conveyed affection for the recipients, such as the “goodnight” message (Grinter and Eldridge 2001). Finally, joking messages were often hard to distinguish, as much of mobile messaging involves kidding and light ribbing. A joking message had to be a clear joke or contain words such as “ha ha” that indicated a joke was being made. All remaining messages were coded as chatting.
Figure 7 shows the message type breakdown for the two messaging conditions, with percentages relative to the total number for each messaging condition. The two were quite similar in terms of relative percentage, although keep in mind that in terms of absolute numbers there were more than twice as many messages sent in the group message condition. In both cases, chatting, arguably the most directly leisure activity, was the dominate usage. Group messaging did yield 10% less chatting, although half the difference was made up in the joking category, which previously had (Grinter and Eldridge 2001) been classified as a sub-category beneath general chatting. A slight increase in experience sharing in the group condition also contributed to the difference in general chatting.
In terms of planning, the amount of coordinating was identical in each messaging condition, although three times as much microcoordination took place in the group condition. The absolute amount of microcoordination was small, however. The percentage of photo sharing and intimate messages was roughly the same.
Figure 7: Breakdown of message types for group-based (above) and one-to-one messaging
Looking at system uses other than messaging, first, an average of 3.75 subgroups were created per study group over the course of the group messaging period. Second, participants created very few Slam events in either messaging condition (MG = .06 events per person per day, MO = .04, ns.).