Conflicts among Net news participants and cultural background

Participants in Internet net news bulletin board discussions can argue with others whom they do not know. The nature of each message in a discussion can be characterized by the frequency with which keywords appear in the message. This incidence or frequency can be summarized as a principal component score. By deriving two characteristic indexes – auto‐correlation coefficients and power spectra – from the principal component scores, we have come up with a novel method. Applying this method to analyze three large net news threads, the index plots revealed two‐generation cyclic fluctuations. Comparing these plots with actual points of conflict obtained by reading the message contents, a fairly good correlation was obtained between the two and it was found that most of the conflicts were among participants with different cultural backgrounds.