INSTITUTE FOR POLICY RESEARCH

With the 2018 midterms approaching and candidates seeking to appeal to voters, do politicians really know what their constituents want?

In the ongoing National Candidate Study, data scientist Christopher Skovron, an IPR postdoctoral fellow, and his collaborators are seeking to understand what politicians believe about public opinion.

“How does public opinion translate into various political outcomes that we care about?” Skovron asked at an April 12 joint-lecture between IPR and Northwestern’s Masters in Public Policy and Administration (MPPA) program. It was the first talk in the “Practical Policies: Real-World Applications of Policy Theories” series, designed to show MPPA students how the theories they learn about in class relate to real-world research.

“This is a great opportunity for IPR faculty to engage with policy practitioners and soon-to-be practitioners in an effort to ensure evidence-based policymaking,” said IPR associate director James Druckman, a political scientist.

“I think it is important to bring the real-life examples to students outside of what we just learn in the classroom,” said Elizabeth Christian, president of the MPPA Student Leadership Council. “Students are better prepared when they can think about topics from multiple viewpoints, and this partnership helps give our students another tool to do that.”

Skovron and his fellow researchers have been collecting data every two years since 2012, surveying more than 5,000 candidates for state legislatures, as well as 1,000 county party leaders from all 50 states. Many candidates were already sitting state legislators when they were surveyed.

“They’re interesting because many of them are operating with really low resources,” Skovron said, noting many state legislature candidates lack access to polling data. “They’re taking votes on legislation that affects all of our lives.”

The researchers provided candidates with issue statements on topics like same-sex marriage, birth control, healthcare, immigration, guns, and abortion. They then compared what the candidates thought their constituents supported to estimates of the true district opinion.

Skovron and his colleagues find that politicians have both inaccurate and biased perceptions of public opinion in their districts. On average, both Republican and Democratic candidates are overestimating how conservative their districts are.

In the 2014 survey, for example, the Democrats were about 5 percentage points off when it came to guessing the level of support for same-sex marriage among their constituents. In other words, they overestimated conservatism in their constituents.

“The Republicans, on the other hand, got it a lot more wrong,” Skovron said, with Republican candidates about 15–20 percentage points off, overestimating how many of their constituents opposed same-sex marriage. That was true from very conservative districts to very liberal districts. 

These findings were reflected in various other issues too, including support for an assault weapons ban. Republicans overestimated the conservative view on this by 30 percentage points on average, according to Skovron.

So what is happening? While Skovron and his collaborators are still working on this issue, he said one contributing factor might be that though Democratic and Republican constituents contact their representatives evenly, Republicans tend to hear from their own party more often.

Although the Republicans seem to have a poorer understanding of true public opinion in their district, Skovron said that could change, as whichever party is in control might change how we perceive things. 

“When we go back and survey this fall, I’m expecting that there may be some differences,” Skovron said.

Christopher Skovron is an IPR postdoctoral fellow. James Druckman is Payton S. Wild Professor of Political Science and IPR associate director.