This article, originally published by Al Zucaro on BocaWatch.org, is preserved for historical purposes by Massive Impressions Online Marketing in Boca Raton.
If there are questions or concerns with the content please e-mail email@example.com.
Well the person that proclaimed Houston’s on the Wildflower site is a “done deal” is at it again. This time Mr. Glenn Gromann is on the public record as saying that the City of Boca Raton’s 2011 Electronic Survey of citizens regarding their preferences for use of the Wildflower site is “by any stretch of the imagination not statistically significant to determine anything.” Councilman Weinroth, on the other hand, recently asked the right question regarding our scoring of that survey, i.e., “[I’m] interested in your statistical confidence interval to validate the results.” This paper explains standard, accepted statistical confidence intervals, also known as margins of error, calculated for our scoring of the City’s survey. We show that Mr. Gromann’s comments are wrong and reveal a pattern of trying to publicly belittle and discredit citizens with information that is incorrect.
The situation regarding the City’s 2011 Electronic Survey is that the scoring of the survey results by the City Staff misrepresents the responses of the citizens. The results of this survey were used as the basis for deciding to pursue an arrangement for a restaurant on the Wildflower site. This is described in more detail in the October 19, 2015 article in the following link.
Mr. Gromann’s comments about this article were made at the October 27, 2015 City Council meeting. View this short video clip if you missed his comments at the meeting.
Using Confidence Intervals to Establish Statistical Significance
The information provided to Councilman Weinroth shows the confidence intervals of our scoring of the City’s 2011 Electronic Survey. The confidence intervals relate to the second line in the following table.
At a 95% confidence level with a population of 91,000 and a sample size of 78, the following confidence intervals exist:
That is, the Park category in line 2 has a confidence interval of 60% ±11% and the Restaurant category in line 2 has a confidence interval of 21% ±9%. The general rule of thumb is that the lead (Park 60%) must be at least twice the margin of error of the second leading category to be significant. So the margin of error on line 2 of above table is 11% and the lead is 60%. Adding 22% (2X11%) to the restaurant result (21%) means that the park result is significant, i.e., 43% vs 60%. These calculations were performed using the mathematics of probability.
A simple way to interpret the results is that there is 95% confidence that if the City’s entire 91,000 citizens voted, between 49% and 71% would favor a park. Additionally, between 12% and 30% would favor a restaurant.
Note that the situation that Mr. Gromann is attacking is the validity of the survey conducted by the City of Boca Raton. This survey was not conducted by the authors of this article or any citizen group. It was designed and conducted by City Staff. The execution of the survey seems fair and unbiased. The question posed to the citizens was open-ended and not leading. All citizens had the opportunity to respond and they were provided three months to do so. Responses were accepted from the City’s Web site and by email and by letter. As noted in the background section, a problem exists with how the City Staff scored the results. The authors of this article, during due diligence, rescored the results based on each citizen’s intent. We then calculated statistical significance based on the fairness of the survey process and the rescored results.
In his comments Mr. Gromann indicates that 150 responses to a polling of 91,000 citizens “is not statistically significant by any stretch of the imagination to determine anything,” yet for four years he accepted the City Staff’s conclusion that a restaurant was preferred based on 61 responses. The timing of his turnaround is curious as it occurred shortly after we wrote an article on the survey scoring issue and the restaurant was announced by the City to be an “undone deal.” It looks like he is making a desperate attempt to prove something, but it is not clear what that is or why.
Another issue with Mr. Gromann’s comments, that is a pattern, is that he waits until all citizens have spoken during the public request interval before he speaks. He then tries to belittle the citizens who have spoken before him in order to create doubt regarding the credibility of their comments. As noted in this article, his comments are either ignorant or untruthful but linger in the minds of City officials and citizens because prior speakers do not have to opportunity to refute his comments. To an extent his tactic overwrites the messages of prior speakers.
It is very time consuming for citizens to perform the quality analysis described in this article. We also have to pay the City for public records. Then we communicate our finding to City Council members via email, but get a quasi-canned response that doesn’t address the issue but says something like “I’ll give it my full consideration.” To make our voices heard we then have to prepare a talk and attend City Council meetings to deliver our message. When City Council members sit silently and allow Mr. Gromann to attack our credibility it is quite discouraging. We then have to write articles like this one to potentially motivate someone to do something.
When Mr. Gromann makes assertions that are untrue he crosses a line. He is either being deliberately dishonest or is ignorant of the truth. There seems to be a pattern of his behaving this way while involved with City business and City leaders have been advised of this, but do nothing. This has to change as it reflects badly on City leadership and is a disservice to the citizens.
This article addressed concerns about Mr. Glenn Gromann’s asserting that the City of Boca Raton’s 2011 Electronic Survey of citizens regarding their preferences for use of the Wildflower site is “by any stretch of the imagination not statistically significant to determine anything.” We rescored the City’s survey results to reflect the intent of each respondent. Using the rescored results, we then calculated confidence intervals to the 95% confidence level using the mathematics of probability. We then calculated statistical significance using the generally accepted rule of thumb that the leading category has to be at least two times the margin of error from the next leading category. The conclusion from this is that the rescored survey results are statistically significant.
We then questioned Mr. Gromann’s behavior regarding his tactics where he attempts to belittle citizens and cast doubt on their credibility using a process advantage, but with no factual information. We then question why the City leadership does nothing about this after many citizen complaints.