More serious flaws detailed in consultant’s work
Three days after the Sarasota County Commission directed its redistricting consultant to work with him, a county resident with expertise in economics and Geographical Information System (GIS) software advised the consultant that one of the resident’s primary concerns is that the consultant’s estimates “include values that are demonstratively outside any reasonable margin of error,” The Sarasota News Leader has learned.
R.N. Collins copied all of the county commissioners on his Oct. 10 email correspondence with Tallahassee-based consultant Kurt Spitzer.
“Out of range errors indicate that either the input data is unreliable or the methodology is flawed, and the estimates should not be used without additional analysis,” Collins added in the email.
Moreover, Collins pointed out, Sptizer’s estimates “cannot be reproduced because the County has not made all of [the] data and transformation files available to the public. Without this data, especially your housing data and transformation files,” Collins continued, “it cannot be determined if substantial differences between your results and those produced by applying your methodology to publicly available housing data are due to bad input data, faulty implementation of your methodology, flawed methodology, or inaccurate publicly available housing data.”
In response to questions posed by the News Leader, Richard Doty, the GIS coordinator and research demographer at the University of Florida’s Bureau of Economic and Business Research (BEBR) — who assisted Spitzer with the figures — wrote in an Oct. 14 email, “The county has our methodology, which is transparent and reproducible.”
Multiple errors Collins said he had discovered in the materials Spitzer provided the county were the focus of a Sept. 29 article in the Sarasota Herald-Tribune written by reporter Zac Anderson.
During public comments on Oct. 7, Collins stressed that he had even more concerns than the article had noted.
Among them was the fact that Collins’ analysis showed an overestimation of the Hispanic population in District 5, which includes part of the City of North Port and the IslandWalk at the West Villages.
According to Spitzer’s material, the estimated 2018 population in Census Tract 2718, Block 1100 is 1,003. Spitzer’s methodology put the number of Hispanic residents there at 394.
Yet, Collins pointed out in his email to Spitzer, the 2010 U.S. Census population for that same block showed three residents, only one of whom was Hispanic.
On Oct. 7, in his public comments, Collins also noted “a systematic underestimation of North County populations.”
In preparing other detailed graphics, Collins used Spitzer’s data in comparison to Sarasota County GIS data to illustrate that Spitzer had wrongly assigned specific population counts to unincorporated parts of the county when the figures should have been included in the cities of Sarasota and North Port and on Longboat Key.
For example, Spitzer placed 95.504 residents in North Port who actually were spread among three unincorporated sections of the county. For a second example, 223.568 Longboat Key residents were assigned to unincorporated areas of the county.
The total population count affected by those mistakes was 1,413.302, Collins determined.
Further, Collins pointed to Census Tract 001000, Block 1007, which encompasses part of the city of Sarasota. Spitzer’s data showed 236.345 residents in that block, but no adults or minorities.
Doty told the News Leader in the Oct. 14 email that he complied all the estimates for Spitzer “in about a week. Our scope and timeline precluded reviewing 11,000+ blocks,” he added, referring to the number in the county.
The County Commission voted 4-0 on May 22 to redraw district lines before the end of this year, as allowed by state law. (Commissioner Christian Ziegler was absent at the time.) It was not until the latter part of June that county staff hired Kurt Spitzer and Associates to prepare the population data. Chair Charles Hines has talked about the tight time constraint for the process.
Anderson also indicated in the Sept. 29 Herald-Tribune article that Doty had communicated to county leaders that it was so difficult to accurately estimate population at the Census block — or neighborhood level — that they should wait for the 2020 Census data to undertake redistricting.
When the News Leader asked Doty whether he had indeed cautioned Sarasota County Government staff that he did not have sufficient time to compile data with the level of detail that he felt would be appropriate for a redistricting initiative, Doty responded, “No, not at all. But I have said that, based on the fact that some of the data we used was at city and county levels distributed to blocks, that there would be blocks that may be higher or lower than what they should be. We do not have accurate data at the block level to determine changes in occupancy, race or ethnicity since 2010, so we have to make some assumptions in how to apply these changes at the block level. There was clearly not enough time to [provide quality assurance for] these data block by block.”
Push and pull
Since the Sept. 29 article appeared, a number of county residents have become more outspoken about the need for the County Commission to pull back from its redistricting initiative this year. Yet, commissioners have continued to contend that the passage in November 2018 of the Single-Member Districts Charter amendment makes it imperative that they redraw district lines before the 2020 election, when three board seats will be open.
The Single-Member Districts amendment requires that a citizen vote only for a commission candidate who lives in the same district as the citizen. Previously, a person has been able to vote for a candidate in each commission race on a ballot in the General Election.
Thus, commissioners say, the district population counts should be as equal as possible, citing the one-person, one-vote standard set by court cases.
However, opponents of redistricting argue that the commissioners are pursuing redistricting only to protect Commissioners Nancy Detert and Michael Moran, whose seats will be open in 2020. Chair Charles Hines is term-limited, so that seat also will be open.
Residents also have complained that the board members simply are trying to prevent election of a Democrat to the commission, as all the board members are Republicans.
The difficulty in the details
On Oct. 7, commissioners asked Spitzer about the assertions of inaccuracies in the data cited in the Herald-Tribune article. They had not heard at that point from Collins, who was among the last of the 10 people to address them that day.
Spitzer explained to them, “As you get higher and higher in terms of the number of blocks, the data becomes more and more accurate. … There might be problems at the block level from different blocks to blocks.”
In his responses to the News Leader’s questions, Doty of BEBR also offered the following details on Oct. 14 in regard to his work:
- “We believe that BEBR’s population estimates at city and county levels are the most accurate estimates available. They are the official estimates for the State of Florida, and every jurisdiction in Sarasota County has agreed that BEBR’s latest estimates are accurate. Based on past performance against decennial census counts, BEBR’s estimates have been historically more accurate than estimates made by the Census Bureau itself. These estimates capture changes in occupancy since the 2010 Census, which of course was conducted in the midst of the Great Recession. It is very important for Sarasota County because that change in occupancy associated with housing units that were built prior to the 2010 Census represent a significant portion of the growth since 2010. “Because this data is only available at city and county levels, we distributed that change in occupancy to the blocks within each jurisdiction so that each jurisdiction matched its 2018 BEBR estimate. For some blocks this may overstate what the likely 2018 block population should be, and for others it may understate it, but it is reasonably accurate at the city, county and district levels. For most of the 11,000+ census blocks in Sarasota County, it should provide a reasonable estimate of population. For some that did not experience an increase in occupancy, or that may have experienced a disproportionately large increase in occupancy, it may not be as accurate.
- “Because much of the data we use is available only at city and county levels, distributing to areas as small as a city block will not always be accurate at that very granular level of geography. The census block level of geography is VERY small, and there are over 11,000 census blocks in Sarasota County. You have to keep that in mind when you talk about evaluating individual blocks. Blocks also contains ‘noise’ (error) intentionally introduced by the Census Bureau to safeguard the confidentiality of the respondents. This is the level of geography requested for this project, but the average error is lower for larger geographic units.
- “Although the new residential units built [— as provided by the Sarasota County Property Appraiser’s Office —] could be summarized by census block, the increase in unit occupancy since the 2010 Census and the changing proportion of Hispanic and African American residents were only available at the city and county levels, respectively. We believe those data are accurate at those levels of geography, but we are distributing those data to tiny individual blocks. There were bound to be blocks with issues, but at the city and district levels, the results should be reasonably accurate. Short of spending months pouring over the 11,000+ blocks with local planners, we believe this is was the best approach to capture city and county level change that was not available at smaller levels of geography.”