In brief
- Whether support and prevention are targeted at the neighbourhood or at the individual institution decides two things: how accurately resources land, and whether a school or nursery recognises itself in the numbers and acts on them.
- For an individual setting, two quantities count: who it actually serves, and the quality of what happens inside it. Neither is captured by an area index. The social environment of the youngest children, on the other hand, is described reasonably well by the CTC youth survey of their older siblings and neighbours.
- First step for a CTC community: where primary-school and early-years catchments coincide, break the survey results down to that level; otherwise use the composition of each institution as the interface between the survey and any funding formula.
- And keep the strategic question open: targeting the most burdened institutions makes sense, but it does not replace a universal base. The two belong together, in proportion to need.
How do you distribute resources so that they reach the children who need them most: measured by the , or by the individual institution? Every European country that funds schools or early-years settings according to social need has to answer this question, and the answers differ. England’s pupil premium is paid to schools per eligible pupil, so it follows the child’s circumstances into whichever school the child attends. The Netherlands has, since 2019, allocated its educational-disadvantage funding for primary schools through a school-level indicator that the national statistics office computes from the characteristics of each school’s pupils. Many other systems, wholly or in part, still work with area-based deprivation indices, because those are what the statistics readily provide.
Behind the funding technicalities lies an old question of social epidemiology: at which level does data-driven prevention recognise need? Communities That Care is addressed by it too. Not one to one: a social index is an allocation instrument, CTC an entire prevention process that measures need, sets priorities and selects, launches and funds programmes. But both stand before the same choice of unit.
The case for looking at the institution
Schools and early-years settings in most European cities do not have a fixed catchment. Where parents can choose, the social composition of a school diverges from that of its surroundings, and the research on school segregation shows that in many European cities schools are more segregated than the neighbourhoods they sit in, with the size of the gap depending on how much choice the system allows (Boterman et al., 2019). Early-years settings, which are chosen even more freely, are no different.
Picture two nurseries in the same neighbourhood, behind the same unemployment rate: one serves almost only children from precarious households, the other mostly children of newly arrived professional families. A purely area-based index gives both the same score and the same funding. The risk has a name: misallocation. An institution-level indicator, built from who actually attends, avoids it.
The argument has honest limits. An index of attendance does not reach the children who never arrive, and across Europe it is precisely disadvantaged families who use early-years provision less (Vandenbroeck & Lazzari, 2014). And more money is not yet better pedagogy: what a setting does with additional resources is decided by its process quality, the daily interactions between staff and children, which no social index measures (Edwards, 2021).
| Unit | Typical data source | Strength | Risk |
|---|---|---|---|
| Area-based (neighbourhood) | Unemployment or benefit-receipt rate of the area | Available everywhere; also says something about families without a place in any setting | Blurred for the individual institution; ecological fallacy; misallocation |
| Institution-based (the individual school or nursery) | Pupil-level administrative data, school-entry health checks, benefit rates matched to home addresses | Reflects the real composition; fairer from the institution’s point of view | Data-intensive; sensitive under data-protection law; reaches only children who attend; does not measure process quality |
An old question: the ecological fallacy
The objection is not an administrative detail but a classic of social epidemiology. From an area characteristic such as a neighbourhood’s benefit-receipt rate one cannot reliably infer anything about individual people; that is the ecological fallacy (Lancaster & Green, 2002). The social structure of a district describes the context validly. It does not say whom a particular institution in that district serves. That is exactly where the institution-level argument is strong.
For CTC this line of thought is familiar. The anonymous youth survey measures and in family, school, peer group and community, using scales that have been validated in the United States and in European adaptations (Arthur et al., 2002; Reder et al., 2024). The whole point of the instrument is that risk and protection vary from one community to the next and that those local levels predict local outcomes, which is why the survey area and the area a coalition acts on must be the same (Hawkins et al., 2004). In practice the results are broken down to small units, by year group, by school and by neighbourhood, and school-level adaptations of the model work at the level of the individual institution by design. CTC, in other words, already zooms in. For the youngest children, whom no youth survey reaches, that picture describes the environment approximately, because the families are the same families. What remains specific to the individual nursery is its composition, who it serves, and its process quality. Neither is supplied by an index over the area.
The real dispute: targeted or universal?
The harder question is not the level of measurement but the strategy. Directing extra support at the few most burdened institutions is intuitive and feels fair. But Geoffrey Rose’s prevention paradox is a warning: more cases usually come from the broad middle of a distribution than from its extreme tail, and whoever picks out only the peaks misses the majority of disadvantaged children, who are spread across the inconspicuous institutions (Rose, 1985). Add to that the price of the label, a stigma that tends to deter people from seeking help rather than encourage it (Rüsch & Thornicroft, 2014). CTC is itself a universal community system, and the case-finding logic of a social index stands in genuine tension with it.
The tension is not resolved by choosing a side, but it is resolved by Michael Marmot’s proportionate universalism: a universal base for everyone, with an intensity that rises in proportion to need (Carey et al., 2015). That is compatible with both positions, with a funding index read as a graded scale rather than a threshold, flanked by universal measures such as access and early family support, and with CTC, which combines a universal system with precisely targeted programmes. Social inequality in children’s learning emerges early and then stays remarkably stable through the school years, as a German cohort followed from infancy to age sixteen has shown (Skopek & Passaretta, 2021). Early support with a universal base and more where more is needed is the approach most likely to meet it.
Why proximity works, and what it costs
There is a second reason the unit matters, beyond accuracy. The closer the data sit to an institution, the more readily it recognises itself in them, and the more readily action follows. School improvement research knows this: data feedback changes practice only once “the data” become “our data” (Geijsel et al., 2010). Aggregated neighbourhood figures are about “the area”; institution-level figures are about “us”. Whether the same holds for early-years settings has not been demonstrated, but it is a reasonable expectation, and a testable one. A CTC community could compare whether an institution moves faster after feedback at its own level than after a bare neighbourhood figure.
That proximity has a price. Matching the home addresses of individual children to small-area benefit rates, as some institution-level indicators do, is delicate under data-protection law, because it brings sensitive characteristics down to a level at which single families can be affected. The same care applies to CTC itself: the more finely an anonymous survey is broken down, the sooner small cell sizes can make individuals identifiable. Accuracy has to be weighed against privacy, and all the more in a prevention system that lives on anonymity and data minimisation. The CTC youth survey reports results only at the level of a school year or an area, never for individuals or classes, for exactly this reason.
Where this leaves us
Two things should be kept open. The view of the institution does not replace the view of the neighbourhood: a nursery can be a point of contact for families in the area even for children who do not attend it, and for that the surroundings count. And an index is no automatism: whether it creates any incentive at all depends on how it is calculated, and most indices rank institutions relative to one another rather than measuring absolute need.
The productive core remains. “Neighbourhood or institution” is a design question for every data-driven prevention effort, not a special problem of early-years funding. For CTC communities it means, in practice: where primary-school and early-years catchments coincide, the survey results can be broken down to that level; where they do not, the composition of each institution is the shared interface between the survey and any social index. And it means naming the limits of each unit openly, which is the least that a community owes the institutions it asks to act on its numbers.
References
Arthur, M. W., Hawkins, J. D., Pollard, J. A., Catalano, R. F., & Baglioni, A. J. (2002). Measuring risk and protective factors for substance use, delinquency, and other adolescent problem behaviors: The Communities That Care Youth Survey. Evaluation Review, 26(6), 575–601. https://doi.org/10.1177/0193841X0202600601
Boterman, W., Musterd, S., Pacchi, C., & Ranci, C. (2019). School segregation in contemporary cities: Socio-spatial dynamics, institutional context and urban outcomes. Urban Studies, 56(15), 3055–3073. https://doi.org/10.1177/0042098019868377
Carey, G., Crammond, B., & De Leeuw, E. (2015). Towards health equity: A framework for the application of proportionate universalism. International Journal for Equity in Health, 14, 81. https://doi.org/10.1186/s12939-015-0207-6
Edwards, S. (2021). Process quality, curriculum and pedagogy in early childhood education and care (OECD Education Working Papers No. 247). OECD Publishing. https://doi.org/10.1787/eba0711e-en
Geijsel, F. P., Krüger, M. L., & Sleegers, P. J. C. (2010). Data feedback for school improvement: The role of researchers and school leaders. The Australian Educational Researcher, 37(2), 59–75. https://doi.org/10.1007/BF03216922
Hawkins, J. D., Van Horn, M. L., & Arthur, M. W. (2004). Community variation in risk and protective factors and substance use outcomes. Prevention Science, 5(4), 213–220. https://doi.org/10.1023/B:PREV.0000045355.53137.45
Lancaster, G., & Green, M. (2002). Deprivation, ill-health and the ecological fallacy. Journal of the Royal Statistical Society: Series A (Statistics in Society), 165(2), 263–278. https://doi.org/10.1111/1467-985X.00586
Reder, M., Runge, R. A., Schlüter, H., & Soellner, R. (2024). The German Communities That Care Youth Survey: Dimensionality and validity of risk factors. Frontiers in Public Health, 12, 1472347. https://doi.org/10.3389/fpubh.2024.1472347
Rose, G. (1985). Sick individuals and sick populations. International Journal of Epidemiology, 14(1), 32–38. https://doi.org/10.1093/ije/14.1.32
Rüsch, N., & Thornicroft, G. (2014). Does stigma impair prevention of mental disorders? The British Journal of Psychiatry, 204(4), 249–251. https://doi.org/10.1192/bjp.bp.113.131961
Skopek, J., & Passaretta, G. (2021). Socioeconomic inequality in children’s achievement from infancy to adolescence: The case of Germany. Social Forces, 100(1), 86–112. https://doi.org/10.1093/sf/soaa093
Vandenbroeck, M., & Lazzari, A. (2014). Accessibility of early childhood education and care: A state of affairs. European Early Childhood Education Research Journal, 22(3), 327–335. https://doi.org/10.1080/1350293X.2014.912895