Regression and Correlation
One thing to note about demographics is that they are not especially helpful in predicting responses to the questions involved in this survey. Regression, another technique to find commonalties between variables, provides an example of this.

Each of the demographic questions (final portion of the survey) were taken in aggregate and analyzed as potentially predicting responses to other attitudinal questions earlier in the survey. In this case, I chose a divisive statement (I chose "Drum corps is changing for the better").

The one demographic variable that best explained the outcome on the "changing for the better" question was the age of respondent. That produced a regression score near nil, or almost no explanatory power. Other variables were entered, but all the variables together only predicted about 10 percent of the reason people think the way they do. Each of the predictors is statistically significant, but the substantive outcome does not allow for good predictive capability. Since drum corps is as personal and internalized as it is in many fans, this is not surprising.

Variables like age and income have little to do with enjoyment of drum corps competitions. This is important for DCI's marketing effort, as it lends credibility to the statement, "Everyone is a potential drum corps fan." Any approach to marketing drum corps must be broad, and must avoid excluding any demographic group.

In terms of fan behavior questions, there are several correlations worth noting. Fans who attend more shows overall are more likely to attend championships for one or more days. Fans who see fewer shows travel shorter distances to see them. Fans who see more shows travel farther and spend more money on souvenirs. These correlations are perhaps intuitive -- the more shows one sees, the more likely one is to have devoted significant energy to either following the drum corps activity or participating in some capacity, as a staff member or as a volunteer.

Demographics
Map of Survey Participants
Discussion
Factor Analysis
Regression and Correlation
Acknowledgements
Part 1: DCW Survey Results