Neil Malhotra: How Politicians Change Their Message to Appeal to Constituents

Image

Senators sometimes reveal different faces to different audiences. | Reuters/Jim Young

If you’ve ever thought politicians have a special talent for telling you what you want to hear, you might be more right than you know. But does presenting a different face to different people ultimately help influence voters?

In short, yes, according to Stanford professor of political economy Neil Malhotra.

In new research, he found that politicians can convince voters that they are on the same side of an issue, even if the reality might be quite the opposite.

“We wondered how legislators actually respond to people who take completely different positions, especially on tough issues where their district might be divided,” Malhotra says. “Maybe you’d expect them to hide their votes, but they don’t. The nuance we find is that they either compensate or reinforce their positions by bringing in unrelated things. And it’s really an effective strategy.”

Bags of Letters

Scholars have suspected that legislators engage in targeted messaging when explaining their actions to constituents, but there has been scant hard evidence, Malhotra notes. Malhotra, along with Christian R. Grose of the University of Southern California and Robert Parks Van Houweling of the University of California, Berkeley, wanted to get a better picture of exactly how these expert communicators varied their messages, especially on tough issues where there might not be clear-cut divides in their constituencies.

In early 2007, around the time that the Senate was debating the Comprehensive Immigration Reform Act, the researchers organized an effort to send each U.S. senator two different letters. One came from a constituent in favor of immigration reform and the other from a constituent opposing immigration reform.

The researchers then compared letters the senators sent in response. They found that the senators were just as likely to respond to someone who disagreed with their position as they were to someone who agreed. In fact, the senators were generally open about discussing their positions and how they voted to constituents on both sides of the issue.

The most salient finding, however, confirmed the researchers’ suspicions: While some senators sent essentially the same letter to both constituents, the majority strategically crafted their messages to include secondary details that suggested agreement, whether the senator held the same views as the constituent or not. That is, a senator might tell a fellow pro-immigration constituent about an amendment she introduced that would allow more green cards for parents of U.S. citizens, but mention to an anti-immigration constituent that she voted for an amendment that would provide more money for border security.

“Thus, without being openly dishonest,” the researchers write, “senators nevertheless convey different faces to different audiences.”

Text Equivalent View Larger

Image

The Impact of Tailored Messaging

The researchers next devised a complementary experiment to test the effectiveness of the legislators’ messages.

Subjects were asked to read either a tailored or an untailored letter, and then were asked whether they agreed with a senator’s position on immigration and how likely they would be to vote for that senator. They were also asked to gauge their overall favorability and, critically, to recall the senator’s vote on the immigration bill, a piece of information that every senator openly discussed in his or her letters.

This last piece provided the most surprising results. Nearly everyone who read untargeted messages identified accurately how senators voted on the immigration legislation. On the other hand, the targeted messages cut this accuracy nearly in half.

In short, by tailoring their messages with compensating language, senators could convince people that they took an opposite position from the one they explicitly stated earlier in the letter.

“I’ve never really seen data this strong before in my life,” says Malhotra, about the readers’ inability to accurately place a senator’s vote on the immigration legislation. “Just by introducing language on other issues like border control, H-1B visas, or STEM [science, technology, engineering and mathematics] graduates, you basically get it down to a coin flip.”

Malhotra suspects these results might explain to some degree how members of Congress can grow more polarized while their constituents remain in the middle.

“These people are good strategic communicators who can potentially take very extreme positions that are out of step with their constituents but then massage them with language,” Malhotra says. “Hopefully our study will encourage others to look at this explanation for polarization as well.”

Looking at the entire legislative apparatus through the lens of these findings can also be clarifying. Think about senators who vote for a bill but then vote against ending a filibuster on that same bill. Now they have the ammunition to make themselves look good to voters on both sides of the issue. “I think congressional procedure reflects a lot of this stuff,” Malhotra says. “You set everything up to help your members take the votes they need to go back home and explain them the right way.”

Better Tailoring with Technology

Malhotra also sees the enormous potential of this message-tailoring technique combined with the sophisticated voter targeting that was so effective for the Obama team in 2012 and is now becoming a common campaign tool.

Political organizations are mining vast troves of data on individual voters — not just their party affiliations but also their likelihood to vote and their stances on specific issues. Meanwhile, advances are being made in microtargeting technology, such as the ability to communicate directly to individual voters through social media or the nascent industry targeting television commercials to specific households.

“The technology is probably not good enough right now, but the reason why companies like Facebook and Google are so valuable is because of the expectation that you can learn a lot about individual people and target really effectively to them,” Malhotra says. “All the senators were given was their position on one issue. Imagine that they are given all the data. Think of how much better the message can be crafted with all that information. This is something that real companies are into, convincing people based on information you have about them. And I would say the politicians are in the same business.”

But beyond politicians, this research could help journalists, whose job is often to explain to voters contradictions in politicians’ voting records, for example.

“I think this might help contextualize some of the stuff they are seeing in Congress — like why this amendment is being proposed, or why this person keeps getting re-elected despite their positions,” Malhotra says. “Informing people about this effective device and getting it out into the open is healthier for democracy.”