Issue 1 2023 and Sherrod Brown 2018 Precinct Dataset and Preliminary Comparison

posts
Author

John Ray

Published

October 1, 2023

Comparing the 2018 Senate race and the 2023 Issue 1 election suggest that, despite The Discourse, Sherrod Brown probably can’t and shouldn’t simply copy the “No on 1” playbook and, while its clear he holds the right position on abortion in Ohio, this is not the end-all, be-all of his path in 2024

On August 8, 2023, Ohio voted on the ballot measure State Issue 1. The ballot measure would raise the share of the vote required to pass amendments to the Ohio Constitution from fifty percent to sixty percent. The measure was proposed in response to another measure, also but separately known as Issue 1, which would protect the individual’s right to “decisions about abortion, contraception, fertility treatment, miscarriage care, and continuing pregnancy.” Thus, a “No” vote on the August 2023 ballot measure was the pro-abortion position, whereas in November a “Yes” vote would be the pro-abortion position.

The successful defeat of Issue 1 in August was widely seen as a bellwether for Democrats’ future prospects in the state, with one account I like summarizing the thinking on this subject nicely:

“Decent chance that this Issue 1 map will be the blueprint for how Sherrod Brown can win next year.”

In 2018 Senator Brown prevailed in the general election with 53.4 percent of the vote in 2018, and “No on 1” prevailed with 57.1 percent of the vote in 2023.

So lets look at the map. The Ohio Secretary of State has the results of the August 8, 2023 special election for Issue 1 and the Redistricting Data Hub has a precinct-level shapefile and some historical election data. The precinct IDs used by the Secretary of State and by Redistricting Data Hub don’t line up perfectly, but they’re pretty close - on an initial pass, about 60 percent of the precinct IDs were identical across data providers, and I simply lined the rest up with a combination of regex’ing and manual checks. Unfortunately, that roughly 40 percent of precincts amounts to about 3,500 precincts whose IDs I had to line up myself, with some help from my friends grepl(), gsub() and, of course, bourbon. (Did I mention I have a tip jar?)

Lets start with No on 1 itself. To the surprise of no one, “No on 1” performed best in a few key areas of the state, highlighted on the map below. “No on 1” won outright in over 80 percent of precincts in the counties of Athens (which contains the town of Athens), Cuyahoga (Cleveland), Franklin (Columbus), Hamilton (Cincinnati), Lorain (Elyna), Lucas (Toledo), Mahoning (Youngstown), Summit (Akron), and Trumbull (Warren). Indeed, in Cuyahoga, Summit, Franklin, and Lucas counties, “No on 1” prevailed in over 95 percent of precincts, including in virtually 100 percent of the “cities proper” within each county. The map also clearly shows “No on 1” wins in various county seats across the state, from Bowling Green to Springfield to Zainesville to Canton.

Lets not ignore the obvious: This map is extremely similar to Sherrod Brown’s map in 2018. The exact same list of big wins applies with a difference of only a few percentage points across Issue 1 and Sherrod Brown’s 2018 Senate race. This shouldn’t be surprising considering “No on 1” 2023 overperformed Brown 2018 by only about 4 percentage points. But while there are no counties in Ohio where Brown won over 95 percent of precincts (he gets so close - so close! - in Cuyahoga, where he won 94.45% of precincts in 2018), Brown handily won the “cities proper” in the exact same places where “No on 1” prevailed as well, taking the majority in virtually every precinct in every major Ohio city in 2018.

And maps are misleading for the reason you know I’m about to state: Despite what you see here, both Sherrod Brown in 2018 and “No on 1” in 2023 won the majority of precincts in the state of Ohio. Brown won an outright majority in fully 52 percent of precincts. “No on 1” won an outright majority in about 62 percent of precincts.

Maps as a form of data visualization have some severe flaws, but I think the chart below shows us a lot. This map subtracts Brown for Senate from “No on 1” across precincts and I’ve shaded the result to highlight places where either Brown for Senate outperformed “No on 1” by over 5 percentage points, where the two votes basically tied, and where “No on 1” outperformed Brown for Senate by over 5 percentage points.

This map shows large patches of “No on 1” overperformance in key parts of the state - a corridor basically stretching from Cincinnati through the suburban ring around but not totally within Columbus up through the suburban ring around Cleveland.

Indeed, plotting “No on 1” overperformance against various Ohio precinct attributes we see, for example, that “No on 1”’s overperformance of Sherrod Brown 2018 is generally higher in denser precincts. But again, lets make sure not to miss the forest for the trees: Overall, “No on 1” did indeed outperform Sherrod Brown 2018 in about two-thirds of Ohio precincts (about 67 percent, or 5,800 of Ohio’s precincts). But the magnitude of that overperformance tends to be slightly higher in denser, more urban parts of the state.

Overall, its hard to miss that “No on 1”’s overperformance of Brown for Senate is higher in the cities. This chart shows the share of precincts broken out by relative performance of each across Ohio’s counties, sorted left to right from least to most populous. In the most populated counties, “No on 1” significantly outperformed Brown in majorities of precincts. In the less populated counties, Brown significantly overperformed “No on 1” in higher shares of precincts.

At the same time, this chart shows that the candidate and the issue position performed about as well as each other in huge shares of most counties. At the risk of letting the data get in the way of some good discourse, “No on 1”’s 57ish percent against Brown’s 53ish percent is hard not to notice but it is not the end-all, be-all of what’s going to happen in 2024. There is plenty of Ohio where you’d rather be Sherrod Brown than “No on 1.”

I also merged in some Census data to see how these observations held up when accounting for various underlying demographic factors. I pulled down Census data available at a combination of the Census Block Group (race, ethnicity, gender, urban/rural status) and County Sub-part (education, income), along with some variables from the Secretary of State file (voter registration rates, DMA designation, past election data). I won’t lower the quality of a good blog post by subjecting the reader to regression coefficients, but messing around with the data I can get my hands on a few stylized facts prsent themselves:

Sherrod Brown 2018 and No on 1 2023 perform better in denser, more diverse precincts. But the candidate and the issue position differ along many different dimensions and we should treat any conclusions about them with those stark differences in mind. But the data also clearly suggest Sherrod Brown can’t just copy the Issue 1 playbook, contrary to what some hot takes may suggest, and I think rather coarse state of the comparison speaks to this reality. There are plenty of parts of Ohio where Sherrod Brown would rather be Sherrod Brown than be “No on 1”. Those areas are disproportionately but not entirely rural, and rarely diverge from “No on 1” by more than a few percentage points. With that in mind, Democrats continue to have a severe rural deficit, and both the candidate and the issue perform worse in rural areas than in the rest of the state. While Brown tends to outperform in rural areas, there is clear room for improvement for Democrats in rural America.

But this is a lot of data to try and mess with. Why not do so yourself? I’ve posted the precinct-level 2018 and 2023 data here (and before you click, did I mention my tip jar?). This data includes Issue 1 and Sherrod Brown performance for each precinct, as well as various precinct IDs to match them across the several different datasets used here:

Feel free to play around with it and let me know what you think. I’m excited to see how this data could be useful to folks gearing up for the big battles ahead in 2024.