How sound data enabled NPI to anticipate a big Democratic victory in the 2022 midterms

One year ago today, I wrote what might just be one of my favorite posts in the his­to­ry of this pub­li­ca­tion: a lengthy elec­toral analy­sis titled: “Why the 2022 midterms in Wash­ing­ton State could look more like 2018 than 2010 or 2014.”

Giv­en that we’ve reached the anniver­sary of that post’s pub­li­ca­tion, it seems like an oppor­tune time to reflect back on it, since it set the stage for the elec­toral analy­sis that NPI offered con­tin­u­ous­ly through­out 2022.

The impe­tus for that Jan­u­ary 28th, 2022 post was a sur­vey con­duct­ed by Elway Research pub­lished by Cross­cut that pur­port­ed to show a “Repub­li­can surge” in Wash­ing­ton State. Our team exam­ined Elway’s data, then com­pared it to our own and that of SurveyUSA/KING5’s. We con­clud­ed that there was no “Repub­li­can surge” and flagged some of the issues with Elway’s sur­vey sam­ples that led us to con­clude they might not be prop­er­ly rep­re­sen­ta­tive of the electorate.

After dis­cussing those issues, I offered the fol­low­ing out­look for the midterms:

We know from decades of research that elec­tions turn on iden­ti­ty and trust. That’s no less true in high­ly polar­ized times. Peo­ple are ulti­mate­ly going to fill in the oval for who they iden­ti­fy with. They won’t be vot­ing on the basis of gas prices, or infla­tion con­cerns, or for­eign pol­i­cy, or some set of issue positions.

Cur­rent events and ide­ol­o­gy do mat­ter and do influ­ence people’s think­ing. They just aren’t the deci­sive fac­tors that dri­ve vot­ing behav­ior. Iden­ti­ty and trust are.

Unless Repub­li­cans can con­vince more Wash­ing­ton vot­ers to trust them, they are going to stay at a dis­ad­van­tage statewide and in a lot of cru­cial­ly impor­tant sub­ur­ban and exur­ban leg­isla­tive dis­tricts. And that would mean that they’re not going to have the kind of pick­ups they had in 2010 or 2014.

Because Democ­rats have most­ly run out of sub­ur­ban and exur­ban pick­up oppor­tu­ni­ties, the par­ty will pri­mar­i­ly be on defense this year, though it will try to go on offense in the redrawn 26th, 10th, 17th, and 42nd Leg­isla­tive Districts.

The stage seems set for a cycle with elec­tion results that will look more like the 2018 midterms than either 2010 or 2014 in Washington.

“The stage seems set for a cycle with elec­tion results that will look more like the 2018 midterms than either 2010 or 2014 in Wash­ing­ton,” I wrote one year ago.

And that’s exact­ly what happened:

• Sen­a­tor Pat­ty Mur­ray was over­whelm­ing­ly reelected
• Rep­re­sen­ta­tive Kim Schri­er secured reelec­tion by a big­ger mar­gin than 2020
• Democ­rats flipped WA-03 with Rep­re­sen­ta­tive Marie Glue­senkamp Perez
• Sec­re­tary of State Steve Hobbs, the first Demo­c­ra­t­ic Sec­re­tary of State in decades, was retained in his posi­tion by Washingtonians
• Vot­ers sent a twen­ty-nine mem­ber Sen­ate Demo­c­ra­t­ic major­i­ty to Olympia, up from twen­ty-eight in 2018 and 2020
• Vot­ers sent a fifty-eight mem­ber House Demo­c­ra­t­ic major­i­ty to Olympia, up from fifty-sev­en in 2018 and 2020

Democ­rats’ leg­isla­tive gains (Sharon Shew­make’s Sen­ate vic­to­ry, Clyde Shavers’ House vic­to­ry) came in the redrawn 42nd and 10th Leg­isla­tive Dis­tricts, two of the four tar­get­ed dis­tricts that I men­tioned in my post.

“The stage seems set for a cycle with elec­tion results that will look more like the 2018 midterms than either 2010 or 2014 in Washington.”

Those words sound prophet­ic now, but at at the time, the 2022 leg­isla­tive ses­sion was­n’t even halfway over, the hor­rif­ic leaked Dobbs deci­sion had yet to be hand­ed down by the Supreme Court, and the August Top Two and Novem­ber gen­er­al elec­tions had not been held. It was ear­ly in the year and we had­n’t even put our first statewide poll of 2022 into the field yet — that would hap­pen in February.

What we did have was plen­ty of sound data from our third and fourth quar­ter 2021 sur­veys. We had polled at mul­ti­ple lev­els — statewide and local — and from those sur­veys, we could see that Democ­rats in Wash­ing­ton were in rel­a­tive­ly good shape despite con­stant talk of a “red wave” that got loud­er as the year went on.

The results of the 2021 local elec­tions, which went real­ly well for Democ­rats and pro­gres­sives over­all, also sug­gest­ed that Wash­ing­ton vot­ers could recom­mit to Demo­c­ra­t­ic gov­er­nance in 2022’s leg­isla­tive and con­gres­sion­al elections.

It isn’t pos­si­ble to know the future and we have often remarked here on the Cas­ca­dia Advo­cate that our team does­n’t pos­sess work­ing crys­tal balls. But it is pos­si­ble to keep an open mind about the future and to break free of group­think. It’s some­thing we’ve tried to do here at NPI ever since our incep­tion in 2003.

Way too much polit­i­cal analy­sis nowa­days seems to be gen­er­at­ed by form­ing a con­clu­sion about what’s going to hap­pen and then work­ing back­wards from there to find evi­dence sup­port­ing that con­clu­sion. (Some­times, no evi­dence is offered, just con­jec­ture by those who like to go by hunch­es rather than data.)

Our approach is dif­fer­ent. We begin by rec­og­niz­ing we do not know the future and that elec­tions are decid­ed by vot­ers, not pun­dits. Rather than assum­ing an elec­tion will fol­low some his­tor­i­cal trend (e.g. the pres­i­den­t’s par­ty always los­es seats in a midterm) or rely­ing on what I fond­ly like to call the pre­dic­tions and rat­ings rack­et, we imag­ine as many out­comes and sce­nar­ios as we can.

Our mantra is any­thing is pos­si­ble.

To ascer­tain what is prob­a­ble, plau­si­ble, and like­ly, we put our faith and con­fi­dence in cred­i­ble data. We have been com­mis­sion­ing our own pub­lic opin­ion research at NPI for ten years now, but we also enthu­si­as­ti­cal­ly fol­low the work of oth­er rep­utable non­prof­its and polling firms that are com­mit­ted to the sci­en­tif­ic method like us, and ask­ing neu­tral ques­tions of rep­re­sen­ta­tive samples.

A lot of polling is done at the nation­al lev­el, spon­sored by big media and nation­al polit­i­cal orga­ni­za­tions. Polls at the state lev­el are rar­er, and polls at the local lev­el are much rar­er. Yet polling at those lev­els can be tru­ly invalu­able, illu­mi­nat­ing the think­ing of vot­ers in places where leg­isla­tive majori­ties are made, and enabling explo­rations of local angles on impor­tant issues. It’s not pos­si­ble to prop­er­ly exam­ine local angles on issues in a statewide or nation­al poll. Con­verse­ly, though, nation­al and state lev­el con­cerns can be part of the mix in a local poll.

In ret­ro­spect, NPI’s expan­sion into local polling in 2021 was eas­i­ly one of the best deci­sions that our staff and board have ever made. It enabled us to dive deeply into issues and con­nect with rep­re­sen­ta­tives of large swathes of the elec­torate here in the Pacif­ic North­west on a reg­u­lar basis.

While Repub­li­cans were pump­ing out bom­bast and try­ing to cre­ate media hype, we were focused on tak­ing the pulse of Wash­ing­ton with neu­tral questions.

Here are three exam­ples of indi­ca­tors that guid­ed our think­ing as last year began.

Voters liked Democrats’ work on key issues

We knew as 2022 began that vot­ers in places like Pierce Coun­ty were very sup­port­ive of the police reform bills and the cap­i­tal gains tax on the wealthy that Democ­rats had passed in the 2021 leg­isla­tive ses­sion — because we made an effort to ask them for their opin­ions. Repub­li­cans kept on insist­ing that sub­ur­ban vot­ers were primed to pun­ish Democ­rats in the midterms for these leg­isla­tive accom­plish­ments, but our polling showed the opposite.

• COVID-19 response: 67% of sub­ur­ban and exur­ban vot­ers in one of our local polls told us they sup­port­ed requir­ing patrons of restau­rants, bars, clubs, and winer­ies to pro­vide proof of vac­ci­na­tion as a con­di­tion of entry. In anoth­er of our local polls, a major­i­ty said they agreed mask-wear­ing should be manda­to­ry in all school build­ings and facilities.
• Police reform: Net sup­port for police reform ideas like ban­ning no-knock war­rants and choke­holds exceed­ed thir­ty-five points in our sur­veys, with super­ma­jori­ties express­ing sup­port. Oth­er police reform and account­abil­i­ty ideas got total sup­port of 74%, 73%, 70%, or 66%.
• Build Back Bet­ter: After the House passed Pres­i­dent Biden’s sig­na­ture leg­isla­tive pro­pos­al (parts of which lat­er became the Infla­tion Reduc­tion Act after nego­ti­a­tions with Joe Manchin) we asked vot­ers what they thought of its planks, like low­er­ing the cost of pre­scrip­tion drugs and invest­ing in elec­tric vehi­cle infra­struc­ture. We found a major­i­ty of sub­ur­ban and exur­ban vot­ers sup­port­ive, with over four in ten strong­ly supportive.
• Cap­i­tal gains on the wealthy: Our statewide polling has long found sup­port for a cap­i­tal gains tax on the wealthy amongst like­ly vot­ers. But our local polling showed that sup­port was also robust in Wash­ing­ton’s sub­urbs and exurbs, despite Repub­li­can rhetoric and fearmongering.

Strong suburban approval ratings for Biden, Inslee, Murray

The extent to which pres­i­den­tial approval rat­ings mat­ter or will help/hurt can­di­dates of the pres­i­den­t’s par­ty in an elec­tion is often debated.

Our team thinks that the use­ful­ness of pres­i­den­tial approval rat­ings has waned in com­par­i­son to pre­vi­ous polit­i­cal eras. But it was nev­er­the­less strik­ing to us that our polling was find­ing rock sol­id approval rat­ings for both Pres­i­dent Biden and Gov­er­nor Jay Inslee in sub­ur­ban and exur­ban Wash­ing­ton even as nation­al reporters were breath­less­ly hold­ing up Biden’s falling nation­al approval rat­ings, which saw a sig­nif­i­cant drop after the U.S. with­draw­al from Afghanistan.

In the places where Ever­green State majori­ties are made, we could see that there was plen­ty of sat­is­fac­tion with Demo­c­ra­t­ic gov­er­nance at the fed­er­al and state lev­els. This sug­gest­ed that Repub­li­cans’ expec­ta­tions of 1994 repeat­ing itself were root­ed in wish­ful think­ing, not an under­stand­ing of cur­rent vot­er sentiment.

For exam­ple, in Sep­tem­ber of 2021, we found that in the City of Sam­mamish, once a solid­ly Repub­li­can city and more recent­ly a bat­tle­ground, Pres­i­dent Biden had an approval rat­ing of 75% and Gov­er­nor Inslee had an approval rat­ing of 71%. Mean­while, in the wider East King Coun­ty exurbs (com­mu­ni­ties to the east of Belle­vue and Kirk­land), we found Biden at 61% and Inslee at 63%.

That sup­port con­tin­ued into 2022. In our 2022 King Coun­ty polling, clos­er to the elec­tion, we found sim­i­lar­ly strong num­bers for Democ­rats, with Pat­ty Mur­ray rat­ed the most high­ly of all in our sum­mer 2022 poll of King Coun­ty vot­ers.

A fierce aversion to today’s Republican Party

Our polling found the Repub­li­can brand is in tat­ters in much of Wash­ing­ton State, with many vot­ers express­ing dis­gust over the par­ty’s wor­ship of Don­ald Trump, the refusal to con­cede the Novem­ber 2020 pres­i­den­tial elec­tion, and the Jan­u­ary 6th insur­rec­tion. In one of our sur­veys, we asked Biden vot­ers if there was any­thing the Repub­li­can Par­ty could do to earn their vote in 2022. Two-thirds sim­ply said no upfront, while the remain­ing third offered com­ments like this:

“Actu­al­ly care about peo­ple and ban Don­ald Trump from ever going near a pres­i­den­tial elec­tion again. He caused an upris­ing, seri­ous­ly the fact that he was not com­plete­ly dis­graced by the par­ty is outrageous.”

“Can­di­dates should be fis­cal­ly con­ser­v­a­tive and have mod­er­ate social views. Far right or extreme con­ser­v­a­tive Repub­li­cans is not what the par­ty needs. No more Trump­ism types!”

“Com­plete­ly gut the par­ty, stop tak­ing bribes from NRA and basi­cal­ly reverse your stance on everything.”

“Decide on a plat­form very much like Ike Eisen­how­er’s and get rid of your lunatic fringe (Boe­bert, etc.). Also, actu­al­ly work with oth­ers in Con­gress rather than sim­ply being a do-noth­ing party.”

“Gen­er­al­ly speak­ing, adopt a plat­form as opposed to plat­i­tudes. I was a reg­is­tered Repub­li­can until 2017. The par­ty is focused on ortho­doxy of dis­cus­sion that revolves around cul­ture war as opposed to any­thing pro­duc­tive. The fact that the Par­ty of a Lin­coln is giv­ing tac­it sup­port to those that open­ly talk of seces­sion and vio­lent rev­o­lu­tion would have the man turn­ing in his grave. I want con­ser­vatism back, not this right wing nation­al­ist stuff.”

“Remove your­self from Trump. Denounce all Q‑Anon con­spir­a­cy the­o­ries. Admit that the elec­tion was not stolen. Cen­sor and remove any par­ty mem­bers that par­tic­i­pates in hate speech, fear lin­ger­ing and/or mis­in­for­ma­tion about Covid and Covid vaccines”

Rather than repent­ing in sack­cloth and ash­es (to bor­row the phras­ing of anoth­er of our poll respon­dents!), Repub­li­cans pro­ceed­ed to put up can­di­dates in 2022 like Joe Kent, who cam­paigned on defund­ing the FBI, arrest­ing Dr. Antho­ny Fau­ci, and can­cel­ing U.S. sup­port for Ukraine in favor of Putin’s “rea­son­able” demands.

(Kevin McCarthy even showed up in Wash­ing­ton to stump for Kent and get his pic­ture tak­en with the man, who oust­ed long­time incum­bent Jaime Her­rera Beut­ler but went on to lose to Demo­c­rat Marie Glue­senkamp Perez.)

At the leg­isla­tive lev­el, Repub­li­cans did try to put up a crop of more rea­son­able can­di­dates. But those folks could­n’t escape the stig­ma of the Repub­li­can brand.

The election results vindicate our research — twice

I observed at the end of that Jan­u­ary 28th, 2022 post that polls are snap­shots in time and men­tioned that we would be going back into the field to get fresh data. We did just that. Not once, but many times. We polled statewide, we polled at the con­gres­sion­al lev­el, we polled at the leg­isla­tive lev­el, and the local level.

Those sur­vey projects allowed us to stay in reg­u­lar con­tact with rep­re­sen­ta­tive sam­ples of the elec­torate, ensur­ing that we could reg­u­lar­ly take the pub­lic’s pulse. Repub­li­cans tried all year long to build a real­i­ty dis­tor­tion field to make their dreams of a “red wave” a self-ful­fill­ing prophe­cy. We were onto their game, and urged reporters not to be tak­en in every chance that we got.

On Top Two Elec­tion Night in August of 2022, the Seat­tle Times’ Dan­ny West­neat rec­og­nized our efforts. These were the open­ing para­graphs of his col­umn:

So much for that rumored big red con­ser­v­a­tive wave.

So much also for the con­spir­a­cy the­o­rists, the elec­tion deniers (most of them, any­way), and the MAGA right-wingers.

All of these things were not far­ing well, at all, in Tuesday’s vote count in the Wash­ing­ton pri­ma­ry. Over­all, vot­ers in this state seemed to be repu­di­at­ing the con­ven­tion­al wis­dom that this would be the first good year for Repub­li­cans around here since 2014.

“Repub­li­can nar­ra­tives have been bust­ed,” tweet­ed the North­west Pro­gres­sive Institute’s Andrew Vil­leneuve, who had been insist­ing for months that local polling did not back the media-fueled notion that there would be back­lash in favor of con­ser­v­a­tives in this state.

Despite their extreme­ly dis­ap­point­ing per­for­mance in Round One, local Repub­li­cans held fast to their dreams of a big vic­to­ry in November.

The ink was­n’t even dry on the returns of the Top Two elec­tion when Repub­li­cans recom­mit­ted them­selves to build­ing their real­i­ty dis­tor­tion field and pitch­ing their “red wave” nar­ra­tive. By late Octo­ber, they were declar­ing that Tiffany Smi­ley had caught up to Pat­ty Mur­ray in the U.S. Sen­ate race and pitch­ing reporters (often suc­cess­ful­ly) to file sto­ries on how the con­test was sup­pos­ed­ly “tight­en­ing.”

Except, of course, it was­n’t. The Repub­li­can-com­mis­sioned polls that showed Smi­ley right behind Mur­ray or tied with her were garbage.

In a series of posts here, I explained over and over again that every cred­i­ble sur­vey (includ­ing our own polling) showed the con­test was­n’t close and reit­er­at­ed our assess­ment that Pat­ty Mur­ray was on track to be com­fort­ably reelected.

As late as a few hours before the ini­tial results were pub­lished, we were still get­ting the occa­sion­al note from a right wing read­er dis­put­ing our analysis.

But not after Elec­tion Night. After the right wing real­i­ty dis­tor­tion field implod­ed and real­i­ty set in, there was just an incred­i­ble qui­et. The Trafal­gar fans who had expressed an inter­est in return­ing to our com­ment threads to talk about the elec­tion results were nowhere to be found. They all vanished.

This Novem­ber research ret­ro­spec­tive pro­vides an excel­lent run­down of our 2022 elec­toral polling and how it com­pared to the actu­al elec­tion results.

Sound data can benefit everyone

At NPI, we have deeply held pro­gres­sive con­vic­tions. Our core val­ues of empa­thy, mutu­al respon­si­bil­i­ty, and free­dom guide our work. But while our advo­ca­cy is sub­jec­tive, we strive to be objec­tive in our research. The pur­pose of our polling isn’t to gen­er­ate num­bers for use in pitch­ing poli­cies — it’s to find out what peo­ple real­ly think about can­di­date elec­tions, bal­lot mea­sures, and a range of issues.

We know that we can only do that if we fol­low the sci­en­tif­ic method and ask neu­tral ques­tions of rep­re­sen­ta­tive sam­ples. So that’s what we do — every time.

As I men­tioned ear­li­er, it isn’t pos­si­ble to know the future. But it is pos­si­ble to gain a bet­ter under­stand­ing of how an elec­tion may unfold with sound data.

Unlike many orga­ni­za­tions that com­mis­sion polling, we share a sig­nif­i­cant num­ber of our find­ings pub­licly, always accom­pa­nied by analysis.

We do this because we believe that sound data can ben­e­fit every­one — espe­cial­ly in an era of ram­pant dis­in­for­ma­tion and extreme polarization.

We could see that Democ­rats were posi­tioned to do very well in 2022 because we and our poll­sters were reg­u­lar­ly talk­ing to vot­ers. Repub­li­cans weren’t wrong that vot­ers were (and are) con­cerned about issues like infla­tion and crime, but our research showed that the midterms weren’t going to turn on fears that Repub­li­cans were plan­ning on exploit­ing (and tried their utmost to).

Rather, as usu­al, the elec­tion turned on who most peo­ple trust­ed. The Repub­li­can Par­ty lost the trust of a lot of vot­ers by devolv­ing into a val­ues-less cult that wor­ships author­i­tar­i­an­ism and rev­els in its embrace of dou­ble stan­dards. The Demo­c­ra­t­ic Par­ty is the only major par­ty left devot­ed to keep­ing the repub­lic. Vot­ers get this and it’s why they recom­mit­ted to Demo­c­ra­t­ic gov­er­nance in 2022.

Your support makes this all possible

None of what we do at NPI would be pos­si­ble with­out the sup­port of our mem­bers and our spon­sors. Much of our work is fund­ed by small dol­lar con­tri­bu­tions and com­ple­ment­ed by sup­port from part­ners who share our com­mit­ment to sound data. We are incred­i­bly grate­ful to all of the peo­ple and orga­ni­za­tions who pro­vid­ed finan­cial sup­port to us in 2021 and 2022. Thank you! We’ll do our best to keep bring­ing you analy­sis of elec­tions and issues that you can rely on to be rig­or­ous and based on the sci­en­tif­ic method.

If you enjoyed read­ing about our work and would like to become a mem­ber of NPI, please fol­low this link to sign up and join our fel­low­ship of sup­port­ers.