Facial Bipartisanship, S.F. HU
Facial bipartisanship: Taiwanese public’s ability to discriminate between the
photos of the candidates of the two dominant political parties in Taiwan
Shun-Fu Hu1 Yu-Chen Chen1,2 Sarina Hui-Lin Chien1,* 1Graduate Institute of Neural & Cognitive Sciences
2Department of Chinese Medicine, China Medical University, Taichung, Taiwan
"Democrats and Republicans can be differentiated from their faces,” discovered Rule and Ambady (2010) in their paper under the same name. Extending their paradigm, the present study aims to delve into the role of age, number of elections attended, education, strategies in
distinguishing among photos of the Democratic Progressive Party (DPP) and Kuomintang (KMT, or Chinese Nationalist Party) candidates in Taiwan, an arguably bipartisan state. The present study consists of two experiments: Experiment 1 recruited 36 university students and Experiment 2, 36 25-55 yr-old adults, both groups gender-balanced. The same set of stimuli (photos of candidates,) were used in both experiments, stimuli controlled in such a way that the number of photos of the two parties, the gender of the candidates, and the number of candidates with glasses, were equated, and that candidates with extensive media coverage were excluded. Participants were asked to guess the party of candidate portrayed in the photo, and to indicate if they had already known to which party the candidate belonged; if they had, their response on that item was excluded in subsequent analyses. The open question “How did you guess?” was asked during an interview after the experiment. Consistent with Rule and Ambady’s study, both university students and adults aged between 25 and 55 “guessed” significantly better than chance (M=.535 (.06) >.50, p<.01). Even after we ruled out the observed response bias favoring KMT (c= -.15(.34) <0, p<.01) statistically, we still found an extant sensitivity, d’=.14 (.34) > 0, p<.01. Although age, education, and number of elections on which participants voted, had no significant effects on overall accuracy of “guessing”, we found an intriguing positive correlation between (r=.36, p=.007) adopting the strategy “face-to-internal-trait inferences” and high accuracy. In sum, KMT and DPP could also be differentiated by faces.
