American Scientist

Whether the conversation is about climate change or reparations, difficult conversations are happening more frequently these days. According to guests Emma Levine and Shereen Chaudhry, both professors of behavioral science at the University of Chicago Booth School of Business, it is possible to navigate difficult conversations without causing harm.

Levine studies the psychology of altruism, trust, and ethical dilemmas. Her main stream of research investigates the tension between honesty and benevolence, and how this influences interpersonal communication.

Chaudhry teaches negotiations to MBA students and studies how people navigate social interactions and relationships. Her research reveals that everyday conversations often carry hidden negotiations over reputation, responsibility, and relationships.

[TRANSCRIPT]

[Music: Wandering by Nat Keefe FADE IN]

[Celia]
Welcome to Wired For This—a deep dive into how we think, believe, change, and connect. In this limited series, we explore the psychology of human behavior and neuroscience—what drives us forward, what holds us back, and how we navigate a world bursting with noise, contradiction, and complexity.
Today, we have not one but two guests: Emma Levine and Shereen Chaudhry. They’re both professors of behavioral science at the University of Chicago Booth School of Business. Dr. Levine studies the psychology of altruism, trust, and ethical dilemmas. Her main stream of research investigates the tension between honesty and benevolence, and how this influences interpersonal communication.
Dr. Chaudhry teaches negotiations to MBA students, and studies how people navigate social interactions and relationships. Her research reveals that everyday conversations often carry hidden negotiations over reputation, responsibility, and relationships.
From American Scientist, I’m Celia Ford, and you’re listening to Wired For This.

[Celia]
We’re talking about difficult conversations, so first I want to ask: what makes conversations difficult?

[Emma]
Conversations like giving negative feedback or delivering bad news or a bad prognosis, or even educating children about hard topics. All of these have this aspect that you fear they could cause emotional harm to the listener. That’s one thing that makes conversations very difficult. You anticipate and worry about that experience of causing harm.

[Shereen Chaudhry]
I also have a narrow view of this.

[Celia]
That’s Dr. Shereen Chaudry

[Shereen]
There’s all kinds of difficult conversations. I think the first thing that comes to people’s minds is political disagreements. But I focus a lot on interpersonal relationships, and specifically conversations about conflicts that you’re having in your life. You’re having a dispute with your spouse. You have to discuss it. Or with a friend or a colleague. Navigating those, especially when there’s blame to go around.
Also cases where maybe you’re managing your reputation or your impression. You want to share some achievements that you’ve had recently. How can you do that and not be perceived as a braggart or a jerk? How can you share information that maybe is self-deprecating, but also not have it hurt your reputation in the long run? Navigating these kinds of tradeoffs.

[Celia]
Shereen, I think this most closely relates to the research you were just talking about. What are some hidden patterns or rules that you’ve noticed when people are following difficult conversations like the ones you just described?

[Shereen]
One thing I’ve been focusing on lately is the balance of blame. There’s a couple of things we’ve observed. For instance, if a person—you open up by blaming your partner for not helping with the vacation planning. They are likely to blame you back for not doing the laundry, which you don’t do any of.
It’s not personal or anything. People are just engaging in an eye for eye exchange, what we call reciprocity. You said something mean to me that threatens my ego. I want to do the same back to you.
But what I find in my research is that’s not the only explanation for this blame game dynamic. Another explanation is if people are trying to set the record straight. If I do think that I am the only one to blame, then I will not blame you back after you blame me. But if I disagree with you—you’re making it seem like you didn’t do anything wrong if you just blame me without taking any blame yourself. I have to correct the record. You get this blame game dynamic simply because people are negotiating the truth.

[Celia]
A couple years ago, Shereen, you published a paper about how one can prove to a skeptical listener that they’re telling the truth, which is really challenging, especially if that person has some beef with you.
How can honest speakers, maybe even subconsciously, convey their sincerity? And on the flip side, how can those strategies be hijacked by people who want to get away with lying?

[Shereen]
That’s a fascinating question. If you are trying to apologize, and you really want this person to believe you, one of the best things you can do is engage in costly signalling. It’s an academic term, but it’s basically, incur some sort of effort or cost that signals to the other person that this is meaningful.
Otherwise it’s what we call cheap talk. You can just say sorry all day long if you feel like it’s not hurting your reputation.
There’s research showing, for instance, that adding a gift to an apology makes it way more effective. Potentially incurring a reputational cost, like a very visible apology by a public figure, is more costly, more believable. The more effort you put into something, the more it signals sincerity. That’s kind of a simple heuristic for signaling sincerity.
But of course it can be used insincerely. The idea is that if you are insincere, you’re less likely to want to incur that cost.

[Celia]
Right. I want to turn to Emma now, because you focus on a very different subset of difficult conversations, where there’s tension between honesty and kindness. Can you walk me through what’s happening in our minds when we face that kind of tension?

[Emma]
Giving bad news, engaging in conflict like Shereen studies, delivering negative feedback, education about racism, bullying, all these things. We want to be honest. We have this abstract value that we should always be honest running through our minds as we approach these topics. But then in the moment we feel and face this very real concern about causing harm to the listener, or causing harm to the relationship.
One thing that’s happening when we’re dealing with these two competing motivations is they’re at different levels of abstraction. We think about honesty as a global moral good, but then when we’re in the presence of another person, the visceral fear, or just discomfort or anxiety, that most people have experienced when actually approaching these conversations, realizing, “I’m gonna blow everything up! This person is going to hate me! This person is going to cry!”—that becomes super intense.
In some recent work we liken this to other self-regulation dilemmas. In the long term, abstractly, I should totally tell my partner that I’m not satisfied with them. We’ll be more happy with each other. We’ll build trust. We’ll fix our problems. We can think in the long term future, broadly, that this is good. We might even think the benefits outweigh the costs. This is a conversation we should have.
But then when faced in the moment with the harm you might cause, you’re kind of shutting down. In that hot state of, I can’t do this, what if they yell at me, I feel so scared—it’s really navigating that.
Just like how you need tools to get you to go to the gym and work up a sweat for that long-term benefit of being healthy, we need tools and tricks to help us overcome this really visceral negative experience of causing harm, of being unkind, to actually be honest.

[Celia]
You’ve studied this in health care workers, who sometimes have to deliver really bad news to very sick people. What does your work tell us about how we can break bad news while balancing honesty with compassion, whether it's a cancer diagnosis or a breakup?

[Emma]
It’s not easy, but some of my best advice actually does come from the health care setting, where we ran these studies with doctors, with oncologists and patients, looking at what people think is acceptable when you have negative news. In this case it was a negative prognosis.

[Celia]
(By “negative,” here, she means bad news. Not “negative” like “it’s not cancer.”)

[Emma]
The one thing that’s interesting, what physicians are tempted to do, and most communicators are tempted to do, is rather than be completely honest and deliver this brutal truth, or err on compassion and give a patient false hope, they might just omit information.
(To be clear, doctors are not regularly withholding information from patients. I have no evidence of that.)
But there are these times of uncertainty where we might not know the full truth, and a physician might think it’s best not to say anything. I don’t want to create unnecessary worry, but I also don’t want to assure them that everything is okay, so I’m just going to wait and say nothing. Patients are like, that’s terrible! The worst thing you can do to me in that position of vulnerability is not give me information. Then I don’t have the benefit of comfort or the information value of a truthful diagnosis or prognosis, the ability to plan.
It was in that setting that we first found that a lot of communicators believe that omission is more acceptable than false hope or other tactics. But recipients actually think the opposite, that omission, not hearing anything, is really dangerous.
One overall takeaway from that—and that extends across contexts. Why do we ghost people instead of breaking up with them? Because we think, I don’t want to hurt their feelings. I don’t want to be completely honest. But I don’t want to keep leading them on. I’ll just ghost them.
Other colleagues have worked on this topic. Recipients hate that. It’s the worst. You left me in the dark. I have no idea what’s going on. You could have been truthful, you could have been nice, and you did neither.
In terms of what we could do better, there are tactics to be kindly or benevolently honest, which requires a lot more work and a lot more planning, but sometimes can be as simple as clearly stating your intentions: I’m letting you know this, because I think you can handle the news, and because I want you to understand where I’m coming from. I want you to prepare for you and your family’s future, whatever the context is. Then proceeding and allowing room for support and conversation. That’s the most costly and time-intensive method, but it’s a way to combine both honesty and kindness instead of avoiding both.

[Celia]
Both of you are principal investigators of the HOPE Lab at the University of Chicago. I think that all seven letters of HOPE Lab make up an acronym, which was very funny to me. Let me see if I got it right. Honesty, Opportunity, Prosociality, Ethics, Language, Affiliation, and Belief. Is that right?

[Celia]
It’s so sneaky that they even got LAB in there.

[Emma]
It didn’t used to be that LAB was an abbreviation for something. It needed to represent our full range of interests.

[Shereen]
I wanted to be part of the lab, and so we had to find a way of getting my research into the acronym.

[Celia]
I would love to know more about what kinds of methods and tools you’re using to study things like honesty and human communication.

[Shereen]
As a first pass, we will design a survey that asks people questions about the situation. Sometimes that means we throw a survey up online and we collect data from people on this platform called Prolific Academic.
But we also do in-person studies. One study that I did was, we recruited real pairs of participants who have real relationships—spouses, roommates, siblings. We recruit people off Michigan Avenue in Chicago. A lot of people come in as pairs.
We have them in separate rooms, and then we ask them to think of a conflict from their life in this relationship, and then to discuss this. We have video recordings. We would just record it over Zoom essentially. Then we have undergraduate students watch those videos, code those videos for different behaviors like blame and so forth.

[Celia]
I’m also just imagining being on a first date with someone, walking down the street, and then a research assistant pulls you aside. Hey, we’re going to do this study where you have to talk about all of your conflicts! Has that ever just completely ruined somebody’s night?

[Shereen]
Oh gosh, I do not know, because I’m thankfully not the one pulling people off the street and breaking up their relationships. But there is research in our department that suggests that’s actually good for a relationship, not bad. Even Emma’s work suggests that.

[Celia]
So you’re setting people up for success in the long term. What about you, Emma?

[Emma]
Our approaches and methods are very similar. So, I also do paired studies with relationship partners and roommates and the like. But these types of things do happen. When I first started at Booth, a couple came to my office and said, we’re here for the study. They’re supposed to come to the lab. The researcher is not supposed to be directly involved. But it was clear that they thought this was a different type of psychology. They were there for therapy. Well, it will be a form of therapy? But yes, you are about to have a difficult conversation with each other. Then they proceeded on their way.
But in addition to doing these lab experiments with couples and real relationship partners, one of the methods I use is what we call “lab in the field.” Similar to what Shereen described, a survey experiment you might write, but focused on delivering it to specific practitioners of interest, like physicians and their patients.

[Celia]
When many people think of difficult conversations, their mind immediately goes to politics—like a classic Thanksgiving dinner fight with that one uncle. How might some of the principles from your research apply when we’re talking with someone we care about, but fundamentally disagree with?

[Shereen]
When it comes to issues like, let’s take climate change, or reparations, I think people do feel as if they are being blamed for something that they feel they’re not responsible for.
If I was giving advice, I would say that people should move away from that. This is not about blaming anyone, or you in particular, for the fact that you have an F-150 truck. Or for what your ancestors did. This is about correcting wrongs. It’s about taking responsibility going forward.
Blame involves the ego, so I would advise. Removing blame from the conversation.

[Celia]
Their colleague at UChicago, Nick Epley, has run experiments demonstrating that politically divisive conversations actually go better if you enter them with positive expectations. Underestimating how much common ground you and that one uncle have can create unnecessary barriers that make conversations harder.

[Emma]
I think there’s still some threat and fear about harm, even in these political conversations. Particularly with people we’re close to. I’m going to anger Uncle Adam because he’s going to think I’m confronting him about his views, when in reality I’m trying to understand them.
Also thinking about, how can we harness maybe pre-scripting what we’re going to say, so we’re not overwhelmed in that moment and fumbling or back off? How can we make our intentions clear? Like Shreen said, I actually really want to understand your viewpoint. I want you to understand mine a bit better too.

[Celia]
I’ve also been thinking a lot about how today, many of us, especially young people, have our deepest conversations through a screen—whether that means texting with a friend, or a chatbot. How might the nature of difficult conversations change when we’re online?

[Shereen] A lot of people say, for instance, that they feel they’re able to hide their emotions more when they’re having these virtual conversations. Especially over email or text, obviously, but even when it comes to Zoom. It’s a more controlled atmosphere. That can be strategically useful in some situations, but you can imagine that it could also mute the connectedness of conversations, and also maybe get in the way of some of the ways that we’re able to make these difficult conversations go better.
I would definitely say, when it comes to navigating emotionally laden situations, you don’t want to be texting.

[Celia]
More from Nick Epley—his work shows that people are terrible at detecting tone over email and text. Nothing compares to the emotional content of someone’s actual voice.

[Shereen]
Also, dialogue. When you’re in person versus email, it’s more of a dialogue versus monologue. Being able to follow up on what someone says has a really big impact on feeling like you share common ground with the other person.
Now that we have ChatGPT, it’s very easy to use it to smooth out our emails, our messages, things like that. The problem with that is there’s all these hidden social signals in our communications that we may not even be aware are there and that we’re picking up on. Now, I think, they’re going to get muted.
Someone is giving you feedback on a paper or a memo or a presentation, and they’ll say small things like, I thought this was really great and so on, you should improve here. If you know that ChatGPT contributed to this email, it now takes the value out of those things almost completely. These small, nice, connective gestures you can have in these communications potentially become less powerful. And so the question is, how is that going to evolve over time?

[Emma]
One thing that I think is interesting about technology mediating communication is potentially these dual effects it has, on the communicator and the listener. It certainly is better from the listener’s perspective to hear a voice, to have that understanding, to feel heard, to feel like I understand your intentions.
But having a conversation face to face is also what makes the emotionality of harm, of relational difficulty, more palpable. This is why people rely on email and text sometimes. The easiest way to be honest is when I’m not facing you. When I’m facing you I’m overwhelmed by the prospect of harming you. Sometimes just writing it out really enables you to get your thoughts out clearly. Ideally we would have that, and then an opportunity to meet in person.
Sometimes then, when I’m actually faced with you, you’re hearing me and you feel better, but now I’m telling all these lies. I’m comforting you. I’m not even finishing with my message because we’re having a better experience, but it’s actually impeding the honesty part of the conversation.

[Celia]
That’s so interesting. I do want to poke at AI more. Do you see AI changing our expectations for authentic human conversations, or our capacity for those conversations? How should we plan on adapting to this new AI-integrated world?

[Shereen]
I do have some sense of how I will respond, looking out, for instance, for student essays that have used ChatGPT. Primarily, a lot of the assignments in negotiation involve just writing. I found myself actually preferring some of the essays that seemed less polished because of that. Could it be that we start to prefer things that seem less polished, have errors, are just more unique in some way? Maybe even less clearly communicating, simply because we’re looking for that authenticity.
Another question is, will that depend on domain? Certainly you want things to be clear when it comes to your professional life, but maybe when it comes to communicating with colleagues on a more personal level, or with friends, we’ll look for more idiosyncrasies, potentially.

[Emma]
I have some optimism about AI. I think it will certainly degrade the meaning of our online communication, but the bright side of that is it might force us to be more reliant on non-online communication. Back to in-person and voice.
But I also think AI has the benefit of offering low-stakes practice in conversation. One thing that makes these conversations so difficult is they’re inherently uncertain. You don’t know if a person is going to blame you back. You don’t know if a person is going to get upset. You’re not prepared for all of these different turns a conversation could take, which leads you to often avoid them.
I got this advice from a colleague recently. I was preparing to have a difficult conversation. She said, have you thought about doing it with AI? Tell ChatGPT, first react really angry! Oh, that’s what anger looks like if I tell this person I’m not pleased with the way they’re working on something.
It allows you to try to navigate some of this uncertainty in practice, what a response could look like and where that could go. I think that’s great, because often we don’t engage in any of this stuff. Because we don’t know how it will go, and we never get the feedback to learn how it could have gone better.

[Celia]
Speaking of being prepared: Sometimes we talk to manipulative or misleading people. What can we do, if we think someone is lying to us?

[Emma] It really depends on what your goals are. Are you willing to walk away from this relationship or this deal or this interaction, or are you trying to salvage it?
If you really think someone is lying and you don’t want to do business with them, or you don’t want to have a relationship with someone who cheated, you should feel free to confront them and be prepared to walk away.
If you’re not prepared to walk away and you want to smooth out the relationship, but hope they get the message and something can change, you have to give them an out, usually. A way that they can interpret what you’re saying that doesn’t feel overly critical or damning. For example, you might want to attribute what they did to incompetence rather than immorality: "I noticed the facts you gave me were this. That doesn’t match up with my preparation for this negotiation. Maybe you should take another look and we should reconvene later?" It gives them the time to think: "Oh, yes, I can say that was my mistake. I’m now going to go back and we’re going to move forward. It’s going to be simple." You can apply that type of insight in other settings as well.

[Celia]
Do power dynamics also affect our roles in challenging conversations?

[Emma]
We’ve all been in situations where we’re painfully aware of how much more difficult it is to engage in a difficult conversation with someone that you perceive as having power over you.
If we think about the types of conversations I was referencing, ones about harm, so far I’ve described something that’s relatively paternalistic. We don’t want to harm someone that we have some influence over, like a patient or our child or a friend.
But when there’s a different power dynamic, such that they have power over you, you’re more often worried about the cost to you, about what will happen to you. What will be the repercussions? That can make a conversation really difficult to engage in.

[Celia] A lot of our listeners are managing research groups, or teaching, or taking care of family. I’m sure plenty of them have had to work through tension from a position of power. What are some of the biggest mistakes that people in power make when they’re trying to share feedback? How can they overcome them?

[Emma]
Back to this idea of managing the tension between kindness or avoiding harm and honesty, there are a lot of mistakes that happen when people are trying to compromise on both of these goals.
One that we already talked about was avoidance: "I won’t be honest, and I won’t be kind. I’ll just escape this conflict altogether." That’s a mistake.
But there’s also other mistakes that people make, like being less than totally truthful. One of my favorite papers in this vein is about this tactic called paltering. This is not my work, but it remains something I cite all the time. Paltering is when you make a truthful statement that is nonetheless misleading.
Let’s say a person gives a presentation. It’s terrible. The gist of the feedback is that it’s terrible. What you say to them is, I really liked the font choice. It’s truthful, but it’s not telling them the gist of your belief. It’s just telling them something that you feel okay about saying, a little bit kind, but not following through on any type of honesty.
That’s something people do all the time. We skirt around the issue. We say something that makes us feel like we didn’t lie, but it’s not delivering to them the feedback they need to hear, the news they need to receive and so on.
Another error we also make is telling what I call prosocial lies. We just tell them lies that feel nice to hear. Actually, that presentation was great! Or absolutely, you’re going to fight this, when that seems unlikely, in the case of illness. I do have work showing that prosocial lies can be beneficial. But they’re in very specific situations, when both parties know what the recipient wants to hear, what their values are, and that there’s no opportunity for them to learn from the truthful feedback.
Often when we’re giving false praise or sugar coating or telling any other type of prosocial lies, it’s a mistake, because there is some truth that they could benefit from hearing.

[Shereen]
There’s research showing that the more power you have, the less executive control you exert in what you say and do. You’re more free. You control yourself less.
But the other thing is that there’s research showing that taking a direct and intense stance against somebody in a conflict can lead to emotional reciprocity, where you get these negative cycles of intense conflict in communication.
Conflict can be great in a team when it makes different ideas come together and you can address each other’s weaknesses in the ideas you have. But when you have really unproductive communication patterns, it’s not productive to have team conflict. Direct and intense communication, like where—I disagree with you. I think you’re wrong. You express it in a way that suggests you’re not flexible to having your mind changed. That can end the conversation.
Maybe the person, if they’re lower status, doesn’t try to challenge you on that. Or maybe they do and it ends the relationship. I can imagine that being more of an issue for people in power if they’re not aware of how their reactions affect other people.

[Celia]
Can you each leave us with one final nugget of advice to carry into our conversations moving forward?

[Emma]
Be honest and kind. It is possible!

[Shereen]
When you’re going into a difficult discussion with someone, maybe about something you think they did wrong, ask yourself first whether you could have contributed to this in any way—or at least whether you think they think you contributed to it in any way.
Because if you can at least acknowledge that, then the conversation, I think, will go much more smoothly, because they’ll be less defensive, feel less of that motivation to correct this narrative about who’s wrong, wanting to protect themselves from being the only person who did anything wrong.

[Celia]
Emma, Shereen, thank you for having this not-too-difficult conversation about difficult conversations. I really appreciate it.

[Emma]
Thanks for having us.

[Shreen]
Thank you. It was fun.

[Celia]
Thank you to Emma Levine and Shereen Chaudhry for joining in on this episode of Wired For This. You can find links to our sources in the episode description.
You’ve been listening to a podcast by American Scientist, published by Sigma Xi, the Scientific Honor Society.
Wired For This is produced and edited by Nwabata Nnani and hosted by me, Celia Ford.
Thanks for listening.