後設認知偏誤威脅心神漫遊研究中思維探針法之建構效度

Abstract

思維探針法是心神漫遊研究中一項重要的研究工具。其施行方式為：在受試者進行作業時，隨機打斷受試者，並詢問受試者當下的思緒狀態為何。透過思維探針法，研究者可以探討心神漫遊前後的行為特徵或認知狀態，或是探討心神漫遊的頻率與其他認知功能之間的關聯性。然而，不是每次給予受試者思維探針時，受試者皆能精準提取自身的思緒狀態。當受試者對思緒狀態的後設覺察低下時，或是當思維探針所問之問題過於困難時，受試者將仰賴自身思緒狀態以外的線索進行回答。我們認為，受試者會有「利用對自身作業表現的主觀評估來回答思維探針」的傾向。若此傾向存在，那麼思維探針法捕捉到的可能就不是受試者的思緒狀態，而是受試者的後設認知偏誤情況。於本研究中，我們利用第二型訊號偵測理論模型量化受試者的後設認知偏誤情況，並探討當受試者覺得自己的作業表現好或壞時，是否有不同的機率回答思維探針中的心神漫遊選項。本研究的三個實驗給予不同思維探針問題，並於實驗中發現：當詢問受試者心神漫遊狀態的意圖性時，受試者會有傾向利用作業表現的主觀評估，回答其心神漫遊狀態的啟動方式，且當受試者的後設覺察越低落時，此傾向則會越強。據此，我們認為後設認知偏誤會威脅思維探針法的建構效度，在思維探針的問題過難，或受試者對自身思緒的後設覺察低下時尤為嚴重。

Thought-probing method is one of the most prominent tools used in mind-wandering studies. This procedure randomly probes participants to halt their ongoing task and report their thoughts at the moment. Researchers can use the thought-probing method to investigate behavioral characteristics or cognitive states before or after mind-wandering episodes, as well as the relationship between mind-wandering episodes and other cognitive functions. However, participants may not accurately retrieve their thoughts every time they receive the probe. When participants' meta-awareness is low or the probe's questions are too difficult, other contextual cues may be used to answer the probe. We hypothesized that participants may respond to the probe according to the subjective evaluations of their own task performance. According to this hypothesis, what is caught by the thought probe is participants' metacognitive bias instead of their genuine mind-wandering episodes. In the current study, in order to study whether different self-evaluations of task performance are accompanied by different degrees of self-reported mind-wandering, we exploit the type 2 signal detection theory model to quantify participants' extent of metacognitive bias. In the three experiments of the current study, we manipulated the questions that were asked in the thought-probing methods. The result of the experiments indicate that when asked about the intention of their mind-wandering episodes, the participants were inclined to base their reports on their self-evaluations of task performance. Moreover, the participants' lower meta-awareness strengthens this tendency. These findings suggest that metacognitive bias threatens the constructive validity of thought-probing methods, especially when demanding probe questions are presented and when participants have insufficient meta-awareness.