Elsevier

Intelligence

Volume 91, March–April 2022, 101618
Intelligence

Strategy Transfer on Fluid Reasoning Tasks

https://doi.org/10.1016/j.intell.2021.101618Get rights and content

Highlights

  • Strategy consistency and transfer across multiple reasoning tasks was assessed.

  • Participants are consistent in what strategy they use across reasoning tasks.

  • Ambiguous reasoning items increased use of response-based strategies.

  • Participants transferred a response-based strategy but it did not affect accuracy.

  • Strategy use appears to be driven by individual differences and the task itself.

Abstract

Strategy use on reasoning tasks has consistently been shown to correlate with working memory capacity and accuracy, but it is still unclear to what degree individual preferences, working memory capacity, and features of the task itself contribute to strategy use. The present studies used eye tracking to explore the potential for strategy transfer between reasoning tasks. Study 1 demonstrated that participants are consistent in what strategy they use across reasoning tasks and that strategy transfer between tasks is possible. Additionally, post-hoc analyses identified certain ambiguous items in the figural analogies task that required participants to assess the response bank to reach solution, which appeared to push participants towards a more response-based strategy. Study 2 utilized a between-subjects design to manipulate this “ambiguity” in figural analogies problems prior to completing the RAPM. Once again, participants transferred strategies between tasks when primed with different strategies, although this did not affect their ability to accurately solve the problem. Importantly, strategy use changed considerably depending on the ambiguity of the initial reasoning task. The results provided across the two studies suggest that participants are consistent in what strategies they employ across reasoning tasks, and that if features of the task push participants towards a different strategy, they will transfer that strategy to another reasoning task. Furthermore, to understand the role of strategy use on reasoning tasks, future work will require a diverse sample of both reasoning tasks and strategy use measures.

Section snippets

Participants

One-hundred-thirty students enrolled in an undergraduate psychology course at Mississippi State University participated for course credit. A target of 55 participants per condition was deemed sufficient based upon work indicating that the correlation between the RAPM and WMC often ranges from 0.3 to 0.6 (Ackerman et al., 2005). To estimate a true correlation of 0.4 or 0.5, sample sizes of 63 and 50, respectively, would be sufficient for detecting the correlation within a stable range of plus or

Study 2

The increase in rate of toggling and proportion of time on the response bank on ambiguous FA problems suggests an increased use of the response options, and may also clarify some of the mixed findings in the task order manipulation in Study 1. Given that there are only five ambiguous items in the original FA task, the lack of sufficient items prompting increased use of the response options may explain why some of the results of Study 1 were significant, whereas other results trended towards a

General discussion

The results across Studies 1 and 2 illustrate that individuals' strategy use, as measured by the rate of toggling between the problem and response bank, is consistent across reasoning tasks, even after accounting for WMC. In addition, different types of problems and tasks orient participants towards different strategies, and participants will transfer those strategies to later reasoning tasks, even in cases when the more effective strategy on the first task is less effective on the second task.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

References (29)

  • P.L. Ackerman et al.

    Working memory and intelligence: The same or different constructs?

    Psychological Bulletin

    (2005)
  • D. Bates et al.

    Parsimonious mixed models

    ArXiv Preprint

    (2015)
  • J.L. Foster et al.

    Shortened complex span tasks can reliably measure working memory capacity

    Memory & Cognition

    (2015)
  • C. Gonthier et al.

    Intraindividual strategy shifts in Raven’s matrices, and their dependence on working memory capacity and need for cognition

    Journal of Experimental Psychology: General

    (2020)
  • Cited by (0)

    View full text