Why averages lie
Ergodicity, idiographic research, and what they mean for evidence-based practice.
In 2004 the Dutch methodologist Peter Molenaar published an article with a title that needed no commentary: A manifesto on psychology as idiographic science. His point was simple and, in our view, far-reaching. The way psychology had been doing its science — by studying groups of people and generalising their averages to the individual patient — was based on an assumption that proved to hold in virtually no psychological system.
That assumption is called ergodicity. It is a term from statistical physics. A system is called ergodic if the behaviour you observe over time in one element shows the same pattern as the behaviour you observe at one moment across many elements. In other words: the average patient and the patient-over-time exhibit the same dynamics.
That would be very practical. Then an RCT showing that treatment X works on average better than treatment Y would translate directly into policy for the individual patient: give X. But the assumption almost never holds. Psychological systems are not ergodic.
The typing-speed example
Molenaar used a simple example that has convinced many. In group studies, typing speed and accuracy are positively correlated: those who type faster type more accurately. Logical — those who type well can be both fast and precise. But when you follow one person over time, and measure her speed and accuracy session after session, the correlation within that person often proves negative: as she types faster, accuracy declines. The group picture and the individual picture tell opposite stories.
This is no exception. In virtually every psychological domain in which it has been measured — mood, behaviour, cognition, personality — one finds the same. Group statistics do not tell us what happens within one person.
"The assumption that the psychological dynamics of individuals equal the dynamics of the group average — that is the assumption of ergodicity. Almost never true."
What this means for evidence-based practice
Here it grows uncomfortable. Nearly our entire evidence-based edifice rests on group studies. RCTs, meta-analyses, treatment guidelines, based on what works on average for average patients. If those averages do not translate to the individual, what then are those guidelines actually worth?
The honest but nuanced answer is: they are not worthless, but neither are they what we sometimes pretend they are. An RCT tells us that in this population, treatment X is on average better than Y. What it does not tell us is in which patient X works, in which Y, and in which neither. It gives us a hypothesis, not a prescription.
That is a more important distinction than it seems. A hypothesis calls for testing on the specific patient. A prescription feels like an obligation, apart from the patient. The difference lies in our freedom to perceive, with our clinical eyes, whether the hypothesis is borne out in this patient.
The idiographic turn
The answer to non-ergodicity is not to abolish group research. It is 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 applies to this patient appear.
The technical tools for such research have grown more accessible in recent years. R-packages such as mlVAR, qgraph and bootnet make it possible to estimate within-person networks. Apps for ESM (PsyMate, Ethica, mEMA) make it feasible to collect data without much investment. What for a long time only researchers could do, clinicians can now do too — though it does call for some training.
What it does in the consulting room
The largest change that idiographic work brings is perhaps cultural. We are used to determining for our patients what is going on — based on diagnoses, based on their account, based on our experience. Idiographic work invites us to investigate that together with them. The questionnaire she fills in twice a day for six weeks is no piece of paperwork. It is her own data, her own story, her own pattern becoming visible.
Many patients experience this as a relief. At last it becomes concrete. At last there is something on paper that represents them, not only a classification of what they "have". And sometimes they discover in it things that neither they nor their clinician would have seen — a relation between social activity and mood that runs the other way than expected, a day of the week when things systematically go wrong, a pattern of flickering just before relapse.
An honest question
Does this break evidence-based thinking? No. Does it make our work harder? Yes. It brings into view a complexity that we could until now partly ignore. But it also gives us new tools. It makes our interventions better targeted. And it gives us and our patients a form of honesty: what we know is what we know of groups, and what we do not know is what is going on in this person — until we investigate it together.
Molenaar's manifesto is from 2004. Almost twenty years later it is only really beginning to sink in. Not as revolution, but as gradual recalibration. Just as it should be, in a complex system.