California’s 9th congressional district wasn’t on anyone’s radar last cycle—and for good reason.
Outspent eight-to-one, no one gave former U.S. Marshall Tony Amador a chance of defeating four-term Rep. Jerry McNerney (D). A weak candidate on paper, Amador raised less than $62,000. He lost by 23 points in the June primary and faced a popular, battle tested Central Valley incumbent who had raised more than
$1 million dollars. Consequently, not a penny from the National Republican Congressional Committee (NRCC), and less than $64,000 in total outside support, flowed to a candidate who ultimately lost by five points.
Amador wasn’t alone. Nine House races in California were decided by nine points or less. Buoyed by lower overall voter turnout, these GOP congressional candidates performed, on average, two percent better than in 2012. The two candidates who received the overwhelming majority of campaign resources did worse.
“Congressional Republicans had a chance to win nine Democratic-held House seats in California and blew every one of them,” wrote Tony Quinn, a veteran California political analyst, three weeks after the midterms. “That they failed in every single congressional race is a testament to the lack of knowledge of the nuances of California on the part of the national party.”
Ouch. I’m not as down on every aspect of the national effort as Quinn. Congressional Republicans don’t really have an easy job of it. While it’s clear there’s a problem, I believe it can be fixed. The solution starts with polling.
Poor survey practices are hurting up-and-coming California Republicans. From San Bernardino to Santa Barbara, seven California Republicans were essentially written off. Those decisions were made based on flawed surveys that committed some cardinal methodological errors. We need to get back to proven methodology by dialing cellphones and not putting a cap on the percentage of completes on cells, stratifying samples, calling in foreign languages, using the voter file, especially in California, wherein the file is robust. If a survey violates any of the above rules, the results shouldn’t be used to make targeting decisions. But polling was just one of the reasons why Republicans failed to pick up House seats in California last cycle.
The NRCC spent real money on only two California races. The committee concentrated 93 percent of its California expenditures on unseating first-term Reps. Ami Bera in CA-7 and Scott Peters in CA-52, both of whom won by 2 points in 2012.
A committee spokesperson recently told The Wall Street Journal that the Bera and Peters races were close and that the committee didn’t invest in the other seven toss-ups because of limited resources, and because “the polling wasn’t there.”
That’s not accurate. The Paso Robles Daily News and Santa Barbara Independent wrote up our mid-October poll in CA-24, which showed our client, Republican Chris Mitchum, leading eight-term incumbent Rep. Lois Capps (D). Our survey, conducted Oct. 15-16, had Mitchum up 41.5 percent to 40.5 percent. Capps’ campaign knew this, too, which explains their effective last-minute shift to aggressive tactics.
Of one particularly misleading ad, which quoted Mitchum as saying “I do not intend to go to Washington to represent the 24th district,” a local newspaper noted, “Capps would not be pulling out such hard-hitting material that uses creative editing to make her point if she didn’t believe she needed to.” Outspent six-to-one, we lacked adequate funds to respond. Election Day results slightly favored Capps and were enough to give her a four-point overall victory.
Our firm polled for several of the Republicans that would be in Congress today had they received even minimal financial support. The polling was there, and it was spot-on. And that includes turnout modeling. For instance, despite a statewide turnout of 42.2 percent—the lowest for a gubernatorial general election in the state’s history—our turnout model predicted 57 percent turnout in CA-24. Final turnout was 56.3 percent.
Voter turnout modeling: One-size doesn’t fit all
There’s a reason why the proposal to split California into six states has gained millions of supporters. Here in California, we’re seeing major fluctuations in voter behavior, on issues, and in turnout based on different geographic regions of the state. Certainly, turnout in California varied greatly by region. Sierra County featured the highest turnout in the state, 73 percent, which was more than double the lowest voter turnout in Los Angeles County with just 31 percent.
Voter turnout modeling has to account for district-by-district shifts. Bogus models that rely on exit polls or self-identified likely voters can make survey results counterproductive because they negatively impact a campaign. Turnout was down, on average, 36 points in the nine competitive congressional races in California. But, again, there were still big variations between districts. CA-24 featured high turnout of 56 percent, compared to just 32 percent in CA-31.
Forget the state’s history
It’s been 20 years since a House Democrat lost reelection in California. There’s no question that Democratic dominance in California congressional races is linked to a financial advantage. In 2014, the nine Democratic candidates in the Golden State’s competitive races benefited from $40 million in total spending, compared to $25 million for the Republicans. Still, the money has not been so lopsided as to justify two winless decades. Money isn’t the problem. It’s how the money is being spent. Flawed research is causing major errors in spending decisions and resource allocation.
No more polls with self-reported voter history
Pollsters need to change how they conduct surveys. Patrick Ruffini, a co-founder and partner at Echelon Insights, put it succinctly: “Rather than relying solely on self-reported vote history, which can lead to vast overestimates of a voter’s likelihood to cast a ballot, the industry has rightly been moving towards listed samples from voter files. These listed samples should further rely on turnout scores developed by modeling firms, rather than cruder measures like whether someone voted in 2010 and 2012.”
Take account of voter registration trends in the district
From 2012 to 2014, voter registration in California declined by 500,000 voters. GOP voter registration fell faster than Democratic, but the results weren’t uniform across the nine competitive congressional districts. In three districts, CD 24, CD 31 and CD 36, the GOP improved the voter registration gap. None of these districts where Republicans were improving received much consideration nationally, either from the NRCC or outside groups. But CD-26, the third-priority of national groups, saw a 3-point shift in registration favoring Democrats from a +5 to a +8 Democratic district. It was the worst trending seat based on voter registration.
Quality research must be tailored to the region
The Central Valley’s anger over water helped Republican Reps. Jeff Denham and David Valadao beat back top Democratic recruits and cruise to reelection. Their strength should have been an indication to look to other Central Valley districts—no matter how weak the fundraising. In CA-16, dairy farmer Johnny Tacherra came with 1,334 votes of defeating Rep. Jim Costa in spite of a significant financial disadvantage.
California’s top-two primary requires replicating the ballot
California’s top-two primary requires a shift in polling strategies. It’s absolutely paramount that polling firms replicate the ballot by including all of the candidates. In the 18-candidate primary Election for CA-33, our candidate, Elan Carr, was able to vault past the presumptive leaders, former Los Angeles City Controller Wendy Greuel (D) and state Sen. Ted Lieu. We found by testing all the candidates—not just the big names—that everyone was getting lost in the clutter. That discovery allowed us to make a last-minute spending change to concentrate on our base.
Carr won the primary because a relatively small group of voters went into the ballot booth intending to vote for him. We didn’t simply rely on GOP voters finding him among everyone else, because we saw that no one would. Now, we need to have a similar influence on GOP decision makers in Washington.
Justin Wallin is COO of Probolsky Research, a full service public opinion research and strategy firm.