The study design is a longitudinal cohort study, with repeated measurements of physiological indicators measured in saliva from a subsample of participants, and social survey data from an entire population cohort.


The two prospective data collection (saliva and social survey) were supplemented with retrospective registry data from an entire population cohort. Registry data on all Icelandic children in the 2004 birth cohort were included in our study.


The social survey was carried out in the entire cohort of children born in Iceland in the year 2004; around 4000 children using a protocol from ICSRA (The Icelandic Center for Social Research and Analysis) that has been used for over 20 years in school-based surveys (Kristjansson et al, 2013; Sigfusdottir et al., 2009).


ICSRA solely utilizes measuring tools that have been validated in the social sciences. Questionnaires were administered to participants in the study at the data collection facility. Research assistants distributed the questionnaires, and participants responded to them in carefully controlled setting enabling them to give confident answers. Upon completion participants sealed the questionnaires in envelopes containing participant identification number.

For reasons of confidentiality, the link between the student’s identification number and their personal information was kept in a secure location, accessible only to a small number of professional data collection staff members who have signed an agreement to maintain strict confidentiality procedures.


Since 1998 the ICSRA has conducted annual population-based cross-sectional surveys among 10-20-year-old youth in Iceland that provides information from 85-90% of the population. The methodology of the survey data collection has been described elsewhere (e.g. Sigfusdottir et al. 2009; 2011; Kristjansson et al. 2008; 2010). Among measures are indicators of family well-being such as conflict, parental support and monitoring, neighbourhood cohesion and social capital. This information will be aggregated to the local community/neighbourhood level for various years in the lives of participants and therefore used to test hypotheses around higher-level variation and supportive functions pertaining exclusively to the environment.