08/01

Two individuals noted that Pine Island township is too narrow. New Haven and Roscoe townships should also be included in the survey. This was already the plan, but not apparent from the survey materials. I will make the change to be more inclusive and note “Pine Island residents and residents along the Pine Island USPS rural route.” These are the residents who will receive the survey mailer.

​

08/03

I put up fliers at City Hall, USPS, Kwik Trip, and the Pine Island Market. I sent survey fliers to all residents along Pine Island USPS' city and rural routes. They are expected to go out the following day. One individual wished to have their data deleted from the survey. I deleted the data and shared screenshots with the individual.

​

08/04

People are responding that they are receiving my survey flier.

​

08/05

Ryan Co. | Google sent mailers to many residents. I did not receive one. Why does this fact make it into the survey log? Because it might impact survey responses going forward. I will look at proportions of responses before and after August 4 to explore whether sentiment shifted as a result of the Ryan Co. | Google mailers.

​

08/06

I put up fliers at Frandsen Bank and Hardware Hank. I put up more fliers at City Hall and the Pine Island Market. Respondent verification has begun. I am verifying names and addresses against the Goodhue County Parcel Viewer and Olmsted County Property Records. As of 08/06, there are 29 responses without a name and address to verify.

​

08/07

I added into the survey, via www.pineislandsurvey.com, three mail-in responses. I noted the date added on the mailed in survey flyers.

​

08/08

I put up fliers at The Pine Island Sports Bar and at the American Legion. I added into the survey, via www.pineislandsurvey.com, seven mail-in responses. I noted the date added on the mailed in survey flyers.

​​

08/10

I added into the survey, via www.pineislandsurvey.com, sixteen mail-in responses. I noted the date added on the mailed in survey flyers. I continue verifying responses. As of 08/10 I have verified 166 resident responses. Someone asked me today whether I use AI to verify records. I do not. I will not upload responses, names, or addresses anywhere. I verify responses against public records by hand.

​

08/11

I continue verifying responses. As of 08/11 I have verified 200 resident responses.

​

08/13

I added into the survey, via www.pineislandsurvey.com, eight mail-in responses. I continue verifying responses. As of 08/13 I have verified 300 resident responses. I have visualized the proportion of responses to the question: Do you support the proposed hyperscale data center in Pine Island? and added to the front page of this website. I had to create another Qualtrics account since my free trial ends in 2 days and I was unable to extend another Strategic Research account trial. The links have all been updated, but the QR code will now go to a survey that is no longer live. I have added an end-of-survey message to that old survey link with a redirect to www.pineislandsurvey.com.

​​

From front page (as the front page text will be updated) Total survey responses: 526. ​THE RESULTS ABOVE WILL LIKELY CHANGE SOMEWHAT. As we get more responses in over the next two weeks, we will see the results shift somewhat. With a current verified response sample of 300, the shifts to the results are expected to be small. ​Response verification continues. ​A large number of responses (n = 64) are missing name/address and therefore cannot be verified.​

​

An individual on Facebook asked a lot of questions. I answered them there. Here are the answers:

1.    What is the quality control points used in verification? Answer: In verification, I use publicly available data from Goodhue and Olmsted counties. I verify that the name and address match county data. All of these publicly available data can be found online.
2.    How many of the 526 were excluded due to the verification process? Answer: I have verified 300 responses. Another 68 cannot be verified because they do not have a full name and address. 32 are excluded because they are a non-55963 address or a duplicate. As of today 08/14 at 4pm we have 579 completed responses. That leaves another 179 to verify.
3.    Are you counting how many are excluded for what metric disqualified it? Answer: Yes. I am recording whether they are excluded for lack of name and address, for having a non-55963 address, for being a duplicate, or for something else. This will be included in the final survey report.
4.    Do you have separate verification metrics for the digital vs mailer survey? Answer: No, I use the same method to verify name and address for the digital vs mailer survey.
5.    What is the % returned vs census data? Answer: We have 579 responses so far. The best census-based figure I find is approximately 4,224 adults (age 18+) living in zip code 55963. That means we have responses from roughly 13.7% of adult residents, halfway through the survey period.
6.    Did you collect a sample size from the target pool that can actually present meaningful numbers? Answer: That’s a difficult question to answer until I know what the final sample size will be. My efforts have been to make sure all residents are aware of the survey and know how to complete it. I sent out survey fliers to all households in zip code 55963. I sent a letter to the editor of the Zumbrota News Record. I sent press releases about the survey to the Zumbrota News Record, the Post Bulletin, and Rochester’s News-Talk KROC. I posted to Facebook and LinkedIn. I have hung up fliers at City Hall, USPS, Hardware Hank, the Van Horn Public Library, Frandsen Bank, the Pine Island Market, Kwik Trip, the American Legion, and the Pine Island Sports Bar. And I rely on word of mouth. Please remind those you know of the survey so that their voices can be heard.
7.    I did receive the mailer - I believe there was a question on it asking for a comment to include in presentation to Council- and if you wanted your name attached. That was a red flag for me on the ability to collect raw data from the target pool, unbiased. Answer: Correct, I ask in the survey if respondents would like to leave a message for our City Council. This message could either be named or be anonymous. When the survey is closed, I will share all those messages with our City Council. What about the question raised a flag for you? The question itself is optional. You do not need to answer it.

​

08/15

Ryan Co. | Google sent mailers to many residents inviting to a Pine Island Community Celebration on 08/19. I did not receive one. Why does this fact make it into the survey log? Because it might impact survey responses going forward. I will look at proportions of responses before and after August 19 to explore whether sentiment shifted as a result of the Ryan Co. | Google mailers and the celebratory event.

​

08/16

I added into the survey, via www.pineislandsurvey.com, eight mail-in responses.

​

08/17

I put up survey fliers at Ag Partners. I posted to Facebook that I will be at the City Council meeting tomorrow, Tuesday, at 6pm, with survey fliers and an iPad for those who would like to submit a survey. An individual on Facebook said they thought this would skew the results because most people who attend City Council meetings are opposed to the Data Center.

​

08/18

I continue verifying responses. As of 08/18 I have verified 404 resident responses. I have updated the results on the main page. I attended the monthly City Council meeting with two iPads and survey fliers and invited people to complete the survey. I received no survey responses before, during, or after the Council Meeting. I was asked by an individual whether I had shared any of the results with my husband Joe. This question was likely in response to Joe's comment in his own speech to the City Council on questions "the majority of residents have about the data center." No, I have not shared any of the survey results with Joe. He sees what everyone else in the community sees: the updates on the Pine Island Facebook page and the updates on this website.

​

08/19

I attended the Ryan Co | Google Celebratory event in Pine Island. I asked people, "Hi! Have you heard about the Pine Island Sentiment Survey?" and handed them fliers. I talked about the survey with ABC6 KAAL, KTTC, Post Bulletin, and MPR radio. The Post Bulletin published an article with a link to this website. I spoke with Peter Fitzgerald of Ryan Co and asked him what his thoughts were on the preliminary results. Peter asked whether I would submit the survey to a "forensic analysis," which means an exploration into whether survey responses were fake, duplicated, or bots. I said yes, but that I will not share the data he asked for. I will share my methods in detail, every step of the validation process I have taken, and the aggregated results. I will not share survey data.

​

08/20

I added into the survey 16 mail-in responses. I continue verifying responses. A number of responses come in from Milton Township in Dodge County. I will explore publicly available Dodge County records to verify these responses.

​

08/22

I added into the survey one mail-in response. I continue verifying responses.

 

An individual asked on Facebook today: "So my husband filled this out three times…. Once again the validity is in question with multiple same-person replies and only Pine Island residents. Please share the stats on multiple replies you have gotten and non resident contributors." My response: "Hi [...] you are right that the survey link is open to multiple responses. I chose to do it this way so that multiple adults in the same household on the same wifi could participate. I remove any duplicates. So far, there are 5 duplicate responses. These were excluded from the results I shared. I also want to clarify that the survey is open to Pine Island residents and nearby township residents. I sent out survey fliers with USPS to all those who live along Pine Island urban and rural postal routes (~2,200 households). So far, there are 39 responses from outside zip code 55963 (responses from Rochester, Oronoco, Mazeppa, Wanamingo, and Zumbrota). These were also excluded from the results I shared."

​

Another individual asked: "I've heard some people say that your survey is inherently wrong because it is such a small sample size. Statistically speaking, do we really need 1k responses for the outcome to be a good indicator of the population? Aren't we way past a statistically significant threshold?" My response: "Hi [...], thanks for asking! Statistically speaking, no, we do not really need 1k responses. At 473 verified responses, we can have a high degree of confidence that the results will hold. With the current sample size, our 95% margin of sampling error is ±3.8 percentage points. That would put the approximate full population estimate at 69.1%-76.7% No. We can also use power analysis to estimate the likelihood that the remaining population would have the opposite view. If we've heard from 473, what is the likelihood that the remaining population of ~3,675 is largely supportive (let's say 70% Yes and 30% No)? The answer is essentially zero.
But larger samples carry more face validity, and 1,000 is a nice and round number 😃 I'd love to see us get there!"

​

From front page (as the front page text will be updated): Total survey responses: 627. ​THE RESULTS ABOVE WILL LIKELY CHANGE SOMEWHAT. As we get more responses in over the next two weeks, we will see the results shift somewhat. With a current verified response sample of 404, the shifts to the results are expected to be small. ​Response verification continues. ​A large number of responses (n = 75) are missing name/address and therefore cannot be verified.

​

08/26

I added into the survey five mail-in responses. I continue verifying responses. In addition to Goodhue and Olmsted counties, I am also checking Dodge County publicly available data.

​

From front page (as the front page text will be updated): Total survey responses: 707. ​THE RESULTS ABOVE WILL LIKELY CHANGE SOMEWHAT. As we get more responses in over the next two weeks, we will see the results shift somewhat. With a current verified response sample of 473, the shifts to the results are expected to be small. ​Response verification continues. ​A large number of responses (n = 88) are missing name/address and therefore cannot be verified.