Good morning.
Wednesday in Whitewater will be mostly sunny with a high of 74. Sunrise is 7:24 and sunset is 5:52 for 10 hours 28 minutes of daylight. The moon is a waning gibbous with 92% of its visible disk illuminated.
On this day in 1962, the Cuban Missile Crisis ends and Premier Nikita Khrushchev orders the removal of Soviet missiles from Cuba. (Other elements of the deal remained secret and an American naval quarantine of offensive weapons to Cuba remained in force until November 20.)
In mid-October, at a meeting of the Whitewater Common Council, the city heard a presentation on a community survey. The document, completed in September and presented in October, appears immediately below.
Skip to PDF contentResidents can review the lengthy survey and come to their own conclusions.
In my own case, an initial question presents itself: is the survey methodology sound?
Yes, it’s sound.
Indeed, fundamental — rational, reliable — principles of sampling are readily available in college mathematics courses, including courses for non-mathematicians. In fact, competitive American high schools offer basic statistical concepts in Algebra II courses or separately in AP Statistics courses for college-bound students. (Of the AP Statistics program and exam: “Selecting methods for collecting or analyzing data; Describing patterns, trends, associations, and relationships in data; Using probability and simulation to describe probability distributions and define uncertainty in statistical inference; Using statistical reasoning to draw appropriate conclusions and justify claims.”) These programs are both pre-college programs; college offerings in these principles are available across our nation.
However, the easiest claim to make is that the survey methodology is intellectually sound. It is. The survey’s creator, the ETC Institute, describes this method:
A seven-page survey, accompanied by a cover letter and a postage-paid return envelope, was mailed to a randomly selected sample of households in Whitewater. The cover letter outlined the survey’s purpose and encouraged recipients to complete and return it by mail or participate online.
To ensure the integrity of the random sample, respondents who completed the survey online were asked to provide their home address, which was matched against the original sample file. Responses from individuals who were not included in the random sample were maintained separately and were not included in the statistically valid results. To further boost participation, ETC Institute conducted follow-up outreach to sampled households and promoted the survey through targeted social media advertisements on platforms such as Facebook and Instagram.
The goal of the survey was to obtain at least 500 completed responses. This target was exceeded, with a total of 512 residents completing the survey. The results for the random sample of 512 residents have a precision of at least ±4.3% at the 95% level of confidence.
This is an appropriate sample size for a survey of a community of Whitewater’s population. (The adequacy of a probability sample does not depend on obtaining responses from some large percentage of the population.) There is nothing novel or controversial about these principles. As I’ve noted, they are taught in introductory statistics courses, including those available to high school and college students.
Former students, who remember a course in college on these topics — including non-mathematicians like me — will recall established formulas to determine the adequacy of a sample size. (The formulas will also bring back, perhaps, difficult memories of studying for exams in the subject. My apologies — watch a few cat videos and you’ll feel better again.) I’ve gone searching online for memory refreshers. Here’s one from Khan Academy (a site that did not even exist when I was in school — we had to use books!): Khan Academy, Determining sample size based on confidence and margin of error. (The online video looks like it’s for AP Statistics prep, by the way.) The sample size ETC took is easily adequate.
These are enough properly gathered responses for a city of our size. Note particularly that, among those identifying as ‘university-related,’ 120 respondents identified themselves as alumni, and 37 as parents or family members of students. Those responses alone account for 157 respondents, none of whom should be excluded from the resident population because of that relationship. While some identified student respondents did answer the survey (22), university-related non-student residents represent a far larger number of respondents than current student respondents. (Contentions along the lines of university affiliation = student = not a resident = discard the response are, themselves, a failure of understanding. My claim is not that the other 22 student respondents are not residents. It’s simply that excluding residents based on their university connection, whatever it may be, would be a failure of an adequate survey method. The survey is, after all, about residents.)
More particularly: excluding university-related residents from a sample would be as erroneous — if not more so by sample size — as excluding those with unfounded claims about survey methodology.
The survey is lengthy; one can find points with which one agrees or disagrees. I find the results neither wholly supportive of city officials nor wholly critical of them. An adequate sample of residents expressed satisfaction with some services while expressing dissatisfaction with other aspects of life here. That’s about what anyone would reasonably expect, and that is what the survey reliably and soundly reveals.
Aerial footage shows aftermath of deadly Spain train collision: