“Do you think low-income Americans pay too much, too little, or a fair amount in taxes?”
This question was part of a tax quiz we published in February. About 7,000 of you answered four knowledge questions, as well as questions about how fair you thought taxes were. And we told you we'd report back with results.
But here's what we didn't tell you.
This was part of an experiment designed by the Tax Policy Center's Vanessa Williamson. We worked with her to implement the quiz, which you can try at the link below:
With Congress slated to debate tax reform in the coming weeks, it may seem like most people are pretty set in their attitudes on taxes — whether they think low-income people pay too much, or whether high-income people pay too little.
But what this research shows is just how pliable this is for many people. Focusing on certain information — like how some states tax groceries — can change how fair we think taxes are, both for ourselves and for others. And with Vox readers, this was especially true of information about the taxes of low-income people.
Here’s how the quiz worked:
Before you took the quiz, you were randomly assigned into one of five groups
Two of the groups received questions that focused on parts of that tax code that benefit low-income people or hurt low-income people.
Two other groups received the same questions, but about high-income people.
Another group didn’t get any questions, because they were the control group.
Then you were asked to answer some questions about tax fairness.
The fairness questions asked about your attitudes toward the taxes of low-income people, high-income people, and yourself.
What the survey revealed was that many people’s views on taxes were pliable — that a certain subset of people were more likely to answer the fairness questions one way or another, based on what kinds of knowledge questions they were exposed to.
For example, 56 percent of people in the control believed low-income people paid too much in taxes:
But of people who got questions that focused on parts of the tax code that suggest low-income Americans pay a lot in taxes, 68 percent believed low-income people paid too much.
Of the people who got questions that suggested the low-income Americans pay little in taxes, less than half believed the same:
The questions weren't explicitly ideological, either. Rather, they were questions like:
How many states tax groceries?
What do the lowest-income Americans pay in state and local taxes?
In addition, this was a group of people who were not only politically engaged enough to read Vox; it was a group that voluntarily took a tax quiz. (I mean, seriously, who does that?) So they probably knew more about taxes than most Americans, and were arguably more set in their beliefs.
Still, a portion of this group has political attitudes that were pliable enough that they could be shifted with just four questions
Now, it isn't a bad thing to change your mind after being exposed to new information. But if you're cynical, this line of research shows how policymakers can draw voters' attentions toward or away from something to accomplish a desired result.
That said, Williamson sees this differently. We have a very rudimentary understanding of how political knowledge correlates with political attitudes. Just because someone believes everyone should receive affordable health care doesn't mean they support policies that accomplish such a thing.
So understanding how knowledge and attitudes interact is crucial.
"This is a really important question for democratic accountability," Williamson said.
What changed attitudes on taxes: focusing on low-income people
In this survey, people's attitudes on taxes for low-income people were much more pliable than their attitudes about taxes for high-income people.
The data at the top of this story illustrates a 19-point swing based on what part of the tax code the questions focused on:
But no matter what treatment readers received about high-income people, their attitudes didn't change much:
So why was this? The data doesn't help us answer this question, but there are some possible explanations:
One is that the respondents have more firmly held beliefs about how higher-income people are taxed. This might be because Vox readers tend to be higher-income people, or perhaps because they tend to be left-leaning, so they already come in with strongly held beliefs about high earners. This might especially be true given the 2016 election, when Bernie Sanders ran an entire presidential primary campaign around how millionaires and billionaires aren’t paying enough in taxes.
It also affects how we feel about our own taxes
We also asked how fair people thought their own taxes were, and again it was information about low-income Americans that made the biggest difference.
But interestingly, the people who thought their taxes were most fair were those in the control group — people who hadn't answered any questions when they were asked about fairness.
So does that mean ignorance is bliss? That we should keep tax policy invisible to keep people happy?
Not really. Williamson cites the gas tax as an example of a relatively invisible tax, since it's baked into the cost of gas. But polling shows it's one of the less popular taxes.
The better takeaways from this result are:
- Maybe the questions were hard, so people responded more negatively after going through the process of answering them.
- Or maybe it drew people's attention to income inequality, which made them feel uncomfortable.
- Or maybe thinking about tax policy makes people feel worse about the subject.
But even with those possible considerations, she adds, "The possibility that ignorance is bliss is worth some theoretical consideration. In an ideal world, taxpayers would be both informed about and accepting of their tax responsibilities. It may be that there is a trade-off between those two goals."
The role of information in political opinion
In some utopia, you would feed information into a citizen, and she would use it to form a coherent political opinion. And if you feed in new information, the citizen would update their views accordingly.
But that's not how it works, and many democratic thinkers have wrestled with the question of how this kind of unreliable citizenry fits into a functioning democracy.
Still, we don't have a good understanding of how knowledge interacts with attitudes. Williamson points out that things like education level or familiarity with how government works isn't a good predictor of whether people have accurate information on policy.
The example Williamson gives: If you're someone who thinks wealthy people should pay more in taxes, you may not know what kind of policies actually achieve that goal. So you may be okay with a higher sales tax, which is actually a regressive tax that hurts low-income people.
That dissonance has ramifications. It's why politicians can lie to us about a new health care policy, tell us it aligns with our attitudes — and fend off truth as an effective antidote.
In short, democracy need a better understanding of the wires connecting truth and an electorate’s attitudes.
Correction: A previous version of this story said Bernie Sanders ran his presidential campaign in 2017, instead of 2016.