Editorials
By Matthys Van Leeuwen, who holds a PH.D in Statistical Analysis
In my previous article, "First Drop Predicts Winners” (Predicting Election Outcomes: A Statistical Lens on King County's Vote Count), I analyzed the initial results from Election Day under the assumption that polls had closed and no further ballots would arrive.
Of course, I was well aware of Washington State's mail-in ballot process, which many view as undemocratic: Millions of ballots continue to be accepted and counted for days—or even weeks—after the first returns are released.
Defenders argue that these ballots must be postmarked by Election Day (November 4, 2025), with the postmark serving as proof of timeliness. Yet, most democracies worldwide enforce two core principles: voter ID requirements and a strict cutoff for ballot acceptance after polls close.
Exceptions, such as for overseas voters, typically require ballots to be mailed well in advance. As an outsider observing this process, I've found it fascinating—and frustrating. Traditional scientific tools and statistical models become largely irrelevant.
In a standard democracy featuring voter ID, paper ballots, and in-person voting, exit polls provide a reliable snapshot: Pollsters sample voters as they leave, apply statistical sampling techniques, and deliver results that are accurate within a margin of error.
Typically, 99% of votes are tallied and certified within eight hours. But in Washington State's mail-in system? Expect daily updates as ballots trickle in, with counts of processed and pending envelopes stacking up like a never-ending backlog.
Imagine being a candidate, glued to your screen, heart pounding with each refresh. The initial Election Day returns—often from early and in-person voters—set the tone: Are you leading? By how much? For moderate or conservative-leaning candidates in King County, a strong start can feel like a false dawn.
I've noticed a chilling pattern over the years, one that should alarm anyone who values democratic integrity, especially in this left-leaning stronghold. Non-Progressive candidates who jump out to a healthy lead often watch it evaporate as the count drags on. Statistical principles suggest the early returns function like an exit poll: The final tally shouldn't deviate dramatically.
Yet it does—often dramatically. A contact once quipped that "progressives vote late.” I was skeptical; after all, these are the folks chanting "vote blue no matter who" from the campaign's outset. There's scant evidence they're chronic procrastinators. Instead, consider it an empirical observation: As ballot batches accumulate, progressive margins swell. Is it strategic delay? Or something more insidious, like exploiting post-Election Day "dumps" to tilt the scales? The question lingers, unsubstantiated, but persistent.
Take three races I spotlighted from Day 1, where moderate challengers initially outpaced their progressive incumbents. Based on historical trends, I suspected they might not hold on:
Kevin Schilling, vying for the 33rd Legislative District Position 1, opened with a 3% edge over incumbent Edwin Obras but now trails by about 3% (with Obras at 50.57% to Schilling's 47.72%, a 641-vote gap and a 6 point reversal as of November 7).
Sue Ann Hohimer, challenging Angelica M. Alvarez for Highline School District Board Director District No. 2, started strong at 2.7% ahead but is now down by roughly 4% (Alvarez leads 51.75% to 47.66%, by 879 votes). This nearly 7-point flip was particularly stark: Alvarez surged with 571 votes in the third day's batch, followed by another 459 vote shift on day four.
As an incumbent, Alvarez's yard signs often pair with those of fellow board member Joe Van, who is challenged by Katie Kresly for District No. 3. Yet, Van remains narrowly behind (49.59% to Kresly's 49.94%, a razor-thin 75-vote margin). Kresly began 3.7% ahead, but unlike the Alvarez-Hohimer contest, this race has shifted by only 3.5% since Election Day.
Why the disparate momentum? Do Alvarez's connections give her an edge in mobilizing late ballots that Van lacks? Based on similar campaigns, you'd expect parallel gains.
Watching this ballot accumulation unfold isn't a thrilling roller coaster—it's more like flying through turbulence during final approach, praying the plane doesn't crash before touchdown.
Can't we all agree it's time for reform: No more ballots accepted after the initial Election Day results? A cleaner, faster process would restore trust and let science—rather than suspicion—guide our elections.
Members Discussion