Menu Close

Lying Because We Care

By Matt Lupoli

Doctoral Candidate | Rady School of Management

Imagine that you have a colleague who will soon be giving a presentation in front of several managers in your organization. A week before she is scheduled to present, this colleague practices her talk with you, and asks for your opinion on her performance. In truth, you thought that the presentation could use significant improvement. However, you are given pause by knowledge that this colleague is dealing with the recent death of a close family member. You can tell that she is in a fragile emotional state, and you feel concerned that giving her negative feedback will only worsen her suffering. In this situation, would you tell your colleague exactly what you thought of her performance?

According to our research, compassion felt towards the colleague could lead individuals to give more dishonest feedback (Lupoli, Jampol, and Oveis, 2017). This work highlights how compassion—that is, the concern for those who suffer and the motivation to enhance the welfare of others—increases prosocial lying, or lying that is intended to benefit others.

When well-intended lies can backfire

We all have experience giving and receiving white (that is, harmless) lies; a familiar example is telling a partner or close friend that they look fantastic in those jeans when you believe otherwise. White lies play an important role in smoothing social interactions–after all, if everyone said exactly what was on their mind 100% of the time, the world would probably be more a frightening place than it already is.

However, not all prosocial, well-intentioned lies are entirely harmless. For example, managers might give subordinates inflated feedback to make them feel good about their performance, at the expense of not giving them constructive criticism which could improve the quality of their work.

Given that prosocial lies can profoundly affect individuals and organizations, and sometimes for the worse, we thought that it is important to ask: What drives people to tell these lies, especially when there are clear benefits of honesty?

Compassion and prosocial lying

As emotion researchers, we naturally turned to our own knowledge base in considering when and why people tell prosocial lies. One particular emotion that we thought would be involved in prosocial lying is compassion. Compassion, or sympathy, drives behavior that is aimed to alleviate suffering and promote the welfare of others, such as giving to charity (Saslow et al., 2013), caring for young (Goetz, Keltner, & Simon-Thomas, 2010), and providing emotional support to coworkers (Lilius et al., 2008). Considering this, it seemed likely that this emotion would play some role in prosocial lying. However, the nature of this relationship was unclear to us.

One possibility we considered is that compassion would decrease prosocial lying, and hence increase honesty. There are two reasons why this could be the case: First, when faced with the opportunity to tell a prosocial lie, compassion may lead individuals to do whatever provides the greatest magnitude of benefits for others. Thus, if the benefits of a hurtful truth clearly outweigh the benefits lying, compassion could then lead individuals to be more honest. Recall the example of giving feedback to a colleague on a talk: if you know that your colleague’s future in the organization is riding on her giving a killer talk, compassion might lead you to tell her exactly where she needs to improve so that she doesn’t lose her job.

Another reason why compassion might increase honesty is because lying can have some pretty nasty consequences. Lies can damage trust, they can trigger retribution, they can damage or even destroy relationships (Boles, Croson, & Murnighan, 2000; Schweitzer, Hershey, & Bradlow, 2006; McCornack & Levine, 1990). And you don’t need to be a social scientist to aware of the harmful effects of lying—most people are unfortunately quite experienced with this. So it’s possible, then, that a lifetime of exposure to the harmful consequences of lying could have spillover effects toward perceptions of prosocial lying, thus making compassionate people averse to lying in general.

On the other hand, we had reason to believe that compassion would increase prosocial lying.

Because compassion involves a heightened sensitivity to the suffering of others, this emotion could focus individuals on the harm inherent in a painful truth. Thus, if lying is seen as a means to prevent or decrease suffering, then compassion might increase this type of lying.

A second reason why compassion might increase prosocial lying is because compassion biases our decision making. One well-known example that illustrates this is the story of Baby Jessica. Jessica McClure was an 18-month-old who captivated media attention nationally after she fell down a well and was trapped for nearly 3 days before being rescued. In the aftermath of this event, Jessica and her family received approximately $1.2 million dollars in donations from around the country. Elsewhere in the world at this time, events such as the Kurdish genocide in Iraq were occurring, which resulted in over 50,000 deaths, but received comparatively little attention and charitable donations (Black, 1993). Mismatches like these between levels of need and levels of aid are not uncommon; you can probably think of other examples where a single, identifiable victim received a seemingly inordinate amount of aid in part because of the compassion this particular person evoked. The take home message of this story is that compassion’s effects on behavior intended to help others are not necessarily calibrated toward promoting the most welfare-enhancing action. Thus, compassion could cause individuals to tell lies in order to prevent perceived harm to others, even when there are benefits of honesty.

To test these competing hypotheses and to determine whether and in what direction compassion influences prosocial lying, we first investigated prosocial lying a context with direct relevance to organizations: providing feedback.

Study 1: The effect of experimentally induced compassion on feedback

In a first experiment, 396 undergraduate participants in the Rady Behavioral Lab were told that they would be paired with another student who had previously written an essay about why they should be admitted to graduate school. Participants then learned that their task would be to evaluate this other student’s essay on several criteria. What they were not told was that in actuality there was no other student, and that all participants would be evaluating the same essay. (The irony of deceiving participants in an experiment about lying isn’t lost upon us. We absolve ourselves with Institutional Review Board approval and by viewing it as prosocial lying in the name of science.)

After participants read the essay and provided their ratings, we informed them that they would also view a message that was written by the essay writer about a recent event which occurred in his or her life. Half of participants were randomly assigned to receive a message that detailed the essay writer’s experience with the recent death of a loved one. This was an emotionally evocative statement that we had previously found to reliably elicit compassion. The other half of participants—those in the neutral condition—read a statement that described this person’s recent trip to the grocery store, which was ordinary if not boring.

Following this experimental induction of either compassion or neutral feelings, we surprised participants by telling them that they would have the opportunity to evaluate the essay again on the same criteria, except this time with the knowledge that the writer would view these evaluations. We explained to participants that each writer could rewrite their essay and submit it to a contest where they could win a prize. To ensure that there were perceived benefits of honesty, we also explicitly said that the information they would provide will help the writer improve his/her essay. Prosocial lying was then defined as the extent to which participants inflated their evaluations upon learning they would be shared with the writer (i.e., shared evaluations minus private evaluations).

What we found fascinated us. First off, the majority of participants (54%) lied to the alleged essay writer. That is, fifty-four percent of participants gave the essay a higher evaluation when they knew that their ratings would be shared with the writer compared to their initial, private evaluations.

But what we cared about most was the effect of the compassion manipulation—that is, would individuals experiencing compassion lie more than those who were feeling neutral?

As it turns out, they did. 64% of participants in the compassion condition provided a higher essay rating when giving the writer feedback, compared to only 45% in the neutral condition—a statistically significant difference. Not only did compassion increase the probability of prosocial lying, but also the magnitude of those lies: those in the compassion condition inflated their quantitative essay ratings to a greater degree than those in the neutral condition. Follow-up questions revealed that these effects were driven by a heightened importance placed on emotional harm in those participants who received the compassion manipulation. This suggests that participants really did lie because they were concerned for the emotional welfare of the fictional essay writer.

Study 2: Trait compassion and feedback

This first study provided evidence that compassion felt towards an individual can increase prosocial lying towards that person. But there are other ways in which this relationship between compassion and prosocial lying might play out in the real world. For instance, it’s possible that just being compassionate in general makes people more likely to tell prosocial lies, regardless of whether compassion is being actively experienced in the moment.

In a second study, we tested this notion. The study used an essay evaluation procedure that closely resembled that of the first experiment. This time, however, we did not elicit a state of compassion amongst participants. Instead, we measured participants’ trait compassion, or how compassionate they are on a daily basis, and examined whether their trait compassion was predictive of prosocial lying to the essay writer.

139 online participants first filled out previously validated questionnaires to assess trait compassion, which asked them to indicate their agreement with statements such as, “I often have tender, concerned feelings for people less fortunate than me,” and “I am a compassionate person.” Then, they engaged in the essay evaluation procedure as before, except there was no message from the essay writer to evoke either compassionate or neutral feelings.

The results of this second study were strikingly similar to those of the first. The degree to which participants inflated their evaluations of the essay when giving feedback to the writer was significantly correlated with their scores on the compassion scales—the more compassionate participants reported that they were, the more they exaggerated their essay ratings. Moreover, those who did give the essay a better rating when giving feedback (i.e., those who lied) had higher levels of trait compassion than those who did not inflate their ratings. Thus, people who were more compassionate in general were more likely to tell prosocial lies.

Study 3: Spillover effects of compassion on prosocial lying

These two studies offered compelling evidence for the positive relationship between compassion and prosocial lying. However, we wanted to go a step further. In life, emotions often have spillover effects on behaviors that are not directed at the source of the emotion. For example, a manager who just endured a stressful traffic jam may bark at employees if his road rage has not subsided, and watching a scary movie might make a person jump at the sound of tree branches tapping on a window. Similarly, in a third study, we examined whether spillover effects of compassion would increase prosocial lying towards a beneficiary that was not the original source of compassion.

Again at the Rady Behavioral Lab, 432 undergraduate students were told that they’d be participating in a study that investigates how visual stimuli influences memory. Specifically, participants learned that would be viewing photos and a video clip, which they would later be asked to recall. Unbeknownst to participants, the photos and video would serve as our manipulation of either compassionate or neutral feelings. Those randomly assigned to the compassion condition viewed a series of slides depicting helplessness, suffering, and vulnerability, followed by a video clip about child malnutrition and starvation. Meanwhile, those in the neutral condition viewed slides of household items and patterns, followed by a video of two men talking in a courtroom.

Next, participants engaged in a task where they were shown arrays of dots dispersed within a square that had a diagonal line dividing it. After a one-second exposure to each square, all participants had to do was indicate whether there were more dots to the left or to the right of the diagonal. Participants were told that for each response of more-on-the-left or more-on-the-right, a small donation would be made on their behalf to a real charity, the Against Malaria Foundation (which was actually true). Crucially, however, we also informed participants that because most people can more easily identify the number of dots on the left side, the charity would be paid ten times more when they reported that there were more dots on the right. Thus, they were incentivized to report that there were more dots on the right (i.e., the response that yielded the higher donation) for the benefit of the charity, even when this was not in fact the case.

The results again pointed to a positive effect of compassion on prosocial lying. On trials in which there were clearly fewer dots on the right of the diagonal (less than 40% of the dots on the right), participants in the compassion condition reported that there were more dots on the right (again, the charity payment-maximizing response) more often than those in the neutral condition. In other words, when there was a conflict between providing an honest response and lying to increase the gains for the charity, those experiencing compassion were more dishonest. When this conflict between honesty and lying for the benefit of the charity was removed—that is, in trials in which there clearly were more dots on the right—participants in both the compassion and neutral condition behaved similarly. These results suggest that compassion for one can indeed lead to prosocial lying for the benefit of another.

Implications for organizations

These findings may be interesting to the casual observer, but why should organizations care? We think that our findings in the context of providing feedback have profound implications for organizations. Giving and receiving feedback is an essential component of nearly all organizations. While it’s not always easy to give, the value of negative feedback is easy to recognize. Without receiving negative feedback, it can be difficult for people to discern what needs improvement, and this could be detrimental to both individual and organizational goals. Our work suggests that compassion, an emotion that is typically considered beneficial to organizations, could result in well-intended but inaccurate feedback.

So should managers then work to get compassion out of organizations? In short, no. We do not doubt that compassion can have positive effects in the workplace, including increased teamwork, extra-role behaviors (i.e., those that help the organization or fellow employees but are outside of one’s job description), and organizational commitment (Barsade & O’Neill, 2014; Grant, Dutton, & Rosso, 2008; Lilius et al., 2008). But given our results, managers may want to think carefully about the organizational contexts in which compassion should be promoted. One possibility is that we should encourage compassion when colleagues ask for help on tasks where they lack expertise or when they’re feeling overwhelmed, but that we should de-emphasize or at least qualify the role of compassion when giving feedback or providing performance reviews. As a personal anecdote, at Rady I have heard several practice talks for high-stakes presentations begin with the speaker requesting of the audience “please don’t be nice.”

We also do not think that being compassionate and giving negative feedback need to be mutually exclusive. One way to productively bring compassion into feedback is to focus on behavior, not the individual him or herself, as criticizing the individual is a sure way to hurt feelings and morale. Employees must also know that the process of arriving at the feedback is fair and just—if you’re going to criticize a person’s performance, make sure that it is clear how or why they’re not meeting the standard, and what they can do to change that. Employees are more likely to express commitment to the organization when they feel that they are cared for in the organization and that their colleagues have good intentions towards them (Grant et al., 2008). In the workplace, part of this means receiving accurate feedback about where one needs improvement.

Limitations and future directions

Our research is not without caveats. For one, we do not claim that compassion increases prosocial lying for all people in all contexts. There may be situations in which compassion has no effect on prosocial lying, or perhaps even decreases prosocial lying. In contexts where most people agree that lying is better than honesty, we would expect most people to lie, regardless of how compassionate they are. For example, it would take a special kind of person to tell a bride that she looks hideous on her wedding day even if that is what’s truly believed. Similarly, if the temporary emotional effects of receiving bad news pale in comparison to what can be gained from the truth, we would not expect the same results. Think of the doctor who diagnoses a patient with a potentially life-threatening disease when life-saving treatments readily available; it is very unlikely that a compassionate doctor (or any doctor for that matter) would lie about the diagnosis. Furthermore, our studies only involved people in the US who did not know each other personally. More research is necessary to determine whether these effects hold cross-culturally and in closer relationships, such as those between friends and family members.

According to the Dalai Lama XIV, “When we are motivated by compassion and wisdom, the results of our actions benefit everyone, not just our individual selves or some immediate convenience.” (Dalai Lama, 1991) While this may be a bit of an oversimplification, we think that the Dalai Lama gets it right when he refers to both compassion and wisdom. Neither compassion nor wisdom alone are sufficient for benefiting others, because we need compassion to be motivated to help others, but we also need wisdom to know how to best help others. And it is through this process of learning about how to balance compassion and wisdom, or emotion and reason, that we will make better decisions for others as well as ourselves.



Barsade, S. G., & O’Neill, O. A. (2014). What’s love got to do with it? A longitudinal study of the culture of companionate love and employee and client outcomes in a long-term care setting. Administrative Science Quarterly59(4), 551-598.

Black, G. (1993). Genocide in Iraq: the Anfal campaign against the Kurds. Human Rights Watch.

Boles, T. L., Croson, R. T., & Murnighan, J. K. (2000). Deception and retribution in repeated ultimatum bargaining. Organizational Behavior and Human Decision Processes83(2),    235-259.

Dalai Lama (1991). Thinking Globally: A Universal Task. EPA Journal: A Magazine on National and Global Environmental Perspectives, 17(4), 51-53.

Goetz, J. L., Keltner, D., & Simon-Thomas, E. (2010). Compassion: An evolutionary analysis and empirical review. Psychological Bulletin, 136(3), 351-374.

Grant, A. M., Dutton, J. E., & Rosso, B. D. (2008). Giving commitment: Employee support programs and the prosocial sensemaking process. Academy of Management  Journal51(5), 898-918.

Lilius, J. M., Worline, M. C., Maitlis, S., Kanov, J., Dutton, J. E., & Frost, P. (2008). The contours and consequences of compassion at work. Journal of Organizational Behavior29(2), 193-218.

Lupoli, M.J., Jampol, L.E., & Oveis, C.(2017). Lying because we care: Compassion increases prosocial lying. Journal of Experimental Psychology: General, 146(7), 1026-1042.

McCornack, S. A., & Levine, T. R. (1990). When lies are uncovered: Emotional and relational outcomes of discovered deception. Communications Monographs57(2), 119-138.

Saslow, L. R., Willer, R., Feinberg, M., Piff, P. K., Clark, K., Keltner, D., & Saturn, S. R.    (2013). My brother’s keeper? Compassion predicts generosity more among less religious individuals. Social Psychological and Personality Science4(1), 31-38.

Schweitzer, M. E., Hershey, J. C., & Bradlow, E. T. (2006). Promises and lies: Restoring violated trust. Organizational Behavior and Human Decision Processes101(1), 1-19.