Even a penny worked. Even a dollar did not.

Five words on a doorstep lifted giving from 28.6 to 50 percent, and swapping the penny for a dollar made the effect disappear. The money never followed.

Canvassers knocked on 84 doors in a middle-income American suburb, collecting for the American Cancer Society. Forty-two households got the standard request for a donation. The other forty-two got the same request with one sentence added to the end.

Even a penny will help.

Twelve of the standard households gave something — 28.6 percent. Twenty-one of the penny households gave, exactly half. And the people who gave under the penny wording did not give less. A dollar was the median and the most common gift in both groups, the mean per donor came out at $1.54 against $1.44, and the difference was small enough that the statistics would not take it seriously. The penny canvass came home with $30.34. The control canvass came home with $18.55.

Robert Cialdini and David Schroeder published that in 1976. It has been quoted at fundraisers ever since, usually without the second experiment, which is the one worth reading.

Even a dollar did nothing

They ran it again with four conditions. Control. "Even a penny will help." A third wording that made pennies acceptable without sounding needy — "we've already received contributions ranging from a penny on up." And a fourth that changed a single word: even a dollar will help.

Control brought in 32.2 percent of households. Even-a-dollar brought in 46.7 percent, which looks like a gain until you test it — then it is not reliably different from saying nothing at all. Even-a-penny brought in 58.1 percent. The wording about other people's pennies brought in 64.5 percent, the best of the four.

Be careful with that dollar result. About thirty households per condition is not much to detect a difference with, and failing to find one is weaker than showing there is none. What it has going for it is that the authors predicted the null in advance and said why.

Now notice what a dollar was. It was the median gift in the first experiment. The sentence that failed is the one that named the amount people were already handing over.

That is the part worth carrying around. The technique is not leniency. Lowering the ask to a sensible floor does nothing measurable. It has to drop far enough that refusing becomes awkward to explain — to the person standing on the step, and rather more to yourself.

Forty-one years later

In 2017 San Bolkan and Stephen Rains gathered every test of the technique they could find and ran the totals. Thirty-four effects, 3,181 people.

The first half of the 1976 result survived. Legitimizing a paltry contribution does raise the share of people who comply, r = .22 — a correlation, not a gain of twenty-two percentage points — which in this literature counts as a real and repeatable thing rather than one lucky afternoon on one street.

The second half didn't. Across the eleven studies that recorded how much people actually handed over — 1,531 donors — the penny sentence pushed individual gifts down, r = −.23. Cialdini and Schroeder got extra donors at no cost per donor. Everybody else, averaged together, got extra donors who each gave less.

The two effects are nearly the same size pointing in opposite directions, and the meta-analysis states the consequence without drama: total amounts raised came out similar to the control conditions.

So what does the sentence buy

What it buys is participation, not money.

Bolkan and Rains tested three explanations and two held up. One is impression management: once a penny has been declared acceptable, "no" stops being a practical answer and starts being a statement about the kind of person you are. The other — perceived need — is simpler still. An organization that will take a penny reads as an organization that is short.

That second explanation is the one Cialdini and Schroeder thought they had ruled out. Their third condition existed precisely to separate the two, and it worked: legitimizing pennies without sounding desperate produced the highest compliance in the study, and the need ratings collected afterwards did not differ between groups. Four decades of accumulated data now say the desperation reading is doing some of the work after all. Two careful analyses, both defensible, disagreeing about one sentence.

Either way, both explanations predict the smaller gift just as neatly as they predict the extra donor. If you gave because refusing had turned awkward, you gave what it took to end the awkwardness. The sentence that opened the door also named the price.

Whether that is a good trade is not a question the data answers. A campaign that needs a number by Friday should probably leave it alone. A campaign that wants a thousand people to have given something once, and to think of themselves afterwards as someone who gives, is buying exactly what it came for.

What I would not take from this

Eighty-four households, one suburb, 1976, a human being physically standing there. Almost every feature of that scene is missing from a donation page. The meta-analysis generalizes it, but it generalizes across studies that are still mostly students and mostly Western.

And the amount effect rests on eleven studies. That is enough to stop repeating the 1976 summary — the sentence raised compliance "while leaving unaffected the size of the donations" — as if it were settled. It is not enough to plan a year of fundraising around.

Where this leaves the map

Good People Map asks for money with five buttons: 5, 10, 25, 50, 100 euro. Underneath them sits a field that accepts any figure down to one euro. So the map does have its penny. It is simply not on a button.

Which is the whole gap the 1976 experiment lives in. The field is what the site will accept. The buttons are what the site is asking for, and the buttons are what a visitor actually reads. The lowest one tells them who the page is for. Five euro says most people. Twenty-five would say something narrower, and it would say it silently, to everyone who was never going to write in and complain.

Worth choosing on purpose, then, rather than inheriting from a template.

On the other side of the same question, whether a bigger match brings in more money, see Matching donations worked. Tripling the match did not.

Sources: Robert B. Cialdini and David A. Schroeder, "Increasing Compliance by Legitimizing Paltry Contributions: When Even a Penny Helps," Journal of Personality and Social Psychology, vol. 34, no. 4 (1976), pp. 599–604 — all figures read from the paper itself. San Bolkan and Stephen A. Rains, "The Legitimization of Paltry Contributions as a Compliance-Gaining Technique: A Meta-Analysis Testing Three Explanations," Communication Research, vol. 44, no. 7 (2017), pp. 976–996. Checked 7 September 2026.