"Thank you so much!" doubled the help. Who ran the experiment now matters.

Eight words of thanks doubled how many people helped again, from 32 to 66 percent. Those two experiments were run by Francesca Gino. What still stands in the famous gratitude study.

Some time before 2009, sixty-nine students at a university in the south-eastern United States were asked to edit a cover letter for a job-seeker called Eric Sorenson. They sent their comments to Eric's Gmail address. Eric did not exist. The reply that came back the next day was written by the researchers, and it came in two versions.

The first said: "Dear [name], I just wanted to let you know that I received your feedback on my cover letter. I was wondering if you could help with a second cover letter I prepared and give me feedback on it." The second said exactly the same, with eight words added after the first sentence: "Thank you so much! I am really grateful."

Of the 34 students who got the plain message, 11 edited the second letter, or 32 percent. Of the 35 who were thanked, 23 did: 66 percent.

That result, published by Adam Grant and Francesca Gino in the Journal of Personality and Social Psychology in 2010 under the title "A Little Thanks Goes a Long Way", has been widely quoted ever since, in management books and on onboarding slides. It is a lovely finding. It is also the kind of finding this site now checks before repeating, and this one comes with a complication that has nothing to do with gratitude.

Four experiments, one idea

The paper has four studies. The second one moved the question to a stranger. Students edited Eric's letter as before and got either a neutral acknowledgement or the thank-you. A day later a different student, Steven Rogoff, emailed asking for help with his own letter. Seven of the 28 unthanked students helped him. Sixteen of the 29 thanked students did, 55 percent against 25. Eric's thanks had spilled over onto Steven.

The third experiment left the lab. At a public university, 41 alumni fundraisers on fixed salaries were split by shift. For 20 of them, the director of annual giving came in person and said: "I am very grateful for your hard work. We sincerely appreciate your contributions to the university." The other 21 heard nothing. In the week before, fundraisers in both groups made about 40 voluntary calls apiece. In the week after, the unthanked ones averaged 41. The thanked ones averaged 63.

The fourth was the quietest. Seventy-nine students edited a letter in the lab, then an actor playing Eric walked in to drop off some forms and chatted about the weather. In half the sessions he added "Thank you for your feedback." Students who heard it spent 22.8 minutes on the next letter. Those who didn't spent 19.8.

What interested the authors was why. They measured whether the thanks made people feel more competent, whether it lifted their mood, and whether it made them feel valued. In the two studies that measured mood, it didn't move at all. Feeling more competent sometimes rose too, but it never explained the extra help. Feeling valued did, in all four studies: the helpers who were thanked felt more strongly that they had made a positive difference to the person and were appreciated by them. On that reading, a thank-you works less as praise and more as a receipt: proof that the help landed somewhere.

The complication

Francesca Gino became one of the best-known behavioural scientists in the world, a Harvard Business School professor who wrote about honesty. In 2021 the researchers behind the blog Data Colada raised concerns about her data, and in June 2023 they laid out, in public, evidence of manipulated data in four papers she had co-authored. Harvard's internal investigation found research misconduct. In May 2025 the university revoked her tenure, the first known case there since at least the 1940s. Gino denies wrongdoing and is suing Harvard, which has countersued. None of the four flagged papers is the gratitude paper.

After the scandal, a group of her co-authors set up a project called Many Co-Authors, where each person who published with Gino describes, paper by paper, who handled the data. For the gratitude paper, Adam Grant's answer is precise. Gino collected and ran experiments 1 and 2, the two with the Eric emails. Grant received SPSS files from her but, in his words, "never had access to the raw data". Gino was not involved in collecting experiments 3 and 4.

Grant also points out something any careful reader of the tables would notice. The two Gino-run studies produced the big effects, help doubling from a third to two-thirds. They also, he notes, used a stronger manipulation and a yes-or-no outcome, where the later studies measured something continuous. He wrote that he was looking into rerunning them. As of his last update, in December 2023, nothing had been published, and I haven't found an independent replication of the Eric experiments since.

Nobody has shown that anything in this paper is wrong. That needs saying clearly. But the headline numbers people quote, 32 against 66, are exactly the part that currently rests on files only one person handled, and that person's other work is now in court.

What is left standing

Take experiments 1 and 2 out and you still have the fundraisers and the actor. A 50 percent jump in calls and 15 percent more time helping. Real effects, just humbler ones, from 41 and 79 people. Samples that small could easily overstate what happens in a bigger group.

The idea is also much older than this paper. Grant and Gino cite a line of studies going back to the 1970s, among them a 1988 experiment with thank-you notes sent to case managers who visited residential clients, and a 1995 study of servers saying thank you and restaurant tips. None of that work passed through the files in question.

So the honest version is smaller than the slide-deck version. A short, specific thank-you probably makes people a little more willing to help again, and the mechanism, feeling that the help mattered, fits a lot of what we already know about why people volunteer and why they stop. Whether eight words really double anything is, for now, an open question.

On Good People Map every good deed has a thank button under it. If you were the person helped, pressing it takes one tap once you are signed in. The research, even in its humbler version, suggests the person on the other side notices, and that it matters most for the next time someone asks them. We wrote about a related finding in People Say Yes About Twice as Often as You Expect: people who ask for help underestimate how glad the helpers are to have been asked. A thank-you closes the same loop from the other end.

Sources

• A. M. Grant, F. Gino, "A Little Thanks Goes a Long Way: Explaining Why Gratitude Expressions Motivate Prosocial Behavior", Journal of Personality and Social Psychology 98(6), 946–955, 2010 — https://doi.org/10.1037/a0017935
• Many Co-Authors project, paper #123, statement by Adam Grant (last updated 26 December 2023) — https://manycoauthors.org/gino/123
• "Harvard Revokes Tenure From Francesca Gino, Business School Professor Accused of Data Fraud", The Harvard Crimson, 27 May 2025 — https://www.thecrimson.com/article/2025/5/27/gino-tenure-revoked/
• "Francesca Gino lost her Harvard tenure. Now she's taking the university to court", GBH News, 16 September 2026 — https://www.wgbh.org/news/education-news/2026-09-16/francesca-gino-lost-her-harvard-tenure-now-shes-taking-the-university-to-court