Sam Wang is at samwang.bsky.social Profile picture
Scientist @Princeton. Unaffiliated voter. Data-based tweeter & frequent lane violator. Substack: https://t.co/63BjUT1yYI Democracy: https://t.co/3MicOP5c7z
6 subscribers
Sep 16 12 tweets 5 min read
With all due respect, I suggest a radical change: instead of viewing polls like a spectator or sport, look at them as a guide to where *you* should act, as a citizen.

No matter what they do today, they will point toward the same actions.

This requires a shift in thinking... In my view, the whole field of poll aggregation has an unhealthy tendency towards oddsmanship. Think of @NateSilver538's background and interests: basically a sports guy, and now we know, a hypergambler.

Instead, let us view polls as citizens. Where can we make a difference?
Aug 21 8 tweets 5 min read
Last night, former First Lady Michelle Obama exhorted Americans to "do something."

You may think your vote only counts if you're in a Presidential swing state. But voters can gain power and strengthen democracy everywhere.

starts with swing states... votemaximizer.org
Image ...and adds many more categories.

Now add places where per-vote power is exceptionally high for another critical question, U.S. Senate control. Princeton Election Consortium voter-power calculations uncover three additional states (shown here in light blue): MT, NE, and FL. Image
Jul 22 10 tweets 3 min read
With the exit of President Biden from the 2024 race, donor money is coming out of the woodwork. There are ways to optimize that with a little math. Here's a simple way to do it, which PEC has been refining since 2008. Here are the core concepts.

First, give where races are close. For the presidency, each vote in Pennsylvania, just a few miles from here, has a thousand-fold more influence than a puny jerseyvote. That's because Pennsylvania is a knife-edge state - and New Jersey isn't.
Apr 3, 2023 14 tweets 6 min read
The most important U.S. election of 2023 is tomorrow. But what happens *after* that?

A dive into Wisconsin state law, and how to overturn the nation's most extreme gerrymander.

My latest, from @princetongerry:
samwang.substack.com/p/sowhat-happe… @ElectoralLab First, the stakes. Wisconsin's legislative map is the most gerrymandered in the nation. No other state comes close.

Partisan bias can be quantified, and the Wisconsin map takes the cake. Statistics from @princetongerry gerrymander.princeton.edu show no other state coming close. Image
Mar 29, 2023 13 tweets 6 min read
Why next week's judicial race in Wisconsin is the most important election in 2023: putting the state's government back in the hands of voters. politico.com/news/2023/03/2… See what I wrote last month about this race. The key is partisan gerrymandering - but also other voting rights and reproductive rights as well.

The two runoff candidates, Janet Protasiewicz and Daniel Kelly, represent opposite sides of these issues: samwang.substack.com/p/the-most-con…
Oct 17, 2022 12 tweets 5 min read
Democracy Moneyball, Part 3: democracy itself.

In the third and last installment of Democracy Moneyball, I consider the most unusual threat of 2022, that of election denialism.

Here I quantify voters' ability by state to mitigate interference in 2024.

samwang.substack.com/p/democracy-mo… The idea of nonpartisan election administration has been taken for granted. But since 2000, disputes over close elections have become increasingly acrimonious. Bush v. Gore was just the tip of the iceberg - and at least led to the bipartisan Help America Vote Act.
Oct 13, 2022 12 tweets 5 min read
Demoracy Moneyball, Part 2: The U.S. House.

samwang.substack.com/p/democracy-mo…

Consequent to the Dobbs decision, national opinion is at the point where it's not possible to be sure which party will control the House in 2023.

But where to put one's time and resources?

A brief dive. First, evidence.

Things have been about the same for 2.5 months. Since August, generic congressional polls have hovered at a bare plurality for Democrats. Republicans have a 2-point advantage because of districting, and so are very slightly favored. election.princeton.edu/the-race-for-c…
Sep 1, 2022 9 tweets 3 min read
Alaskan voters weren't confused about ranked-choice voting. On the contrary, they used the new power of their ballot to make a more complex choice than traditional voting would allow.

A deeper dive (with charmingly hand-drawn visualizations): samwang.substack.com/p/lessons-from… In one view, the ideologically in-the-middle candidate, Nick Begich, didn't get elected. But Alaska politics is known for being more complicated than that. There is math for understanding this.
Sep 1, 2022 5 tweets 2 min read
Let’s not get carried away entirely.

60% of first-choice votes were cast for R’s. Counting that, median overperformance is only 0.7% for D’s compared with 2020.

That said, corresponds to a ~ D+5% margin in November…maybe enough to retain control. An estimated D+3% is needed. Anyway, the variability in special election over/underperformance is *huge*. N=5 is way too small to extract conclusions.

Going back to May, pre-Dobbs, we get N=18, better. Gives a median that is again an overperformance of 1 point.

docs.google.com/spreadsheets/d…
Aug 20, 2022 10 tweets 3 min read
Mitch McConnell flags poor candidate quality as dragging down his party's chances at retaking the Senate.

Is he right? A dive into the data...samwang.substack.com/p/will-an-anti… Earlier this year, Democrats seemed highly likely to lose the Senate. In a 50-50 chamber, even one lost seat could make the difference.

Lots of opportunities for that to happen. 46 safe Democratic seats, 43 safe Republican seats, and 11 more in the midrange.
Aug 5, 2022 8 tweets 2 min read
Median, last 3 Senate polls:
PA D+11%
OH D+6%
NV D+6%
AZ D+5% (50th for D's/51st for R's)
GA D+3%
NH D+3-4%*
WI D+2%*
NC R+3%
MO R+6%*
FL R+6%

So for R's to reach threshold to take the Senate, conditions would have to move 5 points in their direction. (e.g. "meta-margin" = 5%) *only one poll in last month. In case of New Hampshire, the median is calculated across the possible Republican nominees.
Mar 28, 2022 11 tweets 4 min read
Paradoxically, good-government advocates should appeal this ruling to the Supreme Court. The reason? It buttresses the old principle that state courts can rein in the legislature - and that principle is at risk. nytimes.com/2022/03/25/us/… State courts have been asked to intervene in partisan redistricting cases in several dozen states, within the longstanding doctrine of checks and balances at the level of state government.
sloglaw.org/post/litigatin…