With the U.S. Senate set to take up debate on a new health care bill, Cornell researchers asked a simple question.

Does the American public want former President Barack Obama’s health care law repealed and replaced?

The answer is, it depends on how you ask the question.

The team analyzed hundreds of national opinion polls from March 2010, when Obama signed the Affordable Care Act, through the recent presidential election. They wanted to know whether different wording in survey questions would predict support for “Obamacare.”

Support for the law is significantly higher – by about 9 percentage points – when the survey question explicitly mentions “repeal” or “repeal or replace” as an option, they found. The study was published May 4 in Health Communication.

“Given that ‘repeal and replace’ really has been the mantra of Republican lawmakers, it’s interesting that polls mentioning that term don’t show higher support for getting rid of the law. It actually seems to put people in a mindset where they support the existing law even more,” said co-author Jonathon Schuldt ’04, assistant professor of communication.

Schuldt’s co-authors are Kristen Holl ’16, now at Intel, and Jeff Niederdeppe, associate professor of communication.

Republicans have promised for seven years to repeal the Affordable Care Act, under which around 20 million Americans have gained health care coverage, according to The New York Times.

On May 4, Republicans took a significant step toward keeping that promise. That’s when the House approved legislation that would repeal and replace major parts of the Affordable Care Act. The House bill faces significant uncertainty in the Senate, where some Republican senators have rejected it and vowed they will create a new version of the bill.

For any repeal measure to become law, both the House and the Senate must agree on the language.

The research, which began as Holl’s undergraduate honors thesis, centered on the team’s analysis of 376 survey questions that gauged the general sentiment in the U.S. toward the current health care law.

The researchers found these questions in the iPoll databank, a comprehensive collection of U.S. national public opinion surveys archived by the Roper Center for Public Opinion Research at Cornell. The time frame covers 80 months of the health care law’s implementation period. That spans Obama’s re-election and a number of highly contested political races – including the 2016 presidential campaign – in which the health care law was a key issue, and the high volume of opinion polling that occurs with the national election cycle.

Niederdeppe and his colleagues hypothesize that loss aversion, a well-researched concept in economics and psychology, may account for the law’s greater support on questions that include “repeal” or “repeal and replace.” That is, people generally want to avoid a loss more than they want an equivalent gain, he said.

“When people have a law that has expanded health care options for tens of millions in the U.S., talking about taking it away seems to, if anything, increase people’s support for it,” Niederdeppe said.

In the bigger picture, the research also points out that survey questions, depending on the polling organization, can take radically different forms, Schuldt said.

As every good survey researcher knows, there’s no right way to ask a question, he said. One must critically engage with the way survey questions are asked and the organization that’s asking them to get a clearer understanding of public sentiment.

“The punch line is a simple one,” Schuldt said. “The wording of surveys matters more than we think.”