The ``McGurk effect'' demonstrates that visual lip-read information is incorporated into speech perception even when it is conflicting with auditory information. While the McGurk effect has been shown to be robust in English speaking cultures, interlanguage differences between native speakers of Japanese and American English have been found. Native speakers of Japanese tend to show a weaker McGurk effect than native speakers of American English when listening to clear speech of their native language, although these Japanese subjects show a highly increased McGurk effect for stimuli with auditory noise or for stimuli of their non-native language. This auditory reliance of the Japanese may be due to their cultural habit of avoiding looking at a speaker's face. The Chinese who are believed to be similar with respect to the face avoidance also showed a weaker McGurk effect than American subjects. More intriguing about the Chinese subjects was that the subjects who had stayed in Japan longer tended to show a stronger visual effect (r=0.723). These results imply that people who are seriously learning a second language also learn to use any cues including lip-read information to improve their listening comprehension.