Since the computer conference transcripts were shorter, we thought it
worthwhile to check if the differences we have found were affected by the
participation levels. A reasonable measure of participation is a count of
statements made: which we approximate as the sum of +ve and -ve indicators. So
we have plotted the CT ratios versus total counts for each indicator in Figure 7
(for CC) and Figure 8 (for seminars). There is no obvious relationship between
the two for the computer conference discussions. Nearly all the ratios are
strongly +ve. For the seminars, there is a tendency for most indicator ratios
to increase with participation (except perhaps novelty). There was no sign of
the often feared trivialisation of discussion as people talk more. Perhaps none
of the discussions got that excited!
Figure 7. Quality versus participation for computer conference
Figure 8. Quality versus participation for seminars.