Ergodicity (or its absence)
Why a group average rarely tells us what one person needs.
Ergodicity is a term from statistical physics that has, by now, become of considerable significance for psychology — chiefly through what it turns out not to be.
A system is called ergodic if the behaviour one observes over time in one element shows the same pattern as the behaviour one observes at one moment across many elements. In other words: the average patient and the patient-over-time exhibit the same dynamics. That would be practical indeed — group research would translate directly into individual practice.
The problem is that psychological systems are almost never ergodic.
Molenaar's reversal
The Dutch methodologist Peter Molenaar (2004) demonstrated this in a landmark paper. What holds for the group rarely holds for the individual. A well-known example: in group studies, typing speed and typing accuracy are positively correlated — those who type faster type more accurately. But when one follows a single individual at length and measures her speed and accuracy, the correlation within that person often proves negative: as she types faster, accuracy declines. The group picture and the individual picture tell opposite stories.
"The assumption that the psychological dynamics of individuals equals the dynamics of the group average — that is the assumption of ergodicity. Almost never true." — after Molenaar (2004)
For mental health care this is a fundamental problem. Nearly our entire evidence-based edifice rests on randomised controlled trials — group studies showing that treatment X is on average superior to treatment Y. But the group outcome does not tell us which patient benefits from X, which from Y, and which from neither. And within a single person, the relation between factors (sleep and mood, movement and anxiety) may follow a pattern different from the group's.
The idiographic turn
The answer to non-ergodicity is not to abolish group research, but to complement it with research at the level at which we also act: the individual. Idiographic research — one person, many measurements, over time — shows how a specific patient functions, which variables in his or her system relate to one another, and how that relation changes. It is the logical counterpart of the group research we already have.
In practice this means that time-series measurements (ESM, diaries, IROC, repeated questionnaires) are not a luxury or by-product of treatment, but a methodological necessity. Only in such measurements does the dynamics that truly apply to this patient become visible.
The practical implication for the consulting room is mild but important. We may be wary when a study claims "this works for patients like these" — for patients like these do not exist as a category within which validity is guaranteed. We must look at how this specific system responds.