Conflicting Voter Polls Show Different Front Runners For 49th Congressional District
Monday, June 4, 2018
Photo by KPBS Staff
Polls with significantly different results for the 49th Congressional race in San Diego's North County came out in the past week. The results of polls can depend on who paid for them.
Political science professor Carl Luna of Mesa College said voters need to bear in mind who paid for a poll when considering the conclusions.
"You can inadvertently skew results, and they can unintentionally lead people who are being polled to reach conclusions they otherwise might not, if you change the wording. And sometimes, with push polls you can try to lead the voter where you want them to go," Luna said.
"Which is why partisan polls, polls commissioned by a campaign, have to be taken with a grain of salt because they seem to fall out of line with polls done by nonpartisan organisations,"Luna said.
For example, two polls on the 49th Congressional race appeared on June 1, less than a week before Tuesday’s primary election. The first poll, published in Roll Call, was a Democratic telephone poll of 500 likely voters between May 22-24. It was carried out by Tulchin Research and had a margin of error of 4 percent.
That poll placed Democrat Mike Levin in the lead with 17 percent, followed by Republican Diane Harkey with 15 percent. Democrats Doug Applegate, Sara Jacobs and Republican Rocky Chavez came in virtually tied in third place with 12 percent and 11 percent.
The second poll, published online a few hours later by The San Diego Union-Tribune and Channel 10 News, surveyed 700 likely voters from May 29-31. It was carried out by SurveyUSA, and had a margin of error of 4.7 percent.
In that poll, Harkey led with 24 percent, with three Democrats, Applegate, Jacobs and Levin virtually tied for second place with 11 and 10 percent. Republicans Chavez and Kristin Gaspar came in tied for fifth place with 8 percent each.
Jelena Bradic, assistant professor of mathematics at UCSD, said the most likely explanation for such diverging statistical results is what she calls a "biased sample."
"Ultimately it’s all about how the data was gathered and who was included in the data," Bradic said. "Typically, what happens when you observe such large differences, that the polls were gathering what we would call 'subpopulations' rather than the whole population of interests."
Unlike the Republican Party, which endorsed Harkey, the Democratic Party failed to endorse any one candidate. Levin won the most votes, 53 percent, from an endorsement caucus at the Democratic Party Convention in San Diego in February, but he missed the 60 percent threshold needed for a chance at the party’s official endorsement.
While the front-runner is different in the two polls, both polls suggest one Democrat and one Republican may make it through to a runoff on the November ballot. This is important since under California’s open, or “jungle” primary, it is possible that two Republicans or two Democrats could be the top two vote-getters, shutting out the other party in November. Voters can vote for any candidate regardless of party in an open primary.
Both polls put Harkey in one of the top two spots, while Chavez, who was leading in the polls in early May, is not. What is less clear is which Democrat could win enough votes to make the cut.
Luna said voters need to consider the source of all polls, and the margin of error, when considering the conclusions.
"Vote for the candidate that you think will do the best job for you," Luna said. "Polls can be wrong and the purpose of voting is to reflect your intent as a voter."
To view PDF documents, Download Acrobat Reader.