At a New York university copier, adding “because I have to make copies” raised agreement to a small favour from 60% to 93% — for a bigger favour, agreement was 24% with or without the explanation, leaving the redundant explanation no better than asking without a reason.

A request at a copier sounds mundane until the explanation adds nothing: “because I have to make copies.” In a classic field experiment, that redundant phrase accompanied a higher agreement rate for a small favor. For a bigger favor, the advantage disappeared.

Ellen Langer, Arthur Blank and Benzion Chanowitz reported the experiment in a 1978 paper. It involved 120 adults at the Graduate Center of the City University of New York. Researchers asked to copy five or twenty pages first. A favor counted as small when the requester had fewer pages than the person approached, and big when the requester had more.

Agreement was 60% without a reason and 93% with the redundant explanation for small favors; both rates were 24% for big favors. Saying the requester was in a rush produced 94% and 42%, respectively. The larger-favor contrast involving that substantive reason did not reach conventional statistical significance.

The missing condition changes the lesson

The headline includes the part that matters when someone tries to turn this finding into advice: the apparently useful wording had a boundary. Dropping that boundary would change the claim from a result in a particular comparison into something much broader.

Suppose a team tests two versions of a short request and one receives more positive responses. That is useful evidence about those versions in that setting. It does not establish that one word will persuade a different audience to accept a longer, more expensive or more consequential commitment.

The same care applies to a null-looking result. Identical observed percentages show no advantage in that sample. They do not prove that the underlying effect must be exactly zero in every possible setting. The distinction is between describing what happened and claiming a universal rule.

A familiar result deserves its unfamiliar qualification.

People also misjudge the power of asking

Separate research by Francis Flynn and Vanessa Lake, published in 2008, examined people’s predictions about requests for help. Across a series of studies, people underestimated how often others would agree to direct requests.

In Stanford’s account of the work, Flynn described the practical consequence: anticipating rejection can discourage people from requesting assistance that others would actually provide. One study also found that requesters mistakenly expected an indirect approach to be more effective than a direct question.

This addresses a different problem from the copier wording. A person may never reach the point of choosing an explanation because they have already decided that asking would be pointless. Predictions about another person’s response and the response itself are separate things, which researchers can compare.

The communication channel changes the encounter

A later study by Vanessa Bohns and M. Mahdi Roghanizad tested requests to proofread half a page of text across communication channels. The research involved 490 people making requests and 1,490 respondents.

In-person requests elicited more agreement than mediated requests, while voice and video performed better than text-based channels in the reported comparisons. The findings did not establish a clear advantage for video over audio alone. People making the requests underestimated the advantage of asking face to face.

That is relevant to anyone carrying an old persuasion finding into an email campaign or a product interface. A spoken interaction beside a machine includes timing, proximity and another person waiting for an answer. Changing the channel changes the situation, even if the sentence stays the same.

A yes can mean more than one thing

In 2024 research reported by Cornell, Rachel Schlund, Bohns and Roseanna Sommers examined how to make refusal feel easier. Their two studies involved 535 participants and included a request to hand over an unlocked phone.

Giving people specific words they could use to decline increased their reported freedom to refuse more effectively than simply telling them they were allowed to say no. The intervention addressed the difficulty of finding an acceptable response in the moment.

This points to a useful distinction for workplace communication: obtaining agreement and obtaining freely given agreement are different outcomes. Counting only affirmative responses can miss how the recipient experienced the request. The newer studies did not replicate the copier experiment; they asked another question about the social interaction surrounding a request.

Make the explanation do real work

For a manager or product team, a practical approach is to state the request clearly, explain the actual reason and describe what agreement entails. That gives the recipient information with which to make a decision, while giving the requester a more interpretable response.

If wording is tested, the size of the commitment should be part of the design. A request for a minute of attention, an afternoon of work and an ongoing obligation should not be treated as equivalent simply because all three end with the same call to action.

The copier finding is memorable because the explanation was redundant. Its limits are memorable for a better reason: the amount being asked of the other person still mattered.