Less than one hun­dred hours from now, on Tues­day, August 17th, Wash­ing­ton State’s August 2021 Top Two elec­tion will be cer­ti­fied and the results of the qual­i­fy­ing round of this year’s local elec­tions cycle will become final.

With no bal­lots list­ed by King Coun­ty Elec­tions as await­ing tab­u­la­tion, and with almost a month hav­ing now passed since we announced the elec­toral find­ings from our July 2021 sur­vey of the Seat­tle elec­torate, it seems like a good time to revis­it our poll find­ings and exam­ine to what extent they antic­i­pat­ed the out­comes in the actu­al returns tal­lied by King Coun­ty Elections.

Our sur­vey, con­duct­ed by Change Research from July 12th-15th, 2021, end­ed up being the only inde­pen­dent poll with pub­licly released results dur­ing the lead-up to August 3rd, the dead­line to return bal­lots in the elec­tion. While Elway Research did con­duct a poll on behalf of Cross­cut dur­ing that time­frame, it was a statewide poll, not a city poll, and so respon­dents were not asked about Seat­tle races.

At the time we were draw­ing up our 2021 research plans, our team dis­cussed the pos­si­bil­i­ty that there might end up being a total dearth of inde­pen­dent elec­toral polling in Seat­tle this year, espe­cial­ly in the Top Two phase.

We agreed that NPI would step up to ensure that sce­nario would­n’t happen.

Choos­ing to expand our research polling to the local lev­el — build­ing on the statewide polling that we’ve been doing for over half a decade — has been one of the best deci­sions that our orga­ni­za­tion has ever made. I am grate­ful to our staff, our board, and our sup­port­ers for putting us in a posi­tion to answer the call.

As I wrote last month, our region sore­ly needs more cred­i­ble, inde­pen­dent pub­lic opin­ion research. NPI is step­ping up to meet the need. Reli­able data is a very use­ful resource to have in an era of ram­pant mis­in­for­ma­tion and disinformation.

While no poll can pre­dict the future, trust­wor­thy polling can pro­vide a valid snap­shot of what the elec­toral land­scape looks like ahead of an upcom­ing elec­tion, and sug­gest what vot­ers might col­lec­tive­ly decide in that election.

After results roll in, hav­ing trust­wor­thy polling to refer back to gives polit­i­cal observers more con­text to work with when try­ing to inter­pret the data and draw con­clu­sions about what hap­pened and why it happened.

We are now at such a post-elec­tion junc­ture. So let’s dive in and see how the results line up with our polling from last month!

Mayor of Seattle

What the poll found: Of the fif­teen can­di­date field, the poll indi­cat­ed that Bruce Har­rell and Lore­na González were the fron­trun­ners, with 32% (a plu­ral­i­ty) unde­cid­ed. Har­rell placed first in the poll and González sec­ond, with Colleen Echohawk, Jessyn Far­rell, and Andrew Grant Hous­ton trail­ing them.

A visu­al of NPI’s poll find­ing for May­or of Seat­tle, 2021

What hap­pened in the elec­tion: Vot­ers have select­ed Bruce Har­rell and Lore­na González as the Top Two can­di­dates. On Elec­tion Night, Har­rell had a ten point lead over González, a larg­er mar­gin than what he enjoyed in our polling. But in late bal­lots, his lead shrunk and González’s share of the vote grew. The two can­di­dates are now only two points apart, which sug­gests an extreme­ly com­pet­i­tive runoff for May­or in the gen­er­al elec­tion this autumn.

Cur­rent results (final results avail­able after August 17th):

	Bruce Har­rell 69,554 votes	34.02%
	M Lore­na González 65,627 votes	32.1%
	Colleen Echohawk 20,994 votes	10.27%
	Jessyn Far­rell 14,905 votes	7.29%
	Arthur K Langlie 11,362 votes	5.56%
	Casey Sixkiller 6,904 votes	3.38%
	Andrew Grant Houston 5,473 votes	2.68%
	James Don­ald­son 3,212 votes	1.57%
	Lance Ran­dall 2,800 votes	1.37%
	Clin­ton Bliss 1,615 votes	0.79%
	Omari Tahir-Gar­rett 390 votes	0.19%
	Bob­by Tucker 374 votes	0.18%
	Hen­ry C Dennison 346 votes	0.17%
	Stan Lipp­mann 322 votes	0.16%
	Don L Rivers 189 votes	0.09%

Analy­sis: When it came time to make a deci­sion, most vot­ers who were unde­cid­ed went with one of the fron­trun­ners. That’s why both Har­rell and Gon­za­lez are so far out in front of the rest of the pack.

The poll indi­cat­ed Har­rell had the best chance of get­ting through due to hav­ing already gar­nered a strong base of sup­port and sug­gest­ed Gon­za­lez was the most like­ly can­di­date to get the oth­er spot. That’s pre­cise­ly what happened.

At the time the poll returned from the field, Colleen Echohawk was right behind Gon­za­lez, at 10%. Inter­est­ing­ly, that’s the exact same per­cent­age she’s get­ting in the actu­al elec­tion… 10.27%! In ret­ro­spect, it looks like Echohawk peaked around or some­time not long before our poll was in the field. She cer­tain­ly had a chance to con­nect with unde­cid­ed vot­ers, but they end­ed up with the frontrunners.

Jessyn Far­rell and Andrew Grant Hous­ton were the oth­er can­di­dates in the top tier, tied for fourth at 6%, respec­tive­ly. Far­rell did a lit­tle bit bet­ter in the elec­tion, gar­ner­ing 7.29%, while Grant Hous­ton did worse, pulling in 2.68%.

City Attorney

What the poll found: Per­haps no result from our sur­vey turned more heads than our find­ing for Seat­tle City Attor­ney. Our research indi­cat­ed that it was a three-way race between incum­bent Pete Holmes and his two chal­lengers, with Holmes bare­ly ahead and in grave dan­ger of being over­tak­en by both of them. Holmes received 16% sup­port in the sur­vey, while chal­lengers Nicole Thomas-Kennedy and Ann Davi­son each received 14%.

A major­i­ty of respon­dents, 53%, said they were unde­cid­ed, but the low num­ber for Holmes, a three-term incum­bent, indi­cat­ed he was in big trouble.

A visu­al of NPI’s poll find­ing for Seat­tle City Attor­ney, 2021

What hap­pened in the elec­tion: Seat­tle vot­ers end­ed Pete Holmes’ bid for a fourth term by select­ing Thomas-Kennedy and Davi­son as the top two can­di­dates for the posi­tion. Each can­di­date land­ed in the thir­ties in the actu­al elec­tion, vin­di­cat­ing our char­ac­ter­i­za­tion of it being a three-way race. In the ini­tial returns on Elec­tion Night, Davi­son had the top spot, with Holmes in sec­ond and Thomas-Kennedy in third. By Thurs­day after­noon, close to forty-eight hours after Elec­tion Night, Holmes trad­ed places with Thomas-Kennedy, while Davi­son remained in first place. By Fri­day after­noon, Thomas-Kennedy had surged into the top spot, ahead of Davi­son, with Holmes still out of con­tention in third.

Cur­rent results (final results avail­able after August 17th):

	Nicole Thomas-Kennedy 71,234 votes	36.37%
	Ann Davi­son 64,101 votes	32.73%
	Pete Holmes 60,009 votes	30.64%

Analy­sis: As I wrote last week, vot­ers are known for pun­ish­ing com­pla­cen­cy when least expect­ed to. Pete Holmes was over­whelm­ing­ly reelect­ed four years ago with three-fourths of the vote against chal­lenger Scott Lind­say, but this cycle, he could­n’t even man­age to gar­ner half as much sup­port against two opponents.

Pri­or to the release of our poll find­ing, Holmes’ cam­paign had been in what you might call cruise con­trol mode, hav­ing seem­ing­ly not con­sid­ered the pos­si­bil­i­ty of the can­di­date fail­ing to advance to the gen­er­al elec­tion. After the pub­li­ca­tion of our find­ing, Holmes began cam­paign­ing more vig­or­ous­ly, offer­ing cri­tiques of his oppo­nents. But vot­ers dis­sat­is­fied with his per­for­mance backed them any­way, leav­ing him in the same boat Greg Nick­els found him­self in twelve years ago.

City Council, Position #8 (At-Large)

What the poll found: Our research sug­gest­ed that incum­bent Coun­cilmem­ber Tere­sa Mosque­da was on track for reelec­tion against a large field of ten chal­lengers. In the sur­vey, Mosque­da received 26%, with all of her chal­lengers down in the sin­gle dig­its, or at zero. Of the chal­lengers, the poll indi­cat­ed that Kate Mar­tin could have the best chance of end­ing up in the gen­er­al elec­tion along with Mosque­da; Mar­tin received 6%, with no oth­er chal­lenger polling above 2%. A major­i­ty of respon­dents, 55%, said they were undecided.

A visu­al of NPI’s poll find­ing for Seat­tle City Coun­cil Posi­tion #8, 2021

What hap­pened in the elec­tion: Although a huge num­ber of vot­ers pro­fessed them­selves to be unde­cid­ed — in fact, more than in any of the oth­er three races! — most of them even­tu­al­ly decid­ed to back Mosqueda’s reelec­tion.

She cur­rent­ly has 59.36% of the vote, more than any oth­er can­di­date run­ning for city­wide office in Seat­tle this year. How­ev­er, Mosqueda’s oppo­nent in the gen­er­al elec­tion won’t be Kate Mar­tin. In the biggest sur­prise of the Top Two, bridge engi­neer Ken­neth (Ken) Wil­son came out of nowhere to grab the sec­ond place spot, despite hav­ing spent almost no mon­ey and hav­ing been ignored by the mass media. Wil­son cur­rent­ly has 16.21% of the vote, with Mar­tin in third at 11.56%.

Cur­rent results (final results avail­able after August 17th):

	Tere­sa Mosqueda 112,858 votes	59.36%
	Ken­neth Wilson 30,825 votes	16.21%
	Kate Mar­tin 21,969 votes	11.56%
	Paul Felipe Glumaz 10,214 votes	5.37%
	Alexan­der White 2,468 votes	1.3%
	Bob­by Lind­sey Miller 2,433 votes	1.28%
	Jesse James 2,049 votes	1.08%
	Jor­dan Eliz­a­beth Fisher 1,805 votes	0.95%
	George Free­man 1,575 votes	0.83%
	Alex Tsimer­man 960 votes	0.5%
	Bri­an Fahey 884 votes	0.46%

Analy­sis: This was the only race in which our poll did not detect that one of the top two fin­ish­ers was even in con­tention. The sur­vey antic­i­pat­ed Mosque­da might fin­ish in first place (and she has), but it did­n’t reg­is­ter any sig­nif­i­cant degree of sup­port for Ken­neth Wil­son. We think there’s a sim­ple rea­son for that: Ken­neth Wilson’s remark­able surge occurred after the poll had field­ed in mid-July.

Wil­son prob­a­bly ben­e­fit­ed from the neg­a­tive per­cep­tions that a lot of peo­ple in Seat­tle anec­do­tal­ly seem to have of activist Kate Mar­tin, whose “Park My Viaduct” bal­lot mea­sure pro­pos­al fell utter­ly flat with vot­ers four years ago.

Although Mar­tin is out­per­form­ing her stand­ing in the sur­vey, Wil­son was able to eclipse her by offer­ing vot­ers who want­ed an alter­na­tive to Tere­sa Mosque­da a can­di­date they could feel good about sup­port­ing. Mar­tin, metaphor­i­cal­ly speak­ing, has spent years burn­ing bridges with peo­ple across Seat­tle, while Wil­son, lit­er­al­ly speak­ing, has spent years help­ing to get bridges built in and beyond the city.

As men­tioned, Wil­son was in a posi­tion to cap­i­tal­ize on vot­ers’ inter­est in func­tion­ing bridges because he is a bridge engi­neer. He ran on get­ting the high West Seat­tle Bridge reopened to at least one lane of traf­fic in each direc­tion, empha­siz­ing it on his web­site and in his voter’s pam­phlet state­ment. That mes­sage clear­ly res­onat­ed with a sig­nif­i­cant num­ber of voters.

Aside from The Cas­ca­dia Advo­cate and Seat­tle Times colum­nist Dan­ny West­neat, who took the time to talk to the can­di­date, Wilson’s surge has not got­ten much atten­tion from the press. But per­haps that will soon change. And maybe now that Wil­son has got­ten through the elim­i­na­tion round, he’ll make fundrais­ing a pri­or­i­ty so he can afford to make his cam­paign more vis­i­ble to voters.

City Council Position #9 (At-Large)

What the poll found: In the con­test for the city’s oth­er at-large coun­cil posi­tion, which is an open seat this year, our sur­vey found that activist Nikki­ta Oliv­er and small busi­ness own­er Sara Nel­son were the lead­ing can­di­dates. Oliv­er received 26% sup­port in the sur­vey, while Nel­son received 11%. Bri­an­na Thomas, the chief of staff to Lore­na Gon­za­lez, placed third in the sur­vey with 6%.

Half of respon­dents sur­veyed, 50%, said they were undecided.

What hap­pened in the elec­tion: Oliv­er placed first, Nel­son placed sec­ond, and Thomas placed third — the same order as in the poll. Each can­di­date gar­nered more sup­port than in the sur­vey, indi­cat­ing that they all got addi­tion­al sup­port from unde­cid­ed vot­ers. Nel­son did by far the best job of reel­ing in unde­cid­ed vot­ers to her side, surg­ing from just 11% in the sur­vey to 39.49% in the actu­al elec­tion… a huge gain. Oliv­er went from 26% in the sur­vey to 40.16% in the elec­tion, sur­pass­ing Nel­son in late bal­lots (Nel­son held the first place spot on Elec­tion Night). Thomas went from 6% in the sur­vey to 13.42% in the election.

Cur­rent results (final results avail­able after August 17th):

	Nikki­ta Oliver 79,643 votes	40.16%
	Sara Nel­son 78,304 votes	39.49%
	Bri­an­na K Thomas 26,617 votes	13.42%
	Corey Eich­n­er 7,019 votes	3.54%
	Lind­say McHaffie 3,042 votes	1.53%
	Rebec­ca L Williamson 1,645 votes	0.83%
	Xtian Gun­ther 1,403 votes	0.71%

Analy­sis: Nikki­ta Oliv­er was able to build on their already strong base of sup­port to earn the top spot in the gen­er­al elec­tion, while Sara Nel­son’s cam­paign took off and soared in the days lead­ing up to August 3rd. Com­bined, Oliv­er and Nel­son have about eighty per­cent of the vote over­all. They are less than 1,500 votes apart. This race could be incred­i­bly com­pet­i­tive in the Novem­ber gen­er­al election.

On the morn­ing after Elec­tion Night, one Inter­net com­menter argued to NPI that our sur­vey must have been flawed because it did not pre­dict Nel­son’s strong per­for­mance. At the time, Nel­son was ahead of Oliv­er, and this com­menter the­o­rized that the results did­n’t line up because our sam­ple was unrepresentative.

But as we point­ed out to them, their rea­son­ing did­n’t make any sense. As I explained in the intro­duc­tion to this post, polls can’t pre­dict what will hap­pen, they can only sug­gest what might hap­pen. That is an impor­tant distinction.

If our sam­ple had been unrep­re­sen­ta­tive, then the results of the whole sur­vey would have been off, and we would not be see­ing the cor­re­la­tions we’re see­ing between our find­ings and the unof­fi­cial results so far across the dif­fer­ent races.

As with Ken­neth Wil­son, the plau­si­ble expla­na­tion for why Sara Nel­son is doing so much bet­ter in the actu­al elec­tion than in our poll is because her can­di­da­cy caught fire and res­onat­ed with vot­ers dur­ing the home­stretch. Because our poll was in the field before bal­lots dropped, this was­n’t some­thing it could detect. Nev­er­the­less, it did sug­gest that Oliv­er might come in first and Nel­son sec­ond, which is what hap­pened, despite the large num­ber of unde­cid­ed voters.

Final thoughts

About three weeks went by in between when our sur­vey returned from the field and when the first results from the tab­u­la­tion of bal­lots cast by vot­ers became avail­able. Although plu­ral­i­ties and majori­ties of respon­dents to our sur­vey said they were not sure who they were vot­ing for, it was evi­dent that some can­di­dates were in a much stronger posi­tion head­ing into the vot­ing than others.

On August 3rd, we got our first results. While that fresh data was nice to have, it was also some­what incom­plete. KCE data showed that most of the bal­lots cast in Seat­tle were returned in the final hours of the elec­tion, with a major­i­ty of on-time bal­lots hav­ing been sent back in the final twen­ty-four hours. That meant there was a pos­si­bil­i­ty of seri­ous move­ment and some lead changes in the late ballots.

And indeed, that’s what we saw. We now have more com­plete results to pore over, and that data shows that our sur­vey suc­ceed­ed in yield­ing cred­i­ble insights about Seat­tle’s elec­toral land­scape ahead of the Top Two election.

This suc­cess builds on our track record at the statewide level.

In last year’s pres­i­den­tial elec­tion, we were the only orga­ni­za­tion to pub­licly poll up and down the entire statewide bal­lot, ask­ing vot­ers who they were sup­port­ing in each of the can­di­date elec­tions for statewide exec­u­tive posi­tions and State Supreme Court, as well as for Ref­er­en­dum 90, con­cern­ing sex ed. In each con­test, the results lined up with what our poll found, as explained here.

Pri­or to that, our polling for Unit­ed States Sen­ate in 2018 had been on the mon­ey, antic­i­pat­ing Demo­c­ra­t­ic incum­bent Maria Cantwell’s even­tu­al mar­gin of vic­to­ry over Repub­li­can chal­lenger Susan Hutchi­son months in advance.

We are delight­ed to have been able to work with Change Research on this project. Change, which just cel­e­brat­ed its fourth anniver­sary, is a firm that shares our val­ues and com­mit­ment to con­duct­ing rig­or­ous research.

Change did a great job with their mod­el­ing of the elec­torate for this project, ensur­ing that the neu­tral ques­tions that our team at NPI had writ­ten would be asked of a rep­re­sen­ta­tive sam­ple. Those are the two most impor­tant keys to good polling, by the way: neu­tral ques­tions asked of a rep­re­sen­ta­tive sample.

Survey methodology

• Change Research, a Pub­lic Ben­e­fit Cor­po­ra­tion based in Cal­i­for­nia, sur­veyed 617 like­ly August 2021 Top Two elec­tion vot­ers in Seat­tle from Mon­day, July 12th to Thurs­day, July 15th on behalf of the North­west Pro­gres­sive Insti­tute. All respon­dents par­tic­i­pat­ed online.
• Change used tar­get­ed adver­tise­ments on Face­book, tar­get­ed adver­tise­ments on Insta­gram, and text mes­sages sent via the echo19 and/or Scale To Win plat­forms to cell phone num­bers list­ed on the vot­er file for indi­vid­u­als who qual­i­fied for the survey’s sam­ple uni­verse, based on their vot­er file data.
• Regard­less of which of these sources a respon­dent came from, they were direct­ed to a sur­vey host­ed on SurveyMonkey’s web­site. Ads placed on social media tar­get­ed all adults liv­ing in Seat­tle. Those who indi­cat­ed that they were not reg­is­tered to vote were terminated.
• As the sur­vey field­ed, Change used dynam­ic online sam­pling: adjust­ing ad bud­gets, low­er­ing bud­gets for ads tar­get­ing groups that were over­rep­re­sent­ed and rais­ing bud­gets for ads tar­get­ing groups that were under­rep­re­sent­ed, so that the final sam­ple was rough­ly rep­re­sen­ta­tive of the pop­u­la­tion across dif­fer­ent groups.
• The sur­vey was con­duct­ed in Eng­lish, and has a mod­eled mar­gin of error of 4.3% at the 95% con­fi­dence interval.

We’ll be back for the general election!

NPI and Change Research will be join­ing forces again this autumn to sur­vey Seat­tle vot­ers in advance of the Novem­ber 2020 gen­er­al elec­tion. We look for­ward to bring­ing you more cred­i­ble insights about what’s hap­pen­ing in the Emer­ald City before vot­ers make their final deci­sions this autumn.