Evidence registers

The US register of tested programmes for children and young people, run by the Institute of Behavioral Science at the University of Colorado Boulder. It certifies on three tiers. Promising asks for one randomised controlled trial or two quasi-experimental evaluations, with a consistent positive effect and no evidence of harm. Model adds a second high-quality trial and at least one follow-up twelve months or more after the intervention ends. Model Plus adds a replication carried out by researchers from outside the developer’s team. These are the standards of evidence the CTC European Network review applied when it assessed European programmes for Xchange.

An observational design: a group is followed over time and then distinguished by whether it received the programme or not; nobody is assigned. A prospective cohort starts observation in the present and runs into the future, a retrospective one works through a course of events that has already finished. Because the groups can differ from the outset, the risk of hidden confounders stays higher than in a randomised trial.

The working body of a CTC process: practitioners who run schools, youth centres, family services and clubs, together with residents and, ideally, young people themselves. The board reads the community profile, sets priorities and chooses the programmes. It is distinct from the key-leader group, which provides the mandate and the resources. Called Gebietsteam in Germany and community board in the US.

The social-work tradition of enabling residents and local institutions to identify and act on their own needs, rather than having services delivered to them. CTC’s authors place the approach explicitly in this tradition: the community collects its own data, sets its own priorities and owns its own plan. Known as Gemeinwesenarbeit in German-speaking countries.

The randomised controlled trial that established the evidence for CTC. Begun in 2003, it matched 24 small towns in seven US states in pairs and randomly assigned one of each pair to CTC, following more than 4,400 young people from age ten into adulthood. Young people in CTC towns started using substances later, showed less delinquency and violence, and the differences persisted to age 21.

How certain a statement about an effect is, not how large the effect turns out to be. GRADE describes this as the confidence in an estimate of an effect and grades it by how likely further research is to change that confidence. The Green List Prevention judges conclusiveness separately from the study design and makes it depend on group sizes, the number of studies and the presence of findings from the German language area.

The group that does not receive the programme and so shows how the participants would probably have developed without it. That yardstick can never be observed, only estimated; the effect is the difference between what happened and what would have happened without the measure. Without a comparison group it stays open whether a change is down to the programme or to maturation, prior experience and concurrent influences.

Four levels instead of the split between effective and ineffective: a programme is at first merely described, then theoretically grounded, then supported by first indications from practice, and finally by causal evidence. The lower the level, the more freely the evaluation design can be chosen; for a causal statement only designs that allow one come into question. The rating scheme of the Green List Prevention follows this model.

A measure of how large a difference is, independent of the size of the sample. The usual one in prevention is Cohen’s d: the difference between the means, divided by the spread. Unlike a p-value, an effect size says something about the extent of an effect and makes studies comparable with one another; the Green List Prevention therefore asks for it in studies with a control group.

A repeat measurement taken after a programme has ended. It shows whether an effect lasts or was only visible while the measure was running. The standards of the Society for Prevention Research ask for at least one long-term measurement at a reasonable distance from the end of the intervention and treat six months as the sensible minimum; the Green List Prevention uses the same interval and expects at least one positive effect to persist by then.

The online register of effectiveness-tested prevention programmes maintained by the Crime Prevention Council of Lower Saxony (Landespräventionsrat Niedersachsen), with the risk and protective factors each programme addresses. German CTC communities choose tested programmes from it in phase four. The register is in German only.

The people in a community who control budgets, shape policy and can commit their organisation: the mayor or head of the local authority, the police chief, the head of the school administration, and leaders from health, business, faith communities and the media. They give the CTC process its mandate and its resources, they recruit the community board and hold it and the coordinator to account, and they keep the work in place beyond a funding round. The division of labour is the point: the community board does the analysis, the priority setting and the planning, the key leaders make that work possible and act on what it recommends.

A programme is manualised when a written guideline for action settles what happens when and with whom; only that makes it repeatable elsewhere and testable at all. How closely the guideline is followed is part of what decides success: a review of almost 500 studies found effects at least two to three times as large where implementation was careful. The Green List Prevention therefore asks for available manuals, training and instruments of quality control.

The statistical combination of the results of several studies of the same question into one pooled estimate; Gene Glass introduced the term in 1976 for the analysis of a large collection of results from individual studies. Each study is weighted, as a rule by how precisely it measured, and the result is reported with a confidence interval and with a measure of how far the studies disagree with one another. A meta-analysis cannot repair the studies it pools: where the primary evidence is thin, one-sided or drawn from a different setting, the pooled figure inherits all of that.

The geographical unit a CTC process works in. Not an administrative boundary but the area whose residents share schools, services and daily life, and whose practitioners know each other. Choosing it well matters: the youth survey area and the action area must coincide, or the coalition ends up acting on a profile the area does not recognise. Called Sozialraum in German.

An approach that treats adolescence as a period of growth to be supported rather than a set of problems to be prevented, and that works on competence, confidence, connection, character and contribution. It shares its theoretical roots with CTC in the Social Development Model, and CTC’s emphasis on protective factors is where the two meet.

The share of a population that shows a characteristic at one point in time or within a given period, for instance the proportion of fifteen-year-olds who drank alcohol in the past thirty days. Prevalence counts the cases that exist, so it depends both on how many new cases arise and on how long each one lasts; incidence counts only the new cases in a period. A measure that shortens an episode lowers prevalence without changing incidence, which is why prevention studies report the two separately.

A classification by target group, not by timing. Universal offers address a whole population without any individual risk having been established; selective ones address subgroups whose risk is markedly above average; indicated ones address individual young people showing first signs that do not yet warrant a diagnosis. CTC works universally and selectively.

A classification by the point at which one intervenes. Primary prevention starts before a disorder appears and lowers the number of new cases; secondary prevention shortens courses that have already begun, through early detection and treatment; tertiary prevention reduces the impairment a disorder leaves behind. Only primary and secondary prevention programmes enter the Green List Prevention, because that matches the direction of CTC.

Conditions that lower the likelihood of problem behaviour and promote healthy development: strong bonds to family, school and community, clear standards, opportunities to contribute, the skills to use them, and recognition for doing so. In the Social Development Model they are what bonding is made of. Strengthening protection counts as much in CTC as reducing risk.

A design with a comparison group but without random assignment: who receives the programme follows from practice. In the field this is often the only workable option, but it calls for extra precautions such as several points of measurement, statistical control or a stepped roll-out, because the programme and comparison groups can differ systematically.

A study in which participants are assigned to the programme or to the comparison condition at random. Because chance makes the groups comparable on average, a difference at the end is more readily attributed to the programme than to pre-existing differences. The standards of the Society for Prevention Research ask, for a statement of efficacy, for at least two methodologically rigorous studies with consistently positive findings and at least one significant long-term measurement.

Conditions in a young person’s family, school, peer group or neighbourhood that raise the likelihood of problem behaviour such as substance use, violence, offending or leaving school early. Many act on several outcomes at once. CTC measures them locally through the youth survey and asks the community to prioritise the ones that are most elevated.

Aaron Antonovsky’s question of what keeps people healthy, as opposed to what makes them ill. Its central idea, a sense of coherence built from comprehensibility, manageability and meaningfulness, corresponds closely to the protective factors CTC works with. The term is widely used in European health promotion and less so in the US prevention literature.

The developmental theory behind CTC, formulated by Richard F. Catalano and J. David Hawkins. It holds that young people bond to the social units, prosocial or antisocial, that give them opportunities, the skills to use them and recognition for doing so, and that they then adopt those units’ norms. Prevention therefore means creating those three conditions in family, school and community. The Social Development Strategy is its practical form.

A finding is called significant when the data would be sufficiently unlikely on the assumption that there is no difference; the usual threshold is five per cent. The American Statistical Association notes that a p-value measures neither the size of an effect nor the importance of a result, and that scientific or policy conclusions should not depend on whether it falls below a threshold.

A written-out assumption about how a programme works: which steps lead from the measure through intermediate outcomes to the intended goal, and which conditions have to hold along the way. It is developed before implementation together with those involved, and it settles what will later be measured. In the Green List Prevention a theory of change study counts as an evaluation level of one star: it tests the assumed chain of effects, not the effect against a control group.

The question whether a programme developed elsewhere also works here. There is a tension between fidelity and fit: a rigid adoption misses the target group, a free adaptation can remove the very ingredients that worked, and the balance can only be struck deliberately and on the record. The Green List Prevention grades transferability by the number of studies from the German language area, in four steps from given to unclear.

The standardised, anonymous questionnaire that a CTC community puts to school pupils, as a rule between the ages of eleven and seventeen, to measure the risk and protective factors in its own neighbourhoods. Answers are reported for an area or a year group, never for an individual or a class, and the profile they produce is what phase three turns into a short list of priorities. National adaptations translate the scales, drop items that do not apply and re-test them: the German version was confirmed as largely one-dimensional and valid against reported violence, delinquency, substance use and depressive symptoms.