In this article, the author examines the phenomenon of flaming - hostile, uninhibited, computer-mediated communication. His experiment attempts to clarify whether this type of communication is a unique characteristic of the medium, or is dependent on the users' social context. To this end, he examines the level of flaming in four Usenet social newsgroups. The four groups chose were as follows:

• soc.culture.japan
• soc.culture.canada
• soc.culture.indian
• soc.culture.arab

Messages from these groups were analyzed over one week. Each message was first analyzed to determine if it was an uninhibited expression of hostility (1138). The results were categorized as relating to the following issues:

• national and political
• religious
• cultural, social, and ethnic

The results of the experiment demonstrated a considerable difference between the amount of flaming in each group. For example, soc.culture.arab had the least number of postings, but the highest percentage of flaming. If flaming were a result of the medium, the results would have shown an insignificant difference in the percentage of flames of each group. The disparity in the amount of flaming between these groups supports the hypothesis that flaming is a result of social context, rather than the medium.

The author urges further studies be conducted in determining the direction of flames - that is, whether flames are directed at members of a group or individuals perceived as outsiders.

In his conclusions, the author points out that support computer-mediated communication is similar to other forms of communication where "tone and emotional content is affected by social and interaction context, rather than the medium itself." (p. 1141)

This article was an excellent point-study of what has often been considered a unique negative aspect of computer-mediated communication. By studying groups that already have a cultural base, rather than more "virtual" communities, the author places the emphasis firmly on the computer-mediated medium, rather than the "new" online culture. The article was brief, concise, and described the relevant experimental and analysis methods.

The experimental methodology used by the author is an effective example of computer-mediated communication study. By using a readily available source of information, he observed human interactions with a minimal impact on the participants. This type of interaction study provides an efficient way of analyzing human interactions in a virtual community without relying on self-evaluations or the pit-falls rampant in experimental settings.

In addition, this study demonstrates how the medium is not responsible for the level of hostility in online interactions when a cultural basis exists. However, this result cannot be extrapolated to online cultures, whose cultural basis is the media.

This article has 17 references in the reference list.