跨期选择单维占优模型的过程检验

跨期选择需要决策者在眼前与未来的损益之间做出权衡与取舍。主流跨期选择理论认为, 跨期选择是把未来价值折扣到现在, 根据折扣后的价值进行选择的过程; 而单维占优模型则认为决策者把“结果”维度上的差异和“延迟”维度上的差异进行比较(维度间差异比较), 然后根据差异更大的维度(即占优势的维度)进行决策。跨期选择众理论之争无果的原因之一或是研究者未能找到揭示其心理过程的令人信服的证据。为此, 本研究采用过程检验而非结果拟合的方法, 首创了“直观模拟天平”任务, 对维度间差异比较的大小进行测量, 从而为验证单维占优模型提供了支持性的证据。实验1A证明了维度间差异大小中介了选择结果, 被试认为“结果”维度上的差异相对于“延迟”维度上的差异越大, 越倾向于选择延迟长、结果大的选项; 反之, 被试认为“延迟”维度上的差异相对于“结果”维度上的差异越大, 越倾向于选择延迟短、结果小的选项。实验1B证明了维度间差异比较是在选择时实时(而不是在选择后)进行的。实验1C通过调整实验程序消减了共同方法偏差对结果的影响。实验2进一步证明维度间差异比较可以解释众多的跨期选择异象。通过4个实验, 本研究揭示了维度间差异比较构成了跨期选择的重要(尽管可能不是唯一)决策过程, 为支持单维占优模型提供了重要的过程验证证据。

A process test of priority models of intertemporal choice

Intertemporal choice is vital to human survival and development, hence the interest of scientists today in analyzing how it is made. Mainstream theories hold that a discounting process serves as basis for making choices, but other theories forego this process. For example, priority models, such as the tradeoff model and equate-to-differentiate theory, show that individuals tend to compare the difference between dimensions of delay and outcome before deciding in a single dimension when selecting between a smaller, sooner outcome (SS) and a larger, later outcome (LL). To tackle this question and improve our knowledge of the mechanisms underlying human decision making, concrete evidence based on psychological process is required. In the current study, we tested priority models by using a process-oriented method.
A total of 822 college students participated in the study; 238 participated in Experiment 1A, 194 in Experiment 1B, 97 in Experiment 1C, and 293 in Experiment 2. In Experiment 1A, we manipulated the differences of two alternatives (i.e., SS and LL) in terms of monetary outcome (in Problem 1) and delay (in Problem 2). We then asked participants to indicate their preference between SS and LL and to compare the difference on the delay dimension with that on the outcome dimension (we called this process “difference- comparing process”). The aim of Experiment 1A was to examine whether one’s preference was mediated by the difference-comparing process. For Experiment 1B, we altered the procedure used in Experiment 1A (i.e., participants were first asked to conduct the difference-comparing process then indicate their preference) because we wanted to ensure that the difference-comparing process is the online process of selecting rather than the artifact of the experimental procedure. In Experiment 1C, we asked participants to indicate their preference and conduct the difference-comparing process at separate occasions. The aim of Experiment 1C was to address the problem of common-method bias. In Experiment 2, we examined whether the difference-comparing process can explain anomalies cited in previous studies on intertemporal choice.
The results of Experiment 1 showed that the difference-comparing process mediated the preferences of the participants. This finding implied that the participants will more likely select LL when the perceived difference on the outcome dimension is larger than that on the delay dimension, whereas they are likely to select SS when the perceived difference on the delay dimension is bigger than that on the outcome dimension. In addition, we inferred that increasing the magnitude of both outcomes by the same multiplicative constant increases the perceived difference on the outcome dimension, hence the magnitude effect. Adding a constant to both delays decreases the perceived difference on the outcome dimension, hence the common difference effect. Replicating the results of Experiment 1A, Experiment 1B also demonstrated that the difference-comparing process is an online selection process, and Experiment 1C mitigated the concern of common-method bias. The results of Experiment 2 replicated the effects of several existing anomalies, which demonstrates that the difference- comparing process can explain most of such anomalies, except for the sign and speedup/delay effects.
In summary, the results of the four experiments showed that the dimensional difference-comparing is a key process but not the only psychological process of intertemporal choice. The results of our study deepened our understanding of the mechanisms underlying intertemporal choice and added important procedural evidence to current priority models.