Descriptive and Injunctive Norms

What it is

P. Wesley Schultz, Jessica M. Nolan, Robert B. Cialdini, Noah J. Goldstein and Vladas Griskevicius ran the field experiment that makes the distinction matter, on household energy conservation, in which a descriptive normative message detailing average neighbourhood usage produced either desirable energy savings or the undesirable boomerang effect depending on whether the household was already a below-average or an above-average consumer, and adding an injunctive message conveying social approval or disapproval eliminated the boomerang; the household figure that circulates for that study could not be confirmed and is not printed here.

Noah J. Goldstein, Cialdini and Griskevicius's hotel towel experiments are the best-known descriptive-norm demonstration, and in the first, across a hundred and ninety rooms, the descriptive-norm sign produced a reuse rate of 44.1 per cent against 35.1 per cent for the standard environmental appeal, across four hundred and thirty-three observations, while in the second the descriptive-norm conditions combined reached 37.2 per cent against the standard appeal across one thousand five hundred and ninety-five, and the provincial norm, naming guests in this room, reached 49.3 per cent against 42.8 per cent for the other descriptive conditions combined. Those differences amount to about nine rooms in a hundred and then about six, which is worth saying in words because the figures are rarely printed alongside the claim.

In effect

The boomerang is the honest complication and the reason the distinction earns its keep, because a message reporting the average pulls heavy users down and light users up, since the average is a target as well as a description, and attaching approval to the behaviour rather than to the average is what stopped the second half of that happening. Hunt Allcott's scaled version shows what the mechanism does at national volume, as Home Energy Report letters comparing a household's electricity use with its neighbours', across randomised natural field experiments at six hundred thousand treatment and control households, produced a 2.0 per cent reduction in consumption, which the author states is equivalent in effect to a short-run price increase of eleven to twenty per cent. The heterogeneity is the interesting part, because the highest-consumption decile cut usage by 6.3 per cent and the lowest by 0.3 per cent. The design rule that falls out of this is simple: if a descriptive norm is used on a population that straddles the average, the norm will move some of it the wrong way unless something injunctive is attached.

What it does not say

It does not say descriptive-norm messages beat standard appeals, because Gerd Bohner and Lena Erika Schlüter's replication, across two German hotels and nine hundred and twenty-eight guests, found that both standard and descriptive-norm messages increased reuse compared with a no-message baseline, that descriptive-norm messages were not more effective than the standard message, and that effects of proximity were inconsistent across their two studies, so both results belong in the same passage and the honest position is that the descriptive norm's advantage is unsettled. The famous effects are small and the literature they sit in is selected, since Stefano DellaVigna and Elizabeth Linos, comparing a hundred and twenty-six trials covering twenty-three million people run by two nudge units against nudge trials in academic journals, found 1.4 percentage points against 8.7 with about seventy per cent of the gap attributed to selective publication amplified by low power, and every figure above carries that discount. Allcott's 2.0 per cent does not generalise beyond Home Energy Report letters to residential electricity customers, and should not be mixed with the figure from his earlier working paper on a different and much smaller sample. And the two norm types are not cleanly separable in a real message, because a card reporting that seventy-seven per cent of residents do something also implies approval, which means no field experiment in this record manipulates one while holding the other fixed.

---