Social loafing

Why does each person pull less hard on the rope as more hands join the tug-of-war?

Eight people on a rope each pull less than one person on a rope. Ringelmann reported something like that around 1913 — eight producing roughly half of what their individual pulls summed to — though that figure comes from a thin demonstration mostly known through later retellings, so take the direction and not the magnitude.

The obvious reading is that people slack when they can. But a rope hides a second explanation entirely: eight people cannot pull in exactly the same direction at exactly the same instant, so some of the missing force was never withheld, only cancelled. The observation fits a story about motivation and a story about geometry, and the first job is to tell them apart.

Ingham and colleagues did it in 1974 with a blindfold. Participants pulled believing either that they pulled alone or that others pulled behind them; nobody else ever pulled. There was no one to be out of phase with, so coordination loss was zero by construction — and effort still fell in the believed-group condition. What remains once geometry is removed is motivational, and produced by a belief rather than by a crowd.

So people supply less effort when they think it is pooled. Now the real question, since "people slack in groups" restates the finding rather than explaining it: why would effort fall?

The first candidate is economic and respectable. If my effort raises a reward everyone shares, I bear the whole cost and receive a fraction of the benefit, and that fraction shrinks as the group grows. Under-supplying is then the arithmetic working correctly, not a failure of character. Call it free-riding.

It predicts something specific: no shared reward, no loss. Latané, Williams and Harkins had people shout and clap as loudly as they could, alone and in groups, with nothing to win. There was no pooled payoff to divide — and output per person still fell, further in larger groups. Free-riding cannot be the mechanism there, because the quantity it operates on was absent.

What else changed when the group formed? Consider what can be measured. Shouting alone, my contribution is the reading on the meter. Shouting in a group of six, the meter shows a total and my share of it is unrecoverable. My effort has become invisible — not to me, but to anyone who might form a judgement of me.

That gives a second candidate: effort falls when individual contribution cannot be evaluated. The two can be moved independently, which separates them cleanly. Keep the payoff pooled and the group intact, but restore measurement of each person's output. Free-riding says nothing should change; the evaluation account says the loss should largely go. When Williams, Harkins and Latané told participants their individual shouting was being measured within the group, the loafing substantially disappeared.

So there are two distinct causes, separable by which knob you turn, and both are real. Money and grades bring the first into play; any pooled, unmeasurable effort brings the second, whether or not anything is at stake.

Can we break the account? It predicts a limit case that sounds absurd: if the mechanism is about being unevaluated rather than about groups, adding people should sometimes make an individual work harder. That is social compensation — someone who cares about the outcome and expects poor contributions from co-workers raises their own effort above what they would give alone. Karau and Williams's 1993 meta-analysis assembled the boundary conditions: the loss shrinks or reverses when the task is meaningful, when the person believes their contribution is distinct and needed, and when they value the group. An account predicting loafing everywhere is refuted by those reversals; one built on evaluation and dispensability predicts them.

Three refinements tie the pieces together.

First, the neighbours in this collection. The bystander-effect piece turns on diffusion of responsibility — an obligation divided among witnesses — close kin to free-riding but concerning a discrete act rather than a graded quantity of effort. The group-projects piece asks why a group's output can fall below its best member's, which needs an argument about how the pieces combine as well as about effort. My fixed point is narrower than either: the dissociation between the payoff knob and the measurement knob, and the finding that the second does most of the work in the classic experiments.

Second, the vocabulary. "Loafing" implies idleness and describes this poorly. The reductions are modest, unconscious, and reported by participants who believed they had pulled or shouted flat out. What fell was the automatic calibration of how hard is hard enough, which in ordinary life is set partly by being seen.

Third, the honest caveat: the effect is well replicated in direction and unreliable in size. The percentages quoted for output per person at various group sizes come from small laboratory studies of shouting and rope-pulling in the 1970s, and I will not present one as though it generalised to a work team. The meta-analytic moderators — meaningfulness, uniqueness of contribution, group valuation, and reported but contested cultural and gender differences — vary enough that the practical claim is a mechanism with a large error bar, not a coefficient.

• Whether making contributions visible recovers effort or merely moves the slack to whatever remains unmeasured.
• Why the same person loafs on one task and compensates on another, and what distinguishes the two.

Every term the collection defines is gathered in the glossary.