Analysis | A list checked twice: Unwrapping Bloomington city council’s 1,093 votes over 4 years

Two weeks ago, Bloomington’s elected city clerk, Nicole Bolden, called the roll of the nine city councilmembers for their final vote of the four-year term.

The vote was unanimous: Every councilmember voted in favor of the encomiums that were read aloud for each of the five councilmembers who will not be returning to the council in 2024.

Even the final substantive vote, which was taken on a proposal to sell the 3rd Street police station, turned out to be unanimous. All nine councilmembers voted against it.

Even though the 2020-2023 edition of the Bloomington city council will likely be remembered as one of the most divided in Bloomington’s history, unanimous votes were still easily the most common result.

The B Square kept track of the votes in a public Google Sheet in real time as they were taken, meeting by meeting.

Out of 1,093 roll call votes taken by the council for the four year term, 929, or about 85 percent of them, were unanimous. That includes 8–0, and 7–0 votes where someone was absent or abstained, and other similar patterns where there was a zero in either the “yes” or the “no” column.

Of course, the votes where the council was not unified were more memorable than the split votes, which were often preceded by acrimonious debate.

Also likely to lodge in the local collective political memory is a grouping of the council that accounted for some of the splits—”the five” and “the four.”

For the edition of the city council that will officially conclude its service on Dec. 31, 2023, the breakdown of “the five” that will likely be remembered was: Susan Sandberg; Jim Sims; Sue Sgambelluri; Ron Smith; and Dave Rollo.

Making up “the four” were: Matt Flaherty; Kate Rosenbarger; Isabel Piedmont-Smith; and Steve Volan.

Sgambelluri sometimes voted with “the four,” especially early in the four-year term, which made her a kind of swing vote.

She talked about that role the March 26, 2023 League of Women Voters candidate forum for the Democratic Party’s primary election. Sgambelluri described it like this:

Fairly early on in my tenure on council…you started to see patterns in people’s priorities on council and you started to see a group of four that tended to vote similarly and another group of four that tended to vote similarly. And then there was me, OK. That’s fun. All right? So what’s interesting to me is right about that time I had a colleague, pull me aside privately and say: Sue, you have to choose. You have to choose. And I wish I would have come up with a more elegant response. But my semi-elegant response was: Says who? OK? I reject the notion that we can make assumptions and place our colleagues in boxes and make assumptions about what they will and will not vote on or what they do and do not care about. I think there is always opportunity for discussion and for conversation.

As a purely statistical matter, the divide between “the four” and “the five” has at least some merit—but it’s way more nuanced than a simple 5–4 split.

Analyzing voting patterns

To analyze voting patterns, the approach taken here does not identify general topic areas and “score” councilmembers based on their voting record on different issues.

Instead, the idea is to consider each councilmember’s voting record as a mathematical object—a list of ones, negative ones, and zeros.

This approach does not consider the subject matter of the votes at all. Every vote is included, and they’re all weighted the same. For example, the vote on the purchase of Showers West counts the same as a procedural vote on whether to hold a work session at some future date.

Because the analysis is blind to the content of the votes, it might be considered completely objective. It’s just a mathematical computation. That can be viewed as a strength.

At the same time it is a weakness, because the result tells you something about a statistical pattern of votes, but gives no insight into why the patterns look the way they do.

It’s at least a place to start.

A yes is assigned a “1”; a no a “-1” and non-participation a “0”. Non-participation due to absence or abstention are both assigned a “0.” The distance between those mathematical objects, a kind of voting vector, is computed for each pair of councilmembers.

Having a list of distances between every pair of councilmembers is not especially useful for understanding the big picture.

In the same way, having a list of distances between every pair of nine cities would not be useful as a navigational tool. What’s needed is a way to convert those pairs of distances into a map—a two-dimensional plot with nine cities located on it.

A statistical technique called Multidimensional Scaling is a way to transform the distance pairs into a map. That’s the technique that was used to generate the plots below

One chart includes calculations based on all the roll call votes since the current edition of the city council was sworn into office on Jan. 1, 2020.

The other four charts include just one year.

To discern a 5–4 split in the chart for the council’s final year of service would require some squinting.

Even the chart that summarizes the votes for the whole four-year term does not show a neat divide between two groups.

What is evident from the chart is that Sgambelluri and Sims, neither of which will be returning to the council in 2024, formed a kind of statistical bridge between two other, pretty clearly divided groups.

The two remaining groups are: (1) Flaherty, Rosenbarger, Piedmont-Smith, and Volan; and (2) Rollo, Sandberg, and Smith.

In Group 1, Volan’s voting pattern can easily be distinguished from the other three in the group. In the same way, Rollo’s voting pattern sits apart from the other two in Group 2.

Based on their voting patterns, the three other than Volan in Group 1 show the tightest cluster of any three councilmembers. Those three—Flaherty, Rosenbarger, and Piedmont-Smith—are all returning to the council in 2024.

Looking ahead to 2024

One recent indication that the trio of Flaherty, Rosenbarger, and Piedmont-Smith are comfortable collaborating came at the final meeting of Bloomington’s redevelopment commission (RDC).

Flaherty attended the RDC meeting to advocate for approval of the Showers West construction contracts. Neither Rosenbarger nor Piecmont-Smith attended the meeting. Flaherty spoke on their behalf.

Flaherty said, “I’ll also note that I spoke with my colleagues, councilmembers Rosenbarger and Piedmont-Smith. Both of them were supporting votes in the initial decision to purchase the Showers Building, renovate it, and issue the bonds to support that.”

Flaherty added, “All three of us will be continuing into the new term. We all agree this is an important next step, and that our vote last week [to reject the sale of the police station] does not affect our feelings about that.”

The result of the RDC’s meeting was non-action on the Showers West construction contracts.

Added in 2024 to the mix of Piedmont-Smith (District 1), Rosenbarger (District 2), Rollo (District 4), and Flaherty (at large) will be five councilmembers who did not serve on the council this past term. Joining the four from the soon-to-conclude term will be: Hopi Stosberg (District 3), Shruti Rana (District 5), Sydney Zulich (District 6), Isak Asare (at large) and Andy Ruff (at large).

The somewhat convoluted phrasing “who did not serve on the council this past term” in place of “new councilmembers” is a way of acknowledging that Ruff previously served on the council, from 2000 to 2019.

The first meeting of the next edition of the city council is set for Jan. 10 at 6:30 p.m. For readers who want to follow along with the votes from the very start of this group’s service, here’s a link: Shared Google Sheet for 2024 Bloomington City Council votes.