How a question is phrased can drastically alter the answer

How do you get people to tell the truth? According to new research from Wharton, how one frames a question can have a lot to do with it.

In the paper, “Eliciting the Truth, the Whole Truth, and Nothing But the Truth: The Effect of Question Phrasing on Deception,” Maurice Schweitzer, Wharton professor of operations, information and decisions, discovered that subtle tweaks in the way a question is asked can lead to profoundly different answers. His co-authors are Eric VanEpps, a marketing professor at the University of Utah; Jeremy Yip, a management professor at Georgetown University and visiting scholar at Wharton; and Julia Minson, a professor of public policy at Harvard.

Schweitzer and VanEpps recently discussed their research findings on the Knowledge@Wharton radio show on SiriusXM. (Listen to the full podcast above).

An edited transcript of the conversation follows.

Knowledge@Wharton: Before we dig into the research, where did your interest in this topic come from?

Maurice Schweitzer: The initial idea here was to think about, let’s say, a negotiation or an interview, or it could be any number of different contexts, where there is some information that one party wants to get out of the other. [Divulging] that information might put you at a disadvantage. At an interview I might ask you, do you have other [job] offers? In negotiations, I might ask you a similar question or try to get at how much you value things or what your timing is like.

If we were to look at the existing literature and the existing advice people give, it is to say, “Hey, be sure you ask questions.” And sometimes the advice goes one step further and says, “Make sure you ask lots of questions or ask open-ended questions.” But the advice ends there, and there’s no research before now that has informed our understanding of what kinds of questions should we be asking, what kinds of questions might get us closer to the truth than others. That is the work that we embarked on.

Eric VanEpps: You also hear the expression that there’s no such thing as a stupid question, and that has never rung true to me. I’ve always thought there were better questions and there were worse questions. It was interesting in this research to try to tease that apart and determine what kinds of questions convey your own intelligence, what kinds of questions get the answers you are looking for and which kinds of questions don’t.

Knowledge@Wharton: I would think questions are a standard part of the back-and-forth between people.

Schweitzer: You are exactly right. I think questions are an integral part of every communication we have, where we’re constantly engaged and not just stating information but trying to get somebody’s reaction or learn information from people. In this case, what we found is that the different questions that you ask will elicit very different responses.

Knowledge@Wharton: Can a couple of words here and there make that much of a difference?

VanEpps: Yes. It might be the case if you are interviewing somebody for a job that you want scripted questions because you might not realize the subtle differences in the questions you ask and the impact that might have on the answers. In our studies, we script the questions to make sure that we are carefully looking at the differences that are observed. It might be the kind of thing that a practitioner would want to practice as well, which is script and then be careful about how things are phrased. Because even just a few words — putting things in a negative framing or a positive framing — can make a difference.

“There’s no research before now that has informed our understanding of … what kinds of questions might get us closer to the truth than others.”–Maurice Schweitzer

Knowledge@Wharton: In your research, did you ask some questions where the assumption would be negative, some where the assumption would be positive, and others where there was no assumption?

Schweitzer: That’s the framework that we started with, and we found huge differences in the kinds of responses people would give.

VanEpps: If you think there is a problem and you assume the problem, you presume there is a problem. “What is wrong with this car that you are trying to sell me?” — as opposed to asking — “This car doesn’t have any problems, right?” We find that little flip of framing changes the way people respond. People are much more likely to disclose problems when you presume [there is a] problem. But if you say, there is no problem with this car, right? They go, “The air conditioning works great.” They spin right in to telling you what works really well on the car.

Knowledge@Wharton: How much does the knowledge level of the person asking the question play into the response given?

VanEpps: We think questions help to reveal that knowledge. One way you can demonstrate that knowledge is by the questions you ask. If you don’t have much knowledge going into the situation, we don’t necessarily have specific advice for you. We are assuming that you have some knowledge coming into this to determine which questions you should ask. But if you do have that knowledge, don’t just be general and polite. Go ahead and demonstrate that knowledge by being more assertive.

Schweitzer: One of the key ideas that we are talking about here is that questions not only solicit information, they reveal information. They reveal information about our assumptions, about how much knowledge we have, about how assertive we are going to be. And all of those things feed into the responses that we get.

Knowledge@Wharton: What are some of the practical implications of this research?

Schweitzer: Whether it’s for interviews, whether it’s for negotiations or many other contexts, we should really think much more deliberately and carefully about the questions we ask. We should recognize that if we ask a question, we could be revealing information about us that we really don’t want to reveal. So, asking a question could actually be the wrong thing to do.

Knowledge@Wharton: Eric, did you find situations where that revelation of information was more negative than positive?

VanEpps: We had one study where we manipulated whether we asked about a common behavior. We asked whether participants used work time for personal email, which a lot of people do. We also asked, “Do you use work time for online gaming?” That is certainly a less common question; people didn’t engage in that behavior and said there wasn’t anything to disclose. Then they asked: What was the [person asking the question] thinking?

They rated them more negatively; they thought they were less competent and less warm, so there was kind of a backfiring there. If you are asking that presumptuous kind of question — you presume that they use work time for online gaming — people [believe that] you’re overshooting, you’ve missed the mark here, and [they think,] “I trust you less because of that.”

Knowledge@Wharton: It sounds like the activity that you are asking about also affects the person’s response in terms of how much information they want offer.

VanEpps: Now you are focusing them on online gaming, and you didn’t get to any of the other behaviors that they might engage in. Or if you ask the question too generally, it might backfire and they are just going to spin into what they wanted to talk about all along.

Schweitzer: What we found is that the questions guide people in a very specific way. They guide people not only to think about the topic or the level of specificity, but we found that they also communicate information about the presumption. We’re conveying information about our expectations. In the study Eric was talking about, the kinds of revelations we get for things like how much time they spend at work doing personal things, those answers are profoundly different depending on how we ask the question. The truth is out there, but what we get is totally different.

“People are much more likely to disclose problems when you presume [there is a] problem.”–Eric VanEpps

Knowledge@Wharton: What about the motivation for the person who is being asked? Just thinking about the job interview or the negotiation, there’s a financial benefit for that person to present the right answer.

Schweitzer: Let me build on this and take this to a different context. Imagine you are going in for your physical, and the doctor has 40 patients to rush through. The doctor might ask you questions like, “You don’t smoke, right? You know how to lose weight, right?” The doctor might just look at you, make an inference, make a guess, and by asking these affirming questions, they make it more difficult for you to say that you use drugs, you abuse alcohol. It’s a more difficult assertion to say, “I think I’ve got a substance abuse problem.” You are more likely to go along and say, “That’s right.”

The clinician might not have been trained to appreciate how important the question phrasing is in eliciting a different kind of answer. What they might get is just an affirmation. At the end of the day, it does allow them to move more quickly through that busy schedule. There is a lot of pressure for people. When Eric and I were starting this, we were thinking about how we get closest to the truth. There are sometimes cases where I think the person asking the question may not want to get closest to the truth; they may have other motives.

Knowledge@Wharton: How do you get people to understand the importance of question-phrasing if their end goal is efficiency, not information?

VanEpps: Thinking about Maurice’s example of the doctor asking about whether you smoke or use drugs, there are certain kinds of information that are more important. If you use drugs, that is incredibly important. The kinds of questions we ask should probably err on the side of being more likely to find that information out, rather than err on the side of convenience and not finding that information out.

It is incumbent on the question-askers to decide, what is the more important information here? Is it more important for me to find out a problem with the car, or to find out that the car is OK? I know that feels a little bit wibbly-wobbly as a distinction, but recognizing when problems are really important to find out should help guide us to ask questions that focus on those problems.

If there are other problems that don’t matter too much, if you really don’t care too much about the air conditioning in the car or whether the person flosses every day, maybe you can be a little looser with those questions. But when there’s a problem that is really important to discover, then you should be especially careful in how those questions are phrased.

Schweitzer: We know that, for example, drug compliance is very low, even for people who have had heart attacks, even for people who have had heart disease, even for people who have had transplants. The way a clinician asks the question can really guide people to answer very, very differently, and these could have serious health implications.

“There are sometimes cases where I think the person asking the question may not want to get closest to the truth.”–Maurice Schweitzer

Knowledge@Wharton: Can this research can be a teaching tool?

Schweitzer: Absolutely. I think these findings have very broad implications. We have been talking about a doctor’s office, interviews, negotiations, but this is broader than that. It’s certainly even closer to home for us as instructors. In the classroom, if we say, “You don’t have any questions, right?” as opposed to, “What questions do you have?” then the way we approach our students will be profoundly different.

VanEpps: I totally agree about the connections to the classroom. I’ve had friends talk about this in terms of the political moment, what a reporter should be doing when they’re asking questions, that they could be a little bit more direct than general. I think there are broad implications across a lot of settings, including for people buying products. They can be a little more assertive, a little more aggressive in asking questions when they are getting a used car or when they are looking for an apartment.

Schweitzer: One other key finding from our work is that we looked at the interpersonal costs of asking these direct questions. If you presume the problem, the more assertive question-asking didn’t rub people the wrong way. So, it didn’t have the interpersonal costs that we might worry about.

The politeness norms that Eric alluded to guide a lot of the way we interact in our conversations. We often try to be polite, and that might guide us or sort of push us away from being very direct. We found that being direct is OK. People didn’t mind it the way we thought that they would.