How altruistic are you? If you're like most people, your answer might be a lot higher than your estimate of the average person. A 2024 study by Mislavsky and colleagues documented this discrepancy using a simple list experiment, revealing a gap that, under certain conditions, reached 14 percentage points. The finding has sparked a lively debate among behavioural scientists about what self-reports of prosocial behaviour actually measure, and whether our intuitions about human generosity are systematically skewed.

The List That Split Altruism Research

The experiment was elegantly straightforward. Researchers presented participants with a list of 20 prosocial behaviours—donating to charity, volunteering, helping a stranger, and so on. Half the participants were asked: “How many of these acts have you performed in the past year?” The other half were asked: “How many of these acts has the average person performed?” The results showed a consistent pattern: self-reports averaged around 14 behaviours, while peer estimates averaged around 8—a 6-point gap. In a replication that included rarer behaviours, the gap widened to 14 points.

This disparity is not a trivial curiosity. It suggests that people see themselves as substantially more altruistic than they see others, and the size of the gap depends on how the list is constructed. When researchers added infrequent acts—such as donating a kidney or becoming a bone marrow donor—self-reports remained relatively high, while peer estimates dropped further. Participants were willing to claim credit for rare altruistic acts, but assumed that others rarely performed them.

The study, which involved roughly 1,500 adults per experiment, has been cited as a cautionary tale for survey-based research on prosocial behaviour. If self-reports are inflated, then any policy or intervention based on them may be built on a skewed foundation. But not everyone agrees on the source or the severity of the bias.

How the Experiment Worked

The core design was a between-subjects manipulation. One group saw the list and answered for themselves; a separate group answered for “the average American.” The list included behaviours like “donated money to a charity,” “volunteered for a nonprofit,” “helped a stranger carry groceries,” and “donated blood.” Participants simply checked off each behaviour they (or the average person) had done in the past 12 months. The total count was the measure of altruism.

This method avoids some of the ambiguity of open-ended questions, but it introduces its own biases. The list itself primes certain memories. For self-reports, checking off an item may feel like a small confession of virtue. For peer estimates, the same list may evoke a stereotype of the “average person” as less engaged. The researchers controlled for order effects and demographic variables, but the gap persisted across multiple samples.

In a follow-up experiment, the list was expanded to include 30 behaviours, some very rare (e.g., “donated a kidney to a stranger”). Self-reports for the rare items were low in absolute terms, but still much higher than peer estimates. For instance, around 2% of participants reported having donated a kidney, while they estimated that only 0.1% of the average population had done so. This 20-fold difference contributed to the overall 14-point gap.

The researchers also varied the wording: some participants were asked about “the average person of your age and gender,” which narrowed the gap slightly but did not eliminate it. The effect was robust across political affiliation, income, and education levels, suggesting a deep-seated cognitive bias rather than a demographic quirk.

One interesting nuance emerged when participants were asked to estimate the behaviour of a specific person they knew well, such as a close friend. Those estimates were higher than the “average person” estimates but still lower than self-reports. This suggests that the gap is not solely due to the anonymity of the target; even when we know someone well, we may not fully appreciate their altruistic acts, perhaps because many acts occur out of our direct observation.

Where the 14-Point Gap Came From

The gap between self and peer estimates can be explained by several psychological mechanisms. One is the availability heuristic, first described by Kahneman and Tversky: people judge the frequency of events by how easily examples come to mind. Your own altruistic acts are vivid and easily recalled; others’ acts are less visible, so you underestimate them.

A second mechanism is social desirability bias. When asked about their own behaviour, people may inflate their reports to appear virtuous. This is especially true for behaviours that are socially valued, like charity or volunteering. The peer estimate, in contrast, is less subject to this bias because there is no social pressure to portray the average person as especially good—or bad.

Mislavsky and colleagues argued that both self-reports and peer estimates are biased, but in opposite directions. Self-reports are inflated by social desirability and egocentric recall; peer estimates are deflated by the under-representation of others’ good deeds. The true level of altruism likely lies somewhere in between. However, the researchers did not claim to know where.

In a clever additional test, they asked participants to estimate the behaviour of a specific person they knew well (e.g., a close friend). Those estimates were higher than the “average person” estimates but still lower than self-reports, suggesting that the gap is not solely due to anonymity. Even when we know someone well, we may not fully appreciate their altruistic acts. This finding points to a deeper asymmetry in how we access information about ourselves versus others.

Another contributing factor is the “better-than-average” effect, a well-established bias where people rate themselves above the mean on positive traits. For altruism, this effect may be especially strong because the trait is so socially desirable. The list experiment essentially quantifies this effect in the domain of prosocial behaviour, showing that the bias can be as large as 14 percentage points when rare acts are included.

Why Scientists Disagree on What It Means

The finding has generated a healthy controversy. Some researchers argue that the gap reflects genuine differences in altruism: most people are indeed more generous than the average, because the distribution is skewed. A few very altruistic individuals pull the mean up, while the majority cluster near the median. If you are an average person, your self-report may be accurate, and your peer estimate is an underestimate of the true mean.

Others see the gap as largely a measurement artifact. They point out that self-reports and peer estimates are fundamentally different tasks. When you report on yourself, you are performing a memory retrieval task; when you estimate others, you are performing a social judgment task. The two may not be comparable. Moreover, the instructions “have you done X?” and “has the average person done X?” may trigger different response scales, even if the numbers are the same.

A third camp argues that both measures are biased, but the direction of bias is uncertain. Perhaps people overestimate their own altruism due to self-enhancement, but also overestimate others’ altruism due to a “kind world” bias. The net gap could be small or large depending on the context. Mislavsky’s team attempted to resolve this by including an objective behavioral measure in one experiment—actual donations to a charity—but the correlation with self-reports was modest, leaving the question open.

The debate matters for how we interpret survey data on prosocial behaviour. If self-reports are inflated by 40% or more, then studies that rely solely on self-reports may overstate the prevalence of altruistic acts. This has implications for everything from volunteer management to public health campaigns that depend on blood or organ donations.

Consider the case of blood donation. Surveys consistently find that a high percentage of people say they are willing to donate blood, yet actual donation rates are much lower. The list experiment suggests that part of this discrepancy may stem from inflated self-reports. Similarly, in environmental behaviour, people often report high levels of recycling or energy conservation, but objective data tell a different story. The gap identified by Mislavsky and colleagues may be a general phenomenon that extends beyond altruism to any socially desirable behaviour.

What Behavioral Economics Adds

Behavioural economics offers a framework for understanding the gap. The availability heuristic is a well-documented cognitive shortcut: people overestimate the probability of events that are easily recalled. Your own acts are always more available than others’, so you overestimate your relative contribution. This is compounded by the egocentric bias: we tend to notice our own efforts more than those of others, leading to a sense of being underappreciated.

Social norms also play a role. People have a rough idea of what is socially desirable and adjust their self-reports accordingly. But they may not apply the same adjustment to estimates of others. In fact, some studies show that people think others are more selfish than they actually are, a phenomenon called “pluralistic ignorance.” The list experiment may be capturing both self-inflation and other-deflation.

Another concept from behavioural economics is “moral licensing”: after reporting a prosocial act, people may feel licensed to behave less altruistically later. While not directly tested in this study, it suggests that self-reports are not just passive records but can influence future behaviour. This makes it even more important to understand their accuracy.

There is also a connection to the “spotlight effect,” where people overestimate how much others notice their own actions. This could amplify the availability of one’s own altruistic acts, because we think they are more visible than they actually are. Conversely, we underestimate the visibility of others’ acts, leading to a double standard in our mental accounting.

The practical implication is that our sense of being more generous than average may be an illusion, but one that has real consequences. It might motivate us to continue helping, or it might lead to complacency. The gap itself becomes a variable to study, not just a bias to correct.

Practical Takeaways for Surveys and Policy

For researchers, the message is clear: self-reported altruism is likely inflated by roughly 40% relative to peer estimates. This does not mean self-reports are useless, but they should be interpreted with caution. One recommendation is to include both self and peer measures in surveys, as the contrast can reveal the direction and magnitude of bias. Another is to use anchoring techniques or forced-choice formats to reduce social desirability.

For policymakers, the findings suggest that basing volunteer recruitment or donation campaigns on self-reported willingness may lead to overestimates of the available pool. For example, if people say they would volunteer 5 hours a week but actually volunteer 2, a program designed for 5 hours will be understaffed. Using peer estimates as a reality check could provide more realistic benchmarks.

Charities and non-profits might also benefit from framing. If people think others are less generous than they are, highlighting the true prevalence of giving could encourage more donations. Conversely, if self-reports are inflated, people might feel they have already done enough. Understanding the gap can help tailor messages that either correct underestimates or challenge overconfidence.

The list experiment itself is a low-cost tool that any organization can use. Simply asking people to check off behaviours and then comparing self to peer estimates can reveal hidden biases. Similar methods have been used in health behaviour research, where self-reported exercise or diet often diverges from objective measures. The same principle applies to altruism.

One concrete example comes from a study on workplace volunteering. Employees were asked how many hours they volunteered in the past month, and separately to estimate the average for their department. The self-reports were consistently higher than peer estimates, and when actual records were checked, the peer estimates were closer to the truth. This suggests that in organizational settings, peer estimates might serve as a useful benchmark for calibrating self-reports.

What We Still Don't Know

Despite the clarity of the gap, many questions remain. Does the discrepancy predict actual behaviour? A person who claims many altruistic acts may indeed be more helpful, or they may simply be more prone to exaggeration. Longitudinal studies that track self-reports against observed behaviour are needed. Mislavsky’s team plans field experiments where participants are given opportunities to donate or volunteer, and their self-reports are compared to their actions.

Cross-cultural replication is another open question. Most of the studies were conducted with US samples. In collectivistic cultures, where modesty is valued, self-reports might be lower, and peer estimates might be higher, potentially shrinking or even reversing the gap. Early evidence from a few East Asian samples suggests a smaller gap, but the data are limited. For example, a replication in Japan found a gap of only around 4 points, compared to 6–14 in the US. This hints at cultural modulation, but more systematic cross-cultural work is needed.

p>There is also the question of whether the gap is stable over time. If people are made aware of the bias, do they adjust their reports? Some debiasing interventions have been tried, but with mixed results. In one experiment, participants were shown their own gap compared to the average, and then asked to re-estimate. Some reduced their self-reports, but others became defensive and increased them. The gap may be a robust feature of human cognition, not easily corrected.

Another unknown is the role of memory decay. The list asks about behaviours over the past 12 months, but people may forget some acts or misremember their frequency. Forgetting would reduce self-reports, so the true gap might be even larger than observed. Conversely, if people falsely remember doing good deeds, self-reports would be inflated. The net effect of memory on the gap is unclear.

p>Finally, the researchers acknowledge that the list itself constrains what counts as altruism. Some forms of prosocial behaviour, such as emotional support or small daily kindnesses, may not be captured. If those are more common than the listed acts, self-reports might underestimate true altruism. The gap might look different with a more comprehensive list.

Until these questions are answered, the 14-point gap remains a contested signal—a reminder that what we say about our own goodness is not the same as what we see in others. The list experiment has given behavioural science a new tool to probe this discrepancy, but the underlying reality of altruism is still elusive.