After the protest, despite our expressed negative opinions the 9R board was gracious enough to invite us all to sit in on the beginning of the meeting (before they went into closed executive session). I found it to be quite eye-opening. They were discussing the results of a recent survey of DHS students, teachers, staff, and parents that rated 9R's performance. They were brainstorming how to publicly present the survey results. The respondents had registered their opinions on some 16 questions by selecting a number on a "1 to 10" scale. A few board members had already invested a lot of time collating and processing all the data, producing bar charts which showed the variation in responses for each question. This would be an accurate and easy-to-understand way to show the results. To their credit they were willing to do this despite the fact that there were many negative answers as well as many positive ones, although a couple of board members seemed worried that the public might focus on the bad and ignore the good. So far so good, though, and I felt that the board was basically on the right track.

Mary Barter's Proposal
However, Mary Barter was also present at the meeting, and she didn't want to publish the full data from the survey. She proposed that rather than show all of the responses, they use an obscure statistical property called the "mode" to represent the results for each question. What is a "mode," you may ask? I had to look it up, and I've  actually devoted several years to studying the analysis and clear presentation of data. It turns out that the mode is just the survey answer for which the greatest number of votes were cast. In a simple example, if 112 people answered "1" (poor performance) and 122 said "10", then the mode would be 10, and Mary Barter proposed to simply put an "x" in the "10" column for that question and represent that as the survey result. People would erroneously think that everyone who responded to the survey thought that 9R was doing a great job on that issue.

Ms. Barter said she thought that this type of presentation was easier for parents to understand, and that she had used this approach many times in the past when presenting data. I personally found this offensive. First, I resent the implication that Durango parents are too stupid to understand a simple bar chart (we see these charts every day in newspapers, magazines, and TV reports). Second, the mode, while it is in fact a characteristic of any data set, does not faithfully communicate the survey results since it throws away all information except the identity of the most popular answer. Imagine if a teacher surveyed the students in her class, found that 12 had brown hair and 15 were blond, and decided that since the mode was the group of 15 blond students then her entire class must be blond!

How Real Analysts Do It
In practice, real-world survey data results can rarely be described well using any single number. For instance you might be tempted to report the average or "mean" of the results, a common practice for certain types of scientific data sets (where the data forms a classic "bell curve"). But in the example above, using the mean also presents a very distorted picture since the average of 112 people saying "1" and 122 saying "10" would be about 5.7. This type of average is calculated using the formula (112x1 + 122x10) / (112 + 122). Seeing just the average, people would think that the respondents were in agreement on a value of about "6", or what on the 1 to 10 scale would probably be "fair" performance. But really no one at all voted for "6", and there was a great discrepancy in opinions with about half saying "1" and the other half "10." You can only accurately report this type of data using something like a bar chart, where each possible response has a bar whose height is proportional to the number of votes for that answer.

So, I am disappointed that  Mary Barter would try to put a spin on the study results by throwing away almost all of the data. She is highly educated and should certainly know better than to try and use the "mode" to report this type of information. This leads me to think that she is trying to present a simple, rosy picture of 9R when in fact many people are dissatisfied with the administration's performance. How will our school district ever improve if its administration throws away any negative opinions that are expressed? And how many times in the past has other important information been so distorted?

And in Other Opinions...
At the end, the board debated how to list the individual opinions from a "comments" section on the survey. They were worried about a number of comments with vulgar language, as well as some which revealed the identity of individual teachers or students (responses were supposed to be confidential). Some on the board thought these questionable comments should simply be left out of the final report, and they should print only the remaining answers. But I strongly suspect that responses with vulgar language or referring to individuals would most likely be negative in tone. To leave them out would again tip the scales in the favor of positive responses and present a better opinion of 9R's performance than the real comments indicated.

There is an easy way to include problematic comments without upsetting or implicating anyone. Those of us old enough to remember the release of the "Watergate tapes" transcripts recall that President Nixon cursed like a sailor. The editors of the published transcripts just printed "<expletive deleted>" in place of all of the curse words. In our case, we could also include vulgar responses in the format "In my classes, dyslexic students are treated like <expletive deleted>." The gist of the opinion would be accurately shown without printing offensive language. Likewise, responses identifying a particular person could be printed like "I think <my teacher> doesn't pay enough attention to gifted students."

In Conclusion
At the meeting I observed to what lengths Mary Barter and a minority of board members will go to spin the results of a simple opinion survey. Instead of showing the public all of the answers, they want to hide the "bad" data by simply throwing it away. This despite the fact that they spent good money (our money!) to have the survey administered in the first place.

This attempt to mislead the public is but one reason that a large number of us parents, students, and teachers feel that it's time for a change. We feel that the time has come for Mary Barter to move on and for a true new direction for our school district, one where the administration is more open with information and more willing to listen to the public.