華人沒有基本與終極歸因偏誤嗎？對Morris與Peng(1994)的回應

Do Chinese commit neither fundamental attribution error nor ultimate attribution error?

不同於美國受試者傾向特質歸因，Morris與Peng（1994）發現華人無論行為者的國籍為何皆傾向情境歸因。據此，他們認為華人因其集體主義文化背景而無所謂的「基本歸因偏誤」與「終極歸因偏誤」。在本研究中，我們主張區辨兩種不同的情境訊息－－促進性或抑制性情境訊息。敏感於前一種訊息，固然會因「折扣效應」而降低特質歸因的傾向，但注意到情境中存有抑制行為發生的訊息，則有可能增加特質歸因傾向（Jones & Davis, 1965）。因此我們雖同意「華人對情境因素較西方人敏感」，但不認為由此可以推導出「華人傾向情境歸因」的說法。我們以兩個實徵研究說明Morris與Peng（1994）的發現可用樣本特殊性與故事內容不相當等非文化因素來解釋。研究一改進上述研究缺點，平衡國籍因素，並改以留在母國的台灣大學生為受試者重新驗證其說法，但無法重複其發現。研究二則利用台灣在2000年總統大選時，支持與不支持者對候選人敗選結果的歸因，直接檢驗華人對內外團體的歸因傾向是否有所差異。結果顯示不支持者較支持者對候選人的敗選更傾向特質歸因，顯示華人的歸因傾向受到行為者是否為內團體成員的影響。本文對華人歸因傾向與文化差異的意涵有進一步的討論。

Morris and Peng (1994) reported that subjects with Chinese heritage weight situational factors more than dispositional factors when a killing behavior was attributed, regardless the killer's nationality. They therefore argued that Chinese committed neither (1) fundamental attribution error nor (2) ultimate attribution error. The Collectivism/ Individualism distinction of culture was used to explain the results. We pointed out that two types of situational information should be distinguished. One is situational information that might facilitate or explain the occurrence of the target behavior, such as being treated unfairly might facilitate a killing behavior. The other is inhibitory situational information, such as social norm or bystanders' pleading that might prohibit a killing behavior. Previous research has revealed that these two types of situational information might respectively lead to the discounting effect or increase the dispositional attribution. Being more sensitive to situational information therefore would not necessarily lead to the tendency toward situational attribution. Two studies aimed at clarifying the above argument were reported in the current research. Study 1, with local Taiwanese subjects, used the same cover stories as Morris & Peng (study 3, 1994) but equated the story content for in/ out group by exchange the nationality of the killer. The results showed that Taiwanese subjects attributed the killing to dispositional factors much more than to the situational factors as we predicted, particularly for the "student cover story", and there was an interaction between attributional tendency and nationality, opposite to what have reported. We argued that Morris & Peng's findings could be better explained by the specialty of subjects and inequality of content for in/ out group. To further test whether Taiwanese treat in/ out group member differently in attribution, factors that explained the loss of a candidate in Taiwan presidential election of 2000 were weighted by those who did and did not support the candidate in Study 2. It showed that out- group voters weighted dispositional factors more than the in- group voters for the failure of a candidate, which consisted with the so- called ultimate attribution error or, more general, the group- serving bias. Implications for attribution tendency of Chinese and the potential problems of cross culture studies are discussed.