Counterfactual thoughts in complex causal domain: content, benefits, and implications for their function

Abstract

The reliability of previous findings on two crucial aspects of counterfactual thinking, namely the content of counterfactual modifications and their impact on future performance, has been questioned for the frequent use of tasks characterised by simple causal domains, that restrict participants’ possibility to consider a broad range of modifications. To overcome this limitation, we utilised a new experimental task featuring a complex causal domain to investigate such key aspects. The results indicated that participants tend to generate counterfactuals about elements outside their control, especially when presented with a challenging version of the task (Study 1a) and, to a lesser extent, when they receive negative feedback on their performance (Study 1b). Moreover, despite occasional implementations of actions mentioned in controllable counterfactuals, being engaged in counterfactual thinking did not lead to subsequent performance improvements (Study 2). The implications of these findings for the debate on the function of counterfactual thinking are discussed.

Keywords:

• Mental simulation
• counterfactual thinking
• preparatory function
• Functional Theory

Acknowledgements

This research received financial support by the European Union – Next Generation EU – PRIN 2022 (DD 104, 02/02/22) – PNRR – M4 – C2 – INV1.1 – PRIN – Project title: “Reasoning with hypotheses: Integrating logical, probabilistic, and experimental perspectives.” (ID: 2022ARRY9N – CUP E53D23008790006).

Disclosure statement

The authors have no conflict of interest to declare.

Notes

1 We did not register which of the two peripherals (mouse vs. trackpad) participants used to play the game, limiting the possibility to verify if this factor impacted our results. However, we believe this possibility to be remote, as only three counterfactuals among the 1,054 produced in our three studies focused on this aspect. As detailed in the “Coding of Counterfactuals” Section of Study 1a, they were considered as “ambiguous,” leading to their exclusion from the main analyses, because they were not easily categorizable as controllable or uncontrollable.

2 It should be noted that our task bears similarities to the one used in Experiment 1 by Hammell and Chan (2016). However, in their case, the video game was a commercially available product for a gaming console, limiting the authors’ ability to closely control various task features that could influence participants’ counterfactual modifications.

3 The statistical analyses were conducted using R, version 3.6.2 (R Core Team, 2019). Datasets of the three studies and scripts of the analyses performed can be found at https://osf.io/2q8je/. Further details about the samples, the results, and additional analyses can be found in the Supplementary Materials.

4 The decision to employ this narrow concept of controllability is based on the rationale that focusing on elements that cannot be immediately acted upon (e.g., fostering a specific skill through extensive practice) would not contribute to achieving better performance in an experimental setting, where participants typically have only a limited amount of time to prepare for a possible repetition of the task.

5 However, the results concerning these two covariates should be interpreted with caution, as in their case the linearity assumption for the logistic regression was violated.

6 Two participants opted not to disclose information about their gender. Consequently, they were excluded from this analysis as it listed gender among the covariates.

7 The same analysis was repeated excluding participants who scored > 85 points (the value precisely below the top-scoring 10% of easy-version participants in Study 1a) and the results did not change (see Table S8).

8 Alternatives to this control condition were considered but ultimately dismissed due to potential drawbacks. Specifically, prompting participants to reflect in general on their performance might have induced them to engage in causal inferences, a key element for the preparatory effect of counterfactual thinking (Roese & Epstude, 2017), potentially resulting in an underestimation of the positive effects of counterfactual reflection. On the other hand, asking participants to generate thoughts unrelated to their game performance might have excessively distracted them compared to participants in the counterfactual condition, potentially resulting in an overestimation of the positive effects of counterfactual reflection.

9 This question was added to explore whether the beneficial effects of counterfactuals would emerge most evidently among participants who perceive a real opportunity for performance improvement, in terms of both possibility and capability (Roese & Epstude, 2017). Analyses regarding this variable are reported in the Supplementary Materials, but from those it appears that perceived opportunity did not affect the extent to which counterfactual thinking benefited subsequent performances.

10 The negative mood measure was created by means of a principal component analysis (PCA) of the five mood ratings (happiness rating reversed). Both Bartlett’s test of sphericity (p < .001) and Kaiser-Meyer-Olkin measure of sampling adequacy (overall MSA = .81, all item’s MSA ≥ .77) allowed the PCA to proceed. Parallel analysis suggested a structure with a single factor, accounting for 63% of the variance and with all items loading adequately on it (all factor loadings ranging from .73 to .85). The negative mood measure was computed as the resulting factor scores.

11 One participant opted not to disclose information about their gender. Consequently, they were excluded from this analysis as it listed gender among the covariates.

12 In this regression model, the score in the second game was predicted controlling for the score in the first one, making it conceptually equivalent to a test of the improvement between the two games. Because of this, the negative relationship between the second score and the experience with video games could be due to the fact that the more experienced participants scored higher in the first game, leaving less room for improvement in the second one.