but not on non-prequestioned material, but no benefit of integrative prequestions. However, in Experiment 2 when participants were given instructions to seek the answers while reading, integrative prequestions benefited learning of both prequestioned and non-prequestioned material. Individual differences in structure building positively predicted performance, but did not interact with the effects of prequestions.     

Effects of working memory load on long‐term word priming

Three experiments assessed the role of verbal and visuo‐spatial working memory in supporting long‐term repetition priming for written words. In Experiment 1, two priming tasks (word stem completion and category‐exemplar production) were included with three levels of load on working memory: (1) without memory load, (2) memory load that involved storing a string of six digits, and (3) memory load that involved storing a graphic shape. Experiments 2 and 3 compared the effects of a verbal (Experiment 2) or a visual (Experiment 3) working memory load at encoding on both an implicit (word stem completion) and an explicit test (cued recall). The results show no effect of memory load in any of the implicit memory tests, suggesting that priming does not rely on working memory resources. By contrast, loading working memory at encoding causes a significant disruptive effect on the explicit memory test for words when the load is verbal but not visual.