Key study: Yuki et al (2007)

Aim: to see whether people in Japan focus more on the eyes, and people in the USA more on the mouth, when judging emotions.
The idea: in Japan, showing strong emotion openly is often discouraged, so people may read the eyes, which are harder to control. In the USA, emotions are shown more openly, so the mouth may be a clearer cue.
Participants: university students in the USA (Ohio State University) and Japan (Hokkaido University).
Study 1: participants rated emoticons made of different combinations of happy, neutral and sad eyes and mouths, on a 9-point scale from very sad to very happy.
Study 2: the same idea using edited photographs of real faces, with the eyes and mouths taken from different expressions.

Results: Japanese participants gave more weight to the eyes: faces with happy eyes were rated happier by Japanese than by American participants. American participants gave more weight to the mouth: faces with happy mouths were rated happier by Americans. The same pattern appeared with photographs.
Conclusion: culture affects which facial cues people use to recognise emotion. Japanese people focus on the eyes, and Americans on the mouth. This suggests reading facial expressions is partly learned, which challenges the idea that it is completely universal.

Strengths: two studies with different materials (emoticons and photographs) found the same pattern, so the results are more reliable. It used fairly large samples from two cultures, and ratings gave quantitative data that is easy to compare.
Weaknesses: participants were all university students, who may not represent everyone in each country. Emoticons and edited faces are artificial, so ecological validity is low. Combining eyes and mouths from different expressions creates faces that look odd. It compares only two cultures.