The Impact of Adding a Fourth Item to the Traditional 3-Item Remote Associates Test
Saved in:
| Title: | The Impact of Adding a Fourth Item to the Traditional 3-Item Remote Associates Test |
|---|---|
| Language: | English |
| Authors: | Jose A. Diaz, Steven M. Nelson, A. Alexander Beaujean, Adam E. Green, Michael K. Scullin (ORCID |
| Source: | Creativity Research Journal. 2024 36(4):604-616. |
| Availability: | Routledge. Available from: Taylor & Francis, Ltd. 530 Walnut Street Suite 850, Philadelphia, PA 19106. Tel: 800-354-1420; Tel: 215-625-8900; Fax: 215-207-0050; Web site: http://www.tandf.co.uk/journals |
| Peer Reviewed: | Y |
| Page Count: | 13 |
| Publication Date: | 2024 |
| Sponsoring Agency: | National Science Foundation (NSF) |
| Contract Number: | 1920730 |
| Document Type: | Journal Articles Reports - Research |
| Education Level: | Higher Education Postsecondary Education |
| Descriptors: | Test Items, Creative Thinking, Creativity Tests, Performance, Accuracy, Association (Psychology), Reaction Time, Young Adults, Difficulty Level, Foreign Countries, Undergraduate Students, Native Language, English, English (Second Language), Individual Differences, Cognitive Processes |
| Geographic Terms: | United States, United Kingdom |
| Assessment and Survey Identifiers: | Remote Associates Test |
| DOI: | 10.1080/10400419.2023.2200597 |
| ISSN: | 1040-0419 1532-6934 |
| Abstract: | The compound Remote Associates Test (RAT) is a classic measure of creativity. Participants are shown three cue words (sore-shoulder-sweat) and asked to generate a word that connects them (cold). Theoretical views of RAT performance differ in the degree to which they conceptualize performance as depending on automatic spreading activation across semantic networks, strategic generation of bi-associations, and other analytical processes (e.g. executive processes that support fluid intelligence). We tested these views by adding a fourth cue word to determine whether it impaired RAT accuracy (e.g. generation of bi-associations), impaired response times (analytic processes), or improved RAT accuracy without changing response times (e.g. spreading activation). Across four experiments, 551 adults completed 3- and 4-item RAT trials that were matched on linguistic and semantic metrics. Across experiments, adding the fourth word improved accuracy by 27.91%. This performance gain occurred with either modest or no changes to response times or ratings of insight/strategy use. Interestingly, the fourth word predominantly benefited accuracy and response times on difficult trials; on easy trials, the fourth word impaired or did not change performance. The findings suggest that both automatic and strategic/analytical processes contribute to successful RAT performance, with relative dependence on these processes dynamically adapting to the demands of the individual trial. |
| Abstractor: | As Provided |
| Entry Date: | 2024 |
| Accession Number: | EJ1444260 |
| Database: | ERIC |
|
Full text is not displayed to guests.
Login for full access.
|
|
| FullText | Links: – Type: pdflink Url: https://content.ebscohost.com/cds/retrieve?content=AQICAHj0k_4E0hTGH8RJwT4gCJyBsGNe_WN95AvKlDbXJGqwxwH6Ew0m3rsK8QBtMoIrAbJ_AAAA4zCB4AYJKoZIhvcNAQcGoIHSMIHPAgEAMIHJBgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDLEA4fzmdXGNlBekIwIBEICBm5MHT9OPrYaKnSyw9e1I1rG1hK4Y9hCEExfrWXxDmKs7KaYc_zoczepF7wcVEbamx8LcZj8Ozn5wz2AcTiSAT5KN9n1aI9SgnMGiTkgWNo-UGRNGK_DBMEFyyXGKMnW1ekROWY671x1P-PWyqCRL742CzKjH1FQwWiaygj_8wAz4gq9twDa6fbllWbRdjR84x2dIw_DFPCuT_Umg Text: Availability: 1 Value: <anid>AN0180329814;7lo01oct.24;2024Oct21.05:28;v2.2.500</anid> <title id="AN0180329814-1">The Impact of Adding a Fourth Item to the Traditional 3-Item Remote Associates Test </title> <sbt id="AN0180329814-2">Introduction</sbt> <p>The compound Remote Associates Test (RAT) is a classic measure of creativity. Participants are shown three cue words (sore-shoulder-sweat) and asked to generate a word that connects them (cold). Theoretical views of RAT performance differ in the degree to which they conceptualize performance as depending on automatic spreading activation across semantic networks, strategic generation of bi-associations, and other analytical processes (e.g. executive processes that support fluid intelligence). We tested these views by adding a fourth cue word to determine whether it impaired RAT accuracy (e.g. generation of bi-associations), impaired response times (analytic processes), or improved RAT accuracy without changing response times (e.g. spreading activation). Across four experiments, 551 adults completed 3- and 4-item RAT trials that were matched on linguistic and semantic metrics. Across experiments, adding the fourth word improved accuracy by 27.91%. This performance gain occurred with either modest or no changes to response times or ratings of insight/strategy use. Interestingly, the fourth word predominantly benefited accuracy and response times on difficult trials; on easy trials, the fourth word impaired or did not change performance. The findings suggest that both automatic and strategic/analytical processes contribute to successful RAT performance, with relative dependence on these processes dynamically adapting to the demands of the individual trial.</p> <p>Plain Language Summary: Understanding creative achievements in real-world settings requires understanding the cognitive processes that contribute to creativity. One way to do this involves examining performance on standardized creativity instruments across purposefully created conditions. We conducted four experiments in which we modified the semantic information in the Remote Associates Test (RAT), which is one of the more commonly used instruments to assess creativity. We found that as the amount of semantic information provided to participants increased, so too did performance on the RAT. There was, however, a surprising exception: on relatively easy trials, providing more semantic information sometimes hurt performance. These patterns replicated four times using different stimuli sets, designs, and participant samples. Collectively, the findings indicate that multiple cognitive processes are engaged to support creative thinking, with the relative dependence on a given process depending on the difficulty of the problem being solved. Targeting the dynamic nature of automatic and strategic/analytical thinking processes may improve the efficacy of interventions aimed at fostering creative achievements in real-world settings.</p> <p>Creativity is the ability to produce novel and useful ideas (for an overview on approaches to defining creativity, see Plucker, Beghetto, &amp; Dow, [<reflink idref="bib38" id="ref1">38</reflink>]). Higher creative ability is associated with greater success in the classroom (Gajda, Karwowski, &amp; Beghetto, [<reflink idref="bib19" id="ref2">19</reflink>]), in the workplace (Miron-Spektor, Gino, &amp; Argote, [<reflink idref="bib34" id="ref3">34</reflink>]), and in overall well-being (Acar, Tadik, Myers, Van der Sman, &amp; Uysal, [<reflink idref="bib1" id="ref4">1</reflink>]). Creative achievements, though, are the end-result of many cognitive processes (Green, [<reflink idref="bib21" id="ref5">21</reflink>]), including executive control and fluid intelligence (e.g., Frith et al., [<reflink idref="bib18" id="ref6">18</reflink>]; Nusbaum &amp; Silvia, [<reflink idref="bib36" id="ref7">36</reflink>]). An important distinction in creative cognition research has been examining tasks that mostly require divergent thinking versus those that mostly require convergent thinking (Zhang, Sjoerds, &amp; Hommel, [<reflink idref="bib50" id="ref8">50</reflink>]). Divergent thinking tasks require the generation of novel ideas to solve a problem with multiple solutions. By contrast, convergent thinking tasks require identifying the solution to a single-solution problem, typically those in which the solution is not immediately obvious.</p> <p>The current work focuses on convergent thinking in the Remote Associates Test (RAT; Mednick, [<reflink idref="bib33" id="ref9">33</reflink>]). Sarnoff Mednick developed the RAT to assess his associative theory that creativity resulted from forming existing associative elements into new combinations (a view that was historically rooted in Spearman's theory of noegenesis; Fasko, [<reflink idref="bib16" id="ref10">16</reflink>]; Spearman, [<reflink idref="bib43" id="ref11">43</reflink>]). The RAT was developed to elicit domain-general convergent thinking (Ellis, Robison, &amp; Brewer, [<reflink idref="bib15" id="ref12">15</reflink>]), that is, employing knowledge that, in principle, is commonly taught within society (for an alternative to the domain-general view, see Kaufman, Plucker, &amp; Baer, [<reflink idref="bib24" id="ref13">24</reflink>]). Since its development, the RAT has also been used to assess insight, with respondents reporting similar amounts of subjective feelings of insight and reliance on strategies when completing RAT trials (Bowden, Jung-Beeman, Fleck, &amp; Kounios, [<reflink idref="bib10" id="ref14">10</reflink>]; Chein &amp; Weisberg, [<reflink idref="bib11" id="ref15">11</reflink>]).</p> <p>In the original RAT, the trials' stimuli consisted of three cue words and respondents were required to find a fourth word associated with each of the three cue words either by synonymy, semantic association, or forming of a compound word. For example, in a compound RAT trial with the cues <emph>sore-shoulder-sweat</emph>, the best answer is <emph>cold</emph> (i.e., cold sore, cold shoulder, and cold sweat).</p> <p>An advantage of the RAT is that each trial has a single solution, thereby ensuring ease of and uniformity in scoring. Furthermore, the administration is brief enough that it allows for multiple trials for a single respondent (Tik et al., [<reflink idref="bib46" id="ref16">46</reflink>]). A disadvantage of the RAT is that it may not be a so-called process-pure measure of convergent thinking (i.e. also involving some divergent thinking processes; Cortes, Weinberger, Daker, &amp; Green, [<reflink idref="bib13" id="ref17">13</reflink>]); of course, the RAT is not unique in this regard as it would be difficult to support the claim that any instrument capturing a complex psychological attribute has such purity (Loevinger, [<reflink idref="bib29" id="ref18">29</reflink>]). In addition, some work indicates that subjective reports of insight when solving RAT problems actually reflect preparatory control rather than automatic processes (Kounios et al., [<reflink idref="bib26" id="ref19">26</reflink>]). Furthermore, RAT performance is sometimes poor for all but the easy trials (Bowden &amp; Jung-Beeman, [<reflink idref="bib9" id="ref20">9</reflink>]; Chein &amp; Weisberg, [<reflink idref="bib11" id="ref21">11</reflink>]). This notably reduces the dispersion of RAT scores, which subsequently reduces the magnitude at which the RAT scores can predict other phenomena of interest (e.g., academic outcomes, neural correlates; Goodwin &amp; Leech, [<reflink idref="bib20" id="ref22">20</reflink>]).</p> <p>A goal of the current project was to determine whether RAT performance would improve – particularly on difficult trials – if a fourth cue word was added to the typical 3-item trials. In addition to having practical implications, this question is of theoretical interest because accounts of convergent thinking differ in whether they predict that a fourth word would help or harm RAT performance. For example, consider the perspective of the spreading activation model of memory in which semantic knowledge is represented as an organized network of nodes (Collins &amp; Loftus, [<reflink idref="bib12" id="ref23">12</reflink>]) and creativity is determined, in part, by the connectivity across nodes (Benedek et al., [<reflink idref="bib7" id="ref24">7</reflink>]; Kenett, Beaty, Silvia, Anaki, &amp; Faust, [<reflink idref="bib25" id="ref25">25</reflink>]; Marupaka, Iyer, &amp; Minai, [<reflink idref="bib32" id="ref26">32</reflink>]). The more semantic nodes that are activated at a particular time, the more likely activation is to spread to the critical node (i.e., the solution to the problem; see Davelaar, [<reflink idref="bib14" id="ref27">14</reflink>], for similar predictions from the superadditive model that emphasizes a localized semantic search space). A fourth cue word could therefore directly lead to the correct RAT solution by triggering the critical semantic node or indirectly lead to the solution by allowing incorrect solutions to be filtered (Kajić, Gosmann, Stewart, Wennekers, &amp; Eliasmith, [<reflink idref="bib23" id="ref28">23</reflink>]) or otherwise dismissed (Luft, Zioga, Thompson, Banissy, &amp; Bhattacharya, [<reflink idref="bib30" id="ref29">30</reflink>]; Ohlsson, [<reflink idref="bib37" id="ref30">37</reflink>]). Spreading activation is theorized to be an automatic process (rather than a strategic/controlled one), and consistent with this notion, success on RAT trials is often accompanied by the subjective feeling of insight ("aha!"). By the spreading activation view, adding a fourth cue word to the RAT should improve the overall performance because doing so adds additional semantic information.</p> <p>Whereas the spreading activation view posits an additive account of convergent thinking, the bi-associations view posits that convergent thinking requires controlled, multiplicative processing of information. In this multiplicative view, problems are solved in two stages in which people first search for an answer and then test it against the constraints of the trial, repeating this cycle until the answer fits within all constraints (Smith, Huber, &amp; Vul, [<reflink idref="bib42" id="ref31">42</reflink>]). In the context of the RAT, one must generate connections between pairs of unrelated concepts (Benedek, Jurisch, Koschutnig, Fink, &amp; Beaty, [<reflink idref="bib6" id="ref32">6</reflink>]), which would involve generating each of the bi-association pairs of cue words with possible answer words, and strategically evaluating and connecting these associations to determine which answer fits all the constraints. Traditional 3-item RAT trials might be solved with this strategy, but this approach might be less advantageous in a 4-item RAT, which would require more total bi-associate iterations. Because association generation is a controlled process and (presumably) a serial process, adding a fourth cue word would introduce greater chances for errors (i.e., reduce accuracy) and slow overall responding by more than just the time required to read the additional word (~300 ms; Rayner &amp; Clifton, [<reflink idref="bib39" id="ref33">39</reflink>]).</p> <p>A third general view of convergent thinking is that it requires analytic, executive control processes (e.g., Lee, Huggins, &amp; Therriault, [<reflink idref="bib27" id="ref34">27</reflink>]). Prior work has shown that successful RAT performance may be preceded by preparatory activation in frontal regions (Kounios et al., [<reflink idref="bib26" id="ref35">26</reflink>]) and language-based prior knowledge (Tik et al., [<reflink idref="bib46" id="ref36">46</reflink>]). Furthermore, RAT performance is often positively associated with performance on the Raven's Progressive Matrices and other intelligence instruments (e.g., Silvia, [<reflink idref="bib41" id="ref37">41</reflink>]; Taft &amp; Rossiter, [<reflink idref="bib45" id="ref38">45</reflink>]). However, other lines of research have found the RAT to be largely explained by lexical-semantic associative processing with small or no associations to executive processes (e.g., Marko, Michalko, &amp; Riečanský, [<reflink idref="bib31" id="ref39">31</reflink>]). In the current work, the analytical view would predict a fourth cue word to improve accuracy, but at the cost of slower response times (RTs).</p> <p>Based on prior interpretations that the RAT is not process-pure (Cortes, Weinberger, Daker, &amp; Green, [<reflink idref="bib13" id="ref40">13</reflink>]), it seems possible that respondents employ multiple processes – both relatively automatic and strategic/analytical – while solving RAT trials. By this view, adding a fourth word could sometimes help and could sometimes harm RAT performance, with the strategy/process changing as a function of the difficulty level of the trial (Valba, Gorsky, Nechaev, Tamm, &amp; Peel, [<reflink idref="bib47" id="ref41">47</reflink>]). In the RAT, trial difficulty is conceptualized as the semantic distance of the solution word from the individual cue words (e.g., Marko, Michalko, &amp; Riečanský, [<reflink idref="bib31" id="ref42">31</reflink>]). From this view, adding a fourth cue word to trials that have a shorter average semantic distance (easy trials) from the answer word could worsen performance because the extra semantic information would create too many options for bi-association processes. Conversely, adding a fourth cue word to trials with greater average semantic distance (difficult trials) should improve performance because the additional semantic information is necessary for identifying the critical search space in the semantic network (Davelaar, [<reflink idref="bib14" id="ref43">14</reflink>]).</p> <p>We tested these views across four experiments that involved manipulating the addition of a fourth cue word to traditional 3-item RAT trials. The experiments were designed as incremental replications and extensions as follows: In Experiment 1, we aimed to develop an initial set of fourth cue words in which the trials would be matched to the original 3-word trials on average word length, frequency, and forward associative strength. In Experiment 2, we sought to refine the materials generated for Experiment 1 using a broad online sample, inclusive of native English and non-native English speakers. In Experiment 3, the goal was to replicate prior experiments using a sample of only native English speakers, with the methodic addition of including insight and analytic/strategy ratings following each trial. In Experiment 4, we recruited a local sample of university students and utilized a between-subjects design to examine the generalizability of effects. Experiments 1–4 were all approved by the Institutional Review Board at the corresponding author's institution and all participants provided informed consent.</p> <hd id="AN0180329814-3">Experiment 1</hd> <p></p> <hd id="AN0180329814-4">Methods</hd> <p></p> <hd id="AN0180329814-5">Participants</hd> <p>Participants included 110 adults (18–29 years old) who were recruited via the Prolific online research platform in April 2020 (for a similar approach, see Becker &amp; Cabeza, [<reflink idref="bib4" id="ref44">4</reflink>]). An online platform was necessary due to COVID-19 remote-work requirements, but also had the benefit of allowing for samples that were not predominantly comprised of native English speakers (cf. with Experiments 3 and 4, which were solely native English speakers). Given the online platform, we screened for bot-like behavior (e.g., typing non-sense words such as <emph>asgf</emph>), for obvious signs of low effort (e.g., repeated RTs &lt;1 second), and for signs that they misunderstood the task (e.g., typing cue words into the solution word box). Table S1 provides the demographic data for the 78 participants who passed the data quality screening checks.</p> <hd id="AN0180329814-6">Materials</hd> <p>We selected 96 compound RAT trials from the Bowden and Jung-Beeman's ([<reflink idref="bib9" id="ref45">9</reflink>]) norms. We generated a fourth cue word for each trial via an iterative group-consensus process (see Supplemental Appendix A). This process resulted in two lists of 96 RAT stimuli, one list consisting of the original 3-item versions and the other list containing a new, fourth word. The 3-item and 4-item versions were matched on average word length and frequency (Vitrano, Altarriba, &amp; Leblebici‐basar, [<reflink idref="bib48" id="ref46">48</reflink>]) as well as other key linguistic characteristics shown in Table S2 (Balota et al., [<reflink idref="bib3" id="ref47">3</reflink>]). In addition, the 3-item (<emph>M</emph> =.082, <emph>SD</emph> =.11) and 4-item (<emph>M</emph> =.093, <emph>SD</emph> =.11) versions were matched on forward associative strength (i.e., the average strength of the association between the cue words and the answer word; see Nelson, McEvoy, &amp; Schreiber, [<reflink idref="bib35" id="ref48">35</reflink>], norms; <emph>t</emph>(<reflink idref="bib67" id="ref49">67</reflink>) = 1.429, <emph>p</emph> =.158). Therefore, the new fourth item changed the quantity, not the quality (forward strength), of semantic information. Cronbach's alpha values were generally better for 4-item trials than for 3-item trials (Table S3; Taber, [<reflink idref="bib44" id="ref50">44</reflink>]).</p> <hd id="AN0180329814-7">Procedure</hd> <p>Participants first completed a bot check (captcha) to filter out potential automated responses. Then they completed demographic questionnaires and received the RAT instructions. Participants were instructed that they would view a series of words and would need to generate a new word that connects each. They were shown one example item, the solution word, and the solution word associations. Drawing on Bowden and Jung-Beeman's (2003) findings of RAT trial accuracy across multiple time limits, we set a 20-s time limit per trial. Participants were instructed to type a question mark (?) if they did not know the answer and thought they would not be able to generate it before the time limit. The question-mark response was instituted to determine how long participants would exert effort to solve the RAT trial before giving up.</p> <p>During the experimental block, we manipulated the RAT item type within-subjects such that 3-item trials and 4-item trials alternated. Participants completed 30 trials (15 3-item, 15 4-item), which were randomly selected from the 96-item trial bank with the constraint that if the 3-item version appeared, then the matched 4-item version would not appear. We used the Gorilla experiment builder platform to enable RT data collection, based upon the first keypress in each trial (Anwyl-Irvine, Massonnié, Flitton, Kirkham, &amp; Evershed, [<reflink idref="bib2" id="ref51">2</reflink>]).</p> <hd id="AN0180329814-8">Statistical analysis</hd> <p>We used paired t-tests to compare RAT trial type (3-item, 4-item) on accuracy and RTs. RT data were separated for correctly answered trials, incorrectly answered trials, and trials that included a "?" response (26.3% of trials had no response and were excluded from RT analyses). Accuracy-RT correlations were conducted on both the participant-level (e.g., do participants with higher mean accuracy on 3-item trials also have higher mean accuracy on 4-item trials?) and the item-level (e.g., when collapsing data across all participants, are the most difficult 3-item trials also the most difficult when the fourth word is added?). Differences in degrees of freedom were due to some participants having no correctly answered trials for a trial type. Alpha was set to.05, all tests were two-tailed, and statistical analyses were conducted using SPSS version 27. Materials and data are available at Open Science Framework (https://osf.io/wbqzy/).</p> <hd id="AN0180329814-9">Results</hd> <p></p> <hd id="AN0180329814-10">Accuracy data</hd> <p>Accuracy on 4-item trials and 3-item trials strongly correlated in all experiments on the participant-level (see Table S4). Figure 1 shows performance on 4-item trials relative to 3-item trials, <emph>t</emph>(<reflink idref="bib77" id="ref52">77</reflink>) = 1.980, <emph>p</emph> =.051, <emph>d</emph> =.224. Inspection of individual RAT trials suggested a nominal improvement with the addition of the fourth cue word on 55.21% of trials, but a nominal worsening on 36.46% of RAT trials (8.33% showed no change; see Table S5 for these proportions across all experiments). To understand the source of this inter-item variability we examined performance relative to item-level difficulty (Valba, Gorsky, Nechaev, Tamm, &amp; Peel, [<reflink idref="bib47" id="ref53">47</reflink>]). In these item-level analyses (of all 96 pairs), we estimated item-difficulty as accuracy levels on the 3-item version, and then related that original level of accuracy to the item-level difference score between the 4- and 3-item versions (higher values indicating that accuracy improved when the fourth word was added). Figure 2 illustrates that performance on the difficult RAT trials benefitted from adding a fourth word, whereas performance on the easiest trials was hindered, <emph>r</emph>(<reflink idref="bib94" id="ref54">94</reflink>) = −.42, <emph>p</emph> &lt;.001. Based on Figure 1, if performance on the original 3-item RAT was below 25% accuracy, then adding a fourth word improved performance; but, if performance on the original 3-item RAT was above 40%, then adding the fourth word impaired performance.</p> <p>Graph: Figure 1. Mean accuracy across the 3-item and 4-item versions of the RAT in experiments 1–4. Error bars represent 95% confidence intervals.</p> <p>Graph: Figure 2. Scatterplots of the relationship between performance improvement when adding a fourth word (Y axis) and original difficulty level (X axis) in experiments 1–4. The dotted horizontal line indicates the level of performance that would reflect equal performance for 3-item versus 4-item versions. Upper and lower bands represent 95% confidence intervals.</p> <hd id="AN0180329814-11">RT data</hd> <p>Figure 3 shows the distribution of RTs for correctly answered trials. Most of the RAT trials were answered in 4 to 11 seconds (for a density plot, see Figure S1). Participants showed similar RTs across 4-item and 3-item trials when they were correctly answered (<emph>M</emph><subs><emph>4-item</emph></subs> = 7884 ms, <emph>SD</emph> = 3177 ms; <emph>M</emph><subs><emph>3-item</emph></subs> = 7749 ms, <emph>SD</emph> = 3192 ms; <emph>t</emph>(<reflink idref="bib60" id="ref55">60</reflink>) =.284, <emph>p</emph> =.778) or incorrectly answered (<emph>M</emph><subs><emph>4-item</emph></subs> = 9881 ms, <emph>SD</emph> = 4294 ms; <emph>M</emph><subs><emph>3-item</emph></subs> = 9198 ms, <emph>SD</emph> = 3776 ms; <emph>t</emph>(<reflink idref="bib67" id="ref56">67</reflink>) = 1.348, <emph>p</emph> &gt;.182). On trials in which participants responded with a question mark (?), RTs were significantly slower on 4-item trials (<emph>M</emph> = 11268 ms, <emph>SD</emph> = 4725 ms) than 3-item trials (<emph>M</emph> = 9994 ms, <emph>SD</emph> = 4044 ms), <emph>t</emph>(<reflink idref="bib34" id="ref57">34</reflink>) = 2.040, <emph>p</emph> &lt;.05, which may indicate greater effort prior to giving up.</p> <p>Graph: Figure 3. Violin plots of RTs on correct RAT trials in experiments 1–4. Median is marked by a solid line, quartiles are marked by dashed lines.</p> <p>On a participant-level, accuracy was not correlated with RTs (4-item: <emph>r</emph>(<reflink idref="bib68" id="ref58">68</reflink>) =.039, <emph>p</emph> =.752; 3-item: <emph>r</emph>(<reflink idref="bib65" id="ref59">65</reflink>) = −.026, <emph>p</emph> =.837), but on an item-level, greater accuracy was generally associated with faster RTs (4-item trials: <emph>r</emph>(<reflink idref="bib76" id="ref60">76</reflink>) = −.271, <emph>p</emph> &lt;.05; 3-item trials: <emph>r</emph>(<reflink idref="bib75" id="ref61">75</reflink>) = −.188, <emph>p</emph> =.102). Consistent with our finding that item-level difficulty moderated the impact of adding a fourth item on accuracy levels, Figure 4 indicates that the longer the RT of the 3-item RAT version (which we interpret as indicating greater difficulty), the more likely adding a fourth word was to shorten RTs (represented as the item-level difference score in RT between the 4- and 3-item versions), <emph>r</emph>(<reflink idref="bib75" id="ref62">75</reflink>) = −.66, <emph>p</emph> &lt;.001.</p> <p>Graph: Figure 4. Scatterplots of the relationship between response time (RT) improvement when adding a fourth word (Y axis) and original difficulty level (X axis) in experiments 1–4. The dotted horizontal line indicates the level of performance that would reflect equal performance for 3-item versus 4-item versions. Upper and lower bands represent 95% confidence intervals.</p> <hd id="AN0180329814-12">Discussion</hd> <p>Average performance levels on 4-item and traditional 3-item trials were relatively similar in Experiment 1 (<emph>d</emph> =.22). An item-level inspection though suggested that the fourth word may have benefitted difficult RAT trials while also possibly harming performance on easier RAT trials. Relative reliance on spreading activation versus strategic/analytic processes may therefore differ across trials. We aimed to replicate and extend these findings in Experiment 2.</p> <hd id="AN0180329814-13">Experiment 2</hd> <p></p> <hd id="AN0180329814-14">Methods</hd> <p></p> <hd id="AN0180329814-15">Participants</hd> <p>Participants were 202 adults (18–29 years old) who completed the study via the Prolific Online Research Platform in May 2020. Exclusion criteria and quality control screening were the same as in Experiment 1. Table S1 provides the demographic information for the 190 participants who passed the quality control checks.</p> <hd id="AN0180329814-16">Materials and procedure</hd> <p>We revised the Experiment 1 materials to replace the fourth words for 40 trials that were deemed by author consensus to be potentially misleading (Supplemental Appendix A). The goal here was to confirm that trial size (4-item versus 3-item) rather than idiosyncratic characteristics of the words selected to be the fourth items accounted for the response patterns observed in Experiment 1. The lexical characteristics are shown in Table S2. In addition, latent semantic analysis (<ulink href="http://wordvec.colorado.edu/">http://wordvec.colorado.edu/</ulink>) indicated no significant differences in semantic similarity of the cue words to the answer word between 3-item (<emph>M</emph> =.208, <emph>SD</emph> = 0.088) and 4-item trials (<emph>M</emph> =.214, <emph>SD</emph> = 0.080), <emph>t</emph>(<reflink idref="bib95" id="ref63">95</reflink>) = −1.242, <emph>p</emph> =.217.</p> <p>Additional minor changes included: a) adding a second bot check (visual search task), b) adding four example problems to the RAT instructions, c) confirming all new trials were compound RAT trials, and d) adding a 500 ms interstimulus interval (fixation) to introduce brief breaks between stimuli. Statistical analyses were the same as Experiment 1.</p> <hd id="AN0180329814-17">Results</hd> <p></p> <hd id="AN0180329814-18">Overview</hd> <p>The revisions made to the procedure significantly improved the percentage of responders who passed quality control screening from 70.1% in Experiment 1 to 94.1% in Experiment 2, <emph>X</emph><sups>2</sups>(<reflink idref="bib1" id="ref64">1</reflink>) = 31.508, <emph>p</emph> &lt;.001. Most of the patterns observed in Experiment 1 replicated in Experiment 2, typically with larger effect sizes.</p> <hd id="AN0180329814-19">Accuracy data</hd> <p>Figure 1 demonstrates that participants performed significantly better on 4-item trials than on 3-item trials, <emph>t</emph>(<reflink idref="bib189" id="ref65">189</reflink>) = 5.301, <emph>p</emph> &lt;.001, <emph>d</emph> =.385. Yet, like Experiment 1, not all RAT trials showed an improvement. Adding the fourth word worsened accuracy on 35.42% of RAT trials (61.46% of trials improved and 3.12% were unchanged). Even with the updates to the trial sets, this inter-item variability was again attributable to item difficulty. Figure 2 demonstrates that the difficult RAT trials were the most likely to improve in accuracy with the addition of the fourth word, <emph>r</emph>(<reflink idref="bib94" id="ref66">94</reflink>) = −.37, <emph>p</emph> &lt;.001.</p> <hd id="AN0180329814-20">RT data</hd> <p>For correctly answered trials, RTs on 4-item trials (<emph>M</emph> = 8488 ms, <emph>SD</emph> = 3212 ms) initially appeared slower than on 3-item trials (<emph>M</emph> = 7986 ms, <emph>SD</emph> = 3336 ms), <emph>t</emph>(<reflink idref="bib152" id="ref67">152</reflink>) = 1.976, <emph>p</emph> =.050, but this difference was non-significant when accounting for reading time (i.e., relative to 300 ms); <emph>t</emph>(<reflink idref="bib152" id="ref68">152</reflink>) = 0.795, <emph>p</emph> =.428; Rayner &amp; Clifton ([<reflink idref="bib39" id="ref69">39</reflink>]). Similarly, the RT difference on trials in which participants responded with a question mark (?) (<emph>M</emph><subs><emph>4-item</emph></subs> = 13602 ms, <emph>SD</emph> = 3745 ms; <emph>M</emph><subs><emph>3-item</emph></subs> = 12867 ms, <emph>SD</emph> = 3957 ms), <emph>t</emph>(<reflink idref="bib113" id="ref70">113</reflink>) = 2.643, <emph>p</emph> &lt;.01, was also reduced to non-significant when accounting for reading time (<emph>t</emph>(<reflink idref="bib113" id="ref71">113</reflink>) = 1.563, <emph>p</emph> =.121). RTs were significantly slower on 4-item than 3-item trials that were answered incorrectly (<emph>M</emph><subs><emph>4-item</emph></subs> = 12295 ms, <emph>SD</emph> = 4041 ms; <emph>M</emph><subs><emph>3-item</emph></subs> = 10892 ms, <emph>SD</emph> = 4009 ms), <emph>t</emph>(<reflink idref="bib164" id="ref72">164</reflink>) = 4.369, <emph>p</emph> &lt;.001 [even when accounting for reading time; <emph>t</emph>(<reflink idref="bib164" id="ref73">164</reflink>) = 3.434, <emph>p</emph> &lt;.001], and this finding could potentially indicate a reliance on bi-associate generation. If the longer RTs reflected bi-associate generation, we would also expect RTs to positively correlate with accuracy; however, on an item-level, there was a negative correlation for 4-item trials (<emph>r</emph>(<reflink idref="bib91" id="ref74">91</reflink>) = −.276, <emph>p</emph> &lt;.01) and 3-item trials (<emph>r</emph>(<reflink idref="bib90" id="ref75">90</reflink>) = −.326, <emph>p</emph> &lt;.01). On a participant-level, no significant accuracy – RT associations were observed (4-item: <emph>r</emph>(<reflink idref="bib177" id="ref76">177</reflink>) = −.036, <emph>p</emph> =.630; 3-item: <emph>r</emph>(<reflink idref="bib160" id="ref77">160</reflink>) =.190, <emph>p</emph> =.353). Importantly, and mirroring the accuracy by item-level difficulty results, Figure 4 demonstrates that the longer the RT of the 3-item RAT (indicating greater difficulty), the more likely adding a fourth word was to shorten RTs, <emph>r</emph>(<reflink idref="bib90" id="ref78">90</reflink>) = −.73, <emph>p</emph> &lt;.001.</p> <hd id="AN0180329814-21">Discussion</hd> <p>Experiment 2 showed a significant benefit of 4-item trials over traditional 3-item trials in average accuracy, with minimal changes to RTs when accounting for the additional time needed to read the fourth word. These overall patterns generally supported the spreading activation view. However, like Experiment 1, the relative reliance on automatic spreading activation versus analytical/controlled processes appeared to change dynamically across trials. In the next experiment, we extended this line of investigation to a group of only native English speakers and by querying whether correctly answered trials were being solved via subjective insight (Chein &amp; Weisberg, [<reflink idref="bib11" id="ref79">11</reflink>]).</p> <hd id="AN0180329814-22">Experiment 3</hd> <p></p> <hd id="AN0180329814-23">Methods</hd> <p></p> <hd id="AN0180329814-24">Participants</hd> <p>Participants were 201 adults who completed the study via the Prolific platform in July 2020. To be eligible, participants needed to reside in the United States or the United Kingdom, be between the ages of 18–30, and be a native English speaker. Analyses took into account regional spelling differences (e.g. aeroplane vs. airplane). Table S1 provides the demographic information for the 199 participants who passed the quality control checks (99.0%).</p> <hd id="AN0180329814-25">Materials and procedure</hd> <p>Experiment 3 materials were identical to Experiment 2. The procedure was modified such that each correct response was followed by a rating of feelings of insight versus reliance on strategy when solving the trial (1 = "Complete Insight," 2 = "Partial Insight," 3 = "Partial Strategy," and 4 = "Complete Strategy;" Chein &amp; Weisberg, [<reflink idref="bib11" id="ref80">11</reflink>]). Additional minor changes included: a) each trial was displayed for the entirety of the 20-s time limit to ensure that participants would not attempt to rush through trials; b) participants were shown a feedback screen that displayed the cue words, answer word, and associations; and c) to prevent fatigue, participants were given a short break after 15 trials. Statistical analyses were the same as Experiments 1–2, except that the non-parametric Wilcoxon signed-rank test was used to compare insight ratings across conditions.</p> <hd id="AN0180329814-26">Results</hd> <p></p> <hd id="AN0180329814-27">Accuracy data</hd> <p>Overall accuracy levels in this sample of only native English speakers were higher than in prior experiments and all major findings replicated, often with larger effect sizes (Figure 1). Performance was significantly better on 4-item trials than 3-item trials, <emph>t</emph>(<reflink idref="bib198" id="ref81">198</reflink>) = 10.127, <emph>p</emph> &lt;.001, <emph>d</emph> =.718 (Figure 1). Interestingly, even with elevated performance levels in this sample, adding the fourth word was still associated with decreased accuracy on 30.21% of RAT trials (69.79% improved and 0% were unchanged). Replicating the prior experiments, RAT trials that were more difficult showed larger improvements with the fourth word, <emph>r</emph>(<reflink idref="bib94" id="ref82">94</reflink>) = −.337, <emph>p</emph> &lt;.001 (Figure 2). Inspection of the 95% confidence intervals in Figure 2 indicated that the easy trials were unchanged by the fourth word (cf. Experiments 1–2 in which the fourth word worsened performance on easy trials).</p> <hd id="AN0180329814-28">RT data</hd> <p>Participants showed significantly slower RTs to 4-item trials when the trials were correctly answered (<emph>M</emph><subs><emph>4-item</emph></subs> = 6811 ms, <emph>SD</emph> = 1710 ms; <emph>M</emph><subs><emph>3-item</emph></subs> = 6068 ms, <emph>SD</emph> = 1829 ms; <emph>t</emph>(<reflink idref="bib194" id="ref83">194</reflink>) = 4.844, <emph>p</emph> &lt;.001; reading-speed-corrected: <emph>t</emph>(<reflink idref="bib194" id="ref84">194</reflink>) = 2.89, <emph>p</emph> =.004). RTs for providing a question mark did not differ across trial types (<emph>M</emph><subs><emph>4-item</emph></subs> = 13638 ms, <emph>SD</emph> = 2750 ms; <emph>M</emph><subs><emph>3-item</emph></subs> = 13407 ms, <emph>SD</emph> = 2271 ms, <emph>t</emph>(<reflink idref="bib60" id="ref85">60</reflink>) = 0.746, <emph>p</emph> =.459) and RTs on incorrectly answered (<emph>M</emph><subs><emph>4-item</emph></subs> = 11263 ms, <emph>SD</emph> = 3567 ms; <emph>M</emph><subs><emph>3-item</emph></subs> = 10417 ms, <emph>SD</emph> = 2819 ms; <emph>t</emph>(<reflink idref="bib162" id="ref86">162</reflink>) = 2.721, <emph>p</emph> &lt;.01) did not differ after accounting for reading speed (<emph>t</emph>(<reflink idref="bib162" id="ref87">162</reflink>) = 1.757, <emph>p</emph> =.081). On an item-level, greater accuracy was associated with faster RTs on correctly answered 4-item trials (<emph>r</emph>(<reflink idref="bib94" id="ref88">94</reflink>) = −.475, <emph>p</emph> &lt;.001) and 3-item trials (<emph>r</emph>(<reflink idref="bib91" id="ref89">91</reflink>) = −.498, <emph>p</emph> &lt;.001); a similar correlation was observed on a participant-level for 4-item trials (<emph>r</emph>(<reflink idref="bib197" id="ref90">197</reflink>) = −.222, <emph>p</emph> =.002; 3-item trials: <emph>r</emph>(<reflink idref="bib196" id="ref91">196</reflink>)= −.043, <emph>p</emph> =.551). Importantly, the longer the RT on the original 3-item RAT, the more likely adding a fourth word was to quicken RTs, <emph>r</emph>(<reflink idref="bib91" id="ref92">91</reflink>)= −.67, <emph>p</emph> &lt;.001 (Figure 4).</p> <hd id="AN0180329814-29">Insight ratings data</hd> <p>Figure 5a illustrates that participants predominantly reported "Partial Insight" or "Complete Insight" on 3-item trials (70.35%) and 4-item trials (70.08%). A Wilcoxon signed-ranks test showed no significant differences in the mean ratings between trial types (<emph>M</emph><subs><emph>3-item</emph></subs> = 1.97, <emph>SD</emph> =.62, <emph>M</emph><subs><emph>4-item</emph></subs> = 1.94, <emph>SD</emph> =.59; Z = 0.542, <emph>p</emph> =.59). On an item-level, trials with greater accuracy levels were also those more likely to be solved with insight, both in the 4-item sets (<emph>r</emph>(<reflink idref="bib94" id="ref93">94</reflink>) = −.22, <emph>p</emph> &lt;.05) and the 3-item sets (<emph>r</emph>(<reflink idref="bib91" id="ref94">91</reflink>) = −.50, <emph>p</emph> &lt;.001). On a participant-level, the accuracy – insight associations were negligible (4-item: <emph>r</emph>(<reflink idref="bib197" id="ref95">197</reflink>) =.086, <emph>p</emph> =.231; 3-item: <emph>r</emph>(<reflink idref="bib196" id="ref96">196</reflink>) =.077, <emph>p</emph> =.286).</p> <p>Graph: Figure 5. Frequency of insight rating responses in experiment 3 (A) and experiment 4 (B).</p> <hd id="AN0180329814-30">Discussion</hd> <p>Adding a fourth word improved RAT performance, particularly for the more difficult RAT trials. Furthermore, participants reported solving RAT problems with subjective insight more often than relying on strategies (note, however, that process-based interpretations of subjective insight remain debated and could actually reflect preparatory executive control processes: Becker, Wiedemann, &amp; Kühn, [<reflink idref="bib5" id="ref97">5</reflink>]; Bowden &amp; Jung-Beeman, [<reflink idref="bib8" id="ref98">8</reflink>]; Chein &amp; Weisberg, [<reflink idref="bib11" id="ref99">11</reflink>]; Jung-Beeman et al., [<reflink idref="bib22" id="ref100">22</reflink>]; Kounios et al., [<reflink idref="bib26" id="ref101">26</reflink>]). As a final replication and extension, for Experiment 4, we recruited local university students and manipulated trial-type between-subjects rather than within-subjects. Using a between-subjects design is important for identifying which, if any, of the patterns in Experiments 1–3 could be attributable to task switching between 3-item and 4-item trial types.</p> <hd id="AN0180329814-31">Experiment 4</hd> <p></p> <hd id="AN0180329814-32">Methods</hd> <p></p> <hd id="AN0180329814-33">Participants</hd> <p>Eighty-five undergraduate students were recruited via SONA in fall 2020 and received partial class credit for participating. The sample size was determined by an a priori power analysis that was based on the effect size for the primary dependent variable in Experiment 3 (<emph>d</emph> =.718), a two-tailed test, and alpha set to 0.05. This test indicated that 84 participants were needed to achieve.90 power to detect condition effects on accuracy. All participants (100%) passed quality control screening (Table S1).</p> <hd id="AN0180329814-34">Procedure</hd> <p>All materials and procedures were the same as Experiment 3 except that participants were randomly assigned to solve only 3-item trials or only 4-item trials (between-subjects design). The experimental block was 30 trials in length.</p> <hd id="AN0180329814-35">Results</hd> <p></p> <hd id="AN0180329814-36">Accuracy data</hd> <p>Figure 1 demonstrates that participants in the 4-item condition performed significantly better than participants in the 3-item condition, <emph>t</emph>(<reflink idref="bib83" id="ref102">83</reflink>) = 5.175, <emph>p</emph> &lt;.001, and with an effect size exceeding each of the prior experiments (<emph>d</emph> = 1.123). Most RAT trials improved in mean accuracy with the fourth word added (71.88%), though some still worsened in accuracy (23.96%; 4.16% were unchanged). RAT trials that were more difficult in their 3-item version were more likely to improve in accuracy when adding a fourth word, <emph>r</emph>(<reflink idref="bib94" id="ref103">94</reflink>) = −.402, <emph>p</emph> &lt;.001 (Figure 2).</p> <hd id="AN0180329814-37">RT data</hd> <p>RTs were generally quicker in this university student sample (relative to Experiments 1–3). There were no condition differences in RTs for correctly answered trials (<emph>M</emph><subs><emph>4-item</emph></subs> = 6646 ms, <emph>SD</emph> = 1388 ms; <emph>M</emph><subs><emph>3-item</emph></subs> = 6323 ms, <emph>SD</emph> = 1608 ms; <emph>t</emph>(<reflink idref="bib83" id="ref104">83</reflink>) = 0.987, <emph>p</emph> =.327), incorrectly answered trials (<emph>M</emph><subs><emph>4-item</emph></subs> = 11173 ms, <emph>SD</emph> = 2753 ms; <emph>M</emph><subs><emph>3-item</emph></subs> = 10189 ms, <emph>SD</emph> = 2421 ms; <emph>t</emph>(<reflink idref="bib82" id="ref105">82</reflink>) = 1.742, <emph>p</emph> =.085), or trials that included a question mark (?) response (<emph>M</emph><subs><emph>4-item</emph></subs> = 13361 ms, <emph>SD</emph> = 1478 ms; <emph>M</emph><subs><emph>3-item</emph></subs> = 13822 ms, <emph>SD</emph> = 3855 ms; <emph>t</emph>(<reflink idref="bib24" id="ref106">24</reflink>) = −0.360, <emph>p</emph> =.722). Accuracy – RT associations were negligible on a participant-level (4-item: <emph>r</emph>(<reflink idref="bib41" id="ref107">41</reflink>) = −.119, <emph>p</emph> =.457; 3-item: <emph>r</emph>(<reflink idref="bib44" id="ref108">44</reflink>) = −.108, <emph>p</emph> =.486). On an item-level, accuracy and RT showed a significant inverse correlation for 3-item trials (<emph>r</emph>(<reflink idref="bib84" id="ref109">84</reflink>) = −.376, <emph>p</emph> &lt;.001) and 4-item trials (<emph>r</emph>(<reflink idref="bib93" id="ref110">93</reflink>) = −.314, <emph>p</emph> &lt;.01), consistent with Experiments 1–3. In addition, adding a fourth word was most likely to quicken RTs on the most difficult 3-item RAT trials (i.e., those with the slowest RTs), <emph>r</emph>(<reflink idref="bib84" id="ref111">84</reflink>) = −.72, <emph>p</emph> &lt;.001 (Figure 4).</p> <hd id="AN0180329814-38">Insight data</hd> <p>Figure 5b shows that there were no significant differences in mean insight ratings between 3-item trials (<emph>M</emph> = 1.89, <emph>SD</emph> =.50) and 4-item trials (<emph>M</emph> = 2.10, <emph>SD</emph> =.61; Z = 1.805, <emph>p</emph> =.071). Greater ratings of insight occurred on trials with higher levels of accuracy on an item-level in the 4-item condition (<emph>r</emph>(<reflink idref="bib93" id="ref112">93</reflink>) = −.45, <emph>p</emph> &lt;.001; 3-item condition: <emph>r</emph>(<reflink idref="bib84" id="ref113">84</reflink>) = −.19, <emph>p</emph> =.084). This association was not observed at the participant-level (4-item: <emph>r</emph>(<reflink idref="bib41" id="ref114">41</reflink>) =.080, <emph>p</emph> =.618; 3-item: <emph>r</emph>(<reflink idref="bib44" id="ref115">44</reflink>) =.040, <emph>p</emph> =.799).</p> <hd id="AN0180329814-39">Discussion</hd> <p>Experiment 4 replicated the major findings with a between-subjects design: adding a fourth item improved RAT performance, especially for the most difficult trials. Experiment 4 showed that these benefits could be achieved without changing the nature of the task (insight ratings) or incurring RT costs. One potential explanation for RT slowing across some conditions in Experiments 2–3 (though not always exceeding the additional 300 ms needed to read the fourth word) was that alternating trial type (within subjects) led to task-switching costs for the trial type that required the most reading (4-item trials). When comparing the outcomes of Experiment 4 to the preceding experiments, it appears that the benefits of a 4-item RAT for increasing spreading activation across relevant semantic nodes is greatest amongst native English speakers (highly connected semantic networks) and when using a between-subjects design (eliminating task switching costs).</p> <hd id="AN0180329814-40">General discussion</hd> <p>Given the difficulty participants have in identifying connections between three cue words when there are time constraints (Bowden &amp; Jung-Beeman, [<reflink idref="bib9" id="ref116">9</reflink>]), one might reasonably expect that adding a fourth word would only make the RAT more difficult. Yet, participants consistently performed better on 4-item RAT trials than on the traditional 3-item RAT trials, by an average percent increase of 27.91%. The benefit of the fourth item replicated in native and non-native English speakers (Experiments 1–2 vs. 3–4), within- and between-subjects designs (Experiments 1–3 vs. 4), broad online samples and specific university samples (Experiments 1–3 vs. 4), and multiple sets of materials (i.e. iterative changes between each experiment), with the largest effects occurring for native English speakers in between-subjects designs (Experiment 4). Additionally, the fourth word was found to differentially benefit trial performance based on trial difficulty across all experiments. In this section, we consider the processes by which the fourth word influenced convergent thinking.</p> <hd id="AN0180329814-41">Semantic processes contributing to RAT performance</hd> <p>According to the spreading activation view, semantic knowledge is organized within a network of nodes (Collins &amp; Loftus, [<reflink idref="bib12" id="ref117">12</reflink>]), and lexical processing automatically triggers the activation that spreads across associated nodes (Friedrich, Henik, &amp; Tzelgov, [<reflink idref="bib17" id="ref118">17</reflink>]). One interpretation of the current findings, therefore, was that the additional lexical/semantic information (fourth word) increased the likelihood of activation spreading to the critical node, resulting in retrieval of the RAT solution. Reinforcing this interpretation, the effect sizes for adding the fourth word <emph>tripled</emph> from Experiments 1–2 in which non-native English speakers were eligible to Experiments 3–4 in which only native English speakers were recruited (i.e., individuals who have stronger semantic network connectivity for the verbal materials).</p> <p>Other interpretations are also possible, such as the fourth word leading participants to consider and then inhibit incorrect solutions (White &amp; Shah, [<reflink idref="bib49" id="ref119">49</reflink>]) or engage other executive control processes (Lee &amp; Therriault, [<reflink idref="bib28" id="ref120">28</reflink>]). These executive processes could be operating even in cases in which participants report solving trials with self-reported "aha" insight (Kounios et al., [<reflink idref="bib26" id="ref121">26</reflink>]). However, by this analytical/executive view, we would expect to see RT slowing on correct trials to a level that exceeds the time necessary to read the extra word; such slowing was usually not observed, and there was specifically no evidence for slowing in Experiment 4, which used a between-subjects manipulation (thereby eliminating potential task switching confounds).</p> <p>Perhaps more likely was that adding a fourth word helped to constrain the semantic neighborhood, thereby facilitating automatic spreading activation processes or increasing the efficiency of a search process within a localized space (Kounios et al., [<reflink idref="bib26" id="ref122">26</reflink>]). This view converges with Davelaar's ([<reflink idref="bib14" id="ref123">14</reflink>]) superadditive model that a search process can be localized to the intersection of the RAT cue words and to nodes that have strong similarities to the cue words. Interestingly, Davelaar's model theorizes this process to be most effective with weak targets, which was consistent with our finding that the fourth word was most beneficial on relatively difficult RAT trials. However, it is important to acknowledge that in our work, a variable/dimension that was not built into the stimulus set was the likelihood of search intersection. As an example, "sore" and "sweat" may lead to a direct intersection search for "cold" whereas "shoulder" may not lead to such a direct search intersection. Therefore, the RAT may be an imperfect measure of spreading activation because some stimuli may have a greater likelihood of search intersection than others.</p> <hd id="AN0180329814-42">Dynamic processes during convergent thinking</hd> <p>Though average RAT performance consistently improved with a fourth word, some individual trials did not. Performance actually worsened for 23% to 36% of the RAT problems, even though multiple sets of materials were used across experiments. This finding was consistent with the prediction of the bi-associations view that convergent thinking can involve controlled, multiplicative processing of information (Smith, Huber, &amp; Vul, [<reflink idref="bib42" id="ref124">42</reflink>]) and less consistent with analytical processing views that assume performing the RAT is congruent with solving fluid intelligence tasks (e.g., see Lee, Huggins, &amp; Therriault, [<reflink idref="bib27" id="ref125">27</reflink>]). We observed that performance was most likely to worsen when the fourth word was added to relatively easy RAT trials. On such trials, the first cue word often has a short semantic distance from the solution and introducing a fourth word would increase the total iterations required for checking the bi-association pairs (resulting in more opportunities for cognitive errors). Because the fourth word benefitted performance on difficult trials and hurt performance on easy trials, it appears that the relative reliance on automatic spreading activation processes and controlled processes can change dynamically within a single task block.</p> <p>Valba, Gorsky, Nechaev, Tamm, and Peel ([<reflink idref="bib47" id="ref126">47</reflink>]) similarly proposed that RAT strategy changes across relatively easy and relatively difficult trials. Specifically, they argued that easy trials are best solved by removing all weak associations and only identifying strong ones; applied to our current findings, performance on relatively easy RAT trials may have worsened with the fourth word because it was yet another weak association (or at best, a redundant strong association). In contrast to the pattern for easy trials, Valba et al. argued that difficult trials are best solved via continued processing of moderately weak associations. According to this viewpoint, a fourth word would lead to the correct RAT solution by providing an additional source of moderately weak information. Thus, like other areas in the cognitive sciences, the strategy used to perform creativity tasks can fluctuate on a trial by trial basis (e.g., Shelton &amp; Scullin, [<reflink idref="bib40" id="ref127">40</reflink>]).</p> <hd id="AN0180329814-43">Limitations and conclusions</hd> <p>For 60 years, the RAT test has been a useful measure of convergent thinking (Mednick, [<reflink idref="bib33" id="ref128">33</reflink>]). Its advantages have included being objective, domain-general, and brief. Its disadvantages have included reliance on self-reported insight/strategy use, language dependency (cf. Becker &amp; Cabeza, [<reflink idref="bib4" id="ref129">4</reflink>]), and showing low levels of performance on difficult trials. Because difficult trials may involve qualitatively different cognitive processes than easy trials (e.g., Valba, Gorsky, Nechaev, Tamm, &amp; Peel, [<reflink idref="bib47" id="ref130">47</reflink>]), and because floor performance on such trials limits the efficacy of studying the underlying mechanisms and real-world correlates (Goodwin &amp; Leech, [<reflink idref="bib20" id="ref131">20</reflink>]), there is a need to shift-the-scale of performance. The current experiments provide replicable evidence that adding a fourth word to difficult RAT trials improves overall performance, without changing the nature of the task (self-reported insight and RTs). In addition, the current work indicates that multiple cognitive processes contribute to successful convergent thinking, with the relative reliance on relatively automatic versus strategic/controlled processes changing dynamically with the demands of the task.</p> <hd id="AN0180329814-44">Disclosure statement</hd> <p>No potential conflict of interest was reported by the author(s).</p> <hd id="AN0180329814-45">Data availability statement</hd> <p>All data are publicly available at https://osf.io/wbqzy/</p> <hd id="AN0180329814-46">Supplementary material</hd> <p>Supplemental data for this article can be accessed online at https://doi.org/10.1080/10400419.2023.2200597.</p> <p>Correction Statement</p> <p>This article has been corrected with minor changes. These changes do not impact the academic content of the article.</p> <ref id="AN0180329814-47"> <title> References </title> <blist> <bibl id="bib1" idref="ref4" type="bt">1</bibl> <bibtext> Acar, S., Tadik, H., Myers, D., Van der Sman, C., &amp; Uysal, R. (2021). Creativity and well‐being: A meta‐analysis. The Journal of Creative Behavior, 55 (3), 738 – 751. doi: 10.1002/jocb.485</bibtext> </blist> <blist> <bibl id="bib2" idref="ref51" type="bt">2</bibl> <bibtext> Anwyl-Irvine, A. L., Massonnié, J., Flitton, A., Kirkham, N., &amp; Evershed, J. K. (2020). Gorilla in our midst: An online behavioral experiment builder. Behavior Research Methods, 52 (1), 388 – 407. doi: 10.3758/s13428-019-01237-x</bibtext> </blist> <blist> <bibl id="bib3" idref="ref47" type="bt">3</bibl> <bibtext> Balota, D. A., Yap, M. J., Cortese, M. J., Hutchison, K. A., Kessler, B., Loftis, B. ... Treiman, R. (2007). The English lexicon project. Behavior Research Methods, 39 (3), 445 – 459. doi: 10.3758/bf03193014</bibtext> </blist> <blist> <bibl id="bib4" idref="ref44" type="bt">4</bibl> <bibtext> Becker, M., &amp; Cabeza, R. (2022). Assessing creativity independently of language: A language-independent remote associate task (LI-RAT). Behavior Research Methods. doi: 10.3758/s13428-021-01773-5</bibtext> </blist> <blist> <bibl id="bib5" idref="ref97" type="bt">5</bibl> <bibtext> Becker, M., Wiedemann, G., &amp; Kühn, S. (2020). Quantifying insightful problem solving: A modified compound remote associates paradigm using lexical priming to parametrically modulate different sources of task difficulty. Psychological Research, 84 (2), 528 – 545. doi: 10.1007/s00426-018-1042-3</bibtext> </blist> <blist> <bibl id="bib6" idref="ref32" type="bt">6</bibl> <bibtext> Benedek, M., Jurisch, J., Koschutnig, K., Fink, A., &amp; Beaty, R. E. (2020). Elements of creative thought: Investigating the cognitive and neural correlates of association and bi-association processes. Neuroimage, 210, 116586. doi: 10.1016/j.neuroimage.2020.116586</bibtext> </blist> <blist> <bibl id="bib7" idref="ref24" type="bt">7</bibl> <bibtext> Benedek, M., Kenett, Y. N., Umdasch, K., Anaki, D., Faust, M., &amp; Neubauer, A. C. (2017). How semantic memory structure and intelligence contribute to creative thought: A network science approach. Thinking &amp; Reasoning, 23 (2), 158 – 183. doi: 10.1080/13546783.2016.1278034</bibtext> </blist> <blist> <bibl id="bib8" idref="ref98" type="bt">8</bibl> <bibtext> Bowden, E. M., &amp; Jung-Beeman, M. (2003a). Aha! Insight experience correlates with solution activation in the right hemisphere. Psychonomic Bulletin &amp; Review, 10 (3), 730 – 737. doi: 10.3758/BF03196539</bibtext> </blist> <blist> <bibl id="bib9" idref="ref20" type="bt">9</bibl> <bibtext> Bowden, E. M., &amp; Jung-Beeman, M. (2003b). Normative data for 144 compound remote associate problems. Behavior Research Methods, Instruments, &amp; Computers, 35 (4), 634 – 639. doi: 10.3758/BF03195543</bibtext> </blist> <blist> <bibtext> Bowden, E. M., Jung-Beeman, M., Fleck, J., &amp; Kounios, J. (2005). New approaches to demystifying insight. Trends in Cognitive Sciences, 9 (7), 322 – 328. doi: 10.1016/j.tics.2005.05.012</bibtext> </blist> <blist> <bibtext> Chein, J. M., &amp; Weisberg, R. W. (2014). Working memory and insight in verbal problems: Analysis of compound remote associates. Memory &amp; Cognition, 42 (1), 67 – 83. doi: 10.3758/s13421-013-0343-4</bibtext> </blist> <blist> <bibtext> Collins, A. M., &amp; Loftus, E. F. (1975). A spreading-activation theory of semantic processing. Psychological Review, 82 (6), 407 – 428. doi: 10.1037/0033-295X.82.6.407</bibtext> </blist> <blist> <bibtext> Cortes, R. A., Weinberger, A. B., Daker, R. J., &amp; Green, A. E. (2019). Re-examining prominent measures of divergent and convergent creativity. Current Opinion in Behavioral Sciences, 27, 90 – 93. doi: 10.1016/j.cobeha.2018.09.017</bibtext> </blist> <blist> <bibtext> Davelaar, E. J. (2015). Semantic search in the remote associates test. Topics in Cognitive Science, 7 (3), 494 – 512. doi: 10.1111/tops.12146</bibtext> </blist> <blist> <bibtext> Ellis, D. M., Robison, M. K., &amp; Brewer, G. A. (2021). The cognitive underpinnings of multiply-constrained problem solving. Journal of Intelligence, 9 (1), 7. doi: 10.3390/jintelligence9010007</bibtext> </blist> <blist> <bibtext> Fasko, D. (1999). Associative theory. Encyclopedia of Creativity, 1, 135 – 139.</bibtext> </blist> <blist> <bibtext> Friedrich, F. J., Henik, A., &amp; Tzelgov, J. (1991). Automatic processes in lexical access and spreading activation. Journal of Experimental Psychology Human Perception and Performance, 17 (3), 792 – 806. doi: 10.1037/0096-1523.17.3.792</bibtext> </blist> <blist> <bibtext> Frith, E., Kane, M. J., Welhaf, M. S., Christensen, A. P., Silvia, P. J., &amp; Beaty, R. E. (2021). Keeping creativity under control: Contributions of attention control and fluid intelligence to divergent thinking. Creativity Research Journal, 33 (2), 138 – 157. doi: 10.1080/10400419.2020.1855906</bibtext> </blist> <blist> <bibtext> Gajda, A., Karwowski, M., &amp; Beghetto, R. A. (2017). Creativity and academic achievement: A meta-analysis. Journal of Educational Psychology, 109 (2), 269 – 299. doi: 10.1037/edu0000133</bibtext> </blist> <blist> <bibtext> Goodwin, L. D., &amp; Leech, N. L. (2006). Understanding correlation: Factors that affect the size of r. Journal of Experimental Education, 74 (3), 251 – 266. doi: 10.3200/JEXE.74.3.249-266</bibtext> </blist> <blist> <bibtext> Green, A. E. (2016). Creativity, within reason: Semantic distance and dynamic state creativity in relational thinking and reasoning. Current Directions in Psychological Science, 25 (1), 28 – 35. doi: 10.1177/0963721415618485</bibtext> </blist> <blist> <bibtext> Jung-Beeman, M., Bowden, E. M., Haberman, J., Frymiare, J. L., Arambel-Liu, S., Greenblatt, R. ... Kounios, J. (2004). Neural activity when people solve verbal problems with insight. PLoS Biology, 2 (4), e97. doi: 10.1371/JOURNAL.PBIO.0020097</bibtext> </blist> <blist> <bibtext> Kajić, I., Gosmann, J., Stewart, T. C., Wennekers, T., &amp; Eliasmith, C. (2017). A spiking neuron model of word associations for the remote associates test. Frontiers in Psychology, 8, 99. doi: 10.3389/fpsyg.2017.00099</bibtext> </blist> <blist> <bibtext> Kaufman, J. C., Plucker, J. A., &amp; Baer, J. (2008). Essentials of creativity assessment. Hoboken, NJ : John Wiley &amp; Sons.</bibtext> </blist> <blist> <bibtext> Kenett, Y., Beaty, R., Silvia, P., Anaki, D., &amp; Faust, M. (2016). Structure and flexibility: Investigating the relation between the structure of the mental lexicon, fluid intelligence, and creative achievement. Psychology of Aesthetics, Creativity, and the Arts, 10 (4), 377 – 388. doi: 10.1037/aca0000056</bibtext> </blist> <blist> <bibtext> Kounios, J., Frymiare, J. L., Bowden, E. M., Fleck, J. I., Subramaniam, K., Parrish, T. B., &amp; Jung-Beeman, M. (2006). The prepared mind: Neural activity prior to problem presentation predicts subsequent solution by sudden insight. Psychological Science, 17 (10), 882 – 890. doi: 10.1111/j.1467-9280.2006.01798.x</bibtext> </blist> <blist> <bibtext> Lee, C. S., Huggins, A. C., &amp; Therriault, D. J. (2014). A measure of creativity or intelligence? Examining internal and external structure validity evidence of the remote associates test. Psychology of Aesthetics, Creativity, and the Arts, 8 (4), 446 – 460. doi: 10.1037/a0036773</bibtext> </blist> <blist> <bibtext> Lee, C. S., &amp; Therriault, D. J. (2013). The cognitive underpinnings of creative thought: A latent variable analysis exploring the roles of intelligence and working memory in three creative thinking processes. Intelligence, 41 (5), 306 – 320. doi: 10.1016/j.intell.2013.04.008</bibtext> </blist> <blist> <bibtext> Loevinger, J. (1957). Objective tests as instruments of psychological theory. Psychological Reports, 3 (3), 635 – 694. doi: 10.2466/pr0.1957.3.3.635</bibtext> </blist> <blist> <bibtext> Luft, C. D. B., Zioga, I., Thompson, N. M., Banissy, M. J., &amp; Bhattacharya, J. (2018). Right temporal alpha oscillations as a neural mechanism for inhibiting obvious associations. Proceedings of the National Academy of Sciences of the United States of America, 115 (52), E12144 – 12152.</bibtext> </blist> <blist> <bibtext> Marko, M., Michalko, D., &amp; Riečanský, I. (2019). Remote associates test: An empirical proof of concept. Behavior Research Methods, 51 (6), 2700 – 2711. doi: 10.3758/s13428-018-1131-7</bibtext> </blist> <blist> <bibtext> Marupaka, N., Iyer, L. R., &amp; Minai, A. A. (2012). Connectivity and thought: The influence of semantic network structure in a neurodynamical model of thinking. Neural Networks, 32, 147 – 158. doi: 10.1016/j.neunet.2012.02.004</bibtext> </blist> <blist> <bibtext> Mednick, S. (1962). The associative basis of the creative process. Psychological Review, 69 (3), 220 – 232. doi: 10.1037/h0048850</bibtext> </blist> <blist> <bibtext> Miron-Spektor, E., Gino, F., &amp; Argote, L. (2011). Paradoxical frames and creative sparks: Enhancing individual creativity through conflict and integration. Organizational Behavior and Human Decision Processes, 116 (2), 229 – 240. doi: 10.1016/j.obhdp.2011.03.006</bibtext> </blist> <blist> <bibtext> Nelson, D. L., McEvoy, C. L., &amp; Schreiber, T. A. (2004). The University of South Florida free association, rhyme, and word fragment norms. Behavior Research Methods, Instruments, &amp; Computers, 36 (3), 402 – 407. doi: 10.3758/BF03195588</bibtext> </blist> <blist> <bibtext> Nusbaum, E. C., &amp; Silvia, P. J. (2011). Are intelligence and creativity really so different?: Fluid intelligence, executive processes, and strategy use in divergent thinking. Intelligence, 39 (1), 36 – 45. doi: 10.1016/j.intell.2010.11.002</bibtext> </blist> <blist> <bibtext> Ohlsson, S. (2011). Deep learning: How the mind overrides experience. New York, NY : Cambridge University Press.</bibtext> </blist> <blist> <bibtext> Plucker, J. A., Beghetto, R. A., &amp; Dow, G. T. (2004). Why isn't creativity more important to educational psychologists? Potentials, pitfalls, and future directions in creativity research. Educational Psychologist, 39 (2), 83 – 96. doi: 10.1207/s15326985ep3902_1</bibtext> </blist> <blist> <bibtext> Rayner, K., &amp; Clifton, C. (2009). Language processing in reading and speech perception is fast and incremental: Implications for event-related potential research. Biological Psychology, 80 (1), 4 – 9. doi: 10.1016/j.biopsycho.2008.05.002</bibtext> </blist> <blist> <bibtext> Shelton, J. T., &amp; Scullin, M. K. (2017). The dynamic interplay between bottom-up and top-down processes supporting prospective remembering. Current Directions in Psychological Science, 26 (4), 352 – 358. doi: 10.1177/0963721417700504</bibtext> </blist> <blist> <bibtext> Silvia, P. J. (2015). Intelligence and creativity are pretty similar after all. Educational Psychology Review, 27 (4), 599 – 606. doi: 10.1007/s10648-015-9299-1</bibtext> </blist> <blist> <bibtext> Smith, K. A., Huber, D. E., &amp; Vul, E. (2013). Multiply-constrained semantic search in the remote associates test. Cognition, 128 (1), 64 – 75. doi: 10.1016/j.cognition.2013.03.001</bibtext> </blist> <blist> <bibtext> Spearman, C. E. (1930). Creative mind. London, England : Cambridge University Press.</bibtext> </blist> <blist> <bibtext> Taber, K. S. (2018). The use of cronbach's alpha when developing and reporting research instruments in science education. Research in Science Education, 48 (6), 1273 – 1296. doi: 10.1007/s11165-016-9602-2</bibtext> </blist> <blist> <bibtext> Taft, R., &amp; Rossiter, J. R. (1966). The remote associates test: Divergent or convergent thinking? Psychological Reports, 19, 1313 – 1314. doi: 10.2466/pr0.1966.19.3f.1313</bibtext> </blist> <blist> <bibtext> Tik, M., Sladky, R., Luft, C. D. B., Willinger, D., Hoffmann, A., Banissy, M. J. ... Windischberger, C. (2018). Ultra-high-field fMRI insights on insight: Neural correlates of the aha!-moment. Human Brain Mapping, 39 (8), 3241 – 3252. doi: 10.1002/hbm.24073</bibtext> </blist> <blist> <bibtext> Valba, O., Gorsky, A., Nechaev, S., Tamm, M., &amp; Peel, L. (2021). Analysis of English free association network reveals mechanisms of efficient solution of remote association tests. PLoS One, 16 (4), e0248986. doi: 10.1371/journal.pone.0248986</bibtext> </blist> <blist> <bibtext> Vitrano, D., Altarriba, J., &amp; Leblebici‐basar, D. (2021). Revisiting mednick's (1962) theory of creativity with a composite measure of creativity: The effect of stimulus type on word association production. The Journal of Creative Behavior, 55 (4), 925 – 936. doi: 10.1002/jocb.498</bibtext> </blist> <blist> <bibtext> White, H. A., &amp; Shah, P. (2006). Uninhibited imaginations: Creativity in adults with attention-deficit/hyperactivity disorder. Personality and Individual Differences, 40 (6), 1121 – 1131. doi: 10.1016/j.paid.2005.11.007</bibtext> </blist> <blist> <bibtext> Zhang, W., Sjoerds, Z., &amp; Hommel, B. (2020). Metacontrol of human creativity: The neurocognitive mechanisms of convergent and divergent thinking. NeuroImage, 210, 116572. doi: 10.1016/j.neuroimage.2020.116572</bibtext> </blist> </ref> <aug> <p>By Jose A. Diaz; Steven M. Nelson; A. Alexander Beaujean; Adam E. Green and Michael K. Scullin</p> <p>Reported by Author; Author; Author; Author; Author</p> </aug> <nolink nlid="nl1" bibid="bib38" firstref="ref1"></nolink> <nolink nlid="nl2" bibid="bib19" firstref="ref2"></nolink> <nolink nlid="nl3" bibid="bib34" firstref="ref3"></nolink> <nolink nlid="nl4" bibid="bib21" firstref="ref5"></nolink> <nolink nlid="nl5" bibid="bib18" firstref="ref6"></nolink> <nolink nlid="nl6" bibid="bib36" firstref="ref7"></nolink> <nolink nlid="nl7" bibid="bib50" firstref="ref8"></nolink> <nolink nlid="nl8" bibid="bib33" firstref="ref9"></nolink> <nolink nlid="nl9" bibid="bib16" firstref="ref10"></nolink> <nolink nlid="nl10" bibid="bib43" firstref="ref11"></nolink> <nolink nlid="nl11" bibid="bib15" firstref="ref12"></nolink> <nolink nlid="nl12" bibid="bib24" firstref="ref13"></nolink> <nolink nlid="nl13" bibid="bib10" firstref="ref14"></nolink> <nolink nlid="nl14" bibid="bib11" firstref="ref15"></nolink> <nolink nlid="nl15" bibid="bib46" firstref="ref16"></nolink> <nolink nlid="nl16" bibid="bib13" firstref="ref17"></nolink> <nolink nlid="nl17" bibid="bib29" firstref="ref18"></nolink> <nolink nlid="nl18" bibid="bib26" firstref="ref19"></nolink> <nolink nlid="nl19" bibid="bib20" firstref="ref22"></nolink> <nolink nlid="nl20" bibid="bib12" firstref="ref23"></nolink> <nolink nlid="nl21" bibid="bib25" firstref="ref25"></nolink> <nolink nlid="nl22" bibid="bib32" firstref="ref26"></nolink> <nolink nlid="nl23" bibid="bib14" firstref="ref27"></nolink> <nolink nlid="nl24" bibid="bib23" firstref="ref28"></nolink> <nolink nlid="nl25" bibid="bib30" firstref="ref29"></nolink> <nolink nlid="nl26" bibid="bib37" firstref="ref30"></nolink> <nolink nlid="nl27" bibid="bib42" firstref="ref31"></nolink> <nolink nlid="nl28" bibid="bib39" firstref="ref33"></nolink> <nolink nlid="nl29" bibid="bib27" firstref="ref34"></nolink> <nolink nlid="nl30" bibid="bib41" firstref="ref37"></nolink> <nolink nlid="nl31" bibid="bib45" firstref="ref38"></nolink> <nolink nlid="nl32" bibid="bib31" firstref="ref39"></nolink> <nolink nlid="nl33" bibid="bib47" firstref="ref41"></nolink> <nolink nlid="nl34" bibid="bib48" firstref="ref46"></nolink> <nolink nlid="nl35" bibid="bib35" firstref="ref48"></nolink> <nolink nlid="nl36" bibid="bib67" firstref="ref49"></nolink> <nolink nlid="nl37" bibid="bib44" firstref="ref50"></nolink> <nolink nlid="nl38" bibid="bib77" firstref="ref52"></nolink> <nolink nlid="nl39" bibid="bib94" firstref="ref54"></nolink> <nolink nlid="nl40" bibid="bib60" firstref="ref55"></nolink> <nolink nlid="nl41" bibid="bib68" firstref="ref58"></nolink> <nolink nlid="nl42" bibid="bib65" firstref="ref59"></nolink> <nolink nlid="nl43" bibid="bib76" firstref="ref60"></nolink> <nolink nlid="nl44" bibid="bib75" firstref="ref61"></nolink> <nolink nlid="nl45" bibid="bib95" firstref="ref63"></nolink> <nolink nlid="nl46" bibid="bib189" firstref="ref65"></nolink> <nolink nlid="nl47" bibid="bib152" firstref="ref67"></nolink> <nolink nlid="nl48" bibid="bib113" firstref="ref70"></nolink> <nolink nlid="nl49" bibid="bib164" firstref="ref72"></nolink> <nolink nlid="nl50" bibid="bib91" firstref="ref74"></nolink> <nolink nlid="nl51" bibid="bib90" firstref="ref75"></nolink> <nolink nlid="nl52" bibid="bib177" firstref="ref76"></nolink> <nolink nlid="nl53" bibid="bib160" firstref="ref77"></nolink> <nolink nlid="nl54" bibid="bib198" firstref="ref81"></nolink> <nolink nlid="nl55" bibid="bib194" firstref="ref83"></nolink> <nolink nlid="nl56" bibid="bib162" firstref="ref86"></nolink> <nolink nlid="nl57" bibid="bib197" firstref="ref90"></nolink> <nolink nlid="nl58" bibid="bib196" firstref="ref91"></nolink> <nolink nlid="nl59" bibid="bib22" firstref="ref100"></nolink> <nolink nlid="nl60" bibid="bib83" firstref="ref102"></nolink> <nolink nlid="nl61" bibid="bib82" firstref="ref105"></nolink> <nolink nlid="nl62" bibid="bib84" firstref="ref109"></nolink> <nolink nlid="nl63" bibid="bib93" firstref="ref110"></nolink> <nolink nlid="nl64" bibid="bib17" firstref="ref118"></nolink> <nolink nlid="nl65" bibid="bib49" firstref="ref119"></nolink> <nolink nlid="nl66" bibid="bib28" firstref="ref120"></nolink> <nolink nlid="nl67" bibid="bib40" firstref="ref127"></nolink> |
|---|---|
| Header | DbId: eric DbLabel: ERIC An: EJ1444260 AccessLevel: 3 PubType: Academic Journal PubTypeId: academicJournal PreciseRelevancyScore: 0 |
| IllustrationInfo | |
| Items | – Name: Title Label: Title Group: Ti Data: The Impact of Adding a Fourth Item to the Traditional 3-Item Remote Associates Test – Name: Language Label: Language Group: Lang Data: English – Name: Author Label: Authors Group: Au Data: <searchLink fieldCode="AR" term="%22Jose+A%2E+Diaz%22">Jose A. Diaz</searchLink><br /><searchLink fieldCode="AR" term="%22Steven+M%2E+Nelson%22">Steven M. Nelson</searchLink><br /><searchLink fieldCode="AR" term="%22A%2E+Alexander+Beaujean%22">A. Alexander Beaujean</searchLink><br /><searchLink fieldCode="AR" term="%22Adam+E%2E+Green%22">Adam E. Green</searchLink><br /><searchLink fieldCode="AR" term="%22Michael+K%2E+Scullin%22">Michael K. Scullin</searchLink> (ORCID <externalLink term="https://orcid.org/0000-0002-7578-7587">0000-0002-7578-7587</externalLink>) – Name: TitleSource Label: Source Group: Src Data: <searchLink fieldCode="SO" term="%22Creativity+Research+Journal%22"><i>Creativity Research Journal</i></searchLink>. 2024 36(4):604-616. – Name: Avail Label: Availability Group: Avail Data: Routledge. Available from: Taylor & Francis, Ltd. 530 Walnut Street Suite 850, Philadelphia, PA 19106. Tel: 800-354-1420; Tel: 215-625-8900; Fax: 215-207-0050; Web site: http://www.tandf.co.uk/journals – Name: PeerReviewed Label: Peer Reviewed Group: SrcInfo Data: Y – Name: Pages Label: Page Count Group: Src Data: 13 – Name: DatePubCY Label: Publication Date Group: Date Data: 2024 – Name: SourceSuprt Label: Sponsoring Agency Group: SrcSuprt Data: National Science Foundation (NSF) – Name: NumberContract Label: Contract Number Group: NumCntrct Data: 1920730 – Name: TypeDocument Label: Document Type Group: TypDoc Data: Journal Articles<br />Reports - Research – Name: Audience Label: Education Level Group: Audnce Data: <searchLink fieldCode="EL" term="%22Higher+Education%22">Higher Education</searchLink><br /><searchLink fieldCode="EL" term="%22Postsecondary+Education%22">Postsecondary Education</searchLink> – Name: Subject Label: Descriptors Group: Su Data: <searchLink fieldCode="DE" term="%22Test+Items%22">Test Items</searchLink><br /><searchLink fieldCode="DE" term="%22Creative+Thinking%22">Creative Thinking</searchLink><br /><searchLink fieldCode="DE" term="%22Creativity+Tests%22">Creativity Tests</searchLink><br /><searchLink fieldCode="DE" term="%22Performance%22">Performance</searchLink><br /><searchLink fieldCode="DE" term="%22Accuracy%22">Accuracy</searchLink><br /><searchLink fieldCode="DE" term="%22Association+%28Psychology%29%22">Association (Psychology)</searchLink><br /><searchLink fieldCode="DE" term="%22Reaction+Time%22">Reaction Time</searchLink><br /><searchLink fieldCode="DE" term="%22Young+Adults%22">Young Adults</searchLink><br /><searchLink fieldCode="DE" term="%22Difficulty+Level%22">Difficulty Level</searchLink><br /><searchLink fieldCode="DE" term="%22Foreign+Countries%22">Foreign Countries</searchLink><br /><searchLink fieldCode="DE" term="%22Undergraduate+Students%22">Undergraduate Students</searchLink><br /><searchLink fieldCode="DE" term="%22Native+Language%22">Native Language</searchLink><br /><searchLink fieldCode="DE" term="%22English%22">English</searchLink><br /><searchLink fieldCode="DE" term="%22English+%28Second+Language%29%22">English (Second Language)</searchLink><br /><searchLink fieldCode="DE" term="%22Individual+Differences%22">Individual Differences</searchLink><br /><searchLink fieldCode="DE" term="%22Cognitive+Processes%22">Cognitive Processes</searchLink> – Name: Subject Label: Geographic Terms Group: Su Data: <searchLink fieldCode="DE" term="%22United+States%22">United States</searchLink><br /><searchLink fieldCode="DE" term="%22United+Kingdom%22">United Kingdom</searchLink> – Name: SubjectThesaurus Label: Assessment and Survey Identifiers Group: Su Data: <searchLink fieldCode="SU" term="%22Remote+Associates+Test%22">Remote Associates Test</searchLink> – Name: DOI Label: DOI Group: ID Data: 10.1080/10400419.2023.2200597 – Name: ISSN Label: ISSN Group: ISSN Data: 1040-0419<br />1532-6934 – Name: Abstract Label: Abstract Group: Ab Data: The compound Remote Associates Test (RAT) is a classic measure of creativity. Participants are shown three cue words (sore-shoulder-sweat) and asked to generate a word that connects them (cold). Theoretical views of RAT performance differ in the degree to which they conceptualize performance as depending on automatic spreading activation across semantic networks, strategic generation of bi-associations, and other analytical processes (e.g. executive processes that support fluid intelligence). We tested these views by adding a fourth cue word to determine whether it impaired RAT accuracy (e.g. generation of bi-associations), impaired response times (analytic processes), or improved RAT accuracy without changing response times (e.g. spreading activation). Across four experiments, 551 adults completed 3- and 4-item RAT trials that were matched on linguistic and semantic metrics. Across experiments, adding the fourth word improved accuracy by 27.91%. This performance gain occurred with either modest or no changes to response times or ratings of insight/strategy use. Interestingly, the fourth word predominantly benefited accuracy and response times on difficult trials; on easy trials, the fourth word impaired or did not change performance. The findings suggest that both automatic and strategic/analytical processes contribute to successful RAT performance, with relative dependence on these processes dynamically adapting to the demands of the individual trial. – Name: AbstractInfo Label: Abstractor Group: Ab Data: As Provided – Name: DateEntry Label: Entry Date Group: Date Data: 2024 – Name: AN Label: Accession Number Group: ID Data: EJ1444260 |
| PLink | https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=eric&AN=EJ1444260 |
| RecordInfo | BibRecord: BibEntity: Identifiers: – Type: doi Value: 10.1080/10400419.2023.2200597 Languages: – Text: English PhysicalDescription: Pagination: PageCount: 13 StartPage: 604 Subjects: – SubjectFull: Test Items Type: general – SubjectFull: Creative Thinking Type: general – SubjectFull: Creativity Tests Type: general – SubjectFull: Performance Type: general – SubjectFull: Accuracy Type: general – SubjectFull: Association (Psychology) Type: general – SubjectFull: Reaction Time Type: general – SubjectFull: Young Adults Type: general – SubjectFull: Difficulty Level Type: general – SubjectFull: Foreign Countries Type: general – SubjectFull: Undergraduate Students Type: general – SubjectFull: Native Language Type: general – SubjectFull: English Type: general – SubjectFull: English (Second Language) Type: general – SubjectFull: Individual Differences Type: general – SubjectFull: Cognitive Processes Type: general – SubjectFull: United States Type: general – SubjectFull: United Kingdom Type: general – SubjectFull: Remote Associates Test Type: general Titles: – TitleFull: The Impact of Adding a Fourth Item to the Traditional 3-Item Remote Associates Test Type: main BibRelationships: HasContributorRelationships: – PersonEntity: Name: NameFull: Jose A. Diaz – PersonEntity: Name: NameFull: Steven M. Nelson – PersonEntity: Name: NameFull: A. Alexander Beaujean – PersonEntity: Name: NameFull: Adam E. Green – PersonEntity: Name: NameFull: Michael K. Scullin IsPartOfRelationships: – BibEntity: Dates: – D: 01 M: 01 Type: published Y: 2024 Identifiers: – Type: issn-print Value: 1040-0419 – Type: issn-electronic Value: 1532-6934 Numbering: – Type: volume Value: 36 – Type: issue Value: 4 Titles: – TitleFull: Creativity Research Journal Type: main |
| ResultId | 1 |
