Identifiability

What it is

In a group task there are two ways a person's contribution can disappear. It can be mixed into a shared total, so that nobody can say which part was theirs, or it can be visible but never set against anyone else's, so that it carries no information about how the person is doing. Identifiability concerns the first. Comparability, which the researchers treat as the possibility of evaluating one person's output against co-workers', concerns the second.

Harkins and Jackson separated the two. Participants brainstormed in four-person groups. In one manipulation each participant's output was either individually identifiable or pooled with the others'. In a second, crossed with the first, participants were led to believe that their performance either could or could not be compared with their co-workers'. The experiment therefore had identifiable and pooled outputs in both a comparable and a non-comparable setting.

The result was specific. Identifiable outputs produced more ideas than pooled outputs, replicating earlier work on social loafing, but only when participants believed their outputs could be evaluated against others. When outputs could not be compared, identifiability made no difference to the number of ideas. The authors concluded that eliminating social loafing takes both conditions together.

The idea has an earlier history. A 1981 paper by Kipling Williams, Stephen Harkins and Bibb Latane carried the title Identifiability as a deterrent to social loafing: Two cheering experiments, which frames identifiability as something that can deter the effect. This page records the title only, and attributes no finding to that paper.

In effect

The 1985 result changed what identifiability was taken to mean. Being countable is not sufficient when the count has nothing to be measured against. A total that sits alone in a ledger tells a person little about how they are doing, and so gives them little reason to change what they do.

The same pair of conditions appears in the later meta-analysis by Steven Karau and Kipling Williams, which pooled 78 studies of social loafing. They reported that loafing was greatest in studies where a person's output could be evaluated when working alone but not when pooled, and that the difference between working alone and working together largely disappeared where output could be evaluated in both settings. The meta-analysis spans many tasks, and it places the brainstorming result inside a wider pattern.

Karau and Williams account for the pattern with the collective effort model, in which people put in effort when they expect it to lead to a performance, expect the performance to lead to an outcome, and value that outcome. On that account, evaluation restores the first link, because it lets effort be traced to a person. The arrangement of work is then a matter of design, which is the reading that Colin Fisher gives in The Collective Edge when he argues that task design is the most neglected lever available to teams.

What it does not say

The 1985 experiment used one task, brainstorming, in groups of four, so it shows the logic of the two conditions and not how large their effect would be in a workplace or a classroom. The page rests on the experiment's abstract and on the meta-analysis, and it states only what those support.

Identifiability is also not the whole account of effort in groups. The meta-analysis reported that task meaning, the value people place on the group and the uniqueness of each member's input were related to loafing as well, and a single condition should not be read as the cause. A survey of junior-league soccer players by Rune Høigaard and colleagues, which measured perceived loafing and not actual effort, links cohesion and team norms to less perceived loafing without involving identifiability at all.

---