Janz Campaign Poll Shows Nunes Has Reason to Worry

Is Congressman Devin Nunes (R-Tulare) more vulnerable to a challenger in the 2018 election than he has been in past cycles?
According to poll results released by the campaign of Andrew Janz, a Fresno County Deputy District Attorney who is running to unseat Nunes, it could be a highly competitive race.
In a hypothetical matchup between Nunes and a “Democratic opponent,” the poll showed Nunes receiving 50% support, while the unnamed Democrat received 45%.
Compared to the past three general elections in which Nunes received approximately 67%, 72%, and 62% of the vote, these results seem to indicate a steep drop in support in a year when Democratic energy is expected to be high. Even considering the 4.1% margin of error, Nunes’ support appears to be eroding.

Poll: 61% of Independents Favor Democrat

In Janz’ news release, the campaign touts the support of 61% of independents.
In past elections, Nunes has been able to co-opt a fair amount of independent support, usually more than his challengers. A departure from that dynamic in the 2018 election would be interesting, likely signaling a shift in turnout rather than opinion.
Demographic results from the district are also noteworthy.
The generic Democratic opponent garners higher levels of support from Latino voters (70% to Nunes’ 27%) and younger voters (60% of voters 18 to 45 to Nunes’ 34%), according to the campaign poll. As these demographic groups become more prominent within the electorate, Nunes could see his base of support begin to slip.

That effect is not as pronounced in the more northern reaches of Nunes’ district (including Clovis and north Fresno) but that is a topic far too complex for this discussion. However, this type of result will likely become less shocking as the demographic makeup within the district continues to transform.

Looking at Survey Methodology 

I found the information provided by the Janz campaign in the original press release lacking in terms of proving validity and boosting confidence in its accuracy. It was hard to accept such seemingly groundbreaking results without suspecting that there was a bit of tinkering with the sample or with the phrasing of the questions asked of respondents.
However, after reaching out to Janz’s staff, I was provided with a more detailed report showing sampling results. The data improved my own confidence in the accuracy of the campaign poll results. There are a few methodological factors that could be considered favorable to Janz, but not nearly as many as could have been given the scant information provided at first.
Despite taking a fairly representative sample, the poll still weighted the sample slightly toward Democrats. Democratic respondents make up 37% of the sample, but only about 32% of the electorate. This skew can probably be slightly mitigated methodologically by theorizing heightened Democratic registration and turnout in a wave year. However, there are also counterarguments to that theory – namely that midterms are typically slow for Democratic turnout – and, on the whole, it would be a stretch to plan for a 5% jump in their share of turnout even if an increase does materialize.
The sample was, however, representative of the Republican registration levels, meaning the difference that boosted Democratic numbers was drawn from independents (about 2%) and various third parties (about 3%). This improves the validity of the results because Republican measurement is staying consistent.

Another methodological factor that likely benefits Janz was the choice to opt for the “Democratic opponent” phrasing, rather than including his name.
While fairly common practice in the polling industry and not a huge deal, this choice usually results in broader support for the generic candidate than the real candidate would receive because, theoretically, respondents tend to picture their ideal candidate when given that option. Obviously, no candidate is ideal, so this introduces a bit of bias.

Results are Valid, Surprising

So, overall, it can be concluded that these results are essentially accurate, meaning that the race for the 22nd Congressional District could be more competitive than many, including me, had assumed.
The minor flaws of the poll seem to be on par with usual results and do not disqualify its findings.
While Nunes has significantly more fundraising capabilities, name recognition and experience in the field than his challengers, this election will likely prove to be a far greater test to his resources than those won easily in the past.