Faith, spirit and prevention: what really protects young people – religion, spirituality, or the mechanisms behind them?

Religion protects young people against substance misuse, delinquency and mental-health problems. More than 270 studies, several meta-analyses and large longitudinal cohorts say so. In the Communities That Care framework, which communities across Europe use to steer evidence-based prevention, religiosity appears as a protective factor; spirituality does not appear at all. But is the evidence as clear-cut as it looks at first sight? And what does it mean for prevention practice in societies that are increasingly secular and, at the same time, increasingly plural in their religions?

A closer look reveals a research literature that is impressive in scale but more complicated than the headlines suggest, with conceptual blurring, measurement problems and blind spots that carry real consequences for practice.

The evidence: robust, but not simple

The meta-analytic evidence is extensive. Kelly et al. (2015) synthesised 62 studies with a combined 193,656 participants and found consistent inverse associations between religiosity and substance use, illicit drug use and delinquency. Yeung et al. (2009) confirmed protective effects across substance types in a meta-analysis (Zr = −.16, a small but statistically meaningful effect). Russell et al. (2020) replicated the finding specifically for alcohol use (Z = −.21). Yonker et al. (2012), in a meta-analysis of 75 studies with 66,273 young people, reported effect sizes of r = −.17 for risk behaviour and r = −.11 for depression; an r of −.17 means that religiosity accounts for roughly 3 % of the variation in risk behaviour, which is statistically demonstrable and modest. For mental health, Aggarwal et al. (2023) identified, in a systematic review with meta-analysis of 25 high-quality longitudinal studies, a protective effect of spiritual wellbeing against depression (r = −.153), although that particular finding rests on two studies.

Individual longitudinal studies paint a consistent picture too. Chen and VanderWeele (2018) found, in the Growing Up Today Study (N = 5,681–7,458 depending on the outcome), prospective protective effects of regular attendance at religious services against drug use and possibly against starting to smoke. Petts (2009) followed 2,472 young people in the NLSY79 dataset over ten years and showed that stable trajectories of religious participation went with less delinquency. Kim-Spoon et al. (2012) showed that the quality of the parent–child relationship moderates whether the intergenerational transmission of religiosity is associated with fewer symptoms of psychological distress in adolescents: parental religiosity benefits young people only where the parent–child bond is good.

So much for the strength of the findings. Three complications deserve particular attention.

Complication 1: religiosity and spirituality are not the same thing

In the research literature, religion and spirituality are often treated as synonyms or bundled together as “R/S”, a practice that blurs conceptual differences. Pargament (1999) defined spirituality as the search for the sacred and understood it as the core of religion, not its opposite. He warned explicitly against the popular tendency to narrow religion to the institutional and to idealise spirituality as the purely individual. Zinnbauer et al. (1997) showed empirically that 74 % of respondents identified as both spiritual and religious; the popular dichotomy of “spiritual but not religious” describes only a minority. Hill et al. (2000) cautioned against loading spirituality with positive and religion with negative connotations.

The empirical distinction yields instructive findings. Good and Willoughby (2014) examined the two constructs separately in a longitudinal study of Canadian adolescents. Their central result: institutional religiosity – attending services, organised religious activity – was specifically and robustly associated with lower substance use, prospectively as well as concurrently. Personal spirituality – experiences of transcendence, spiritual beliefs – showed broader associations with general wellbeing but did not predict change in substance use over time. Vitorino et al. (2018) found in a Brazilian sample that the group high in spirituality but low in religiosity reported higher quality of life but also significantly higher anxiety, possibly because spiritual searching without the social support structures of a religious community can generate existential insecurity.

The implication: institutional religiosity and personal spirituality work through different mechanisms. The first works mainly through social integration, behavioural norms and prosocial networks; the second through meaning-making, coping with stress and identity work. Whoever lumps the two together cannot identify the mechanisms at work.

Complication 2: measurement problems and tautologies

Koenig (2008) identified a fundamental methodological problem. Many widely used spirituality instruments contain items that already measure mental health, such as “I feel inner peace” or “my life has meaning”. Correlate such scales with wellbeing and you produce tautological findings: an association that exists only because the same thing has been measured twice.

The philosopher Thomas Metzinger (2024) has criticised the vagueness of the concept of spirituality from another angle. What passes for “spirituality” in popular discussion, he argues, often remains conceptually unclear. Metzinger holds that contemplative practices such as meditation constitute an epistemic practice in their own right – a form of conscious knowing that works independently of religious belief systems but cannot, precisely for that reason, be reduced to stress reduction or self-regulation. Sharpened for prevention research, the criticism runs: the field sometimes mistakes the vehicle (religion, spirituality) for the active ingredient (the concrete psychosocial mechanisms behind it).

This objection is not a dismissal of religious experience. But it forces a clarification: when prevention research speaks of “spirituality as a protective factor”, it has to make sure it is not simply measuring the absence of mental-health problems and calling it their cause.

Complication 3: for whom does the protective factor work, and for whom does it not?

One finding in the evidence deserves particular attention because it shows the limits of supposedly universal protective factors. Rostosky et al. (2007) used the nationally representative Add Health dataset and followed adolescents over six years into young adulthood. Their result: each additional unit of religiosity lowered the odds of binge drinking by 9 %, of cannabis use by 20 % and of cigarette smoking by 13 % (odds ratios of 0.91, 0.80 and 0.87) – but only among heterosexual young people. Among sexual minorities the protective effect disappeared entirely. The authors read this as a sign that, for young people who meet rejection or stigma in religious communities, the protective effect of those communities is not merely absent but may reverse.

The finding underlines that protective factors are not context-free properties. They work within social relationships, and where those relationships are marked by exclusion or contempt, a protective factor can become a risk factor. For prevention programmes that want to use religiosity as a resource, that is a non-negotiable caveat.

Religiosity in the CTC model: present, but marginal

The Communities That Care system records religiosity in its youth survey as a protective factor in the peer and individual domain, operationalised primarily through the frequency of attendance at religious services. Spirituality is not measured as a construct in its own right. And religiosity’s position is peripheral: when Feinberg et al. (2007) aggregated the CTC scales, they identified seven to eight empirically coherent factor clusters, and religiosity fitted none of them. Kim et al. (2015) examined fifteen protective factors addressed by the CTC system in the Community Youth Development Study; religiosity was not among them.

The prevention researcher Andreas Beelmann has repeatedly stressed that prevention research must concentrate on factors that interventions can change (Beelmann, 2022). From that perspective religion and spirituality are awkward targets: school-based prevention programmes cannot, and should not, make young people more religious. Beelmann has also documented, in his social-developmental model of radicalisation, that religious identity is not only protective but can under certain conditions become a risk factor, for intergroup prejudice, for separation and, in the extreme, for radicalisation (Beelmann, 2022). A protective factor that turns into a risk factor depending on context demands more differentiated treatment than a single scale in a youth questionnaire.

Bridging the gap: mechanisms, not packages

How can these tensions be resolved? Not by an either–or, neither by uncritically adopting the protective-factor narrative nor by rejecting religious resources wholesale, but by shifting the gaze: away from the packages “religion” and “spirituality” and towards the mechanisms that underlie them.

The research has identified those mechanisms: self-regulation and impulse control; social integration in prosocial communities; an internalised moral orientation (already captured in the CTC model as “belief in the moral order”); intergenerational religious socialisation within the family (Regnerus, 2003; Petts, 2009); and the quality of the parent–child relationship as the condition that moderates the transmission of religious values (Kim-Spoon et al., 2012). All of these mechanisms can be located in religious or spiritual contexts, but none of them has to be. A young person who finds social belonging, clear behavioural norms and adult role models in a congregation benefits from the same protective processes as a young person in a sports club, in a well-run youth group or in a secular mindfulness programme.

For much of Europe this matters a great deal. Hodapp and Zwingmann (2018), in the first meta-analysis of studies from German-speaking countries – 67 studies with 119,575 participants – found a weighted average correlation between religiosity or spirituality and mental health of only r = .03, which is to say practically nil, and lower than the figures from the United States. That is not a deficit of those societies. It indicates that, in a secularised society, the protective mechanisms are supplied through other structures: clubs and associations, the welfare state, all-day schooling.

Implications for prevention practice

What follows from all this for community prevention?

First: the international evidence for protective effects of religion and spirituality is real and replicated. Ignoring it would be as wrong as over-generalising it. For communities working with CTC, this means that religious communities are potential partners – not as providers of interventions, but as places where protective processes may already be under way.

Second: measuring “religiosity” in the CTC youth survey by attendance at services is crude. It captures neither the personal significance of faith nor spiritual orientations beyond organised religion. A more differentiated measure, one that reflects Europe’s increasingly plural religious landscape, would be desirable, together with culturally sensitive instruments that take the tautology problem into account.

Third: any use of religious protective factors in prevention has to reflect on whom it works for and whom it does not. The finding that religiosity shows no protective effect for non-heterosexual young people, and may do harm, is not a footnote; it is the test case for the assumption of universality.

And fourth, more fundamentally: prevention science should not argue about whether religion is “good” or “bad”. The more productive question is which psychosocial mechanisms religious and spiritual contexts activate, and how those mechanisms can be made accessible to the young people who have no religious background, who experience religious communities as excluding, or who grow up in a society where institutional religion no longer plays a central role for the majority.