Adolescence is a developmentally sensitive period, when individuals are more likely to engage in risky behaviours, such as substance use (e.g. alcohol, illegal drugs, tobacco) (Hall et al., 2016). Substance use during adolescence can have a significant impact on life trajectories and is associated with adverse outcomes including mental health problems, involvement in crime, poor educational attainment, and adult substance misuse) (Englund et al., 2013; Lynne-Landsman et al., 2010; Taylor et al., 2017). Factors associated with an increased risk of adolescent substance use include a higher level of sensation seeking (Quinn & Harden, 2013), parental substance use (Yule et al., 2018), low parental monitoring (Rusby et al., 2018), low socioeconomic position (SEP) (Karriker-Jaffe, 2011; Patrick et al., 2012), and conduct problems (Heron et al., 2013).

There is also some evidence that parental religiosity plays a role in young people’s substance use. Parental religiosity has been found to be a protective factor against substance use in several studies (Hayatbakhsh et al., 2014; Hoffmann, 2023; Jankowski & Zamboanga, 2019; Mak, 2018) but one other study found no association (Parsai et al., 2010). There are limitations in the previous literature that limit understanding of the association between religiosity and substance use. The studies often rely upon single item measures of religious belief, such as church attendance or the importance of religion in the parent’s life. These measures are unable to differentiate between those who have no religious belief, weak religious belief, or are who agnostic. Single items cannot capture the multifaceted nature of religious belief and could distort differential associations with substance use. For example, previous studies found that religious importance has a weaker association with substance use than church attendance (e.g. Kelly et al., 2015; Livne et al., 2021). These limitations suggest using methods that are able to differentiate between qualitatively different groups of religious (non)belief, such as latent class analysis in the present study.

Much of the earlier research is conducted on US samples (e.g. Foshee & Hollinger, 1996; Hoffmann, 2023; Kim-Spoon et al., 2014; Merrill et al., 2005). The US has a relatively high level of belief in God (approximately 82%; Duffy et al., 2023) compared to other countries such as the UK (approximately 49% (Duffy et al., 2023)). Furthermore, even within those who may have a belief in god in the UK, they may not necessarily identify with a religion (Woodhead, 2017), or being religious (Oake, 2021). There is evidence of more benefits to the individual (e.g. social support, inclusion, lower likelihood of prejudice Chuah et al., 2016; Gebauer & Sedikides, 2021; Ng & Gervais, 2016; Stavrova & Siegers, 2014)) associated with being religious within a religious majority social group, compared with being non-religious in a religious majority social group. Some examples of differences in US and UK samples can be seen in recent work examining mental health outcomes, which found that being religious was a risk factor for some offspring mental health outcomes, such as self-harm, oppositional defiant disorder, and depression (Halstead et al., 2023, 2024), which is the opposite of the majority of the US literature. It is possible that between country differences may extend to the association between religious belief and substance use. Therefore, we have chosen to examine a UK sample.

Many studies in religious belief research obtain their samples opportunistically, often in university student populations (Koenig, 2012), and examine cross-sectional associations (Kelly et al., 2015). The use of opportunistic sampling in university populations student (i.e. younger, better educated participants) can introduce selection bias based on SEP factors (Howe et al., 2013; Munafò et al., 2018), which are predictive of substance use (Heron et al., 2013). Reverse causality is possible in cross-sectional studies, which has been observed previously in the case of criminal behaviour predicting religion (Heaton, 2006). In the context of the present study, offspring substance use may introduce issues to the family environment (such as navigating possible drug dependency, changes in behaviour due to substance use), which may impact a parent’s religious belief, or ability to practice it. Furthermore, we intend to address the issue of the use of student populations by examining a sample taken from the general population.

The current study, based on data from a large birth cohort, examines prospective associations between different patterns of maternal religious belief, and adolescent offspring substance use (alcohol, cigarette smoking, cannabis use, other drugs use, such as amphetamines, heroin, and cocaine) at age 17.5 years. We seek to examine whether maternal religious (non) belief is associated with subsequent substance use by their offspring, when controlling for confounders and examined longitudinally.

The data were collected as part of the ‘Teen focus 4’ clinic, when the offspring were approximately 17.5 years old (mean = 17.82, SD = 0.456).

Analyses were adjusted for the potential confounding effects of parental occupational social class, maternal age, maternal educational attainment, economic hardship/problems, maternal anxiety and depression, maternal adverse childhood events (reported retrospectively) and child sex as assigned at birth. See Appendix 2 for more information.

In comparison to the complete case sample, the imputed sample had higher proportions of alcohol use, cannabis use, drug use and smoking. They also had higher proportions of low SEP indicators (social class, education, financial problems/hardship) (See Table 1 for more information).

In the complete case sample, compared to offspring of agnostic mothers, offspring of atheist mothers had higher odds of cannabis use and other drugs use [OR 1.33(1.01, 1.75) and OR (1.68 (1.12, 2.51), respectively], but these associations were attenuated in the fully adjusted model [OR 1.16 (0.87, 1.55) and OR 1.50 (0.99, 2.27), respectively]. There was also weak evidence of an association between offspring of highly religious mothers and higher odds of other drug use [OR 1.47(0.94, 2.30)] and lower odds of smoking weekly [OR 0.60 (0.36, 1.00)].

See Table 2 for the results of the models using imputed data in the imputed sample, compared to the offspring of the agnostic mothers, the offspring of atheist mothers had greater odds of cannabis and other drug use [OR 1.32 (1.05, 1.66) and OR 1.41 (1.02, 1.95), respectively]. There was also weak evidence of offspring of highly religious mothers and offspring of atheist mothers having lower odds of smoking [OR 0.68 (0.45, 1.02) and [OR 0.72 (0.50, 1.04), respectively].

We found evidence in a large UK birth cohort sample that, compared to the offspring of agnostic mothers, offspring of atheist mothers were more likely to have smoked cannabis or tried other drugs, and the offspring of highly religious mothers were less likely to smoke weekly. These associations were robust to the inclusion of maternal SEP indicators, maternal mental health, previous substance use, and stressful life events and were demonstrated within both the complete case and the imputed analyses.

Our findings align with most of the studies that examined the association between religious belief and adolescent substance use in both the UK and US, with greater religious belief associated with lowers odds of smoking (Hussain et al., 2019; Nunziata & Toffolutti, 2019; Whooley et al., 2002) and lower religious belief associated with greater odds of substance use (King et al., 2013; Moscati & Mezuk, 2014). However, these studies examined an individual’s own religious belief, rather than the impact of their parent’s religious belief. In adolescence, offspring and parental religious belief may be very similar (Leonard et al., 2013), and the associations we see in the present study may be attributable to this similarity. Future studies with both measures of parental and offspring religious belief available to them may be able to explore this possibility.

Higher levels of parental monitoring and control by religious parents (Kim & Wilcox, 2014; Kim-Spoon et al., 2014) may lead to offspring being less likely to try cigarettes, for fear of detection. Religious parents may be less likely to use substances themselves, so their offspring may follow the example set by their parents and avoid drugs (Kendler et al., 2003, 2015). Lower levels of parental control and monitoring (Kim & Wilcox, 2014), as well as a more open-minded (Gurney et al., 2013), less traditionalist attitude (Saroglou, 2010; Saroglou et al., 2004) may lead to offspring of atheist parents feeling comfortable with trying cannabis or other drugs. The moral views of parents may also be passed on to their offspring, informing their decision-making (Chen et al., 2024), which in the case of children of religious parents may include abstinence related decisions.

There is also the possibility that social factors may explain the associations (Ford & Hill, 2012; C. Smith, 2003). Offspring of religious parents may be more likely to have a religious social group (Stroope, 2012) and community services which may emphasise ‘Christian’ values such as abstinence (Rosansky & Rosenberg, 2019), whereas the offspring of atheist parents may rely upon secular sources of community and socialising, which may not have the same emphasis on abstinence as religious groups. Given that much substance use is based on peer influence from those engaged in risky behaviours (Allen et al., 2012; Henneberger et al., 2021), being less exposed to these influences may reduce the risk to offspring of religious parents.

The use of latent class analysis allowed us to capture qualitatively different types of belief, rather than relying on single item measures. The latent classes differentiated between highly and moderately religious classes as well as atheist and agnostic classes. We used multiple imputation to address bias due to missing data, as seen in the complete case sample, where the group analysed had higher prevalences of measures of low SEP than the original cohort.

However, there are limitations of the present study that should be considered when interpreting the findings. The ALSPAC cohort is predominantly White, and those who identify as religious are mostly affiliated with Christianity. This means that the current results may not generalise to non-white, non-Christian religious populations. The study is also limited in making inferences about the time point in which the data was collected (1990–2007). The relationship that young people in the UK have with substance use has changed over time, notably experiencing a significant drop in substance use (based on lower numbers of young people accessing treatment services) (Young People’s Substance Misuse Treatment Statistics 2021 to 2022, 2023), In addition, the proportion of the population who believe in God (Duffy et al., 2023) has changed since the data were collected (e.g. 62% of people believed in God in 1999, when the religion data were collected, compared to approximately 49% in 2021). Whether the associations found in this study hold in the present day would be a valuable exploration. Finally, we did not have access to measures of the offspring’s own religious belief. Future research should examine the role of the young person’s own religious belief, which may provide an insight into the mechanisms between parental religious beliefs and offspring substance use.

We found, when compared to the offspring of agnostic mothers, the offspring of atheist mothers had higher odds of using cannabis or trying other drugs, and the offspring of highly religious mothers was less likely to smoke weekly. These results provide a novel insight into the role of parental religiosity in adolescent substance use in an understudied population, using methods that overcome limitations of the previous literature. Future work should explore the role of parental religious belief independently of offspring religious belief, as well as exploring mechanisms that explain the association, such as parenting variables and social groups.

We are extremely grateful to all the families who took part in this study, the midwives for their help in recruiting them, and the whole ALSPAC team, which includes interviewers, computer and laboratory technicians, clerical workers, research scientists, volunteers, managers, receptionists, and nurses.

The UK Medical Research Council and Wellcome Trust (Grant ref: 217065/Z/19/Z) and the University of Bristol currently provide core support for ALSPAC. This publication is the work of the authors, and they will serve as guarantors for the contents of this paper. A comprehensive list of grants funding is available on the ALSPAC website (http://www.bristol.ac.uk/alspac/external/documents/grant-acknowledgements.pdf). This project was made possible through the support of grants from the John Templeton Foundation (ref no. 61356 and 61917). The opinions expressed in this publication are those of the author(s) and do not necessarily reflect the views of the John Templeton Foundation.

Please see the ALSPAC data management plan which describes the policy regarding data sharing (http://www.bristol.ac.uk/alspac/researchers/data-access/documents/alspac-data-management-plan.pdf), which is by a system of managed open access. Data used for this submission will be made available on request to the Executive (alspac-exec@bristol.ac.uk). The datasets presented in this article are linked to ALSPAC project number B4512, please quote this project number during your application.

The steps below highlight how to apply for access to the data included in this study and all other ALSPAC data:

1. Please read the ALSPAC access policy (http://www.bristol.ac.uk/media-library/sites/alspac/documents/researchers/data-access/ALSPAC_Access_Policy.pdf) which describes the process of accessing the data and samples in detail and outlines the costs associated with doing so.

2. You may also find it useful to browse our fully searchable research proposals database (https://proposals.epi.bristol.ac.uk/?q=proposalSummaries), which lists all research projects that have been approved since April 2011.

3. Please submit your research proposal (https://proposals.epi.bristol.ac.uk/) for consideration by the ALSPAC Executive Committee. You will receive a response within 10 working days to advise you whether your proposal has been approved.