Child Ind Res

DOI 10.1007/s12187-014-9285-z

Family, School, and Community Correlates

of Children’s Subjective Well-being: An International
Comparative Study

Bong Joo Lee & Min Sang Yoo

Accepted: 26 November 2014

# Springer Science+Business Media Dordrecht 2014

Abstract The primary purposes of this study are twofold: to examine how family,
school, and community factors are related to children’s subjective well-being; and to
examine the patterns of the relationships between family, school, and community
variables and children’s subjective well-being across nations. We use the data from
the pilot study of the International Survey of Children’s Well-Being for our analysis.
We use multiple regression and multilevel methods in the study. We find that family,
school, and community lives all significantly affect the levels of children’s subjective
well-being. We also find that family, school, and community lives of children are
important predictors of subjective well-being even after controlling for the country-
specific cultural and contextual factors. We find that the economic variables of GDP
and inequality are not significant factors predicting children’s subjective well-being.
Rather it is the nature of children’s relationships with immediate surrounding environ-
ments, such as frequency of family activities, frequency of peer activities, and neigh-
borhood safety, are most consistently related to the levels of children’s subjective well-
being across the nations.

Keywords Subjective well-being . Family . School . Community

1 Introduction

There has been a growing interest in studying subjective well-being of children around
the globe in recent years. Most of the early studies of children’s subjective well-being

B. J. Lee (*)
Department of Social Welfare, Seoul National University, Gwanak-gu, Gwanak-ro 1, Seoul 151-746,
South Korea
e-mail: bongjlee@snu.ac.kr

M. S. Yoo
Social Welfare Research Center, Seoul National University, Gwanak-gu, Gwanak-ro 1, Seoul 151-746,
South Korea
e-mail: yms04@snu.ac.kr
 B.J. Lee, M.S. Yoo

primarily focused on measurement issues (Adelman et al. 1989; Huebner 1991, 1994).
The key issue for these studies was whether and how one can measure children’s
subjective well-being. Later studies began measuring and reporting subjective well-
being of children. However, they largely focused on a particular aspect of well-being
(such as health, family, school, etc.) based on small samples (Casas et al. 2007; CRN
1996; Pollard and Lee 2003).
More recent studies began collecting and reporting data on multiple domains of
subjective well-being using large samples at the national level (Ben-Arieh and Shimoni
2014; Casas et al. 2013b; Lee et al. 2013). With national-level data, a few studies
attempted to compare the level of children’s subjective well-being across nations
(Bradshaw et al. 2013; Martorano et al. 2013). These studies compared the national
averages of selected subjective well-being indicators across nations and provided
rankings. While these kinds of study contributed to providing new insights into
understanding children’s well-being across nations, a key concern has been whether
national averages can be compared directly and rankings based on such comparisons
are meaningful. A major criticism has been that ‘national averages’ of subjective well-
being from different national and cultural contexts cannot be compared directly
(Vittersø et al. 2005). In fact, recent studies provided some evidence to suggest that
comparisons of the averages of children’s subjective well-being measures across
nations are not reliable. They found that comparisons of correlations and regression
coefficients can be supported, but not means (Casas 2011; Casas et al. 2013a).
The primary purpose of this paper is to contribute to a better understanding of
variations in children’s subjective well-being by comparing family, school, and com-
munity correlates of children’s quality of life across nations. Our main research
questions are twofold: (1) How are family, school, and community factors related to
children’s subjective well-being?; (2) How do the patterns of relationship between the
family, school, and community factors and children’s subjective well-being vary across
nations? In order to pursue these research questions, we first examine the effects of
family, school, and community factors on subjective well-being pooling data from 11
countries at the individual level using multiple regression methods. Second, we exam-
ine the relationships separately by the countries. Lastly, we use hierarchical linear
model to examine the effects of family, school, and community factors on children’s
subjective well-being at the individual level while controlling for country-level
variables.

2 Background

Studies on the subjective well-being of children and adolescents found reliable indi-
vidual differences (Gilman and Huebner 2003; Park 2004). There are several explana-
tions for the variations in children’s subjective well-being.
Positive psychology has emphasized the joint importance of personality, the social
environment, and circumstances in determining levels of subjective well-being
(Helliwell 2003; Gilman and Huebner 2003; Diener et al. 1999). Personality has been
treated as the most powerful predictor of subjective well-being since people react
differently to the same circumstances depending on their personality traits such as
optimism and extroversion. Studies show that people evaluate the circumstances based
Correlates of Children’s Subjective Well-being: An International Study

on their unique expectations, values, and previous experiences (Helliwell and Putnam
2004; Diener et al. 1999).
Even though personality has been reported to have strong effects on subjective well-
being, social environments such as family, school, and community around children can
also affect their lives. A series of studies highlighted the importance of family, school,
and community influences on children’s well-being. For example, Ash and Huebner
(2001) found that adolescent life satisfaction was linked to a variety of ongoing life
experiences in the family and school environments.
Family is one of the key factors to predict one’s subjective well-being (Joronen and
Astedt-Kurki 2005). Adolescent life satisfaction is influenced by their family charac-
teristics (Henry 1994). Positive family experiences influence children’s subjective well-
being and they are identified as even stronger predictors of life satisfaction than peer
experiences (Gilman and Huebner 2003; Dew and Huebner 1994). These studies
demonstrated that good relationships with a close person buffer the negative impact
to life satisfaction.
On the other hand, there is evidence to suggest that socioeconomic status and family
structure have limited effects on life satisfaction. If basic needs are met, income level or
material conditions lose their explanatory power (R. Cummins 2000; Kahneman et al.
2006). Vandewater and Lansford (1998) used U.S. national sample data and tested
subjective well-being differences between the married-never divorced group and the
divorced-not remarried group. Parental conflict influences children’s well-being; how-
ever, no well-being differences were found between family structure groups.
School environments in which children spend a significant portion of their time
affect children’s subjective well-being. In particular, peer relationships have strong
effects on children’s subjective well-being. Pleasant relationships like talking and
playing with peers have positive effects while unpleasant relationships such as bullying
have negative effects on subjective well-being (Huebner et al. 2004; Nickerson and
Nagle 2004).
Unsafe environments or adverse life events reduce children’s life satisfaction.
Studies focusing on children’s subjective well-being at the community level found that
neighborhood characteristics and safety are the key factors influencing their well-being
(Coulton and Korbin 2007; McDonell 2007).
Some researchers have attempted to examine whether subjective well-being varies
across countries. Cross-cultural studies show that value differences, as well as differ-
ences in environmental factors, explain such differences in life satisfaction across
various countries. There is some evidence that implies that culture modifies how people
feel about their life. Oishi et al. (1999) examined cross-cultural differences in predictors
of life satisfaction among 39 nations. They concluded that need and value system
modify life satisfaction at the country level. In a study with a similar outcome variable
conducted among children, Park and Huebner (2005) pointed out that there are cross-
cultural differences in the perceived levels and correlates of life satisfaction.
A few studies attempted to explain whether social policies at the national level make
differences in subjective well-being among children. The results are mixed. Bradshaw
et al. (2013) compared children’s subjective well-being across European nations, and
they found that there was an association between subjective well-being and all the other
objective domains at the country level such as material, health, education, behavior and
housing environment. Klocke et al. (2013) used multi-level analysis to examine the
 B.J. Lee, M.S. Yoo

effects of country-level differences such as economic growth and youth unemployment

on children’s subjective well-being while controlling for variations at the individual and
school levels. They concluded that the macro (country) variables do not explain how
children’s subjective well-being varies across countries. However, they found that the
micro (family) and mezzo (school) levels explain the variation.

3 Method

3.1 Data

For this study, we used data from the pilot study of the International Survey of
Children’s Well-Being (ISCWeB). ISCWeB is a worldwide research project on chil-
dren’s subjective well-being. Data collection for the pilot study took place between
winter 2011 and winter 2012 in 14 different countries in three age groups (8, 10, and
12 years old). The sample was based on school in all countries. Because the survey was
done as a pilot study, most of the countries employed convenience sampling method.
Only 5 countries, including Chile, South Africa, South Korea, Spain, and England had
a representative sample with random sampling methods. Because most data are col-
lected based on convenience sampling methods, it should be noted that the analysis
presented in this paper are limited in generalizability.
We use the 12 year old dataset for this study.1 Only 11 countries participated for the
12 year old group surveys. The countries include: Algeria (Oran County), Brazil (Rio
Grande do Sul State), Chile, England, Israel, Romania (Cluj County), South Africa
(Western Cape), South Korea, Spain (Catalonia), Uganda (Eastern Uganda), and United
States (South Dakota). In each of the country dataset, we selected those who answered
their age is 12 years old.2
Table 1 shows the sample population included in this study. Data from a total of
12,077 children were used in the study. The sample size ranged from 122 in Algeria to
3,685 in Spain.

3.2 Variables

The key dependent variable of the study is the subjective well-being of children. We
use the General Domain Satisfaction Index (GDSI) developed by Casas et al. (2013b)
as the measure. The composite indices were initially developed for Spanish children
based on various existing life satisfaction instruments. The GDSI is a life satisfaction
scale with eight domains as shown in Table 2. The GDSI includes the Brief
Multidimensional Student Life Satisfaction Scale (BMSLSS with a modified 11-pont
scale) (Seligson et al. 2003) and the Personal Well-being Index-School Children
(Cummins and Lau 2005; Tomyn and Cummins 2011; Casas et al. 2013a). The

1
The 12 year old dataset gave the largest number of countries and the most complete set of variables for the
analysis. Only 10 countries participated in the survey of 8 year olds. The main dependent variable used in the
study is GDSI. The 12 year questionnaire had the most complete set of GDSI scale.
2
Because the classes in which the majority of the children are in the targeted age groups were selected as
sampling units for the survey, the actual age of some children in the selected classes was not 12.
Correlates of Children’s Subjective Well-being: An International Study

Table 1 Sample size by countries

Algeria Brazil Chile England Israel Romania South South Spain Uganda USA Total
Africa Korea

122 506 422 601 723 1,070 1,002 2,602 3,685 1,035 309 12,077

GDSI also adds an additional list of items on satisfaction with different domains in
children’s lives.
Measuring children’s subjective well-being through their own evaluation of life is a
complicated matter. Children’s day-to-day life experiences, for those of school age

Table 2 Domains and items of General Domain Satisfaction Index

Domain Item

GDSI Domain 1: Satisfaction with family and home The house or flat where you live?
The people who live with you?
All the other people in your family?
Your family life?
GDSI Domain 2: Satisfaction with material things How satisfied are you with all the things you have?
GDSI Domain 3: Satisfaction with interpersonal Your friends?
relationships The people who live in your area?
Your relationships with people in general?
GDSI Domain 4: Satisfaction with local area The local police in your area?
The outdoor areas children can use in your area?
The area where you live in general?
GDSI Domain 5: Satisfaction with health Your health?
How you are dealt with when you go to the doctors?
GDSI Domain 6: Satisfaction with time How do you use your time?
management What do you do in your free time?
GDSI Domain 7: Satisfaction with school The school you go to?
Other children in your class?
Your school marks?
Your school experience?
GDSI Domain 8: Personal Satisfaction The freedom you have?
The amount of choice you have in life?
The way that you look?
Yourself?
Your self-confidence?
How safe you feel?
With the things you want to be good at?
What may happen to you later in your life?
How you are listened to by adults in general?
Doing things away from your home?
 B.J. Lee, M.S. Yoo

particularly, come from many different contexts, such as, family, school, and commu-
nity. The dimensions of how children feel about their life are also diverse such as
material things, health, and relationships. Because child subjective well-being is a
composite of life experiences in many contexts and dimensions, researchers have argued
that using a one-item scale such as the Overall Life Satisfaction measure as the only
indicator of children’s subjective well-being has limitations (see, for example, Casas
et al. 2013b). In line with those criticisms, this study uses the GDSI, a multi-domain
composite scale measure of children’s subjective well-being, as the dependent variable.
One important limitation of the GDSI is that it is a relatively new measure and not
fully tested yet. However, increasingly more studies in different cultures have used the
GDSI in recent years (for example, see Casas et al. 2013b in Spain, Schütz 2014 in
Brazil, Grigoras 2013 in Romania, and Lee et al. 2013 in South Korea). The recent
evidence of increasing use of the GDSI in international contexts provides a good reason
of using the GDSI in an international comparative study. It should also be pointed out
that a one-item scale might be more susceptible to cultural biases in measuring
subjective evaluation of life satisfaction. Thus, in order to avoid such problem using
a one-item scale, we decided to use the GDSI. Because three recent studies using the
GDSI in Brazil, Romania, and South Korea, used the same data as this study, using the
GDSI gives a better chance of comparing the results across the countries.
The GDSI scale used as the dependent variable was calculated in the following way.
First, we calculated an average score from 11-point items in each domain. Next, we
calculated the final GDSI score by averaging the averages from the eight domains.
As shown in Table 3, the independent variables include family, school, and community
factors that were shown to be related to children’s subjective well-being in previous studies.
Family life variables include frequency of family activity, home safety, and access to
material resources. Frequency of family activity was assessed by the question “how
often in the past week have you spent time doing the following things with your
family?” There were three activities presented: talking together, having fun together,
and learning together. Four response were offered: “Not at all (0),” “Once or twice (1),”
“Most days (2),” and “Every day (3).” The independent variable is the mean score of
the three items. The values range from 0 to 3. Home safety was measured by
questioning how much the children agree with the statement, “I feel safe at home.”
Five responses were offered: “strongly disagree (0),” “disagree (1),” “neither disagree
nor agree (2),” “agree (3),” “very much agree (4).” The values range from 0 to 4. Access
to material resources was measured by the question, “which of the following things do
or don’t you have?” There were 4 items offered: “clothes in good condition to go to
school in,” “access to computer at home,” “access to internet,” and “mobile phone.” We
added up the 4 items, and the values range from 0 to 4.
School life was measured by frequency of peer activity, experience of bullying, and
school safety. Frequency of peer activity was assessed by the question “how often in the
past week have you spent time doing the following things with your friends apart from
at school?” There were 3 activities presented: talking together, having fun together, and
meeting to study. Four responses were offered: “Not at all (0),” “Once or twice (1),”
“Most days (2),” and “Every day (3).” The resulting mean values range from 0 to 3.
Respondents were also asked about the experience of bullying including being hit and
excluded by others in the school. Each variable was coded as 1 when there was such an
experience and as 0 with no such experience. School safety was measured by
Table 3 The family life, school life and community measures

Domain Variable Item Response Range Note

Family Frequency of family activities1 Talking together Not at all(0)—Everyday(3) Mean of 3 items (range 0~3)
Having fun together
Learning together
Home safety Feel safe at home Strongly disagree(0)—Very much agree(4) 1 item (range 0~4)
Access to material resources Clothes in good condition to go to school in Yes(1), No(0) Sum of 4 items (range 0~4)
Access to computer at home
Access to Internet
Mobile phone
School Frequency of peer activities2 Talking together Not at all(0)—Everyday(3) Mean of 3 items (range 0~3)
Having fun together
Learning together
Correlates of Children’s Subjective Well-being: An International Study

Bullying (hit) Hit by other children in school Yes(1), No(0) 1 item (range 0~1)
Bullying (excluded) Left out by other children in class
School safety Feel safe at school Strongly disagree(0)—Very much agree(4) 1 item (range 0~4)
Community Areas to play in neighborhood There are enough places to play or to have a good time Strongly disagree(0)—Very much agree(4) 1 item (range 0~4)
Neighborhood safety Feel safe when I walk around in the area I live in Strongly disagree(0)—Very much agree(4) 1 item (range 0~4)

1) cronbach’s alpha of frequency of family activities is .722 in pooled dataset

2) cronbach’s alpha of frequency of peer activities is .636 in pooled dataset
 B.J. Lee, M.S. Yoo

questioning how much they agree with the statement, “I feel safe at school.” Five
responses were offered: “strongly disagree (0),” “disagree (1),” “neither disagree nor
agree (2),” “agree (3),” “very much agree (4).” The values range from 0 to 4.
Children’s perceptions about community where they live were measured by the
questions about areas to play and neighborhood safety. Areas to play was measured by
how much they agree with the statement, “In my area there are enough places to play or
to have a good time.” Five responses were offered: “strongly disagree (0),” “disagree
(1),” “neither disagree nor agree (2),” “agree (3),” “very much agree (4).”
Neighborhood safety was measured by questioning how much they agree with the
statement, “I feel safe when I walk around in the area I live in.” Five responses were
offered: “strongly disagree (0),” “disagree (1),” “neither disagree nor agree (2),” “agree
(3),” “very much agree (4).” The values of each variable range from 0 to 4.
The control variables include family structure and gender. The Family Structure
variable was coded into 4 categories: two parent family, mother-only family, father-
only family, and others. In addition, gender was used as a control variable because
gender differences have been reported in various subjective well-being studies. Male
was coded as 1 and female was coded as 0.
In this study, we also employ 4 macro variables at the country level. They
are GDP per capita, public spending on education (% of GDP), child mortality
rate under 5, and the inequality level measured by the GINI index. These
variables are used to measure the economy (GDP and GINI) and the country’s
policy effort for improving education and health of children (public spending
on education and the child mortality rate). Table 4 shows the definitions and
sources of the country-level variables.

3.3 Statistical Method

We use regression and multilevel analysis methods. There are 3 steps for our analyses.
First, we analyze factors affecting children’s subjective well-being using the pooled
data from the 11 countries. In the analysis, we examine the effects of family, school,
and community factors while controlling for the effects of country dummy variables.
Second, we examine the correlates of children’s subjective well-being separately for
each country. Third, we use a hierarchical linear model (HLM) to analyze how country
level variables affect children’s subjective well-being while controlling for the other
factors at the individual level. Because children are nested within countries, the basic
data structure for our analysis is hierarchical. When ignoring the nesting of individuals
into countries, the estimated standard errors will be too small and the risk of Type I
error inflated (Raudenbush and Bryk 2002; Olsen and Dahl 2007). We used HLM
version 7 software for the analysis.

4 Analysis Results

4.1 Descriptive Analysis Results

Table 5 shows the distribution of various measures of subjective well-being included in

the survey by country. The results show that children’s subjective well-being varies
Table 4 Definitions and sources of country-level variables

Variable Definition Source Year

GDP per capita GDP per capita is gross domestic product divided by midyear World Bank 2012 Except Israel(2011)
population. GDP is the sum of gross value added by all
resident producers in the economy plus any product taxes and
minus any subsidies not included in the value of the products.
It is calculated without making deductions for depreciation of
fabricated assets or for depletion and degradation of natural
resources. Data are in current U.S. dollars.
Public spending on Public expenditure education as % of GDP is the total public World Bank Algeria(2008), South Korea, Romania(2009), Israel,
education, total expenditure (current and capital) on education expressed as a Brazil, Spain, South Africa, England, USA(2010),
(% of GDP) percentage of the Gross Domestic Product (GDP) in a given Chile(2011), Uganda(2012)
year. Public expenditure on education includes government
spending on educational institutions (both public and private),
education administration, and transfers/subsidies for private
entities (students/households and other privates entities).
Correlates of Children’s Subjective Well-being: An International Study

Child Mortality under 5 Under-five mortality rate is the probability per 1,000 that a newborn World Bank 2012
baby will die before reaching age five, if subject to current
age-specific mortality rates.
Inequality (GINI Index) GINI index measures the degree of inequality in the distribution CIA (The World Factbook) Algeria(1995)We could not find more recent GINI data
of family income in a country. The index is calculated from the for Algeria., Spain, South Africa(2005), USA(2007),
Lorenz curve, in which cumulative family income is plotted. U.K., Chile, Uganda(2009), Israel(2008), South Korea,
Romania(2011), Brazil(2012)
 B.J. Lee, M.S. Yoo

Table 5 Distribution of subjective well-being indices (mean, standard deviation)

GDSI SLSS BMSLSS PWI SC7 PWI SC9 OLS

Algeria 7.95(1.41) 2.71(.88) 8.29(1.59) 8.06(1.40) 8.04(1.40) 7.97(2.87)

Brazil 8.33(1.21) 2.71(.75) 8.51(1.34) 8.57(1.30) 8.50(1.26) 8.38(2.10)
Chile 8.20(1.29) 2.75(.72) 8.33(1.50) 8.55(1.52) 8.46(1.45) 8.00(2.57)
England 8.24(1.31) 2.49(.36) 8.25(1.60) 8.31(1.42) 8.28(1.41) 7.81(2.20)
Israel 8.59(1.19) 3.16(.72) 8.72(1.34) 8.99(1.31) 8.80(1.33) 8.61(2.09)
Romania 8.87(1.05) 3.02(.70) 9.03(1.16) 9.10(1.07) 9.03(1.07) 8.67(1.91)
South Africa 8.14(1.24) 2.63(.81) 8.37(1.47) 8.17(1.43) 8.19(1.36) 8.19(2.55)
South Korea 7.31(1.44) 2.26(.69) 7.48(1.57) 7.41(1.62) 7.30(1.58) 7.04(2.31)
Spain 8.96(.91) 3.30(.71) 9.07(1.14) 9.34(.94) 9.19(.96) –
Uganda 6.64(1.42) 1.92(.66) 6.62(1.85) 6.76(1.53) 6.89(1.47) 7.24(2.85)
USA 8.53(1.27) 2.91(.77) 8.61(1.48) 8.64(1.41) 8.55(1.42) 7.87(2.20)
Average 8.16(.68) 2.72(.40) 8.30(.70) 8.35(.75) 8.29(.69) 7.98(.53)
Cronbach’s α .893 .793 .720 .832 .859 –
Skewness(S.E) −1.041(.02) −.431(.02) −1.169(.02) −1.152(.02) −1.07(.02) −1.16(.03)

1) GDSI, general domain satisfaction index, 11-point, 29 items (Casas et al. 2013b) SLSS, student life
satisfaction scale, 5-point, 7 items (Huebner 1991); BMSLSS, brief multidimensional student life satisfaction
scale, 11-point, 5 items (Seligson et al. 2003); PWI-SC7, personal well-being index school children, 11-point,
7 items (R Cummins and Lau 2005); PWI-SC9, personal well-being index school children, 11-point, 9 items
(Tomyn and Cummins 2011); OLS, overall life satisfaction (11-point), 1 item
2) Overall Life Satisfaction is not asked in Spain
3) Cronbach’s alphas in each country: Israel (.832), Uganda (.800), Brazil (.855), Spain (.824), South Africa
(.827), South Korea (.903), Romania (.879), Algeria (.815), England (.892), USA (.903), Chile (.822)

across countries. Spain has the highest level of well-being in the five measures. On the
contrary, Uganda and South Korea are at the bottom consistently.
Table 5 also shows Cronbach’s α for each measure. All measures of subjective well-
being show reasonably high level of Cronbach’s α. The GDSI has the highest level of
Cronbach’s α at .893. PWI SC9 and SC7 had similarly high levels of Cronbach’s α. We
also calculated Cronbach’s α for each country’s data in order to check the measure is
reliable within each country sample. The results are presented in note 3 in Table 5. As
the results show, Cronbach’s α for all countries used in the study was above .8
indicating a high reliability.
Since we use GDSI as the key dependent variable, we wanted to make sure there are
reasonably strong correlations between GDSI and the other subjective well-being
measures. Table 6 shows the correlations among the various measures. The results
show that GDSI is strongly correlated with the other measures, ranging from .62 with
OLS to .92 with PWI-SC9.
The results in Table 6 show that the GDSI is more highly correlated with other
similar measures consisting of sums of domain satisfaction scores such as BMSLSS,
PWI-SC7, and PWI-SC9 than it is with the context-free subjective well-being measures
of SLSS and OLS. It is not surprising to see higher correlations with BMSLSS, PWI-
SC7, and PWI-SC9, because the GDSI include the items from the three existing scales
when it was developed. It should be also pointed out that the lowest level of correlation
Correlates of Children’s Subjective Well-being: An International Study

Table 6 Correlation matrix of subjective well-being indices

GDSI SLSS BMSLSS PWI-SC7 PWI-SC9 OLS

GDSI 1
SLSS .700* 1
BMSLSS .894* .638* 1
*
PWI-SC7 .894 .678* .782* 1
* *
PWI-SC9 .921 .681 .818* .974* 1
OLS .621* .571* .577* .606* .621* 1

*Correlation is significant at the 0.01 level (2-tailed)

with OLS suggests the GDSI actually captures some concept of subjective well-being
different from a one-item life satisfaction measure.
Table 7 and Table 8 present the descriptive statistics of the independent variables
used in the study, at individual-level and country-level, respectively. At both levels, one
can see that there are substantial variations across countries. It should be noted that
school bullying experiences were not asked in Spain. In South Africa, the family
structure question was not asked.
Table 9 shows a correlation matrix between the study’s independent variables and
the subjective well-being measures included in the survey. The results show that the
correlations between the independent variables and the OLS, SLSS, and GDSI are in
similar patterns. One can also see that the correlations between the family, school, and
community factors and children’s subjective well-being are stronger than the ones with
the national-level variables.

4.2 Regression Analysis Results

Table 10 presents the results from multiple regression analyses using the pooled sample
data. Models 1 to 5 include all 11 countries in the analyses. Those models exclude
family structure and bullying variables because South Africa data did not have the
family structure variable and Spain did not have the bullying variable. Model 6 with all
available variables includes only 9 countries in the analysis, excluding Spain and South
Africa.
In general, boys tend to have higher levels of subjective well-being than girls.
Children living with both parents, as shown in Model 6, also tend to have higher level
of subjective well-being than those in other types of family structure.
Model 2 adds family life variables to the model. Family factors explain around 40 %
of variation in children’s subjective well-being. The more family activities a child has,
the higher subjective well-being a child reports. The number of things a child has and
home safety are also positively related to children’s subjective well-being.
In Model 3, we find that school life variables explain an additional 9 % of the
variations of children’s subjective well-being. School safety and frequency of peer
activities are all important factors influencing children’s subjective well-being. In
Model 6, we also find that being involved in bullying as being excluded by others
and hit by others decrease subjective well-being.
Table 7 Descriptive statistics of individual level variables by country

Variables Israel Uganda Brazil Spain South Africa South Korea Romania Algeria England USA Chile Total

Family Frequency of family activities 2.08 2.06 1.89 2.34 2.10 1.53 1.98 2.29 1.98 1.95 2.09 2.02
Access to material resources 3.39 0.88 3.62 3.70 2.74 3.81 3.73 2.21 3.87 3.61 3.44 3.37
Home safety 3.88 2.98 3.30 3.79 3.51 3.39 3.68 3.55 3.59 3.56 3.62 3.56
School Frequency of peer activities 2.00 1.59 1.54 2.07 1.87 1.61 1.75 1.92 1.99 1.68 1.61 1.83
School safety 3.10 3.20 3.20 3.45 3.19 2.70 3.05 3.27 3.04 3.21 3.32 3.15
Bullying(hit) 0.38 0.24 0.18 – 0.37 0.13 0.35 0.26 0.34 0.25 0.18 0.25
Bullying(excluded) 0.21 0.11 0.30 – 0.47 0.05 0.41 0.36 0.40 0.45 0.31 0.25
Community Areas to play in neighborhood 2.93 2.17 2.37 3.03 2.77 2.15 2.80 2.50 2.61 2.86 2.13 2.63
Neighborhood safety 3.20 2.25 2.41 3.11 2.21 2.10 3.04 2.78 2.73 2.97 2.83 2.68
Demographics Gender (male) 0.50 0.51 0.46 0.48 0.46 0.42 0.50 0.49 0.43 0.43 0.58 0.47
Both parents (yes) 0.86 0.77 0.57 0.80 – 0.89 0.86 0.89 0.76 0.59 0.70 0.74
Mother only (yes) 0.10 0.14 0.33 0.16 – 0.07 0.11 0.06 0.21 0.31 0.24 0.13
Father only (yes) 0.00 0.04 0.04 0.01 – 0.03 0.01 0.02 0.02 0.07 0.02 0.02
Others (yes) 0.01 0.05 0.05 0.00 – 0.01 0.02 0.01 0.01 0.02 0.04 0.01
n 723 1,035 506 3,685 1,002 2,602 1,070 122 601 309 422 12,077
B.J. Lee, M.S. Yoo
Table 8 Descriptive statistics of country-level variables by country

Variables Israel Uganda Brazil Spain South Africa South Korea Romania Algeria England USA Chile

GDP per capita 33250 547 11340 28624 7508 22590 9036 5348 39093 51749 15452
Public spending on education, total (% of GDP) 5.6 3.3 5.8 5 6 5 4.2 4.3 6.3 5.6 4.1
Child mortality under 5 4 69 14 5 45 4 12 20 5 7 9
Inequality (GINI Index) 39.2 44.3 51.9 32 63.1 31.1 33.2 35.3 40 45 52.1
Correlates of Children’s Subjective Well-being: An International Study
Table 9 Correlation matrix of independent variables and various subjective well-being scales

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

1. Frequency of family activities 1

2. Access to material resources .010 1
3. Home safety .285** .222** 1
4. School safety .296** .007 .238** 1
5. requency of peer activities .346** .122** .173** .201** 1
** ** ** **
6. Areas to play in neighborhood .269 .117 .210 .238 .238** 1
** ** ** **
7. Neighborhood safety .279 .126 .276 .293 .229** .421** 1
** ** ** **
8. Bullying(hit) .003 −.060 −.025 −.103 −.025 −.003 −.023* 1
** ** * **
9 Bullying(excluded) .054 −.003 .005 −.037 −.020 .036 .016 .262** 1
** ** ** ** ** ** **
10. GDP per capita .047 .493 .207 .030 .150 .125 .169 −.015 .009 1
11. Public spending on education −.012 .412** .135** −.013 .104** .093** .011 .055** .147** .539** 1
** ** ** ** ** ** ** ** *
12. Child mortality under 5 .039 −.719 −.232 .029 −.092 −.082 −.150 .059 .020 −.744** −.416** 1
13. Inequality (GINI Index) .040** −.352** −.104** .040** −.057** −.032** −.107** .098** .197** −.436** .250** .620** 1
14. OLS .361** .104** .261** .310** .229** .260** .285** −.054** −.024* −.017 .062** −.016 .102** 1
** ** ** ** ** ** ** ** * ** ** ** **
15. SLSS .440 .299 .386 .353 .290 .347 .404 −.039 .021 .265 .145 −.292 −.133 .571** 1
** ** ** ** ** ** ** ** ** ** ** **
16. GDSI .482 .322 .436 .440 .359 .443 .483 −.050 .015 .245 .189 −.285 −.083 .621** .700** 1

**Correlation is significant at the 0.01 level (2-tailed)

B.J. Lee, M.S. Yoo
Table 10 Multiple regression model

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6

B BETA S.E. B BETA S.E. B BETA S.E. B BETA S.E. B BETA S.E. B BETA S.E.

Demo- graphics Gender (male) .056* .020 .026 .076*** .026 .020 .134*** .047 .019 .055** .019 .018 .040* .014 .017 .082*** .027 .023
Family-Both parents (reference)
Family - Mother only −.182*** −.041 .034
Family - Father only −.206** −.022 .070
Family - Others −.179* −.017 .079
Family Frequency of family activities .785*** .404 .014 .564*** .290 .014 .466*** .240 .014 .421*** .217 .013 .400*** .201 .017
Access to material resources .342*** .261 .010 .336*** .256 .009 .300*** .228 .008 .147*** .112 .012 .138*** .112 .016
Home safety .517*** .263 .015 .406*** .207 .014 .321*** .164 .013 .270*** .137 .013 .276*** .147 .016
School Frequency of peer activities .287*** .138 .015 .199*** .096 .014 .190*** .091 .013 .152*** .073 .017
*** *** ***
School safety .400 .277 .010 .313 .216 .010 .308 .213 .009 .329*** .224 .012
Victim of bullying(hit) −.111*** −.030 .030
Victim of bullying(excluded) −.326*** −.085 .032
*** ***
Correlates of Children’s Subjective Well-being: An International Study

Community Areas to play in neighborhood .186 .163 .008 .164 .144 .008 .190*** .158 .010
Neighborhood safety .221*** .186 .008 .184*** .155 .008 .186*** .147 .010
Country South Korea (reference)
Israel .449*** .074 .040 .511*** .100 .045
***
Uganda −.542 −.106 .049 −.519*** −.119 .061
Brazil .679*** .095 .045 .806*** .134 .050
Spain .566*** .182 .026 – – –
Romania .884*** .176 .034 1.013*** .235 .039
*
Algeria .092 .006 .087 .189 .016 .095
England .309*** .047 .042 .472*** .085 .047
Table 10 (continued)

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6

B BETA S.E. B BETA S.E. B BETA S.E. B BETA S.E. B BETA S.E. B BETA S.E.

USA .580*** .064 .055 .754*** .100 .062

***
Chile .314 .040 .049 .424*** .065 .054
South Africa .386*** .074 .037 − − −
Constant 8.172*** 3.585*** 2.634*** 2.652*** 3.301*** 3.323***
R square .000 .394 .485 .553 .600 .592
Adjusted R square .000 .394 .485 .552 .599 .591
Changed R square .009*** .373*** .093*** .075*** –

1)+p<0.1, * p<0.05, ** p<0.01, *** p<0.001

2) Model 6 includes only 9 countries, excluding South Africa and Spain
3) Regression models conducted using SPSS 18
B.J. Lee, M.S. Yoo
Correlates of Children’s Subjective Well-being: An International Study

Community factors are added to the model in Model 4. Community factors explain
yet another 7 % of the variation in children’s subjective well-being. The results show
that the more children can play and feel safe to walk around in their neighborhood, the
higher children’s subjective well-being is.
Model 5 presents country fixed effects model where country dummies are intro-
duced to the analysis. We find significant country level effects (except for Algeria)
while controlling for the other variables. These results indicate that there is much
country-specific variation in children’s subjective well-being over family, school, and
community factors. However, we also find that including the country dummies do not
change the effects of the other variables significantly. In turn, these findings indicate
that the family, school, and community factors are still important predictors of chil-
dren’s subjective well-being even after controlling for the country-specific culture and
context factors. We also find that the country specific context factors explain only
limited portion of the total variation of subjective well-being. The individual factors of
demographics, family life, school life, and community explain about 55 % of the
variation, whereas the country dummies explain only about 5 % of the variance.
Table 11 shows the results from regression analyses run separately for each country.
Again, in general, we find that family, school, and community lives of children are all
significant predictors of subjective well-being across countries. Among the variables,
frequency of family activities, frequency of peer activities, and neighborhood safety
have the most consistent effects on children’s subjective well-being across the
countries.
While the number of things a child has is positively related to the level of subjective
well-being in most countries, it is not in two South American countries included the
study, in Brazil and Chile. Uganda is the only country where home safety is not a factor
influencing subjective well-being. Algeria is the only country where areas to play in the
neighborhood has no effect. Being a victim of being excluded from other children at
school has more consistent negative effect on children’s subjective well-being across
countries than being a victim of being hit by other children.
Comparing how the independent variables are related to the subjective well-being
across countries show some interesting results. Boys reported higher level of subjective
well-being controlling for the other factors in Brazil and South Korea than girls.
However, Israeli boys reported lower subjective well-being compared to their counter-
part girls. There was no difference between the genders in the other countries. Children
living in a mother-only family show lower level of subjective well-being than those
living with two parents in Israel, Uganda, Spain, South Korea, England, USA, and
Chile. However, living with mother only was not a factor lowering children’s subjective
well-being in Brazil, Romania, and Algeria.

4.3 Hierarchical Linear Model Analysis Results

The aim of this section is to examine how both individual-level and country-level
factors influence children’s subjective well-being.3 The main research question can be
formulated as two-level hierarchical linear models. At level-1, the units are children and

3
Although the ISCWeB data provide school identification information of the child, 3-level model was not
employed because information on school characteristics was not collected in the surveys.
Table 11 Multiple regression Model by countries (standardized coefficient with standard error)

Israel Uganda Brazil Spain South Africa South Korea

BETA S.E. BETA S.E. BETA S.E. BETA S.E. BETA S.E. BETA S.E.

Demographics Gender (male) −.121*** .066 −.037 .071 .096** .074 −.015 .024 −.006 .060 .097*** .041
Family - Both parents (reference)
Family - Mother only −.047 + .109 −.054* .103 −.012 .080 −.034** .032 – −.046** .076
Family - Father only −.022 .496 −.076** .185 −.068* .188 .002 .103 – .002 .117
Family - Others −.025 .288 .002 .171 −.087** .169 .005 .167 – −.003 .170
Family Frequency of family activities .168*** .052 .207*** .055 .357*** .053 .259*** .022 .306*** .045 .201*** .029
Access to material resources .104*** .030 .085** .064 .036 .049 .070*** .020 .141*** .028 .055*** .037
Home safety .169*** .090 .035 .033 .065* .051 .126*** .027 .050* .039 .248*** .029
School Frequency of peer activities .139*** .053 .115*** .058 .133*** .052 .132*** .026 .148*** .041 .078*** .027
School safety .266*** .030 .186*** .036 .200*** .041 .214*** .016 .127*** .028 .302*** .022
Victim of bullying(hit) −.020 .071 .024 .084 −.027 .098 – −.055* .063 −.023 .064
Victim of bullying(excluded) −.097*** .086 −.057* .120 −.176*** .085 – –.117*** .063 −.051*** .101
Community Areas to play in neighborhood .198*** .029 .234*** .027 .117*** .032 .119*** .011 .177*** .025 .173*** .021
Neighborhood safety .174*** .034 .102*** .029 .210*** .034 .195*** .012 .175*** .022 .136*** .023
Constant 3.223*** 3.784*** 4.881*** 4.456*** 4.781*** 2.061*** 4.858*** 3.107*** 3.164*** 3.809*** 4.541***
R square .495 .374 .561 .406 .463 .536 .424 .545 .610 .618 .419
Adjusted R square .486 .366 .549 .405 .457 .534 .417 .490 .601 .601 .400

Romania Algeria England USA Chile

BETA S.E. BETA S.E. BETA S.E. S.E. BETA S.E.

Demographics .011 .051 −.104 .200 −.004 .074 .097 .059 .102
Family - Both parents (reference)
−.020 .079 −.021 .407 −.044 + .084 .106 −.088* .119
B.J. Lee, M.S. Yoo
Table 11 (continued)

Romania Algeria England USA Chile

BETA S.E. BETA S.E. BETA S.E. S.E. BETA S.E.

.017 .023 .078 .618 .035 .241 .181 −.087* .328

−.026 .197 .064 1.094 −.027 .376 .369 −.081* .264
Family .193*** .040 .229** .158 .256*** .053 .068 .227*** .076
.169*** .039 .237** .077 .073** .091 .070 −.005 .044
.156*** .043 .263** .133 .195*** .050 .073 .168*** .068
School .047 + .037 .135+ .144 .066* .070 .068 .044 .069
.235*** .026 .162* .100 .191*** .041 .057 .183*** .057
−.093*** .056 −.054 .239 −.085** .081 .118 −.074+ .134
−.108*** .054 −.030 .228 −.182*** .077 .107 −.080* .112
Community .114*** .022 .071 .066 .191*** .034 .050 .108** .036
.157*** .028 .162* .080 .164*** .038 .056 .265*** .041
Constant
R square
Adjusted R square
Correlates of Children’s Subjective Well-being: An International Study

1)+p<0.1, * p <0.05, ** p<0.01, *** p<0.001

 B.J. Lee, M.S. Yoo

each child’s outcome is represented as a function of a set of individual characteristics.

At level-2, the units are countries (Raudenbush and Bryk 2002). This model explicitly
specifies that the children are nested within each country. If we do not take into account
the fact that the children are nested within the countries in our analysis using the
ordinary liner regression, the estimates of the standard errors of the coefficients will be
biased because the assumption of independence of the Level-1 units cannot be met
(Garson 2013; Olsen and Dahl 2007; Raudenbush and Bryk 2002).
We employ a random-intercept model in which only the intercept parameter in the
level-1 model is assumed to vary at level 2. This model allows us to investigate how
characteristics of country are related to the level of children’s subjective well-being
even after controlling for family, school, and community variables in the same model.
This model also allows us to examine the patterns of the relationships between family,
school, and community variables and children’s subjective well-being after controlling
for the country-specific cultural and contextual factors.
Table 12 shows the results from hierarchical linear models fitted to the data. Models
1 to 3 include all 11 countries in the analysis excluding school bullying and family
structure variables because these variables were not surveyed in Spain and South
Africa.4
Model 1 is the unconditional model, where only the mean differences of children’s
subjective well-being are modeled. The unconditional Model is the simplest possible
hierarchical linear model which is equivalent to a one-way ANOVA with random
effects (Raudenbush and Bryk 2002). The intra correlation coefficient (ICC=.241)
represents the proportion of the variance in outcome between countries. The results
also show that 24.1 % of the variance is over countries, and the remaining 75.9 % at
individual level.
In Model 2, we include individual-level variables in the analysis. Similar to the
results from multiple regressions, all individual-level variables are significantly related
to children’s subjective well-being in expected directions. Individual-level predictors
explained 43 % of the within-country variance in children’s subjective well-being.
In Model 3, the country-level variables are included in the analysis to examine
which country level characteristics explain variation in children’s subjective well-being.
We find that the economic variables of GDP and GINI are not significant predictors of
children’s subjective well-being. Public spending on education is also not significant.
Rather we find a significant effect of the rate of child mortality under 5 predicting
children’s subjective well-being (p<.1). The child mortality rate is negatively related to
subjective well-being. The results indicate that the overall health status of children at
the country level has something to do with children’s subjective well-being. Country-

4
It should be noted that our effective sample size for multilevel analysis is relatively small. A small number of
groups could lead to biased outcomes in the multilevel analysis (Maas and Hox 2005). For example, small
sample size could lead to a situation where standard error are biased downward and researchers overstate the
level of their significance tests. However, the literature on multilevel analysis gives varying rules of thumb in
terms of the minimum number of groups, ranging from 10 to 30 or even more (Stegmueller 2013). There
are also several studies that used a small number of groups (Normand and Zou 2002; Steenbergen
and Jones 2002). Furthermore, A Baysian approach, which we employed in this analysis, could be
robust and yields considerably conservative tests when using a small number of groups in
multilevel analysis (Stegmueller 2013).
Correlates of Children’s Subjective Well-being: An International Study

Table 12 Hierarchical linear model results

Unconditional Model Conditional Model

Model 1: Null Model 2: Micro Model 3: Micro and

Model level predictors Macro level predictors

Fixed Effect B SE B SE B SE
Intercept 8.160*** .195 8.184*** .120 8.203*** .102
Individual-level variables
Gender (male) .040 .065 .004* .017
* ***
Frequency of family activities .421 .024 .421*** .013
Access to material resources * .150*** .020 .048*** .011
Home safety* .270*** .085 .270*** .013
Frequency of peer activities* .190*** .025 .190*** .013
School safety* .307*** .045 .308*** .009
Areas to play in neighborhood* .164*** .026 .164*** .008
Neighborhood safety* .184*** .010 .184*** .008
Macro-level variables
GDP per capita* −.000 .000
Public spending on education, .083 .172
total (% of GDP) *
Child mortality under 5* −.015+ .008
*
Inequality (GINI Index) .011 .015
Variance Components
Within country (Sigma2) 1.442 (75.9 %) .822 (84.8 %) .821 (87.9 %)
Between country (Tau) .459 (24.1 %) .147 (15.2 %) .113 (12.1 %)
Total 1.901 (100.0 %) .969 (100.0 %) .934 (100.0 %)
Proportion variance explained
Within country 43.00 43.31
Between country 23.13
N children 12,056 12,056
N countries 11 11

1)+p<0.1, * p<0.05, ** p<0.01, *** p<0.001

2) Between country variance in Model1 is intra-class correlation coefficient (ICC = .241).
ICC ¼ ττ 0000 ¼ :459þ1:442
:459
¼ :241
3) Hierarchical linear model conducted using HLM 7
4) HLM software programs use an empirical Bayes estimation strategy (Raudenbush and Bryk 2002)
5) Spain and South Africa are excluded due to missing values in model 4 (bullying was not asked in Spain;
family structure was not asked in South Africa)
6) Proportion variance explained at level 1= bτ 00 ðunconditionalÞ− bτ 00 ðconditional modelÞ (Raudenbush and Bryk 2002),In
bτ 00 ðunconditionalÞ
Model 2, proportion variance explained at level-1 is calculated by = 1:442−:822 1:442 ¼ :430
2 2
7) Proportion variance explained at level-2= σ ðunconditional Þ− σ ðconditional model Þ
σ2 ðunconditional Þ (Raudenbush and Bryk 2002)
Tau already changed by adding individual variables in Model 2, we use tau from model 2. Hence, in Model3,
proportion variance explained at level-1 is calculated by = :147−:113
:147 ¼ :231
8) * Grand mean-centered
 B.J. Lee, M.S. Yoo

level predictors explained 23 % of the between-country variance in mean levels of

children’s subjective well-being.
The HLM results also show that within country variance is much bigger than
between country variance. In the conditional model, country-level variables explain
about 23 % of the variance, while individual-level variables account for about 43 % of
the variance. Similar to multiple regression results, the HLM analysis also show that
individual-level variables explain almost half of the variance in children’s subjective
well-being.

5 Discussion and Conclusion

The primary purposes of this study are twofold: to examine how family, school, and
community factors are related to children’s subjective well-being; and to examine the
patterns of the relationships between family, school, and community variables and
children’s subjective well-being across nations.
We find that family, school, and community lives all significantly affect the levels of
children’s subjective well-being. We also find that there is significant country-specific
variation in children’s subjective well-being over family, school, and community
influences. However, we find that family, school, and community lives of children
are important predictors of subjective well-being even after controlling for the country-
specific cultural and contextual factors.
When we examine the effects of family, school, and community factors across
nations, we find that frequency of family activities, frequency of peer activities, and
neighborhood safety are most consistently related to the levels of children’s subjective
well-being across the countries. HLM analysis also reveals that a good portion of
within country variation in children’s subjective well-being is explained by family,
school, and community factors considered in the study. We find that the economy
variables of GDP and inequality at the country level are not significant factors
predicting children’s subjective well-being. We also find that the public spending on
education is not related to children’s subjective well-being. Rather we find some
evidence to suggest the overall health status of children, represented as the rate of
child mortality under 5, is related to children’s subjective well-being.
The significance of this study is to further our understanding of correlates of
children’s subjective well-being. Most previous comparative studies of children’s
subjective well-being were performed using country-level aggregate data except for
very few studies (for example, see Klocke et al. 2013). This study uses individual-level
data, taking advantage of the nested nature of the data across nations. This allows one
to better differentiate within country and between country variances. It should also be
noted that this study uses the data from a wide range of the different cultures across the
world including countries from Europe, Asia, North America, South America, and
Africa. Most previous studies only considered comparisons in a particular geographic
region, most notably in Europe. This study adds diversity to the existing literature.
Another advantage of this study is the use of multi-dimensional index measure
of children’s subjective well-being. GDSI is a relatively new measure, which
has begun to be validated in recent studies (Casas 2011; Casas et al. 2013b for
the studies using the measure).
Correlates of Children’s Subjective Well-being: An International Study

While these are the strengths of the current study, we also need to point out some
important limitations of it. First of all, the data used in the analysis are from a pilot
study where most of the country data were collected using convenience sampling
method. This limits our ability to generalize the findings presented in the study.
Secondly, it should be pointed out that the analysis includes a very limited number of
countries. This might have implications for finding only a limited amount of between
country variance in the analysis. More research is needed to better understand the
between country variation in children’s subjective well-being with representative data
from more countries. Thirdly, the country-level variables used in the study are still very
rudimentary. They might not capture the cultural and contextual dimensions of nations
that affect the levels of children’s subjective well-being. More effort is needed to
develop better macro measures to enhance our understanding of between country
variations in children’s subjective well-being. Fourthly, this study uses the GDSI as
the key dependent variable, which is a relatively new and not fully tested measure. Any
study trying to use a relative new measure as a key variable could run into the reliability
problem of the new measure. 5 However, the recent evidence of increasing use of the
GDSI in various international contexts provides at least a reasonable rationale for using
the GDSI in an international comparative study like this study. Lastly, cross-sectional
data provide us with a snapshot of the population at a single point in time but do not
reflect changes over time, so a longitudinal study is needed for a better understanding
of children’s well-being.
Nonetheless, the findings of this study have a number of important implications.
Findings of consistent effects of family, school, and community factors across nations
confirm again that the quality of children’s relationships with their immediate environ-
ments matter. Social relationships with family, peers, and community regardless of
cultural and contextual differences across nations has a strong influence on the subjec-
tive well-being of children. These findings point out that social capital matters for
children’s subjective well-being (Helliwell and Putnam 2004).
Consistent with recent findings of Klocke et al. (2013), this study also finds that
macro factors such as economy (such as GDP and inequality) and spending on
education are not significant factors affecting children’s subjective well-being. Rather
the relationships children have with family, school, and community matter more for
their subjective well-being. This finding suggests that children’s subjective well-being
is affected by more immediate surroundings of children’s lives than macro societal
factors.
The significance of this study is to fill the gap in knowledge on correlates of
children’s subjective well-being around the globe, which is a relatively new research
topic. The strength of this article is to try to do so using individual-level analysis with
international comparative data taking into account of the nested nature of the data.

5
In the analysis, we also examined context-free measures such as OLS (overall life satisfaction) and SLSS
(student life satisfaction scale). In the multiple regression model with pooled dataset, SLSS had similar results
with GDSI but the r-square was decreased from .591 to .495. In the same analysis using OLS, which was not
asked in Spain, it also has similar outcome with GDSI, but explanatory power decreased by almost half to
.266. Despite explanatory power differences from scale to scale, we conclude that family, school, and
community variables affect children’s subjective well-being consistently using different measures of subjective
well-being. The results with either SLSS or OLS are available from the authors upon request.
 B.J. Lee, M.S. Yoo

Finally, we should caution, as always, against drawing any direct causal inferences
from our analysis. Our study is based on observational data. Any analysis, including
ours, using observational data risks model misspecification problems.

References

Adelman, H. S., Taylor, L., & Nelson, P. (1989). Minors’ dissatisfaction with their life circumstances. Child
Psychiatry and Human Development, 20(2), 135–147.
Ash, C., & Huebner, E. S. (2001). Environmental events and life satisfaction reports of adolescents a test of
cognitive mediation. School Psychology International, 22(3), 320–336.
Ben-Arieh, A., & Shimoni, E. (2014). Subjective Well-being and perceptions of safety among Jewish and
Arab children in Israel. Children and Youth Services Review, 44, 100–107.
Bradshaw, J., Martorano, B., Natali, L., & de Neubourg, C. (2013). Children’s subjective well-being in rich
countries. Child Indicators Research, 6(4), 619–635.
Casas, F. (2011). Subjective social indicators and child and adolescent well-being. Child Indicators Research,
4(4), 555–575.
Casas, F., Figuer, C., González, M., Malo, S., Alsinet, C., & Subarroca, S. (2007). The well-being of 12-to 16-
year-old adolescents and their parents: Results from 1999 to 2003 Spanish samples. Social Indicators
Research, 83(1), 87–115.
Casas, F., Bălţătescu, S., Bertran, I., González, M., & Hatos, A. (2013a). School satisfaction among
adolescents: Testing different indicators for its measurement and its relationship with overall life
satisfaction and subjective well-being in Romania and Spain. Social Indicators Research, 111(3), 665–
681.
Casas, F., Bello, A., González, M., & Aligué, M. (2013b). Children’s subjective well-being measured using a
composite index: What impacts spanish first-year secondary education students’ subjective well-being.
Child Indicators Research, 6(3), 433–460.
Coulton, C. J., & Korbin, J. E. (2007). Indicators of child well-being through a neighborhood lens. Social
Indicators Research, 84(3), 349–361.
CRN (1996). International Comparative Survey: A child's view of what a family should be.
Cummins, R. (2000). Personal income and subjective well-being: A review. Journal of Happiness Studies,
1(2), 133–158.
Cummins, R., & Lau, A. (2005). Personal wellbeing index–school children. Victoria: Deakin University.
Dew, T., & Huebner, E. S. (1994). Adolescents’ perceived quality of life: An exploratory investigation.
Journal of School Psychology, 32(2), 185–199.
Diener, E., Suh, E. M., Lucas, R. E., & Smith, H. L. (1999). Subjective well-being: Three decades of progress.
Psychological Bulletin, 125(2), 276.
Garson, G. D. (2013). Fundamentals of hierarchical linear and multilevel modeling. Hierarchical Linear
Modeling: Guide and Applications Sage Publications Inc, 3–25.
Gilman, R., & Huebner, S. (2003). A review of life satisfaction research with children and adolescents. School
Psychology Quarterly, 18(2), 192.
Grigoras, B. A. (2013). The subjective well-being of children. Doctorial dissertation: Babes Bolyai University,
Cluj Napoca, Romania.
Helliwell, J. F. (2003). How's life? Combining individual and national variables to explain subjective well-
being. Economic Modelling, 20(2), 331–360.
Helliwell, J. F., & Putnam, R. D. (2004). The social context of well-being. Philosophical Transactions-royal
Society of London Series B Biological Sciences, 1435–1446.
Henry, C. S. (1994). Family system characteristics, parental behaviors, and adolescent family life satisfaction.
Family Relations, 447–455.
Huebner, E. S. (1991). Initial development of the student's life satisfaction scale. School Psychology
International, 12(3), 231–240.
Huebner, E. S. (1994). Preliminary development and validation of a multidimensional life satisfaction scale for
children. Psychological Assessment, 6(2), 149.
Huebner, E. S., Suldo, S. M., Smith, L. C., & McKnight, C. G. (2004). Life satisfaction in children and youth:
Empirical foundations and implications for school psychologists. Psychology in the Schools, 41(1), 81–
93.
Correlates of Children’s Subjective Well-being: An International Study

Joronen, K., & Astedt-Kurki, P. (2005). Familial contribution to adolescent subjective well-being.
International Journal of Nursing Practice, 11(3), 125–133.
Kahneman, D., Krueger, A. B., Schkade, D., Schwarz, N., & Stone, A. A. (2006). Would you be happier if you
were richer? A focusing illusion. Science, 312(5782), 1908–1910.
Klocke A., Clair A., & Bradshaw J. (2013). International variation in child subjective well-being. Child
Indicators Research, 1-20.
Lee B. J., Kim S. S., Ahn J. J., Yoo J., Yoo M. S., Choi C. Y., et al. (2013). What does composite well-being
index of children tell us about Korean children’s quality of life? : Save the Children Korea
Maas, C. J., & Hox, J. J. (2005). Sufficient sample sizes for multilevel modeling. Methodology: European
Journal of Research Methods for the Behavioral and Social Sciences, 1(3), 86.
Martorano B., de Neubourg C., Natali L., Bradshaw J., & UNICEF (2013). Child Well-being in Economically
Rich Countries: Changes in the first decade of the 21st century: UNICEF Office of Research.
McDonell, J. R. (2007). Neighborhood characteristics, parenting, and children’s safety. Social Indicators
Research, 83(1), 177–199.
Nickerson A. B., & Nagle R. J. (2004). The influence of parent and peer attachments on life satisfaction in
middle childhood and early adolescence. In Quality-of-Life Research on Children and Adolescents (pp.
35-60): Springer.
Normand, S. L. T., & Zou, K. H. (2002). Sample size considerations in observational health care quality
studies. Statistics in Medicine, 21(3), 331–345.
Oishi, S., Diener, E. F., Lucas, R. E., & Suh, E. M. (1999). Cross-cultural variations in predictors of life
satisfaction: Perspectives from needs and values. Personality and Social Psychology Bulletin, 25(8), 980–
990.
Olsen, K. M., & Dahl, S.-Å. (2007). Health differences between European countries. Social Science &
Medicine, 64(8), 1665–1678.
Park, N. (2004). The role of subjective well-being in positive youth development. The Annals of the American
Academy of Political and Social Science, 591(1), 25–39.
Park, N., & Huebner, E. S. (2005). A cross-cultural study of the levels and correlates of life satisfaction among
adolescents. Journal of Cross-Cultural Psychology, 36(4), 444–456.
Pollard, E. L., & Lee, P. D. (2003). Child well-being: a systematic review of the literature. Social Indicators
Research, 61(1), 59–78.
Raudenbush S. W., & Bryk A. S. (2002). Hierarchical linear models: Applications and data analysis methods
(Vol. 1): Sage.
Schütz, F. F. (2014). Bem-estar em crianças de diferentes configurações familiares e em acolhimento
institucional. Brazil: Master’s these, Universidade Federal do Rio Grande do Sul.
Seligson, J. L., Huebner, E. S., & Valois, R. F. (2003). Preliminary validation of the brief multidimensional
students’ life satisfaction scale (BMSLSS). Social Indicators Research, 61(2), 121–145.
Steenbergen, M. R., & Jones, B. S. (2002). Modeling multilevel data structures. American Journal of Political
Science, 218–237.
Stegmueller, D. (2013). How many countries for multilevel modeling? A comparison of frequentist and
Bayesian approaches. American Journal of Political Science, 57(3), 748–761.
Tomyn, A. J., & Cummins, R. A. (2011). The subjective wellbeing of high-school students: Validating the
personal wellbeing index—school children. Social Indicators Research, 101(3), 405–418.
Vandewater, E. A., & Lansford, J. E. (1998). Influences of family structure and parental conflict on children’s
well-being. Family Relations, 323–330.
Vittersø, J., Biswas-Diener, R., & Diener, E. (2005). The divergent meanings of life satisfaction: Item response
modeling of the satisfaction with life scale in Greenland and Norway. Social Indicators Research, 74(2),
327–348.