1 Introduction

1.1 The Importance of Taking Children’s Subjective Well-Being (SWB) into Account

The United Nations Convention on the Rights of the Child (UNCRC) affirms that both family and State should give attention to the following aspects of children’s lives: health and safety, material security, education and socialization, and the sense of being loved, valued and included in families and societies (McAuley and Rose 2010). Even in rich countries, income inequality and unequal opportunities among children increased (Hudson and Kühner 2016; Toczydlowska and Bruckauf 2017), resulting in the need to put them at the heart of equitable progress; leave no child behind; improve data collection on violence, early childhood, migration and gender; use these data to improve policy in national contexts; and commit to sustainable development (Brazier 2017). Guided by the 12th article of the UNCRC and its General Comment No. 12 (CRC/C/GC/12 2009) regarding the right of the child to be heard, this may involve asking children ‘how they are’ and taking them also into account when making decisions.

In the local context of Barcelona city, the Childhood and Citizen Focus 2017–2020 blueprint, which is a mid-term childhood policy framework, includes ‘The Children Have Their Say Programme’ (Barcelona-City-Council 2017a). Taking the interest and utility of knowing children’s SWB as a starting point (Bradshaw 2015; Casas 2016; Casas and Rees 2015), the purpose of the programme is to analyse children’s SWB using the responses given by a probabilistic sample of Year 5–6 primary school students (mean age = 10.74) in 2017 to an adapted preliminary version of the third wave of the International Survey of Children’s Well-Being (ISCWeB), promoted by the Children’s Worlds project.

Examining children’s strengths, assets, and abilities, allows to identify the factors of a positive developmental trajectory and value the core elements necessary for children’s SWB to prosper (Pollard and Lee 2003). Children’s SWB is generally defined as an umbrella concept, which includes the cognitive and affective evaluations that children make about their lives, the circumstances affecting their lives and the context in which they live (Savahl 2017). Authors such as Diener define SWB based on life satisfaction, positive and negative affect, while Cummins and Veenhoven have underlined the influence of affective components on children’s SWB (Casas 2018). Life satisfaction, positive and negative affect have substantial loadings on a latent SWB factor, which supports the generalizability of a hierarchical structural conceptualization of SWB (Busseri 2018).

According to the optimism bias, when people of any age are asked about their lives in a survey, they typically report high evaluations, but children tend to display even higher scores than adults, since optimism decreases with age (Casas 2011). Therefore, the question is whether ‘children share high common SWB simply because they are children’ or whether and how, in spite of being children, they differ in their appraisals of their lives; that is to say, whether there are different reasons for being more or less satisfied (Ben-Arieh et al. 2017). Children’s SWB is essential for promoting positive development, but also for identifying risks, and it seems to be influenced by housing, safety, bullying, school achievements and social interactions (Dinisman et al. 2015). The freedom to choose and self-perception (Lee and Yoo 2017), but also social support, friendliness and choice about time use, have all been reported as important factors contributing to children’s SWB (Bradshaw and Rees 2017).

1.2 Prior results from the Children’s Worlds Project

An overall analysis of the Spanish and Catalan samples in the Children’s Worlds project also permits to expose that religion explained little of children’s variance in SWB (Kosher and Ben-Arieh 2017), and that children who lived in a welfare state reported higher satisfaction with regard to material resources and safety (Uyan-Semerci et al. 2017). Gross Domestic Product and inequality did not appear to be relevant factors predicting children’s SWB (Lee and Yoo 2015), and sociodemographic characteristics explained a low percentage of the variance (Dinisman and Ben-Arieh 2016). However, lower material resources and higher social exclusion were related to lower children’s SWB (Castellá et al. 2015, Gross-Manos 2017). Economic and material resources explained some variation in social exclusion (Crous and Bradshaw 2017), being deprivation negatively related to psychological well-being (Crous 2017).

From the first wave (2011/2012) of 14 participating countries, it was found that children’s context of living, the ways they spend their time and their life evaluations are important to their SWB (Rees and Dinisman 2015). Life changes had a negative effect; that is, moving house, living in another country, and changing area, school or the people they live with (Montserrat et al. 2015). The strongest predictors of children’s SWB were relationships, school, gender, and quality of neighbourhood (Lawler et al. 2010). The nature of their relationship with their immediate context, such as frequency of family activities, peer activities, and neighbourhood safety, was the most strongly related to children’s SWB (Lee and Yoo 2015). Greater perceived participation was related to higher children’s SWB, and there were variations by gender (Grigoraş et al. 2018). Girls who perceived having greater participation in family decisions displayed higher scores (González-Carrasco et al., 2015a). And children could have different points of view: some were very satisfied overall, others displayed low satisfaction with outdoor activities, or different levels of satisfaction with health and safety (Yoo and Ahn 2017).

From the second wave (2013/2014) of 21 participating countries, it was found that in the cases of low SWB, girls’ SWB was driven by relational factors such as satisfaction with peers, whilst for boys it was by school (Kaye-Tzadok et al. 2017). Regarding family, children living with both parents evaluated their family and SWB higher than those living with a single parent or in separated families (Dinisman et al. 2017). Regarding education, satisfaction ‘with other children in your classroom’ was important to children’s SWB, while satisfaction ‘with your life as a student’, ‘your relationship with teachers’, ‘your school marks’ and ‘things you have learned at school’ displayed separate, but also important contributions (Casas and González-Carrasco 2017). Children’s school SWB also decreased by age and depends on how teachers and other children in school treat them, and how safe a child feels at school (Kutsar and Kasearu 2017). Children reporting having been bullied displayed lower SWB, and this was related to age, gender and deprivation (Bradshaw et al. 2017). Regarding activities outside school, although their relationship with children’s SWB was not analysed specifically, it was found that playing sports, watching television and using computers were more common among all children, although girls spent more time helping out at home, doing homework and reading (Rees 2017b).

For the Spanish findings, in the first wave, according to Dinisman and Rees (2014), the 12-year-old sample scored very high on most children’s SWB scales. The Spanish mean for the Students’ Life Satisfaction Scale (SLSS) was 8.42, while the overall mean for all countries was 7.73. Of the domain-based cognitive SWB indicators also used by ‘The Children Have Their Say Programme’, they were scored between 8 and 9 points, the highest scores being satisfaction with ‘health’, ‘people you live with’, ‘friends’ and ‘things you have’, and the lowest with ‘freedom’. Children scoring highest in SWB lived in semi-urban contexts, were born in Spain, had not repeated any school year, lived in a one-family household, had two adults at home with a paid job, had parents with at least secondary education, and had more material and cultural belongings at home. They reported that they had never worried about money, thought other people treated them well, felt greater personal safety, felt heard, reported doing daily activities together with their family, did physical exercise every day, had been told that children have rights, had experienced fewer important recent changes in their lives, and felt their time was well organized (Casas et al., 2012a, 2013a). Living with two parents was related to higher satisfaction with school, interpersonal relationships, leisure time, health and self-perception than living with one parent or ‘in care’ (Dinisman et al. 2012). And, in the second wave, according to Rees and Main (2015), for the 10–12-year-old sample, SLSS was 9.02, while the overall mean for all countries was 8.92. The sample had a lower percentage of children with low SWB (2.80%) compared to other country samples, but a medium percentage with high SWB (56.80%). Of the domain-based cognitive SWB indicators also used by ‘The Children Have Their Say Programme’, they were scored between 8 and 9 points, the highest scores being satisfaction with ‘all the things you have’, ‘health’ and ‘people you live with’, and the lowest with ‘life as a student’ and ‘classmates’.

1.3 Rationale for Understanding Children’s SWB in the Local Context

Positive psychology, ecological theory, social psychology, and sociology are useful for understanding children’s SWB (Ben-Arieh et al. 2014). From positive psychology, it is understood that children’s SWB and happiness are accomplished through a pleasant (enjoying relationships and hobbies), good (being engaged in activities) and meaningful life (being part of something), which is the consequence of socialization and identity processes (Holte et al. 2014). And ecological theory recognizes that children are active participants who create subjective and meaningful representations of their experiences in microsystems, that is, the immediate contexts that most influence development (Garbarino 2014). However, children can minimize contextual factors and overlook ecological levels of analysis (mesosystems, exosystems, and macrosystems), because ‘like a fish swimming in water, we take the contexts of our lives for granted’ (Kloos et al. 2012). Likewise, social psychology focuses on the social representation of childhood, but also on children’s social problems or needs from beliefs, attitudes, stereotypes, perceptions, evaluations and aspirations (Casas et al. 2014). And sociology applies the concept of generation to children to link childhood with structural and relational approaches (Gaitán 2014; Qvortrup 2014), while children’s SWB is also included in some societal indicators systems (Domínguez-Serrano and del Moral Espín 2018).

The ISCWeB provides the opportunity to analyse children’s evaluations and perceptions of their own lives through general and specific survey questions, such as their satisfaction with life or their feelings and emotions, and their satisfaction with their immediate context. Systematic questions regarding different contexts are used to identify interpersonal relationships, and also the activities they are involved in. Moreover, to be coherent with previous results, gender, birth/origin, family affluence and type of family are known. And to be rigorous with the setting of the research, neighbourhood income is added as a substitute for Gross Domestic Product, while information about schooling is also explored (perceived specific needs because it is an inclusive educational system, and type of school and location).

Therefore, the aim of this article is to analyse children’s SWB according to domain-based indicators of the cognitive component of SWB and indicators of the affective component of SWB, as well as indicators of interpersonal relationships in different contexts and activities outside school, besides sociodemographic and educational characteristics. It is necessary to identify actions that may have an impact on children’s socialization and identity processes in their living spaces in order to promote favourable debates and appropriate policies. The research questions are listed below.

  • To what extent do domain-based cognitive indicators contribute to children’s SWB?

  • To what extent do positive and negative affect indicators contribute to children’s SWB?

  • To what extent do different interpersonal relationship indicators contribute to children’s SWB?

  • To what extent do indicators of activities outside school contribute to children’s SWB?

  • To what extent do sociodemographic and educational characteristics contribute to children’s SWB?

Considering the prior literature review, the overall hypothesis is that the higher the scores of the aforementioned indicators, the higher the levels of SWB (or the opposite for negative affect). For sociodemographic characteristics, a lower SWB could be expected among girls, children born abroad, those not living with both parents and those living in low-income neighbourhoods or with families with low affluence.

2 Method

2.1 Sampling Design

A school-based sampling method was employed because schooling is compulsory and 97% of children are at school (Government-of-Catalonia 2017a). The target population was children in Years 5–6 of primary school in Barcelona city in 2017, a population of 27,285 students (Government-of-Catalonia 2017b). The Children’s Worlds project analyses Years 3, 5 and 7, and ‘The Children Have Their Say Programme’ Years 5–6 in order to adjust the sample to the final years of primary education in Barcelona city. Ceded by the Barcelona Education Consortium in September 2017, the sample framework was the list of schools with Year 5–6 students in Barcelona city, which included the number of students and class groups (26,890 students; 324 schools). The 2% of children in privately-funded schools were not included due to a lack of access to provisional enrolment. Firstly, for purposes of representation, the sample framework was stratified by the ten municipal districts, type of school (subsidized, and general state), and size (one, two, and three or more class groups per year). Secondly, when a school was randomly selected, if it had one or two class groups, they were directly selected, and when there were three or more, two of them were randomly selected. All students in each class group were included (mean of 25).

Each stratum, under simple random selection, has a confidence level of 95% and a 5% margin of error, assuming a maximum variance of p = q = 50%. The sample size was initially increased because of non-response error (3%) and conglomerate homogeneity (Deff = 1.2) assumptions. A similar number of children were selected by municipal district using a non-proportional strategy (subgroups of around 400). The sample error was 1.4% for the whole sample. The intended sample size was 3940 children from 50 schools, but two extra random schools were added because the observed non-response error during the fieldwork was higher than initially assumed. Of the 50 initial schools, 40 are the first random selection, 8 the first substitute, and 2 the second substitute. Of the extra schools, there are two of the five selected. The number of children considered in the fieldwork was 4237 (52 schools) and the number of surveyed children was 3971 (52 schools). That is, the 93.72% of the children considered in the fieldwork were surveyed. The children’s participation was inclusive and without any explicit exclusion criteria (no response was principally due to justified school absences). Sampling weights were accurately calculated using the TIMSS and PIRLS procedures because their sampling design is very similar (Joncas and Foy 2011).

Data collection took place in the schools from January to April 2017 (one school in December 2016 for logistic reasons). In accordance with legal regulations, family consent was not required because the research was promoted by a public administration in compliance with its competences, although families, teachers and children were well informed (Organic-Law 1996, 2011). The Barcelona Education Consortium sent a first institutional message to schools asking for collaboration, and ethical standards were followed (Graham et al. 2013). There were no payments for fieldwork schools and surveyed children (benefits included participatory workshops in 2018, where the class groups interpreted the survey results and make proposals to improve children’s SWB). A coordinated stable fieldwork team of four interviewers (three in each session) conducted the sessions considering the technical aspects of online surveys administered to groups of children in schools, but also guaranteeing active listening, reading comprehension, and emotional accompaniment. They explained that the purpose was to know ‘how they are’ and that their collaboration was voluntary, that they could not answer questions they did not want to, and that all their responses were anonymous because their names and surnames were not requested.

2.1.1 Sample Characteristics

Age, Gender and Birth/Origin

The mean age of participants was 10.74 (SD = .678), of whom 49.43% were girls, and 50.57% boys, this being the same proportion as the target population. 90.23% were born in Catalonia/Spain (for 75.13% both parents were born in Catalonia/Spain), and 9.77% were born in another country (for 75.42% both parents were born in another country). The birth/origin of 8.56% was not reported, and this subgroup was not used for statistical analysis.

Type of Family

76.66% of participants lived in a family nucleus with two parental figures, 5.05% with one parent, and 18.35% in two family nuclei of different characteristics (frequency between the two nuclei and parental figures varied). Children living with one parent saw other people in the family not living with them more frequently. Note that type of family was not reported by 6.03%, and this subgroup was not used for statistical analysis.

Neighbourhood Income and Family Affluence

Neighbourhood income was a constructed variable from the intersection between ‘What neighbourhood do you live in?’ and the ‘Family Income Index’, an indicator of the mean income level of residents in the 73 neighbourhoods of Barcelona city (Barcelona-City-Council 2017b). The fieldwork team identified all the neighbourhoods after the children orally informed them of the street name (no postal address was registered due to data protection). There were small differences between the surveyed sample and the target population because ‘neighbourhood income’ was not a sampling variable in the school-based sampling method, and the number of children selected by districts was non-proportional (Table 1).

Table 1 Distribution of neighbourhood income
Full size table

Family affluence was based on the sum of the responses to six items: number of automobiles, whether they had an individual bedroom, number of computers, number of bathrooms, they had a dishwater, and number of holidays abroad last year (Boyce et al. 2006; Currie et al. 2008). A total between 0 and 4 (9.81%) was considered low family affluence; from 5 to 9 (67.53%), middle family affluence; and from 10 to 13 (22.66%), high family affluence. Family affluence was not reported by .96%, and this subgroup was not used for statistical analysis. Family affluence is related to whether there are income due to work at home and the frequency with which the child is concerned about money, as well as the material goods available to the child (home internet access, two pairs of shoes, school supply, or sports equipment). The mean difference in family affluence according to neighbourhood income is significant at the .01 level (children from neighbourhoods with low income: M = 6.98, SD = 2.15; middle: M = 7.78, SD = 2.05; and high: M = 8.99, SD = 1.94).

Perceived Special Needs

22.31% of the participating children perceived they had at least one personal difficulty or illness, or, in other words, special needs (Government-of-Catalonia 2015). 3.91% expressed ‘having a major visual or auditory difficulty, or some type of malformation (wearing glasses not included)’, 7.81% ‘having a learning difficulty (for example, dyslexia, attention deficit, hyperactivity)’, and 14.15% ‘having a long-term illness (for example, diabetes, an allergy, asthma, growth delay, epilepsy)’. Who reported perceiving special needs had headaches, stomachache, back pain or insomnia more frequently in the previous six months.

Type of School and Location

39.98% of children attended subsidised schools, and 60.02% to general state schools, being the same proportion than the target population. In relation to school location, the home of 63.38% was in the same district where children were studying, of 24.01% was in a different district, and of 12.61% was outside the city.

2.2 Conceptualization

2.2.1 Children’s SWB

This was measured using the Children’s Worlds Subjective Well-Being Scale (CW-SWBS), based on the SLSS (Huebner 1991). This scale measures context-free subjective well-being through 11 point-scales, where 0 means ‘do not agree at all’ and 10 ‘totally agree’. Confirmatory Factor Analysis using the original items did not fit well enough [χ2 = 1072.611, df = 14, CFI = .951, RMSEA = .138 (.131–.145)]. When excluding items 6 (‘I enjoy my life’) and 7 (‘I am happy with my life’), the resulting reduced version displayed good fit indexes [χ2 = 21.107, df = 3, CFI = .999, RMSEA = .039 (.024–.055)], so this version is the one used here. The CW-SWBS-5 showed an excellent internal consistency (α = .916), and a similar mean and distribution for its items. The mean was 8.97 (SD = 1.34), the median 9.4, and the 5th percentile 6.2 (Table 2). Regarding validity, the CW-SWBS-5 showed an acceptable correlation of .689 with the Overall Life Satisfaction single-item indicator, and a moderate correlation with ‘I feel positive about my future’ (.559) and ‘How satisfied are you with what may happen later in your life?’ (.451).

Table 2 Descriptive analysis of the CW-SWBS-5
Full size table

2.2.2 Domain-Based Cognitive SWB

This was measured using 15 domain-based indicators of the cognitive component of SWB taken from the Personal Wellbeing Index School Children, PWI-SC (Cummins and Lau 2005), the Brief Multidimensional Students’ Life Satisfaction Scale, BMSLSS (Seligson et al. 2003), and others from the Children’s Worlds project (for instance, Casas et al., 2012b; Casas et al. 2013b). These items use 11 point-scales, where 0 means ‘not at all satisfied’ and 10 ‘totally satisfied’. The highest means are for satisfaction with ‘health’, ‘things-have’, ‘people-live-with’, ‘house’, ‘friends’ and ‘things-learned’, and the lowest with ‘free-time’, ‘freedom’, ‘life-as-student’ and ‘heard-by-adults’. Note that the median for all the items was 9 or 10 (with the exception of ‘free-time’), which means that more than the half were very satisfied; and that the 5th percentile was usually 6 or less, which means that being none or little satisfied was infrequent (Table 3). Exploratory factor analysis was developed first (the unweighted least squares method was used to extract the factors, one factor being selected by parallel analysis, and the rotation process was orthogonal). The initial factor explained 32.31% of total variance, and when the principal component analysis was applied with 8 fixed factors based on correlations with varimax rotation, they explained 76.97% of variance. These were coherent with the contribution of each item to the initial factor and the microsystems defined in ecological theory, which include psychological and health aspects, family and house/material, other children and school, as well as time use and neighbourhood. Although the factors included coherent domains, the constructed variables were not used in the regression models because each factor only had three, two or one indicator/s.

Table 3 Descriptive analysis of indicators of domain-based cognitive SWB (sorted by mean)
Full size table

2.2.3 Affective SWB

This was measured by means of 6 indicators of the affective component of SWB taken from Russell’s Core Affect theory (Russell 2003) and used in the Children’s Worlds project. These items use 11-point scales, where 0 means ‘not at all’ and 10 ‘all the time’ during the previous two weeks. Positive affect items showed the highest means, and more than half reported being very happy and full of energy, while more than a quarter considered themselves to be very calm. In relation to the negative responses, no more than a quarter said they were very stressed, bored or sad (Table 4). Exploratory factor analysis was conducted. The initial factor explained 23.10% of the total variance, and two factors by means of principal components analysis, 55.63%. These were coherent with the contribution of each item to the initial factor and core affect theory. The constructed variables were not used in the regression models because each factor only had three indicators.

Table 4 Descriptive analysis of indicators of affective SWB (sorted by mean)
Full size table

2.2.4 Interpersonal Relationships

These were measured by means of 21 indicators used in the Children’s Worlds project related to interpersonal relationships between the child and family, other children, in school, and in neighbourhood (5 point-scale, where 0 means ‘I do not agree’ and 5 ‘I totally agree’). Means were always above 3.4, that is, more than half agreed somewhat, a lot or totally, whereas less than a quarter usually did not agree or agreed only a little. Note that where ‘No response’ and ‘Do not know’ were above 4%, this coincided with a mean of below 4 (Table 5). Reliability analysis was developed, and items on interpersonal relationships had very good internal consistency (α = .870). See appendix for a view of how the items were worded.

Table 5 Descriptive analysis of interpersonal relationships indicators (sorted by mean)
Full size table

2.2.5 Activities Outside School

This was measured by means of 12 indicators used in the Children’s Worlds project related to the frequency with which children participate or are engaged in activities outside school (6 point-scales, where 0 means ‘never’ and 6 ‘every day’). Note that participants were asked about ‘frequency of participation’ but not about ‘time use’ (the amount of time spent on particular activities). More than half reported that they did homework, watched television and spend time with family at least five days a week; and more than half did sports, played outside, did home chores, used social media, played electronic games and took care of family members at least three days a week. More than a quarter reported doing nothing, doing extra-classes and working with the family twice a week or less (Table 6). According to Rees (2017a), this type of question set could have some gaps and omissions, so continuous revision is recommended. See the appendix for a view of how the items were worded.

Table 6 Descriptive analysis of activities outside school indicators (sorted by mean)
Full size table

2.3 Data Quality and Analytical Strategy

The questionnaire was administered online through school computers (68.17% of the cases) or electronic tablets (31.87%). Only 5.31% ended up using the paper version due to Internet connection. The initial presentation was made to the whole group, introducing the first block of questions, and the survey was completed with individualized support by the fieldwork team. The questionnaire was available in Catalan (80% preference) and Castilian Spanish (20%). The group sessions lasted one hour (more than 97.5% were able to finish on time, although one hour is rather a long time for this age). The teachers viewed the fieldwork team’s treatment of the children as very satisfactory (mean of 4.5 on a five-point scale).

The collected data are considered trustworthy, departing from the fact that children are competent informants when the instruments are appropriate and the practices respectful with their contextual situations (Casas et al., 2013c; González-Carrasco et al., 2015b; Rodríguez-Pascual 2017). In post-data collection, only nine questionnaires were excluded because less than 40% of the items were responded to (analysed sample = 3962). In the ‘No response’ and ‘Do not know’ options, the children did not express their opinion and these were not included in the data analysis, this leads to different ‘gl2 indicators’ in the multiple regression models. For interpretation purposes, 11-point scales were transformed into three categories according to data distribution (‘Very satisfied’ corresponded to 9–10 points, which are the mode and median; ‘Quite a lot satisfied’ to 7–8, which have the same quantity of answers; and ‘None or little satisfied’ to 0–6, where the number of responses diminishes).

In accordance with Hair et al. (1999), after descriptive analysis, exploratory and confirmatory factor analysis and principal component analysis of the different indicators, multiple regression analysis was used as a way of analysing the relationship between children’s SWB as the criterium indicator and different sets of predictor indicators (domain-based cognitive SWB, affective SWB, interpersonal relationships, activities outside school, and sociodemographic and educational characteristics). It is analysed whether the mean of children’s SWB varies according to the sets of predictors. Estimation was by means of the stepwise method, that is, the indicator that explains most variance was identified, and the rest were added until there were no additional significant variables (statistical significance at .01, multicollinearity was not observed and there were no influential cases). The adjusted R squared was used to contrast the hypothesis that the variation explained by the model is higher than that explained by the mean. To interpret the results, indicators with major change in R squared were taken into account (practical significance). Note that different sets of predictor indicators were used in the analysis because they have different meanings and the aim was to identify their specific contribution; otherwise, some indicators could cover the critical contribution of others. Moreover, since data for quantitative variables do not have a normal distribution, as well as using an initial multiple regression analysis due to there being little difference and it being habitual, the continued use of logistic models is also recommended (Ferrer-i-Carbonell and Frijters 2004).

3 Results

With regard to indicators of domain-based cognitive SWB (Table 7), a model with 13 of the 15 indicators contributes to 55.07% of the CW-SWBS-5 scale’s variance. The indicators with the highest contribution are ‘heard-by-adults’ (adjusted R squared: .271) and ‘safety’ (adds .104). The adjustment addition of ‘time-use’, ‘body’ and ‘house’ ranges from .052 to .028; and of ‘classmates’, ‘life-as-student’ and ‘people-live-with’ from .017 to .010. Although ‘freedom’, ‘things-have’, ‘friends’, ‘health’ and ‘free-time’ modify the adjusted R square from .006 to .002, their practical contribution is not evident. The domains with no statistical contribution are ‘things-learned’ and ‘neighbourhood’.

Table 7 Contribution of domain-based cognitive SWB to children’s SWB
Full size table

Concerning the affective SWB indicators (Table 8), a model with the six affect indicators contributes to 42.75% of the CW-SWBS-5 scale’s variance. The strongest contribution is with ‘happy’ (adjusted R squared: .387). The adjustment addition of ‘sad’ and ‘full-of-energy’ ranges from .015 to .014. Although ‘calm’, ‘bored’ and ‘stressed’ modify the adjusted R squared from .009 to .001, their practical contribution is not evident.

Table 8 Contribution of affective SWB to children’s SWB
Full size table

Concerning interpersonal relationships (Table 9), a model with 12 of the 21 indicators contributes to 39.42% of the CW-SWBS-5 scale’s variance, the most significant being ‘We have a good time together in my family’ (adjusted R squared: .165), along with ‘I have enough friends’, ‘My parents give me enough freedom’, and ‘I feel safe when I walk around in the area I live in’ (adds between .093 and .031).

Table 9 Contribution of interpersonal relationships to children’s SWB
Full size table

Concerning activities outside school (Table 10), a model with 4 of the 12 indicators contributes to 11.52% of the CW-SWBS-5 scale’s variance, ‘Relaxing, talking or having fun with family’ being the most significant (adjusted R squared: .097).

Table 10 Contribution of activities outside school to children’s SWB
Full size table

Finally, concerning sociodemographic and educational characteristics (Table 11), living in two family nuclei with a greater frequency of a single parent, low family affluence, perceived special needs, not living in one family with two parental figures, not being born in Spain and the parents neither, or low neighbourhood income, all contribute to lower CW-SWBS-5 (4.76% of the CW-SWBS-5 scale’s variance). This low percentage means that there are very few children with these characteristics who are not sufficiently satisfied with their lives.

Table 11 Contribution of sociodemographic and education characteristics to children’s SWB
Full size table

4 Discussion

Taking into account the optimism bias, the question is whether ‘children share high common SWB simply because they are children’ or whether and how, in spite of being children, there are different reasons to be more or less satisfied (Casas 2011, Ben-Arieh et al. 2017). Through the analysis developed here, children’s SWB measured by CW-SWBS-5 would appear to be related to both domain-based cognitive SWB and affective SWB (55.07% and 42.75% of variance explained, respectively). Children’s SWB is higher when domain-based cognitive evaluations have higher scores; being satisfaction with ‘heard-by-adults’, ‘safety’, ‘time-use’, ‘body’ and ‘house’ the ones that most contribute. Children’s SWB is also related to affect indicators; being ‘happy’ the one that most contributes. In addition, children’s SWB seems to be related to interpersonal relationship indicators (39.42%), and the ones with the highest contributions are ‘We have a good time together in my family’, ‘I have enough friends’, ‘My parents give me enough freedom’, and ‘I feel safe when I walk around in the area I live in’. Children’s SWB seems to be low but statistically related to indicators of activities outside school (11.52%); being ‘Relaxing, talking or having fun with family’ the one that most contributes, and to sociodemographic and educational characteristics (4.76%), while none of the latter makes any important contribution. Further research could add some age-adapted psychological and social well-being scales based on previous theories (Ryff and Keyes 1995, Keyes 1998) for a better comprehension of children’s SWB.

The results for children’s SWB in Barcelona city in 2017 are similar to those found by different authors using the Children’s Worlds project databases, but with some new and suggestive findings. Previous results suggested that children’s SWB was influenced in different countries by housing, safety, bullying, school achievements and social interactions (Dinisman et al. 2015), and in relation to domain-based cognitive SWB, by the freedom to choose and self-perception, but also social support, friendliness, and choice about time (Bradshaw and Rees 2017; Lee and Yoo 2017). The analysis carried out here has identified satisfaction with ‘heard-by-adults’ (also in Casas et al., 2012a, 2013a), ‘safety’, ‘time-use’, ‘body’ and ‘house’ to add. In fact, being satisfied with ‘heard-by-adults’ and ‘safety’ may have a similar significance to self-perception, while ‘time-use’ may be similar to freedom to choose and choice about time, whereas housing may still make some type of contribution to children’s SWB. In relation to affective SWB, which is part of SWB (Casas 2018, Busseri 2018), this analysis shows that happiness makes the highest affect contribution to children’s SWB. It is important to recognize that children who are satisfied with their lives are feeling happy, and that children may feel sad, bored or stressed without subsequently making a negative evaluation of their lives. Further research could focus on affective SWB in context with school lives and their interpersonal relationships, as well as body and health.

As regards interpersonal relationships, frequency of family and of peer activities and neighbourhood safety were identified as being important (Lee and Yoo 2015). In this analysis, ‘We have a good time together in my family’, ‘I have enough friends’, ‘My parents give me enough freedom’, and ‘I feel safe when I walk around in the area I live in’ displayed a significant contribution to children’s SWB, and all of them showed associations with family, other children and neighbourhood. This suggests that apart from family, the other children play an important role in socialization and identity processes of children, as well as feeling free and safe in the place they live, especially in school. In relation to activities outside school, it was found that sports, watching television and using computers games are the most frequent activities (Rees 2017b; Rees and Dinisman 2015), while doing homework and talking with family can be added according the analysis. Mainly, just as Lee and Yoo (2015) posited, what is striking is that only spending time with family seems to have a critical contribution to children’s SWB.

In relation to sociodemographic and educational characteristics, it was found that, although exosystems and macrosystems have a low statistical relation with children’s SWB, being deprived of economic and material resources could be critical (Castellá et al. 2015; Crous 2017; Crous and Bradshaw 2017; Dinisman and Ben-Arieh 2016; Gross-Manos 2017; Lee and Yoo 2015). In Barcelona city in 2017, a child who lived in a neighbourhood with low income tended to express a lower SWB than children who lived in richer neighbourhoods. And children from less affluent families, or, in other words, who are deprived, tend to express a lower SWB than those from a more affluent family. As observed in previous Spanish samples (Casas et al., 2012a, 2013a; Dinisman et al. 2012), children with foreign origins tend to express a lower SWB than those without. And children who do not live in a family nucleus with two parental figures also tend to express a lower SWB (Dinisman et al. 2017).

Three results can be highlighted that have not been reported in previous articles related to the Children’s Worlds project: (1) children with perceived special needs tend to be less satisfied, meaning that more targeted analysis is needed for this group, starting with a systematic review of the literature; (2) type of school and location seem not to be related to overall children’s SWB, although this relationship could be analysed in domain-based cognitive SWB indicators, such as self-perception related to school and time (Bradshaw et al. 2017; Casas and González-Carrasco 2017; Kutsar and Kasearu 2017); and (3) at 10–12-years-old, gender seems to be unrelated to overall children’s SWB, although domain-based cognitive SWB indicators could manifest gender differences (González-Carrasco et al., 2015a; Grigoraş et al. 2018), and it must be considered whether gender differences exist among other ages or types of children (Lawler et al. 2010; Montserrat et al. 2015; Kaye-Tzadok et al. 2017; Yoo and Ahn 2017).

Some strengths have been identified to promote positive development (Pollard and Lee 2003), as well as some children’s problems or needs (Ben-Arieh et al. 2014; Casas et al. 2014). By way of support for policy and the development of services (McAuley and Rose, 2010), as observed by the pattern in the Spanish samples (Dinisman and Rees 2014; Rees and Main 2015), there is a priority margin for improvement in some domain-based cognitive SWB indicators (Table 2). That is, since a vast majority of children are very satisfied with specific indicators, the small percentages of those who are not are valuable because they indicate children in vulnerable situations (Brazier 2017). This is the case with ‘health’, or ‘things-have’, ‘people-live-with’ and ‘house’, or ‘friends’ and ‘things-learned’. What is more, other evaluated domains have an even greater margin for strategic improvement, as the percentage of ‘very satisfied’ responses moves away from 100% (‘heard-by-adults’, ‘life-as-student’, ‘freedom’ and ‘free-time’). In addition, domain-based cognitive SWB indicators with reciprocal statistical correspondence suggest potential improvement effects in these domains (Table 3). For instance, between ‘life-as-student’ and ‘things-learned’; between ‘health’ and ‘body’; between ‘freedom’, ‘heard-by-adults’ and ‘safe’; between ‘classmates’ and ‘friends’; and between ‘time-use’ and ‘free-time’. It is also worth noting that the data suggest life satisfaction and aspirations have different meanings (Table 2).

By improving their satisfaction, domain-based cognitive SWB indicators with the highest contribution to children’s SWB could have a great impact on their SWB (Table 7). This is the case when it comes to satisfaction with ‘heard-by-adults’, ‘safety’, ‘time-use’, ‘body’ and ‘house’. Thus, following the 12th article of the UNCRC and their General Comment No. 12 (CRC/C/GC/12 2009), about the right of the child to be heard, that children feeling adequately heard by the adults around them could be the first step to giving attention to children’s lives and improving their well-being and, indirectly therefore, their security. Moreover, when children are satisfied with their life, they are happy and have a more positive perception of specific interpersonal relationships such as ‘We have a good time together in my family’, or activities outside school, such as ‘Relaxing, talking or having fun with family’ (Table 8, 9, 10). That is the reason why we should all take adequate account of the quality and quantity of time that parents spend with their children. And by extension, the closest people in their lives.

On the other hand, although not a determining factor, despite children’s optimism bias, it is more likely that children express lower SWB when (a) living in two family nuclei with a greater frequency of a single parent, (b) their family is less affluent, (c) they have perceived special needs, (d) they do not live in a family nucleus with two parental figures, (e) they were not born in Spain and neither were their parents, or (f) their neighbourhood income is low (Table 11). Therefore, promoting better evaluations of cognitive domains (especially feeling adequately heard by adults) and enjoying relationships (especially with family members) both could improve all children’s lives. Furthermore, activities outside school, and in particular sociodemographic and educational characteristics, both could be used to identify children in vulnerable situations. Further research must focus on SWB of children affected by social inequalities.

Note that data was collected through an adapted preliminary version of the third wave of the ISCWeB using a probabilistic sample of Year 5–6 students. The findings can be generalized to the target population (Year 5–6 students in Barcelona city in 2017) due to sampling design, and could be a reference for exploring children’s SWB in a society similar to Barcelona city. The findings cannot be assumed to pertain to younger children, because they could have different needs or problems, and neither to adolescents because they are also in another phase of their education. Given that a self-response questionnaire was used, it is recommended that other sources also be taken into account when making decisions for improving children’s lives. This was the first time the survey has been used with a specific sample in Barcelona city, and data collection should be carried out periodically to find out whether there are changes over time. This is a cross-sectional study and no cause-effect relationship can be established. It could be timely and consistent to use the questionnaire as a research technique for assessing social or educational interventions based on quasi-experimental designs. Finally, it makes sense to state that if we want to gain an in-depth understanding of children’s SWB, it is also necessary to use other research techniques that improve their reflection and participation.