以變形臉孔典範探討台灣成人本種族知覺登錄優勢與他種族分類偏誤

Exploring the Own-Race Face Encoding Advantage and the Other-Race Face Categorization Bias in Taiwanese Adults: Using a Morphing Face Paradigm

Levin（1996）提出本種族辨認優勢與他種族分類優勢相對論點，目前的證據多以白種人為主，跨文化的研究證據較少。本研究採用Walker與Tanaka（2003）的變形臉孔典範，設計兩個實驗來檢驗台灣成人是否呈現種族優勢相對論。實驗一為臉孔區辨作業，測試成人區辨亞洲人和白種人目標臉孔與混合亞洲人和白種人之變形臉孔的能力。實驗二為種族分類作業，測試成人分類混合亞洲人和白種人臉孔之不同比例亞洲人變形臉孔的種族。實驗一的結果顯示，對亞洲人目標臉孔僅需23%的變形即有一半機率可辨別兩張臉孔間的不同，而對白種人目標臉孔則需32%的變形才有一半機率能辨別出不同，支持本種族的知覺登錄優勢假說。實驗二的結果顯示，參與者對兩個種族的變形臉孔愈接近混合比例各半的情況下分類的反應時間愈長；此外，參與者傾向將50%的亞洲成分臉孔（即亞洲人臉和白種人臉比例各半）分類為白種人，臉孔需達55.71%以上亞洲成分才有一半機率會被分類為亞洲人，顯示出他種族臉孔的分類偏誤。整體而言，本研究以台灣成人參與者之結果支持本種族辨認優勢與他種族分類優勢的相對論點。

The own-race recognition advantage and the other-race categorization bias have been robustly demonstrated (Levin, 1996). Using a morphing face paradigm similar to Walker and Tanaka (2003), the present study designed two experiments to test whether the own-race encoding advantage and the other-race classification advantage exist in Taiwanese participants. In Experiment 1: Same/different face discrimination task, we examined the participants' perceptual discriminability for Asian-parent and Caucasian-parent conditions with swathes of morphed images of Caucasian and Asian faces of the same gender. In Experiment 2: Race categorization task, the participants were asked to categorize (Asian or Caucasian) a series of faces that were morphed by different proportions of Asian- and Caucasian-parent faces (in 10% increment or decrement). The results of Experiment 1 showed that participants (N = 20) had a higher rejection rate for Asian-parent condition than that of Caucasian-parent condition. Furthermore, fitted with a sigmoidal function, the participants' data exhibited a significantly lower discrimination threshold for the own-race condition (23% for Asian vs. 32% for Caucasian), supporting the own-race encoding/discrimination advantage. The results of Experiment 2 revealed that the response time increased as the morph level moved from the two end points (0%, 100%) to the middle value (50%). Moreover, the 50% morphing face (physically half Asian and half Caucasian) was judged to be more like a “Caucasian” face, indicating the other-race categorization bias. In sum, our results lend support for the relativity hypothesis showing the own-race recognition advantage and the other-race categorization bias in Taiwanese participants.