The Same Sign, Different Situation

The field evidence for social-norm messaging is most revealing when one message family is carried into several real settings. This comparison focuses on what the “Be like Toni” intervention changed in restrooms, a train-station café, and a bakery, and on what the contrast can—and cannot—tell us about context dependence. The practical judgment is not that a sign works in some places and fails for mysterious reasons; it is that attention, opportunity, competing motives, and the form of the decision help determine whether a norm becomes an observed action.

Three calm settings made the nudge look portable

Mundt and colleagues tested a simple visual intervention in five field experiments involving 1,163 people. The intervention used a stick figure named Toni and the familiar “Be like Bill” social-media format as its visual grammar. The wording paired a descriptive norm—what Toni does—with an injunctive cue that Toni is environmentally responsible, socially considerate, or otherwise “cool.” The sign then gave a concrete instruction: use one paper towel, go without a plastic lid, or wear a mask. Participants generally did not know they were taking part in an experiment, and the comparison was between a nudge condition and a no-intervention control.

The first test took place in a women’s restroom at a German university learning center. The behavior was paper-towel use after handwashing, a choice that could be made immediately at the dispenser. Among 179 observed users, the nudge condition used fewer towels than the control condition. The reported effect size was d = 0.48, and the difference was statistically significant. This is a direct behavioral result in a real setting, not merely an intention to conserve resources. It is also a low-friction result: the intervention did not ask people to buy something, rearrange a schedule, or accept a visible health risk. It asked them to take fewer sheets at the moment of use.

The second test kept the behavior and the message family largely constant while moving to a women’s restroom at a German Christmas market. This produced a more heterogeneous group in age and educational background than the university sample. Among 183 observed users, those who encountered the sign again used fewer paper towels, with a smaller effect size of d = 0.32. The replication matters because it reduces the chance that the first result was a one-location accident. The smaller effect matters just as much: moving the intervention changed more than the address, and the result did not travel with exactly the same strength.

The third test returned to the university restroom and changed the wording rather than the behavior. It compared a prescriptive message—Toni takes one towel; take one—as well as a proscriptive message—Toni takes more than one; do not do that—with a no-message control. Both messages reduced paper-towel use relative to control. The prescriptive comparison produced d = 0.46 and the proscriptive comparison d = 0.40.