Available online 17 April 2024

Author links open overlay panel, , , Abstract

Human society rests on communicated information, much of which is shared without an expectation of reward. We suggest that, like other forms of prosociality, this type of information provision is fueled by gratitude. To reflect the fact that information differs in some ways from other goods, we call this form of gratitude epistemic gratitude. In a first experiment (all preregisterered, with US participants), we show that participants are more grateful for information that provides more benefits, at a greater cost to the sender, that was sent intentionally, and gratuitously. Experiment 2 shows that information shared with a large audience generates less gratitude in individual audience members. Experiment 3 shows that information that can be further passed on to others elicits more gratitude. In the supplementary materials, we also report a series of inconclusive experiments testing whether gratitude increases when an initially doubted piece of information is confirmed, and whether participants think others communicate in a way that maximizes gratitude in the audience. In conclusion, we speculate on the consequences of epistemic gratitude—in particular, which type of information is more likely to elicit epistemic gratitude—for diverse cultural phenomena, from personalization in marketing to rumor diffusion.

Section snippetsGratitude and prosociality

Gratitude is a moral affect that is a likely essential ingredient of social life, and in particular of prosocial behavior–behavior that benefits a receiver at an immediate cost for a provider (for review, see, McCullough et al., 2001). When people find themselves on the receiving end of a prosocial action (under conditions specified presently), they feel grateful. This feeling of gratitude has two consequences. First, it motivates the recipient of the prosocial act to adjust their estimate of

Epistemic gratitude

In the experimental literature, prosocial acts have typically been operationalized as goods (e.g. money) or services (e.g. performing a task in one's stead). However, prosocial acts can also take the form of the provision of information. We benefit when someone gives us a useful tip, or teaches us a new skill. It might seem as if the provision of information doesn't entail a cost, but in fact it often does; either an opportunity cost (we might be doing something else instead of transmitting the

The current experiments

The current experiments have several goals. The main goal is to test whether feelings of gratitude are evoked in the same context in the case of information–epistemic gratitude–as for other prosocial acts. In so doing, we test predictions about the factors that affect feelings of gratitude that have, to the best of our knowledge, never been addressed (in particular, gratuitousness), or that have only been tested in a piecemeal manner (benefit, cost, and intentionality). In a series of

Experiment 1

Experiment 1 tests whether feelings of gratitude triggered by the provision of information vary in line with the four factors suggested to influence feelings of gratitude by McCullough et al. (2008), leading to the four following hypotheses:

H1: Participants will feel more grateful towards the source that provides them with information leading to more benefits.

H2: Participants will feel more grateful towards the source that expended more effort to find or provide them with the information.

H3:

Participants

A power analysis showed that for a binomial test with an expected frequency of 0.65, an α-level of 5% and a power of 95%, we would need a minimum of 147 answers per dependent variable (excluding the answers that express no gratitude). Given that we planned to exclude “not gratitude” responses (see below), we collected data from 185 participants (90 female, 91 male, 1 other, M age = 36.58) from the US, UK and Ireland using the Prolific platform.

Materials and procedure

On the Prolific platform, participants were

Results and discussion

For each dependent variable, we first removed those participants who answered “I feel grateful towards neither” and performed a binomial test (H0 = 0.5) on answers aggregated across all cover stories. All reported analyses are two-tailed.

The frequencies of choices in each condition are depicted in Fig. 1 (the percentages reported here exclude the “grateful to neither” answers). In the Redundancy Condition, we find an effect of order: participants chose the first source as the one they were more

Experiment 2

As noted above, it's easier to benefit several people at once with information as with most other goods. However, if a source provides information to many people at once, the cost they pay to do so is less diagnostic of their feelings towards us, as we do not know whether they would have bothered providing the information only for us. This leads to the following hypothesis:

H5: Increases in the size of the audience to whom the message is directed, that stands to benefit from the information,

Participants

A power analysis showed that for a z-test on two proportions, with an expected proportion for choosing the first source in the neutral condition set to 0.80 and the expected proportion of choosing the same in the audience size condition set to 0.55, with an α-level of 5% and a power of 95%, we would need a minimum of 178 answers (excluding the answers that express no gratitude). Given this, we chose to collect data from a total sample of 200 participants (100 in each condition). Our final

Results

We removed the “I feel grateful towards neither” responses (N = 9 in the Public condition, N = 13 in the Private condition) and performed a Z-test on proportions of the aggregate answers across cover stories between the two conditions (two-tailed). Confirming our predictions, our results showed that the proportion of participants more grateful to the first source was significantly different between the two conditions: 86.05% in the Private Condition vs. 34.07% in the Public Condition

Experiment 3

One of the potential benefits of information is that we can pass it on to others, who might in turn feel grateful to us–whereas doing so is typically more difficult with other goods that need to be consumed for the benefits to accrue (e.g. money). In our next experiment, we wanted to test whether people feel more grateful for information they can later “reuse” in their social circle. We made the following prediction:

H6: People will feel more grateful towards someone who gives them a piece of

Participants

A power analysis showed that for a z-test on two independent proportions, with an expected proportion for choosing the second source in the audience condition set to 0.75 and the expected proportion of choosing the same in the no audience condition set to 0.5, with an α-level of 5% and a power of 95%, we would need a minimum of 190 answers (excluding the answers that express no gratitude). Given this, we chose to collect data from a total sample of 200 participants (100 per condition). Our

Results

Like in the previous analyses, we removed those answers that indicated no gratitude (N = 12 in the Reputational Benefit condition, N = 13 in the No Reputational Benefit condition) and analyzed the data using a Z-test on proportions of the aggregated answers. Confirming our hypothesis, the results showed that the proportion of participants more grateful to the first source was significantly lower (43.18%) when the second source's advice lead to reputational benefits, than when the second

General discussion

Experiment 1 showed that participants felt more grateful to those who had provided them with information that was (i) more beneficial (to the recipient), (ii) more costly to acquire (for the source), and that was delivered (iii) intentionally, and (iv) gratuitously, without expectation of immediate return. In Experiment 2, participants felt less grateful to sources that addressed a larger audience, presumably because they felt that the information was not transmitted for their benefit in

Conclusion

Human societies could not function if people did not engage in prosocial behavior with no immediate benefit expected in return. This includes the prosocial provision of information, from informal advice to on-the-job training. Gratitude fuels prosocial behavior: it motivates the recipients of prosocial behavior to reciprocate, and the providers of prosocial behaviors to keep providing. We suggest that epistemic gratitude plays an analogous role for the transmission of information.

In a series of

Funding

This research was funded by “An Evolutionary and Cultural Perspective on Intellectual Humility via Intellectual Curiosity and Epistemic Deference” from the John Templeton Foundation to HM, ANR-17-EURE-0017 to FrontCog, ANR-10-IDEX-0001-02 to PSL, and ANR-21-CE28-0016-01 to HM.

CRediT authorship contribution statement

Mia Karabegovic: Writing – review & editing, Writing – original draft, Supervision, Methodology, Formal analysis, Data curation, Conceptualization. Léo Wang: Methodology, Conceptualization. Pascal Boyer: Writing – review & editing, Writing – original draft, Methodology, Conceptualization. Hugo Mercier: Writing – review & editing, Writing – original draft, Supervision, Resources, Project administration, Methodology, Conceptualization.

Declaration of competing interest

None.

References (49)M. Berriche et al.Internet users engage more with phatic posts than with health misinformation on Facebook

Palgrave Communications

(2020)

P. BoyerDeriving features of religions in the wild

Human Nature

(2021)

J. BrightThe social news gap: How news reading and news sharing diverge

Journal of Communication

(2016)

J.R. Carey et al.A test of positive reinforcement of customers

Journal of Marketing

(1976)

S. Chandra et al.Personalization in personalized marketing: Trends and ways forward

Psychology & Marketing

(2022)

R.B. CialdiniInfluence: The psychology of persuasion

(1993)

N. Claidière et al.Argumentation and the diffusion of counter-intuitive beliefs

Journal of Experimental Psychology: General

(2017)

I.B. CohenRevolution in science

(1985)

A. CoppockGeneralizing from survey experiments conducted on mechanical Turk: A replication approach

Political Science Research and Methods

(2018)

M. Dockendorff et al.Argument evaluation and production in the correction of political innumeracy

Thinking & Reasoning

(2024)

P.S. Dodds et al.Human language reveals a universal positivity bias

Proceedings of the National Academy of Sciences

(2015)

P.L. HarrisTrusting what you’re told: How children learn from others

(2012)

L.K. John et al.Shooting the messenger

Journal of Experimental Psychology: General

(2019)

J.J. Jordan et al.Uncalculating cooperation is used to signal trustworthiness

Proceedings of the National Academy of Sciences

(2016)

View full text

© 2024 Elsevier Inc. All rights reserved.