What a Leadership Assessment Should Measure

What a trait score predicts on its own

Timothy Judge and Cindy Zapata pooled 125 studies of personality and job performance in 2015. Before testing anything about the jobs, they estimated how well each of the five broad traits predicts performance across all of them. Conscientiousness came first, at a corrected correlation of .21. Emotional stability and extraversion were near .12 and .11, and agreeableness and openness were close to zero.

An average across 125 studies hides the variation that matters. The same paper found that all five traits predicted performance better where the job left people free to decide how to do the work. Several also did better in jobs that called on that particular trait: extraversion where the work needs social skill, openness where it needs new ideas. So a trait score is half of a measurement, and the job is the other half.

Measure the conditions where the work is done

P×E measures two conditions of a role. Threat is how serious the consequences are when something goes wrong. Stability is whether rules, priorities and relationships hold steady long enough for practiced methods to keep working. High or low on each gives the four arenas: Chaotic, Crucible, Clouded and Cultivated.

Industry is too coarse a unit for either one. The US Department of Labor’s O*NET database scores occupations from 0 to 100 on how serious an uncorrectable mistake would usually be. In a single hospital, emergency medicine physicians score 98 and medical and clinical laboratory technologists 75, while medical and health services managers score 35.

O*NET has a measure for threat and none for stability. Stability has to come from the people doing the work, and a senior leader and a front-line employee may describe the same unit differently. When they do, record both answers.

A description can feel accurate whether or not it is

In 1949 Bertram Forer gave 39 psychology students a personality test and, a week later, a written sketch of each one’s personality. He asked them to rate how well it described them, from 0 to 5. The average was above 4. Every student had received the same sketch, which Forer had put together from a newsstand astrology book.

Feeling recognized by a report is weak evidence that the report is right. A personality measure describes recurring tendencies. Judgment, technical skill and the ability to learn a new role need other evidence: work the leader has done, decisions they made, and people who watched them under similar demands.

What the report should hand back

An assessment is finished when it gives the leader something to test. Each P×E recommendation names one behavior to try, the measure that will show whether it held, and a date to review it. It also states, before the leader starts, the signal that would show the plan is wrong. A recommendation cannot be produced without a stated limitation.

The reasoning runs in five steps, and each is labelled by where it comes from. Two are measured: the leader’s profile and the condition of the team. Three are hypotheses: the behavior, how the team will respond, and the result that should follow in the work.

P×E also computes a single fit score, and it does not show it. No validated coefficient for leader-arena fit exists, and a single number would invite readers to rank people by it.

Where an assessment stops

An assessment narrows the uncertainty about a leader in a role. The work then supplies better evidence, and the reading should change when it does. P×E supports reflection and development. It is not a validated hiring, promotion, or performance-prediction tool.