Chapter 5 Results
5.2 Findings
5.2.1 Do rules affect behaviour? If so, are the changes in behaviour
5.2.1.2 Live Venue Overview
Computer logs were analysed for interaction with the rule module for teams both with and without preset rules. The logs were also used to trace individual
participant behaviour associated with the firing of the preset Give_Feedback rule. The individual behaviours were set against the contextual behaviour of other team members. Participant comments that applied to the rule module or specific rules, whether in on-line discussions or survey questions, provided a second data stream. Team managers from the live venues acting as a third data source, were asked for general observations and specifically about intra-team conflicts arising from participant behaviour.
All teams’ participants were physically and culturally close to their fellow
members. Venues 1, 3, 4 and 5 met face-to-face to sort out task-related problems. This was sufficient to assume that they knew what the appropriate norms of team behaviour might be. This meant that there may be no need to clarify or monitor behaviour of the virtual team, and hence no need to use the rule module.
Issues that arose within the school groups were sorted out primarily by teachers or “in the corridor” by a team member. Two issues arose that required teachers to step into the team performance, one of abuse and another of level of participation. One instance of “flaming” involving offensive language directed at another individual (venue V5) and another of verbal bullying that belittled another member’s ability (venue V1). The second issue addressed in a face-to-face manner was one of irregular contribution by one member (venue V1).
A computer solution to the first of these two issues would rely on content analysis at a level of sophistication beyond the prototype rule module (but worth
considering as a feature of future versions). Irregular contribution, however, could have been monitored by the rule module.
V1_M6 suggested that a rule be adjusted to accommodate the attendance behaviour of another member, and asked if any of the team knew how to create rules. V1_M6 assumed that attendance measured contribution. There actually were a number of rules about attendance, (Poor_Attendance,
Poor_Attendance_Tell_Me, section 3.5.1) but none that named poor attendees to the whole team. The opportunity for creating a suitable rule, based on attendance or some other set of measures was lost, because neither the leader in this team nor the participant wanting the rule, made a request to the researcher. The existing attendance rules failed to change behaviour in this instance.
Participants from venue V1 were asked after the trial how often they were tempted to create a rule and to comment. Five of the six participants responded. All respondents reported that they were tempted to create a rule (four selecting the second choice, “1-2 times” and a fifth, V1_M6, the last option, “7+” times). They all noted not acting on this temptation, one because he would “rather decipher binary code” and three because they did not know how to do it. V1_M6 didn’t know how to make rules but did make a cursory attempt to use the tutorial. The discussion types posted by individual members were extracted and aligned with the firing of the Give_Feedback rule, so that changes in the posting of messages classified as “feedback” might be identified. Discussion types were also
provided for the whole team, grouped in threads. Indications of peer influences on behaviour assisted in understanding the effect that rule firing had on behaviour.
Feedback posted in Discussions (%)
20.9 20 25 19.2 1.3 0 5 10 15 20 25 30 V1 V2 V3 V4 V5 Venue P e rc e n ta g e
Figure 18. The proportion of discussion posts that were labelled “feedback” from each venue. Venue 5 comprised five separate teams. None of the venue 5 teams had the Give_Feedback rule, while the other 4 venues did have this rule. “Feedback” was one of 8 categories available for each discussion post.
The comparison of venues strongly supported the proposition that the presence of the rule Give_Feedback was responsible for the high proportion of discussion posts being labelled as “feedback” (Figure 18), although there may have been something inherent in venue 5 other than the rule, which accounted for the lack of “feedback” postings. This would have to have been at the venue level rather than the team level as 5 teams operated within the venue. Further, the probability of a post being “feedback”, given random classification, should be 12.5%, rather than occurring 20% of the time. Choosing from a list of classifications will not
necessarily result in a true random choice, as an individual may tend to choose from the beginning, the end or the middle. Individual tasks, general choices like “other” and team habits, all have the potential to provide variation from
“random”. What is notable from the result is the relatively consistent outcome from a variety of venues. The rule Give_Feedback appears to be the likely cause
for this cross-case result. Certainly detailed examination of another data stream was warranted. Did the rule influence behaviour at an individual level?
Behaviour change subsequent to rule firing also strongly suggested that the rule altered member behaviour. The following evidence comes from a close analysis of classification behaviour as it related to firing of the Give_Feedback rule for the venue V1 team. Analysis from one team at this depth is sufficient to reinforce the findings from the focus groups and the outcomes from the venue V2 team.