Family Connection and Flourishing Among Adolescents in 26 Countries

We used data from the third wave of the ISCWeB, a survey of children’s well-being, daily activities, and time-use that was conducted between 2016 and 2019 across 35 countries. The detailed survey methods are described elsewhere and summarized here.^32–34  Teams of investigators in each country administered separate questionnaires to 8-, 10-, and 12-year-olds. The questionnaires were developed in English, translated into the languages of participating children, and then backtranslated. To reach survey respondents, investigators used random sampling of mainstream schools across their country or within specific region(s) of their country. However, because the full sampling frame of potential respondents was not established, the participation rate was not reported. Each team received the appropriate ethical approval for the survey, all children provided informed consent, and parents gave active or passive consent for their children to participate.

For our analysis, we used the publicly available, deidentified data provided by the ISCWeB investigative team. We used data only from those who completed the 12-year-old questionnaire because it was the only questionnaire containing the items we used to measure our outcome (flourishing). Nine countries were not included in our analysis: 5 did not survey the 12-year-old age group, 2 did not include all the survey items we used to measure our exposure (family connection), and 2 did not have public data available on age or gender. The ISCWeB had 41 125 respondents to the 12-year-old questionnaire in these 26 countries. Within sampled schools, specific grades were targeted to receive the 12-year-old questionnaire. We restricted our analysis to the 11- to 13-year-old respondents to the 12-year-old questionnaire to decrease variability around the target age of 12 years, leaving a sample of 39 286 adolescents.

The flourishing score was based on 6 close-ended survey items. Each item was aligned with a dimension of Ryff’s Psychological (eudaimonic) Well-Being Scale^16,35 : self-acceptance (“I like being the way I am”), environmental mastery (“I am good at managing my daily responsibilities”), positive relations with others (“People are generally friendly towards me”), autonomy (“I have enough choice about how I spend my time”), personal growth (“I feel that I am learning a lot at the moment”), and purpose in life (“I feel positive about my future”). On an 11-point Likert-type scale, anchored at 0 (“not at all agree”) and 10 (“totally agree”), adolescents were asked to indicate their level of agreement with each item. Those with complete data on at least 5 items were included in the analysis, and we calculated a mean flourishing score (range 0–10) from the available items. The reliability and construct validity of the scale has been established.^36,37  In our sample, the internal consistency (Cronbach’s α) of the flourishing score items was .84. To assess our study aim in a manner that was both statistically valid and interpretable, we used a binary measure for flourishing (scores >8), with the cut point based on qualitative studies from the ISCWeB investigative team.³⁸ 

We created a family connection score using 5 survey items. Each item asked about a dimension of connection in the adolescent’s home context: care (“There are people in my family who care about me”), support (“If I have a problem, people in my family will help me”), safety (“I feel safe at home”), respect (“My parent(s) listen to me and take what I say into account”), and participation (“My parents and I make decisions about my life together”) (data provided by the International Survey of Children’s Well-Being project team, July 13, 2021). On a 5-point Likert-type scale from 0 (“I do not agree”) to 4 (“I totally agree”), adolescents were asked to indicate their level of agreement with each item. Those with complete data on at least 4 items were included in the analysis, and we calculated a mean family connection score (range 0–4) from the available items. The internal consistency (Cronbach’s α) of the family connection score items in our sample was .79.

Our analyses included 7 covariates, which we considered as potential confounders of the association between family connection and flourishing. These variables, all based on self-report by the child, included gender (girl or boy), age (whole years), household structure (living with mother or stepmother and/or father or stepfather and coded as living with both, either, or neither), and country. We also included 3 variables related to the child’s economic circumstances: material resources, family financial worry, and food sufficiency. The material resources variable was a count (0–8) of 8 items children reported (yes or no) that they had: clothes in good condition, 2 pairs of shoes in good condition, enough money for school trips and activities, internet at home, equipment and things needed for sports and hobbies, pocket money or money to spend on yourself, mobile phone, and equipment and things needed for school. The survey items used for the family financial worry variable (“How often do you worry about how much money your family has?”) and the food sufficiency variable (“Do you have enough food to eat each day?”) each had the response options of “never,” “sometimes,” “often,” or “always.”

Our analytic sample included 37 025 of the 39 286 (94.2%) adolescents with completed surveys after we excluded 2260 adolescents who had missing data on either the exposure (family connection) and/or outcome (flourishing), and one who had missing data on sampling stratum (Supplemental Table 3). We conducted statistical analyses with Stata/MP version 15.1 (Stata Corp), and we used the Stata “svyset” command, with the ISCWeB variables caseweight and stratum (school), to account for the complex sampling design. All reported percentages were weighted.

We first computed the mean (95% confidence interval [CI]) of the family connection score and the prevalence (95% CI) of flourishing across levels of the covariates. We then used a logistic regression model to examine the association between flourishing (binary dependent variable) and family connection score, while controlling for all 7 potentially confounding covariates. For these regression analyses, 6217 cases (16.8%) of the analytic sample were missing data on one or more of 5 covariates (gender, family structure, material resources, family financial worry, or food sufficiency) (Table 1). Missing data for these covariates were imputed³⁹  using sequential regression imputation⁴⁰  to create 20 imputed data sets. We then ran logistic regression models on the imputed datasets and reported model parameters that were aggregated across datasets.⁴¹ 

We analyzed the family connection score in 2 ways. To facilitate interpretation of our findings, we first used family connection as a categorical independent variable with 5 levels (<2.5, 2.5 to <3.0, 3.0 to <3.5, 3.5 to <4.0, and 4.0) and used the lowest level of family connection score (<2.5) as the reference group. Regression-based margins, standardized to the distribution of covariates in the study population, were used to estimate the adjusted prevalence (95% CI) of flourishing at each level of family connection. In a logistic regression model with all 7 covariates, we also used the family connection score as a continuous independent variable and estimated the probability (95% CI) of flourishing across the entire range of family connection scores. In secondary analyses, we examined the association between family connection score and flourishing in each country. We also examined the association with flourishing defined using a different cut point (score >9).

Of the 37 025 adolescents included in this analysis, the mean (SD) age was 11.9 (0.6) years, and 51.4% were girls. Family financial worry was reported to occur “always” or “often” by 24.2% of adolescents, and 6.3% reported “never” or only “sometimes” having enough food to eat each day (Table 1). The mean (SD) family connection score was 3.4 (0.7), and adolescents were distributed, as follows, across 5 levels of increasing score: <2.5 (11.2%), 2.5 to <3.0 (8.8%), 3.0 to <3.5 (24.2%), 3.5 to <4.0 (25.1%), and 4.0 (30.7%) (Supplemental Fig 2). The prevalence (95% CI) of flourishing was 65.8% (65.3–66.3). The mean (SD) flourishing score was 8.3 (1.7), and adolescents were distributed, as follows, across 6 levels of increasing score: 0 to 4 (2.7%), >4 to 6 (8.0%), >6 to 7 (8.7%), >7 to 8 (14.7%), >8 to 9 (24.6%), >9 to 10 (41.2%) (Supplemental Fig 3).

Family connection scores were similar across age levels and between boys and girls, but the prevalence of flourishing was higher in boys and those who were younger (Table 1). The highest family connection scores and the highest prevalence of flourishing were for those adolescents who reported never having family financial worry, always having enough food, or living with both parents (Table 1).

The prevalence of flourishing increased in a graded manner as the level of family connection increased (Table 2 and Fig 1). After controlling for gender, age, family structure, material resources, family financial worry, food sufficiency, and country, the prevalence (95% CI) of flourishing increased across the 5 levels of increasing family connection score: 34.9% (33.3–36.5), 45.0% (43.2–46.8), 58.2% (57.2–59.3), 72.6% (71.6–73.5), and 84.3% (83.6–85.1), respectively. The adjusted prevalence of flourishing was 49.4 (95% CI 47.6–51.2) percentage points higher among those with the highest level of family connection (4.0) compared with those with the lowest level of family connection (<2.5). Secondary analyses of the association between family connection score and flourishing within each country showed similar graded associations across countries as in the pooled analysis (Supplemental Table 4). The graded association between family connection and flourishing was also seen when we used a different cutpoint to define flourishing (score >9) (Supplemental Table 5).

Based on cross-sectional survey data obtained from over 37 000 adolescents residing in 26 countries, we showed that higher levels of family connection were associated with a greater prevalence of flourishing. To our knowledge, this is the first population-based study of adolescents showing the association between a measure of relational health (family connection) and flourishing (assessed as eudaimonic well-being). This cross-sectional association cannot be interpreted as a causal relationship. However, the association was graded and strong, accounted for potential confounders assessing relative socioeconomic disadvantage and adversity, and was present in a range of countries across Europe, Asia, Africa, and South America. The psycho-social-biologic mechanisms supporting the causal links between adult-child connection and later flourishing arise from a well-established body of research on attachment⁴²  and mammalian evolutionary biology.^43,44 

Aside from studies using ISCWeB data,^36–38,45  we are not aware of any other population-based studies of adolescents that include a measure of flourishing, like the one used here, based on Ryff’s framework of eudaimonic (psychological) well-being.¹⁶  As reviewed by Nahkur and Casas,³⁷  investigators have examined the psychometric properties of Ryff’s scale of psychological well-being among diverse groups of adolescents. We know of 7 studies that have used Ryff’s scale of psychological well-being to examine correlates of flourishing among adolescents,^46–52  but none of these studies have examined family connection. Frameworks of positive youth development include constructs such as purpose, relationships, competence, and identity.^53–56  Although these constructs are aligned with our conceptualization of flourishing, the body of work on positive youth development does not explicitly assert the developmental goal of eudaimonic well-being.

The cross-sectional design of this study has inherent limitations. We cannot make causal inferences, and we cannot exclude reverse causality or common rater bias as possible explanations for our findings. While the ISCWeB was international in scope, it did not collect nationally representative samples in participating countries. A school-based sampling frame was used in each country, but this frame excluded children not enrolled in mainstream schools, and the scope of the study made it infeasible to calculate survey-wide response rates at the school or student levels. The family connection and flourishing measures have limitations. Although the measures of family connection and flourishing had adequate internal consistency, the measurement of these multidimensional constructs in adolescents is still evolving. We are not aware of any studies of adolescents in which responses to the 6-item flourishing measure used in the ISCWeB were compared with responses from Ryff’s full scale of psychological well-being. Only a single item was used to assess each of the 6 dimensions of flourishing.¹⁶  Furthermore, adolescents’ understanding of the flourishing items and/or their response style for these items may differ by country; these differences may be due to cultural factors, including language, and warrant some caution when comparing levels or correlates of flourishing among countries.³⁷ 

Family connection, as operationalized here and in our other recent work,^14,15,31  is consistent with the definition of relational health in the context of family: safe, stable, and nurturing relationships that children and adolescents experience with their parents or primary caregivers.¹  The importance of family connection is not just to develop resilience to adversity but to promote flourishing, with or without adversity. Future research should address the more applied question of how to enhance and sustain adult-child connection in the face of social forces that favor disconnection.^57,58  These studies could also examine the potential role of children’s connections with nonparental adults in their schools, faith-based institutions, and communities to promote children’s flourishing,⁵⁹  as well as the role of health care innovations, such as the family-centered pediatric medical home,⁶⁰  in supporting adult-child connection. Research in this area will benefit from obtaining data directly from children and adolescents, using emerging measures of both family connection and flourishing.¹⁰  While parents have aspirations for their children’s well-being (who children are currently) and well-becoming (who children will be in the future), children may also hold present and future aspirations for themselves. For adults to provide the safe, stable, and nurturing relationships that allow children to flourish, adults must understand children’s perspectives on both connection and flourishing.

As for enhancing adult-child connection, children’s perceptions that they are safe and seen by an adult may be as important as children’s reports of what they do with adults. For example, whether adverse childhood experiences are classified as traumatic and impair later health and functioning is not only about the events and circumstances but also about their enduring emotional impact.⁶¹  Similarly, whether positive childhood experiences promote lifelong flourishing may depend more on the emotional impact for children of consistently feeling safe and seen by an adult (ie, the emotional climate), rather than on the particular activities the child shares with an adult. This possibility could be empirically evaluated in studies using psycho-physiologic measures of social safety.⁶²  If supported by further study, efforts to increase positive childhood experiences might benefit from helping adults understand how best to make children feel safe and seen (ie, a focus more on being than doing).⁶³  This process may even involve interventions in which the adults themselves experience being safely seen,⁴²  so that they can provide this gift to children.⁶⁴  As summarized by the evolutionary neurobiologist C. Sue Carter, “Without positive relationships, especially in early life, humans fail to flourish, even if all of their basic needs are met.”⁶⁵ 

The data used in this publication come from the third wave of the Children’s Worlds project: An international survey of children’s lives and well-being (www.isciweb.org). The views expressed here are those of the authors. They are not necessarily those of ISCWeB.