🚨 New paper (with Kasey Rhee & Nico Studen). We use a new within-precinct design to isolate how ideology affects vote choice holding turnout fixed, analyzing 3.4M precinct observations across state & fed elections (2016-2022). tldr: Ideological moderation affects vote shares, but not by much. 🧵⬇️ - ThreadSky

adambonica.bsky.social • 18 days ago

🚨 New paper (with Kasey Rhee & Nico Studen). We use a new within-precinct design to isolate how ideology affects vote choice holding turnout fixed, analyzing 3.4M precinct observations across state & fed elections (2016-2022).

tldr: Ideological moderation affects vote shares, but not by much. 🧵⬇️

Comments

nharper.bsky.social•18 days ago

📌

wsandholtz.bsky.social•17 days ago

📌

rathshaven.bsky.social•17 days ago

"We argue that inconsistent findings on the electoral penalty stem, in part, from challenges in disentangling its two distinct pathways: how it sways voters (persuasion) and how it affects voter turnout."
So your ideology is so horrid you can't sway voters... disenfranchisement is the only option.

jalapena.bsky.social•17 days ago

📌

americancrisis.bsky.social•17 days ago

📌

trashawaythrow.bsky.social•17 days ago

📌

kimzimm.bsky.social•18 days ago

📌

bradleveck.bsky.social•18 days ago

📌

weshallprevail.bsky.social•17 days ago

SHARE EVERYWHERE!
Find your protests! Make your posters! Take some friends! WE CAN DO THIS!!! It takes all of us! "If we don't fight like hell, we won't have a country anymore"! Sounds eerily familiar doesn't it! TAKE OUR COUNTRY BACK FROM THE CORRUPT OLIGARCHS NOW!

https://www.fiftyfifty.one

wmay.bsky.social•17 days ago

So, Adam, I think most would say DW-NOMINATE is wrong about AOC+the squad. You seem to be aware, since you include the GGUM scores that were made to address that problem.

But it's not clear how much your composite score fixes that. As a sanity check, can we see where it places the squad?

409frank.bsky.social•10 days ago

Great work.

adambonica.bsky.social•18 days ago

2/ Key finding: The "electoral penalty" for candidate ideology exists but is modest. A standard deviation shift toward moderation yields only a 0.6 percentage point gain in vote share due to persuasion effects.

laraputnam.bsky.social•16 days ago

Reading w/interest! Am I right in understanding that midpoints are contest-specific, not candidate specific? so eg in races below Pres has further right midpoint than state rep, even though Dem state rep is to right of Dem President: the difference in midpoint is shaped by GOP opponent's position—?

svateboje.bsky.social•16 days ago

i have no real business butting in, but it seems like the midpoint is a just that between the two candidates in the contest. so, yes each midpoint is shaped by the GOP oppo. for state rep, D shifts right wrt to prez contest, but R shifts left. but maybe i am not following your question.

laraputnam.bsky.social•16 days ago

My understanding matches yours, but therefore I'm confused by the use of the term "candidate midpoint" eg in figure title below, & by discussion framed as reflecting "candidate ideology" or "candidate's ideological extremity" rather than something like "contest's ideological terrain"

laraputnam.bsky.social•16 days ago

See my high tech effort to think this thru👇 Given very interesting dataset they have compiled, if they seek to understand importance of candidate moderation/extremism they could have used eg absolute value of distance from precinct (or national) candidate median. I'm trying to understand that choice

svateboje.bsky.social•15 days ago

long way to go but getting a better grasp on what the paper is doing.

i am not confident w comparing aggregated spatial distances over time spans of decades (more so w added DIME stuff).

set that aside. is house race a good proxy for that local ballot-wide picture?

if so, why avg mid pt? +

svateboje.bsky.social•16 days ago

ah. i am following you now.

rmkubinec.bsky.social•17 days ago

Cool paper! I'd like to see some models without the controls as there are some concerns about colliding, i.e., the controls causing each other (is third party entrance endogenous to fundraising totals, for example)

danhicks.bsky.social•18 days ago

98% R^2? Are the precinct fixed effects themselves doing most of the predictive work?

adambonica.bsky.social•18 days ago

3/ To put this in perspective: If a Democrat shifted from AOC-level progressive to Manchin-level moderate they'd gain only ~1 pct point via persuasion—about half the effect of incumbency advantage. This doesn’t account for turnout effects, which appear to be more consequential (more on that soon!).

teerup.bsky.social•18 days ago

Somehow that has become the frequent behavioral effect of association with the maga leader. Even though it remains fairly common among those without a spine, it should never be deemed as normal.

adambonica.bsky.social•18 days ago

4/ Our within-precinct design has several advantages:

1️⃣ Isolates persuasion from turnout effects by examining how a fixed electorate responds to different ideological configs

2️⃣ Controls for unobserved precinct-level characteristics and ballot composition effects

3️⃣ Estimating contextual effects

adambonica.bsky.social•18 days ago

5/ To illustrate, voters in Tinicum PA saw the same 6 contests on their ballot, each with different sets of candidates, creating natural variation in the "treatment." This allows us to observe how the same group of voters responds to different ideological matchups while holding turnout fixed.

adambonica.bsky.social•18 days ago

6/ We also introduce a new composite ideology measure that combines multiple established metrics (CFscores, DW-NOMINATE, DW-DIME, Hall-Snyder, GGUM, Shor-McCarty) using multiple over-imputation. This measure strongly correlates with existing measures of ideology, both within and across party.

adambonica.bsky.social•18 days ago

7/ Since our within-precinct design enables us to examine heterogeneity across contexts, we can directly test how the ideology-vote relationship varies by office type, incumbency status, and precinct partisan composition.

Comments

Posting Rules

Reply