On difference between direct-response method and strategy method in decision-making: behavioural and neural evidence in a reward-punishment game

AbstractDecision-makers may follow either the direct-response method or strategy method. While the two elicitation methods are theoretically equivalent, it is uncertain whether they lead to the same outcomes in practice. To explore this issue, we conduct an experiment based on a reward-punishment game under both methods, in which functional magnetic resonance imaging (fMRI) is used to provide neural evidence for uncovering the underlying behavioural mechanisms. The results show that both the reward and punishment rates are significantly higher under the strategy method compared to the direct-response method. We develop behavioural models to explain these differences and identify the ownership effect, strategic thinking, and counterfactual thinking as potential drivers. Comparison of neural activity between the two treatments demonstrates that during reward/no-reward decisions, the direct-response treatment accrues stronger activation in the bilateral anterior insula, indicating that decision-makers overweigh reward cost due to ownership effect, which consequently leads to reduced willingness to impose rewards. Conversely, during punishment/no-punishment decisions, the strategy treatment causes stronger activations in the bilateral anterior insula and inferior parietal lobule, implying that decision-makers perceive strong unfairness due to counterfactual thinking, which further results in more punishment decisions. Moreover, our findings reveal that strategic thinking influences decision-making during long-term interactions.Keywords: Behavioural ORelicitation methodreward-punishment gamefMRIcounterfactual thinkingownership effect AcknowledgementsThe authors thank the CBIR (Center for Biomedical Imaging Research) of Tsinghua University for technical support.Disclosure statementNo potential conflict of interest was reported by the authors.Notes1 No statistical procedures were used to predetermine the sample size, but our sample size for one treatment is not smaller than similar fMRI experiments in previous publications (Grecucci et al., Citation2013; Sanfey et al., Citation2003; Zhao et al., Citation2016).2 We conduct regression analyses of the senders’ and receivers’ decisions in the direct-response and strategy treatments, respectively. The coefficients of time trend are insignificant at the level of p = 0.05.Additional informationFundingThis research is supported by the National Natural Science Foundation of China under grant 72271136.