Safety, Perceptions of Good Treatment and Subjective Well-Being in 10- and 12-year-old Children in Three Countries

Satisfaction with safety and satisfaction with how adults listen to children and how they take what they say into account are the most important satisfaction domains that contribute to children’s subjective well-being (SWB). However, there is still more to know about what contributes to both domains. Little is understood about their mediating effects on the safety perceptions of being cared for and supported in terms of children’s SWB. Age and country are also relevant variables in this equation that suffers from a lack of information. Therefore, this article attempts to shed light on these questions by using the third wave of the Children’s Worlds data set that covers Spain, Brazil, and Chile and focuses on the 10- and 12-year-old age group. Results show that perceived contexts (home, school, and neighborhood) in each country are very important for assessing satisfaction with personal safety, while having parents who listen and take children’s opinions into account is very important for SWB in all contexts. The importance of the effects of feeling safe on SWB increases from late childhood to early adolescence, with its indirect effects being much more important than direct effects. Most children do not perceive to be listened to by teachers or do not perceive that what they say is taken into account by their teachers, which does not turn out to be very relevant for SWB in any of the three countries. Despite existing relationships between all of the variables analyzed, there are differences depending on the country and age group, with a common relationship observed between some of them.


Introduction
Background Satisfaction with safety is part of Cummins and Lau (2005) Personal Well-Being Index-School Children (PWI-SC) (Cummins & Lau, 2005, 3 rd edition) jointly with other six items regarding satisfaction, each one corresponding to different quality-of-life domains (standard of living, health, life achievements, personal relationships, personal safety, community connectedness, and future security). According to the authors, these seven items are theoretically embedded as representing the first level deconstruction of the global question 'How satisfied are you with your life as a whole? (Vaqué, 2014). This question serves as a measure of SWB homeostasis both in adults and children (Cummins, 2014), being a mechanism controlled and regulated by automatic neurological and psychological processes. The ultimate purpose of the SWB homeostasis is to maintain a normally positive sense of well-being.
The Convention on the Rights of the Child (CRC) (UNICEF, 1989) plays an "instrumental role in setting the research on issues related to children and serving as a foundation for frameworks for data collection and analysis" (Mekonen & Tiruneh, 2014, p. 2,496). Different perceptions and evaluations of children regarding their rights have shown to be correlated with their SWB (Casas, 2017;Casas et al., 2018). For example: Children's general perceptions of adults in their country or their neighborhood respecting children's rights (Casas et al, 2022), or adults listening to them and taking what they say into account , children's perceiving that their participation rights are respected in the family, in the school or in the municipality (Casas, 2017), children's own assessment on whether they are well-treated at home and school (Casas, 2017).
Many of the frequently used SWB measurement instruments do not fully address many aspects that are core to children's rights (Lundy, 2014), such as satisfaction with how adults listen to children and take what they say into account. Like in the case of human rights (Chapman, 1996;Tisdall, 2015), research on children's rights has tended to focus more on the non-fulfilment of rights, and thus on its negative impact, rather than on the positive effects of their fulfilment or in the promotion of their rights (Casas, 2018). For example, Barsch (1993) observed that international research on the right to freedom is mostly focused on the lack of "objective" freedom in many countries, than in the subjective feelings of freedom by citizens. This article is an attempt to link the research on some children's rights, especially the right to participate and being listened to, with that of child subjective well-being.

3
Safety, Perceptions of Good Treatment and Subjective Well-Being… However, little is known about the relationships between satisfaction with safety and satisfaction with how adults listen to children and take what they say into account, and what contributes to these two satisfaction variables themselves, specially the latter one. Their potential mediating role between perceptions of safety and of been listened to and taken into account and children's SWB, is also in need to be explored.
From the scientific literature we know that being cared for, supported and respected are important for children's SWB (see Ahmed et al., 2021, and the review by Lippman et al., 2014, on positive and protective factors in adolescent well-being). It is likely that adults perceived as caring, supportive, and respectful are frequently also perceived by children as being listened and being taken into account by them. It has already been demonstrated that being heard is an important prerequisite for children perceiving that their rights are respected by adults and for developing selfconfidence (af Ursin & Haanpää, 2018). We hypothesize that perceptions of being cared for, supported, and respected will contribute to satisfaction with how adults listen to children and take what they say into account.
Safety is also connected with perceptions of being cared for as illustrated by Ahmend et al. (2021), since the participant children aged 8 to 12 qualitatively defined safety in terms of being loved, respected, and cared for by significant adults and being able to do things without restrictions. This explains why feelings of insecurity are closely linked to an inferior children's agency and autonomy (Steckermeier, 2019) while jeopardizing the rights recognized in the UN Convention on the Rights of the Child (1989). Steckermeier (2019) also states that children conceptualize safety under a local rather than socio-political perspective, which explains the approach we used, considering that perceived safety in different life contexts (e.g., home, school and neighborhood) will contribute to overall satisfaction with safety as observed by González-Carrasco et al. (2018) with samples from Spain, Algeria, Israel, and South-Africa and in 16 countries by Steckermeier (2019). These latter two studies were carried out with data belonging to the second wave of the ISCWEB international project also known as Children's Worlds (www. isciw eb. org). So far, it has collected three waves of data from different countries with the objective of having solid representative data on children's lives, daily activities, time use, and, particularly, their own perceptions and evaluations of personal well-being.
The third wave dataset used here included 12 items that are of interest for the analysis pursued in this research with 10-and 12-year-old age groups (henceforth referred to as 10yo and 12yo). These 12 items are related to children's feelings of safety, perceptions of respect, being listened to, cared for and supported, and their evaluation of satisfaction with two aspects of life (satisfaction with safety and satisfaction with how adults listen to children and take what they say into account). It also includes the items of a SWB psychometric scale (the Children's Worlds Subjective Well-Being Scale, CW-SWBS). Casas and González-Carrasco (2021) have pointed out the excellent fit and cross-country comparability of the five-item version of this scale (the CW-SWBS5).
We have chosen three countries (Brazil, Chile, and Spain) because previous analyses of these same three countries showed that satisfaction with safety contributed most to overall life satisfaction ; however, satisfaction with how adults listen to children and take what they say into account was not measured at that time. Besides, according to the UN Office in Drugs and Crime's International Homicide Statistics database, these three countries had very different crime rates in 2020 (measured through intentional homicides per 100,000 people), combined with a different time-trend from 2019 (Spain: 0.64, a decrease of 10.54%, Chile: 4.84, an increase of 22.25%, and Brazil: 22.4, an increase of 7.51%), which provides more reasons to compare them. A further reason is that these three countries ranked very differently in terms of children's SWB according to the results of the third wave of the Children's Worlds project. Specifically, Spain had the 5 th highest mean of 35 (quartile 1) for the 10-year-olds group and the 7 th of 30 (quartile 1) for 12-year-olds, Chile the 13 th (quartile 2) and 21 st (quartile 2), and Brazil the 19 th (quartile 2) and 29 th (quartile 4), respectively.
As a matter of contextual information and according to the UNICEF Brazil 2019a report, there had been a considerable progress in the fulfilment of children's rights with child mortality at that time showing a historical drop and a decrease in the rate of out-of-school children. However, important challenges persisted such as multi-dimensional poverty and violence. In the case of Chile, children were also among the four priorities of the Government agenda (UNICEF Chile, 2019b). However, 2019 was marked by violent protests and a dramatic increase of immigration, mainly from Venezuela. Children being institutionalized remained a major challenge since two out of 11 large institutions were closed and another one in the process of being closed.
Finally, in Spain, advances were made in relation to children's poverty but it still remains among the European countries with the highest children's poverty rates and the highest drop-out from school rates (Save the Children, 2019). Alone immigrant children were not only being left behind but even criminalized publicly while stronger economic investment in helping families was especially needed.
More information of the three countries' characteristics can be found in the country reports for the 3 rd wave of the Children's Worlds project (https:// isciw eb. org/ thedata/ publi catio ns/ count ry-repor ts/ count ry-repor ts-of-the-third-wave-2016-2019/). Determinants of children's SWB are sensitive to age González-Carrasco et al., 2020;González-Carrasco et al., 2017a, b;Kim et al., 2019) as are their safety perceptions (González-Carrasco et al., 2018), which is why separate analyses of the relationships between variables will be carried out for the 10yo and the 12yo.

Aim
The aims of this article are: (a) to explore the contributions of feeling safe in different contexts and of being cared for, supported, and respected by relevant people (parents/care givers and other family members, teachers and friends) to the SWB of children aged 10 and 12 in Brazil, Chile and Spain; (b) to analyze whether satisfaction with safety and satisfaction with how children are listened to and how they are taken into account by adults are factors that mediate the influence of feeling safe, being cared for, supported, and respected on the SWB of these children; (c) to 1 3 Safety, Perceptions of Good Treatment and Subjective Well-Being… analyze whether the relationships between all explored variables differ depending on the country and age group.
Additionally, three hypotheses will be tested regarding how the mediation effects work: (1) Feeling safe in different contexts (i.e.: home, school, and the area where the child lives) will contribute positively to satisfaction with safety, (2) feeling being cared for, supported, and respected will contribute positively to satisfaction with how adults listen to children and take what they say into account, (3) the indirect effect of feeling safe in different contexts on SWB, through satisfaction with safety, will be stronger than its direct effect. As the indirect effect of feeling being cared for, supported. and respected on SWB, through satisfaction with how adults listen to children and take what they say into account will also be stronger.

Data
The data set used in this study was taken from the third wave of the International Survey of Children's Well-Being (ISCWeB), partially founded by the Jacobs Foundation, and collected between 2016 and 2019 through a questionnaire. It is composed by 35 countries and 128,184 children, making it the world's largest database on children's subjective indicators, which is why it has been chosen for this article. Two different versions of the questionnaires were used; one for the 10yo and 12yo age groups and a different one for the 8yo age group, which had fewer items and a different format. An international committee supervised the design of data collection in order to guarantee appropriate representativeness of the data for each region or country. More details on the data collection procedure in each country can be obtained from the project webpage: https:// isciw eb. org/

Sample
A representative sample of a region or large cities was obtained from each of the countries as indicated in Table 1. The final sample consisted of N = 3,790 children from the 10yo group and N = 3,780 from the 12yo group. 1 3 Data collection was organized according to class groups at school. The overall mean age of each age group was slightly above the selected age. Age distribution was not exactly the same in all countries, as might be expected given that children were at the start of the school year in Spain (Northern Hemisphere) and near the end of it in Brazil and Chile (Southern Hemisphere) (Table 1).

Measures
The third wave ISCWeB questionnaires included the following items of agreement/ disagreement on a 0 to 4 scale: Three items on safety: 'I feel safe at home'; 'I feel safe at school'; 'I feel safe when I walk in the area I live in'. Seven items on perceptions of respect, being listened to, cared for, and supported: 'My parents/carers listen to me and take what I say into account'; 'My friends are usually nice to me'; 'My teachers listen to me and take what I say into account'; 'There are people in my family who care about me'; 'My teachers care about me'; 'If I have a problem, people in my family will help me'; 'If I have a problem, I have a friend who will support me'.
It also included two items on satisfaction, measured on a 0-10 scale ranging from 'Totally disagree' to 'Totally agree'. They are satisfaction with how safe you feel and satisfaction with how you are listened to by adults in general.
The Children's Worlds Subjective Well-Being Scale: The new CW-SWBS is an improved version of the context-free multi-item SWB scale used in the second wave. Improvements were made using advices from children in different countries, who were asked to suggest new wordings where items were not clearly stated. This was more common in countries where Indo-European languages are not spoken. It was scored on an 11-point scale from 'Totally disagree' to 'Totally agree'. Its psychometric properties have been cross-country analyzed by Casas and González-Carrasco (2021), who demonstrated that a five-item version, excluding the item I like my life displayed better fit and cross-country comparability. Therefore, this CW-SWBS5 version is used in the present study.

Procedure
The data collection took place in the school setting. Approval was obtained from an appropriate ethics committee in each country before the survey was applied. The data set used for this analysis had been cleaned and prepared by the international project before making the database available. This process included identifying and excluding cases with high proportions of missing data, and identifying and excluding cases with systematic response patterns. The latter point is of particular relevance to this article given that some authors, such as Cummins and Lau (2005), recommend excluding from the analysis respondents who score at the top or bottom of the scale for all items in the PWI-SC (Personal Well-Being Index-School Children version) because consistent answers at the extremes tend to be unreliable. However, this presents problems with this approach, since that it automatically excludes anyone who expresses complete satisfaction with all aspects of their life covered in one instrument.
Many children at these ages seem to be extremely satisfied with their lives, even if that does not mean they are satisfied with all aspects or dimensions in their life measured by other instruments. According to Rees and Main (2015), it is possible to adopt a broader approach when using Children's Worlds data sets because the questionnaire includes several different sets of items. Therefore, uniform response patterns were identified for five different sets of items in the data set and cases were excluded if they exhibited such patterns for more than one of the five-item sets (Rees & Main, 2015).
The sample used in this paper was cleaned a second time. Specifically, from all the individuals in the Children's Worlds data set from the 3 studied countries, N = 218 (5.44%) 10-year-olds and N = 225 (5.62%) 12-year-olds were excluded from the data base because they had 3 or more missing values in the items belonging to the CW-SWBS5. Their distribution was similar by country and gender, and no other characteristic was identified for the excluded subgroup. From the remaining individuals still N = 218 (5.44%) 10-year-olds and N = 175 (4.37%) 12-year-olds displayed some missing value in the items here used, and these scores were submitted to Full Information Maximum Likelihood (FIML) estimation for regression imputation as implemented in AMOS25 software, which takes into account all variables in the model. FIML is considered an appropriate and commonly used approach for estimating missing data in structural equation modeling, since it produces unbiased estimates under MAR (missing values at random) and MCAR (missing completely at random) conditions (Cham et al., 2017).

Data Analysis
First, descriptive statistics for the variables analyzed were calculated. Then, using a structural equation modelling (SEM), a model relating the 10 items on safety, perceptions of respect, being listened to, cared for, and supported to the CW-SWBS5 latent variable was designed and separately tested for each age group. The comparability of the data across the three countries was checked using multi-group SEM (MGSEM). The AMOS 25 software with maximum likelihood estimation was used. Because subjective/psychological well-being data usually departs from normality, data were handled in structural equation models by means of the bootstrap method to compute standard errors.
After a good fit and comparability of data were established, a new model was designed using the pooled sample, including the two satisfaction items mediating between the ten items on safety, perceptions of respect, being listened to, cared for and supported, and the SWB indicator (the CW-SWBS5 latent variable). All direct and indirect effects were included in the model in order to analyze which relationships were significant. After excluding the non-significant relationships, a final model was tested and results were analyzed using MGSEM with constrained loadings and intercepts in order to check for metric and scalar invariance. Metric invariance allows for a meaningful comparison of correlations and regressions, while scalar invariance allows for a meaningful comparison of the latent means (Coenders et al., 2005). When any constraint is added to a model, a change in any fit index of more than 0.01 is not considered acceptable (Chen, 2007;Cheung & Rensvold, 2002).
The fit indices used were the CFI (comparative fix index), the RMSEA (root mean square error of approximation) and the SRMR (standardized root mean square residual). We concluded that results higher than 0.950 for CFI and below 0.05 for RMSEA and SRMR are excellent, in accordance with Arbuckle (2010) and Byrne (2010).
Squared multiple correlations (SMC) were calculated with each model because they indicate how accurately each variable is predicted by the other variables in the model (Arbuckle, 2010;Byrne, 2010). Additionally, the remaining per cent variance is accounted for by its single-factor error. If error was attributable to measurement error only, we could say that the estimated reliability of the variable is the value displayed for each SMC variable. Therefore, each SMC value is an estimate from the lower band of reliability relating to its variable (Arbuckle, 2010;Byrne, 2010). Table 2 shows descriptive statistics for the items used in this research with the pooled sample. Details regarding each country are displayed separately in Tab s1 (supplementary material). Among the 10-year-olds, the lowest scores for the items of feeling safe, and perceptions of being respected, being listened to, cared for, and supported corresponded to I feel safe when I walk in the area I live in, while the highest were for I feel safe at home and There are people in my family who care about me, the mean for satisfaction with safety being higher than that for satisfaction with being listened to. In the case of the 12-year-olds, the lowest mean was for My teachers care about me and the highest for I feel safe at home. Within this age group, the mean for satisfaction with safety was also higher than that for satisfaction with being listened to.

Descriptive Results
Tab s2 (supplementary material) shows Pearson correlations between the studied variables. All of them were statistically significant and moderate for both the 10yo and the 12yo groups. The highest correlation for the 10yo corresponded to My teachers care about me and My teachers listen to me and take what I say into account (0.554) followed by satisfaction with how you are listened to by adults in general and satisfaction with how safe you feel (0.481). For the 12yo group, the highest correlations were My teachers listen to me and take what I say into account and My teachers care about me (0.656) and satisfaction with how you are listened to by adults in general and satisfaction with how safe you feel (0.571).

3
Safety, Perceptions of Good Treatment and Subjective Well-Being…

SEM Including 10 Items
A model that links the 10 items on safety, perceptions of respect, being listened to, cared for, and supported to the CW-SWBS5 latent variable was separately tested for each age group (Models 1 and 5 in Table 3; Fig. 1). Fit was excellent for both models, as well as for the same models analyzed as multi-group by country (Models 2 and 6 in Table 3). The same model with constrained loadings (Models 3 and 7) and with constrained loadings and intercepts (Models 4 and 8) still showed excellent fit. The decrease with each additional constraint was < 0.01, supporting metric and scalar invariance, i.e., comparability of all statistics among the three countries.
In Tab s3 (supplementary material) detailed results are offered for Model 4, and in Tab s4 (supplementary material) for Model 8 using bootstrapping. The most noteworthy results are: Regression weights of the item My teachers care about me of the SWB indicator are not significant for either age group in any of the three countries. If I have a problem, I have a friend who will support me is non-significant for the 12yo in all three countries, but only in Chile is non-significant for the 10yo. My teachers listen to me and take what I say into account is non-significant for both age groups in Chile and Brazil, but it is significant in Spain. There are people in my family who care about me is non-significant for the 10yo in Spain. I feel safe when I walk in the area I live in is non-significant for both age groups in Brazil, whereas My friends are usually nice to me is non-significant only for the 10yo.
The largest regression weights for the CW-SWBS5 latent variable for the 10yo are observed for I feel safe at home in Brazil and Spain and for If I have a problem,  The explained variance (measured by the SMC means) of the latent variable of the 10 items analyzed here is 26.9% in Spain, 38.8% in Brazil, and 43.9% in Chile for the 10yo group, and 42.7% in Spain, 49.2 in Brazil, and 62.2% in Chile for the 12yo group.

SEM Mediation Models Including the Two Satisfaction Variables
Next, a new model was created using the pooled sample, including the two satisfaction items mediating between the 10 items on safety, perceptions of respect, being listened to, cared for and supported, and the SWB indicator (the CW-SWBS5 latent variable). All direct and indirect effects were included in the model in order to identify which relationships were significant (Fig. 2).
Seven of the tested relationships for the 10yo and eight for the 12yo were nonsignificant. Four of them were common to both age groups. First, I feel safe when I walk in the area I live in, My teachers listen to me and take what I say into account, and My teachers care about me did not display any direct significant effects on the CW-SWBS5 latent variable. Second, My teachers care about me did not show any significant direct effects on Satisfaction with how safe you feel.
The other non-significant relationships were age-group specific. My teachers listen to me and take what I say into account and There are people in my family who care about me did not display any significant effects on Satisfaction with how safe you feel for the 10yo. There are people in my family who care about me did not display significant effects on Satisfaction with how you are listened to by adults in general.
If I have a problem, I have a friend who will support me did not display any significant direct effects on the CW-SWBS5 latent variable nor on Satisfaction with how safe you feel or on Satisfaction with how you are listened to by adults in general for the 12yo. I feel safe at school did not show any significant effects on Satisfaction with how you are listened to by adults in general.
After excluding these non-significant relationships from each age-group model, the new models were tested and their results analyzed using MGSEM by country with constrained loadings and intercepts in order to check for metric and scalar invariance (i.e., cross-country comparability of the data; Fig s1 to s6 in supplementary material).
The fit of all models was excellent and both metric and scalar invariance was supported for the two age groups, meaning that we can meaningfully compare all statistics across the three countries (Table 4). Detailed results are displayed in Tab s5 to s7 (supplementary material). Some remarkable results are described next.
The effects of Satisfaction with how you are listened to by adults in general and Satisfaction with how safe you feel on the SWB latent variable are important and significant in all countries for both age groups, ranging from 0.196 to 0.352. For the 10yo, explained variance (measured by SMC means) ranges from 0.391 in Spain to 0.571 in Chile for the CW-SWBS5 latent variable, from 0.160 in Spain to 0.298 in Chile for Satisfaction with how safe you feel, and from 0.228 in Brazil to 0.391 in Chile for Satisfaction with how you are listened to by adults in general. Therefore, all explained variances are larger in Chile than in the other two countries, suggesting  .020 that this is an excellent model for this country and that fewer additional variables are likely needed to improve the Chilean model. The contribution of Satisfaction with how you are listened to by adults in general to the SWB latent variable is higher in Brazil and Chile than the contribution of Satisfaction with how safe you feel to the same latent variable for the 10yo, while in Spain it is the other way around. The contribution of Satisfaction with how safe you feel to the SWB latent variable is higher only in Brazil, while the contribution of Satisfaction with how you are listened to by adults to the same latent variable is higher in Chile and Spain for the 12yo. These results suggest that SWB is sensitive to age and to the different contexts where children live.
The item with the largest significant contribution to Satisfaction with how you are listened to by adults in general, for both the 10yo and 12yo, is My parents/carers listen to me and take what I say into account in all three countries, which is higher in every case for the 12yo group. However, the highest contribution to Satisfaction with how safe you feel varies by country and age group. In Brazil I feel safe when I walk in the area I live in makes the highest contribution for both age groups. In Chile it is I feel safe at school for the 10yo and I feel safe at home for the 12yo, while in Spain it is I feel safe at home for the 10yo, and I feel safe at school for the 12yo. These results suggest the different perceived contexts in each country are very important to evaluate satisfaction with personal safety, while having parents who listen and take their child's opinions into account is important for SWB in all contexts.
I feel safe at home displays significant direct (> 0.1) and indirect effects through Satisfaction with how safe you feel on the SWB latent variable for the 10yo in all three countries whereas it is only significant in Spain through Satisfaction with how you are listened to by adults in general. The direct effects for the 12yo are much lower (non-significant in Brazil), while indirect effects are much more notable (> 0.1) and significant in all three countries. This particularly applies to indirect effects through Satisfaction with how you are listened to by adults in general, suggesting early adolescents more strongly relate safety to being listened to and taken in to account by adults than 10yo children do.
I feel safe at school displays significant direct effects on the SWB latent variable in all three countries, but they are larger than 0.1 only in Spain for both age groups, while indirect effects mainly through Satisfaction with how safe you feel are significant and noticeable in the three countries and both age groups. Indirect effects of this item through Satisfaction with how you are listened to by adults in general are only significant in Brazil and Spain for the 10yo.
I feel safe when I walk in the area I live in shows no significant direct effects on the SWB latent variable, but significant indirect effects mainly through Satisfaction with how safe you feel in the three countries and both age groups as well as through Satisfaction with how you are listened to by adults in general in Chile and Spain for both age groups.
My parents/carers listen to me and take what I say into account displays significant direct effects on the SWB latent variable only in Spain for the 10yo and in all three countries for the 12yo. This item also shows significant indirect effects though Satisfaction with how safe you feel in Chile and Spain for the 10yo and in the three countries for the 12yo, suggesting the importance of its effects on SWB increases from late childhood to early adolescence and that its indirect effects are much more important than direct effects.
My friends are usually nice to me shows significant direct effects on the SWB latent variable only in Chile for the 10yo and in the three countries for the 12yo, but they are > 0.1 only in Spain, while indirect effects are significant in Chile and Spain for both age groups through Satisfaction with how safe you feel.
The item My teachers listen to me and take what I say into account shows only significant indirect effects on the SWB latent variable through Satisfaction with how you are listened to by adults in general for the 10yo in the three countries, but only in Spain for the 12yo and through the two mediating variables, suggesting most children do not perceive to be listened to by teachers or for teachers to take what they say into account, something being not very relevant for their SWB.
The direct effects of the item There are people in my family who care about me on the SWB latent variable are significant in Brazil and Chile for the 10yo and in Chile and Spain for the 12yo. Indirect effects are only significant in Chile for the 12yo through Satisfaction with how you are listened to by adults in general.
My teachers care about me only displays significant indirect effects on the SWB latent variable through Satisfaction with how you are listened to by adults in general in Brazil and Spain for the 12yo, suggesting most children do not consider to be cared for by teachers, something not perceived as very relevant for their SWB.
If I have a problem, people in my family will help me also displays significant direct effects on the SWB latent variable in Brazil and Chile (> 0.1) for the 10yo and in all three countries for the 12yo (> 0.1), with significant indirect effects only in Chile for the 10yo and in the three countries for the 12yo, except for in Spain where Satisfaction with how safe you feel is the mediator variable.
If I have a problem, I have a friend who will support me displays significant small direct effects on the SWB latent variable in Brazil and Spain, but only significant indirect effects through Satisfaction with how safe you feel in Spain for the 10yo, and no significant effects for the 12yo in any of the countries. This suggests that most children do not consider they would be supported by friends if they have a problem, something not perceived as very relevant for their SWB.
Finally, implied means of the endogenous observed variables with constrained loadings and intercepts for the final models are detailed for each country in Tab s7 of the supplementary material, with the lower and upper bootstrap bounds and significance. Scores are the highest in Spain and the lowest in Brazil with Chile in between for all items. The mean scores for Satisfaction with how safe you feel are higher than for Satisfaction with how you are listened to by adults in general in the three countries and for both age groups. Table 5 summarizes the results of the standardized regression weights displayed in Tab s3 and Tab s4 (supplementary material) for the three safety perceptions and the seven indicators of perceptions of being respected, supported, and cared in relation to the SWB latent variable.

Discussion
Regarding the first aim of this study, the direct contribution that the three items on feeling safe in different contexts and the seven items on being cared for, supported, and respected by different people make to an SWB indicator (the CW-SWBS) among children aged 10 and 12 in Brazil, Chile, and Spain is significant in most cases, with some exceptions. The first result therefore lends support to the first hypothesis formulated. Furthermore, perception of safety in the surrounding area and satisfaction with safety contributes to predicting higher levels of satisfaction with life as a whole to a greater extent than scores on the Students' Life Satisfaction Scale (SLSS) with samples from Spain, Algeria, South Africa, and Israel pertaining to the second wave of the Children's Worlds project (González-Carrasco et al., 2018). Using a mediation model, Varela et al. (2020) also observed a direct effect of perceptions of safety at home and in the community and an indirect effect through neighborhood satisfaction on Chilean adolescents' life satisfaction.
The direct contribution to SWB made by items related to home and schoolrelated safety is significant and rather homogeneous in all countries and both age groups. Only one item on safety (I feel safe when I walk in the area I live in) is non-significant in one of the countries, Brazil, for both age groups. All other exceptions refer to the items on being cared for, supported, and respected. My teachers care about me does not display significant direct effects for any age group in any of the three countries, suggesting children do not consider being cared for by teachers, and its contribution to SWB, to be irrelevant. If I have a problem, I have a friend who will support me is non-significant for the 12yo in all of the three countries, but significant for the 10yo group in Brazil and Spain, suggesting friend support is more expected and more important for SWB in late childhood than among early adolescents, at least in these two countries. Kim et al. (2019) found that relationships in this context have weightier implications for children's self-assessed SWB as they grow up. Although My teachers listen to me and take what I say into account displays non-significant effects in Chile and Brazil for both age groups, they are significant in Spain. There are people in my family who care about me is non-significant for the 10yo in Spain. My friends are usually nice to me shows non-significant effects for the 10yo in Brazil.
Therefore, some country diversity has been found with regard to the influence that some perceptions about being cared for, supported, and respected exerts on SWB, as well as some differences attributable to age. However, My parents/carers listen to me and take what I say into account displays consistently significant and notable effects on SWB in all countries that increase with age. If I have a problem, people in my family will help me also displays significant increasing-with-age effects in all countries. In a qualitative study conducted in Pakistan, Ahmed et al. (2021) observed that participating children aged 8 to 12 living in different rural and urban settings had constructed their safety and insecurity perceptions in relational and community contexts in which parents, siblings, friends, teachers, and neighbors played an important role, which potentially explains the country differences found in the present study.
With regard to the second aim of this study, when Satisfaction with how safe you feel and Satisfaction with how you are listened to by adults in general are added to the models, the direct contribution of each of these two items to SWB is higher than that of any of the other items examined in all three countries and both age groups. Consequently, we analyzed whether these two items mediate the effects of the other variables on SWB, and tested two hypotheses on how they do so.
Feeling safe at home, at school, and in the area where the child lives, measured with three different items, significantly contribute to overall satisfaction with safety for both age groups, although its contribution varies notably by country, probably reflecting different meanings of safety for Spanish, Chilean, and Brazilian children. The importance of feeling safe at school, home, and in the neighborhood has also been reported previously (Ben-Arieh & Shimoni, 2014;Lee & Yoo, 2015;Steckermeier, 2019), showing that environments perceived as unsafe have important negative consequences for children's SWB (Ben-Arieh et al., 2009) and health (Côté-Lussier et al., 2015). Feeling cared for, supported, and respected contribute to satisfaction with how adults listen and take the child into account, giving support to the second hypothesis, although the contribution of each of the seven studied items differs by country and age group, likely reflecting cultural differences regarding to what extent children participate in decisions concerning their own lives (see, for instance, González et al., 2015;Grigoraş et al., 2018;Lloyd & Emerson, 2017, for more details on the connection between greater children's perceived participation and a higher SWB).
I feel safe when I walk in the area I live in does not display direct effects on SWB in any of the three countries or age groups, but does display significant and noticeable indirect effects through Satisfaction with how safe you feel for both age groups in all three countries. This result is coherent with the third hypothesis presented. Surprisingly, this item also displays significant indirect effects on SWB through Satisfaction with how you are listened to by adults in general for both age groups in Chile and Spain. That is not the only case, since the other two items on safety also display indirect effects on SWB through Satisfaction with how you are listened to by adults in general, suggesting that there may be more opportunities for children to be listened to by adults in safe contexts. I feel safe at home displays significant indirect effects in the three countries for the 10yo and in Chile and Spain for the 12yo, while I feel safe at school shows significant indirect effects in Brazil and Spain only for the 10yo.
For both age groups in all three countries the contribution of having parents who listen to Satisfaction with how you are listened to by adults in general is significant and notable. However, some of the other items do not display significant effects depending on the country and age group.
The effects of having supportive friends or caring teachers on Satisfaction with how you are listened to by adults in general does not reach significance in any of the three countries for the 10yo. Having supportive family members or good friends is not significant in Brazil and Spain, and having caring teachers is not significant in Brazil alone. Other aspects that do not achieve significance for the 12yo include having teachers who listen-in Brazil and Chile, having caring family membersin Brazil and Spain, having caring teachers-only in Chile, and having good 1 3 Safety, Perceptions of Good Treatment and Subjective Well-Being… friends-only in Brazil. These results suggest that feeling cared for, supported, and respected may have different meanings for children in different countries and that the feelings may change with age.
Other unexpected results show that some of the items related to being cared for, supported, and respected by different people have significant indirect effects on SWB through Satisfaction with how safe you feel. Perceiving that My friends are usually nice to me indicates a significant indirect contribution in Chile and Spain for both age groups, while perceiving supportive friends also does so in Spain for the 10yo. The same applies to children perceiving that parents listen to them in Chile and Spain for the 10yo and in all three countries for the 12yo, and perceiving that the family is supportive in Chile for the 10yo and in Chile and Brazil for the 12yo. These results suggest that in some contexts and situations feelings of being cared for, supported, and respected by different people does contribute to an overall feeling of safety. This result has been confirmed in previous studies (see, for instance, Ahmed et al., 2021), since supportive relationships shape children's perceptions of safety regardless of the specific events they are confronted with (Smith & Pollak, 2021). Supportive relationships can even exert a protective role when it comes to situations that clearly compromise children's safety, such as bullying (Miranda et al., 2019).
Therefore, as expected, we can state that Satisfaction with how safe you feel and Satisfaction with how you are listened to by adults in general are inter-related satisfaction domains, displaying moderate correlations (Tab s5 and s6 of the supplementary material). This finding is related to the third aim of this study. Some of the relationships between all the variables we have analyzed with our data differ depending on the country and age group, as described above. Although this was to be expected to some extent, considering the differences between countries discussed in the background section, the existence of common relationships between some variables has also become evident. While children being listened to and children's participation in decisions concerning their daily life is crucial for their agency in general, it is also highly determined by their social position (Ahmed et al., 2021), which in turn is modulated by age and social inequalities. This helps explain both commonalities and divergences among countries and the differences by age observed here, as well as the age sensitivity of SWB indicators reported previously in the literature González-Carrasco et al., 2020;González-Carrasco et al., 2017a, b;Kim et al., 2019).

Conclusions
Our results demonstrate once again that children's SWB is sensitive to age and to the different contexts where they live. Since SWB is a relational phenomenon inscribed in a dynamic of relational exchange where cultural and social practices are unique and particular, according to specific contexts (White, 2017), cross-country comparison, as has been done here, becomes fundamental. Accordingly, perceived contexts in each country are also very important for evaluating satisfaction with personal safety, while having parents who listen and take their child's opinions into account is very important for SWB in all contexts. This observation reinforces the interest on 1 3 converging the research on children's rights (especially the right to participate and being listened to) and children's subjective well-being, as it is done in this article.
Children not feeling safe at home, school or in the surrounding area poses a challenge to both their rights and their SWB since perceptions of security enable taking autonomous decisions and provide a sense of connectedness (Fattore et al., 2019). The latter also contributes to mature agency (Ahmed et al., 2021;Steckermeier, 2019).
The importance of the effects of feeling safe on SWB increases from late childhood to early adolescence, and its indirect effects are much more important than its direct effects. Early adolescents more strongly associate safety to being listened to and taken in to account by adults than 10yo children. Such a result seems to connect to the expectation of not remaining invisible in adults' eyes, as often happens during childhood.
One surprising and worrying result is that most children do not perceive to be listened to by teachers or for teachers to take what they say into account, which is not perceived as very relevant for their SWB in any of the three studied countries. Being heard at school is a necessary condition for children's right to an education in this context (af Ursin & Haanpää, 2018), so this result further proves the need for obtaining adequate knowledge about children's rights from the children themselves.
Future studies could include gender as a variable of interest when analyzing cross-nationally the relationships between satisfaction with safety and how adults listen to children and take what they say into account and children's SWB, since some differences have been detected previously for safety perceptions, generally in favor of boys (see, for instance, Sarriera et al., 2020). They could also be contrasted with objective indicators in line with the work of Uyar-Semerci et al. (2017) who observed that differences in safety perceptions by gender are influenced by the "safety welfare context for children", i.e., objective conditions regarding the provision of health, education, and more at the country level, with girls showing lower SWB in low safety contexts. Further steps could also include exploring the connections among the two studied satisfaction domains and children's agency through satisfaction with the freedom children have. The latter is an indicator that is already included in the Children's Worlds project and considered crucial for children's SWB as well (see, for instance, Steckermeier, 2019). Pending challenges include exploring satisfaction with safety and how adults listen to children and take what they say into account among younger children, given the importance of safety perceptions for children's development (Katz et al., 2017). Triangulating children's, parents', and teachers' perspectives of these two indicators would also be of interest, building on the work of Ben-Arieh et al. (2009).