Content uploaded by Anne Konu
Author content
All content in this area was uploaded by Anne Konu on Aug 17, 2021
Content may be subject to copyright.
School Effectiveness and School Improvement 0924-3453/02/1302-187$16.00
2002, Vol. 13, No. 2, pp. 187±200 #Swets & Zeitlinger
Evaluation of Well-Being in Schools ± A Multilevel
Analysis of General Subjective Well-Being
Anne I. Konu, Tomi P. Lintonen, and Ville J. Autio
Tampere School of Public Health, University of Tampere, Finland
ABSTRACT
The aim of this study was to explore the variation of general subjective well-being on individual
and school levels. The classroom survey data of the School Health Promotion Survey (SHPS)
were gathered in 1998 (N39 886) and in 1999 (N47 455) in different parts of Finland.
The subjects were 8th and 9th graders from 458 secondary schools. Well-being was measured
using the General Subjective Well-being Indicator (GSWI). The variation was studied using
linear multilevel modeling. One percent of the variation in GSW occurred at the school level.
The great majority of subjective well-being variation occurred between pupils.
INTRODUCTION
Adolescents' ill-being is one of the main concerns of the millennium
(Navarrate, 1999). Ill-being at school can be measured in terms of the
prevalence of depression, for instance. This, however, implies a rather narrow
focus on severe disorders and accordingly on a minority of adolescents; the
well-being of the majority of adolescents is equally important.
Lazarsfeld, Guttman, Suchman, and others (McDowell & Newell, 1996)
developed the ®rst generation of psychological well-being measures in the
post-war years. Since then, well-being has been measured by a number of
researchers using various instruments, including The Life Satisfaction Index
(Neugarten, Havighurst, & Tobin, 1961), The General Health Questionnaire
Address correspondence to: Anne Konu, Tampere School of Public Health, FIN-33014
University of Tampere, Finland. Tel.: 358 3 215 6172. Fax: 358 3 215 6057. E-mail:
anne.konu@uta.®
Manuscript submitted: March 27, 2001
Accepted for publication: January 8, 2002
(Goldberg, 1978), The General Well-being Schedule (Dupuy, 1984), and The
Oxford Happiness Inventory (Argyle, Martin, & Crossland, 1989). These
scales are mostly concerned with personality disorders, distress and psycho-
logical well-being, and deal with areas such as happiness, life-satisfaction and
morale. Bowling (1991) and McDowell and Newell (1996) have reviewed
health and well-being scales in more detail.
According to Veenhoven (1991) overall subjective well-being can be under-
stood in terms of life satisfaction, contentment and hedonic level, while
different aspects of subjective well-being include self-appraisals like job
satisfaction, self-esteem and control belief. By life satisfaction, Veenhoven
(1991) means ``the degree to which an individual judges the overall quality of
his life-as-a-whole favourable'' (p. 10).
Allardt (1989) cross-tabulates his concept of well-being with the dichotomy
of objective and subjective indicators. Allardt divides the other axis into the
categories of having, loving and being. Having (material and impersonal needs)
refers to dissatisfaction±satisfaction, loving (social needs) to unhappiness±
happiness, while in the being (needs for personal growth) category the
question is whether a person experiences alienation or personal growth. The
subjective aspect refers to asking people to express their attitudes and to
convey their subjective perceptions of their living conditions.
Huebner and colleagues (Huebner, 1991; Huebner, Gilman, & Laughlin,
1999; McCullough, Huebner, & Laughlin, 2000; Terry & Huebner, 1995) have
concentrated on the construct of subjective well-being among children and
adolescents. They use the Students' Life Satisfaction Scale (SLSS) and see
subjective well-being among children and adolescents as a three-component
construct: global life satisfaction and positive and negative affect (Huebner,
1991; McCullough et al., 2000).
Well-being in the school context has been studied by Knuver and Brandsma
(1993), Samdal (1998) and Opdenakker and Van Damme (2000). Knuver and
Brandsma (1993) studied both cognitive and affective outcomes in school.
Affective outcomes referred to attitudes the student had towards school and
learning. They measured school well-being by Stoel's (1980) questionnaire
about pupils' experience (positive or negative) of their school and its
organisation, their teacher and their classmates. Schools' cognitive and
affective outcomes were relatively independent (Knuver & Brandsma, 1993).
Samdal (1998) measured subjective well-being with one single item: ``In
general, how do you feel about your life at present?'' Student support,
adequate expectations and teacher support were the most important predictors
188 ANNE I. KONU ET AL.
of subjective well-being. Opdenakker and Van Damme (2000) used a well-
being questionnaire consisting of eight indicators: well-being at school, social
integration in the class, relationships with teachers, interest in learning tasks,
motivation towards learning tasks, attitude to homework, attentiveness in the
classroom, and academic self-concept. The study indicated some school char-
acteristics (e.g., instruction and knowledge acquisition) to be effective for both
achievement and well-being while other characteristics affected more promi-
nently either one. Overall, the results showed that the relative in¯uence of classes
and schools on achievement was much higher than the in¯uence on well-being.
In this study, we used the General Subjective Well-being Indicator (GSWI)
which is based on the Raitasalo (1995) modi®ed short version of the Beck
Depression Inventory (Beck & Beck, 1972; Beck, Steer, & Garbin, 1988). The
GSWI consists of 13 items of general perceptions of life: positive mood,
future orientation, success, satisfaction, global self-esteem and speci®c self-
esteem (appearance), social orientation, decision-making, sleeping, energy,
appetite and anxiousness.
On the one hand, health and well-being are often seen as personal states. On
the other hand, the context in which people spend a considerable amount of
time may have a signi®cant impact on their health and well-being (Hawe,
1998). Until recently, the context-level effects have been studied mainly in the
educational ®eld by comparing pupils' achievements between schools
(Arnold, 1992; Goldstein, 1995; Young & Reynolds, 1996), rarely so in public
health (Hawe, 1998; Rice & Leyland, 1996).
Multilevel models (hierarchical linear models) offer the possibility to study
data taking into account its natural hierarchical structure; for example, schools,
classes and pupils. These models provide the possibility to partition the variation
of the studied phenomenon into variation attributable to individual factors and to
higher level factors (Goldstein, 1995; Rice & Leyland, 1996).
The aim of this study was to explore the variation of school children's general
subjective well-being on pupil and school levels using the 13-item GSWI to
establish how the phenomenon is divided between the individual and the context.
METHODS
Data Collection
The data were collected as a part of The School Health Promotion Survey
(SHPS). The 16-page classroom survey covers numerous areas of pupils'
MULTILEVEL STUDY OF PUPILS' WELL-BEING 189
health and lifestyle and has been carried out in Finland annually since 1995.
The data were collected in the same geographical areas every other year. Self-
administered questionnaires were used in classrooms supervised by the
pupils' own teacher. The mailing envelopes containing the questionnaires
were sealed in the presence of pupils to ensure con®dentiality. The data used
in this present study were collected in April 1998 (N43 085) in the eastern
part of Finland and in April 1999 (N50 282) in the southern and western
parts of Finland. The respondents were eighth and ninth graders, aged 14.3±
16.2 years. In Finland, the compulsory education consists of 9 school years,
starting at the age of 7. The respondents came from the last two classes of
compulsory education. Usually, students continue their studies after ninth
grade either in upper secondary school or in vocational education.
The rate of valid responses among registered pupils in these areas (Statistics
Finland, 1999) during the surveys was 61% in 1998 and 74% in 1999; 6.5% of
the cases were further excluded because respondents came from schools with
under 70 pupils (3.0%), from special schools (1.0%) or from schools with less
than 15 respondents in each gender/grade group (2.5%). This was done for two
reasons. First, too few respondents from a school may not be representative for
the whole school. Second, it has been argued that the size of the partitions in
multilevel analysis has to be adequate (Rice & Leyland, 1996). Forty-six
percent of the ®nal data (N87 341; 458 schools) were from the year 1998
and 49% of the respondents were females.
The cases with more than 4 missing items out of 13 comprising the
explanatory variable (GSWI) were excluded (1.4% of the data) from the
analysis. This procedure did not change the results of the analysis.
Measurement and Analysis
The dependent variable is a score, the GSWI, which is based on Raitasalo
modi®ed (Raitasalo, 1995) 13-item Beck Depression Inventory (RBDI). Raita-
salo added an introductory question to each statement area. In addition, a
positive option was added to each item and the order of the other statements was
reversed. Each positive statement scores 1, which means that the maximum
score is 13. True well-being and signs of severe depression cannot occur at the
same time, so the GSWI score was set to zero if the RBDI depression score
exceeded 15 points. This operation was applicable to 1.7% of the respondents,
most of whom (80%) scored only one or two points on the GSWI scale. The
questions included in the GSWI scale were internally consistent: the Kuder-
Richardson (KR20) estimate (Black, 1999, pp. 282±286) was 0.76.
190 ANNE I. KONU ET AL.
In two samples of pupils in 1998 in eastern Finland (N43 085), and in
1999 in the southern and western part of Finland (N50 282) the GSWI
distributions were similar; the mean was 4.6 and standard deviation 3.0 both
years.
The independent variables (Table 1) on the individual level were categorized
as background (5 variables), school conditions (19), social relationships in
school (10), means for self-ful®lment in school (12), and health status (4). The
school-related categories were based on the conceptual model of well-being in
school (Fig. 1; Konu & Rimpela
È, in press). Gender, grade, family structure,
unemployment and guardian's education were used for adjustment purposes.
Variables were dichotomised and aggregated (except gender and grade) to
school level by using the percentage of each variable's positive option. The
school-level variables were centred for statistical program use. A total of 50
individual-level and 48 school-level variables were tested for relations with
subjective well-being using linear multilevel modeling (Goldstein, 1995; Rice &
Leyland, 1996). First, only the dependent variable was included in the model to
®nd out the distribution of variation between individual and school levels.
Second, variables from each of the four categories (school conditions, social
relationships, means for self-ful®lment, and health status) of the School Well-
being Model were included separately in the regression model adjusted for
gender, grade, family structure, unemployment, and guardian's education. Only
variables that had been signi®cant were included in the ®nal model (35
individual and 10 school-level variables). In the regression model the estimation
method was Iterative Generalised Least Squares (IGLS; Goldstein, 1995).
MLwiN (beta version) statistical program package was used for analysis.
RESULTS
Distribution of the general subjective well-being indicator was close to normal
(M4:6, SD 3:0), with those not being well accumulated at score zero
(Fig. 2). There were clear gender differences in the pupils' GSWI; the mean
for boys was 4.93 (95% CI 4.90±4.96) compared with 4.29 (95% CI 4.26±
4.31) for girls. The variation among boys was greater (SD 3:2) than among
girls (SD 2:8). Among eighth graders the mean was 4.54 (95% CI 4.51±
4.56) and among ninth graders 4.69 (95% CI 4.66±4.72). The mean of the
school averages of pupils' GSW was 4.6 (SD 0:4) with 10% percentile at
4.1 and 90% percentile at 5.1 (Fig. 3).
MULTILEVEL STUDY OF PUPILS' WELL-BEING 191
Table 1. Individual-Level Independent Dichotomous Variables, Answering Options and the
Proportions of Positive Responses.
Category Variable Option (alternative) %
Gender male (female) 51
Grade 9th (8th) 50
Parents' unemployment no (father, mother or both) 68
Family structure nuclear family (other) 75
Guardians' education 12 or more years (less than 12 years) 45
Pupils have a home room yes (no) 60
Locker where to keep things yes (no) 31
Number of schedules during a term many (one) 90
School lunch is a relaxing break yes (no) 80
The amount of school work just right (too much or too little) 54
Crowdedness disturbing school work no (yes) 83
Noise disturbing school work no (yes) 73
Lighting disturbing school work no (yes) 82
Ventilation disturbing school work no (yes) 36
Temperature disturbing school work no (yes) 39
Dirtiness disturbing school work no (yes) 64
Inappropriate desks complicating
school work
no (yes) 47
Facilities disturbing school work no (yes) 54
Restlessness disturbing school work no (yes) 71
Haste disturbing school work no (yes) 58
Violence disturbing school work no (yes) 79
Risk of accidents disturbing school
work
no (yes) 83
Getting to meet school nurse easy (dif®cult) 87
Good work atmosphere in class yes (no) 54
Being bullied at school no (yes) 66
Teachers interested in how pupil
is doing
yes (no) 28
Problems working in teams no (yes) 94
Problems getting along with school
friends
no (yes) 95
Having bullied at school no (yes) 56
Pupils enjoy being together yes (no) 72
Problems getting along with teachers no (yes) 81
Teachers treat pupils fairly yes (no) 53
Pupils' views are respected in
school development
yes (no) 46
Con®dentiality in school health
service
satis®ed (not satis®ed) 78
Social relationships School conditions Background
192 ANNE I. KONU ET AL.
Most of the GSW variation occurred at the individual level, only 1%
occurring at the school level (school-level estimate 0.086 with standard error
0.009; individual-level estimate 8.901 with standard error 0.043). The ®nal
model with 35 variables from the individual level and 10 variables from the
school level accounted for 19% of the individual-level variation and 65% of the
school-level variation. The ®nal model estimates and their standard errors are
shown in Table 2. The Konu and Rimpela
È(in press) School Well-being Model
divides these variables into four categories: school conditions, social relation-
ships, means for self-ful®lment, and health status. At the individual level,
variables belonging to the means for self-ful®lment category corresponded
most with GSW variation (see also Konu, Lintonen, & Rimpela
È, in press).
DISCUSSION
Researchers in education have used multilevel approaches in studying the
effects of schools on pupils' achievements (Arnold, 1992; Goldstein et al.,
Table 1. (Continued).
Problems ®nding a personal way to
study
no (yes) 75
Problems with tasks requiring
personal activity
no (yes) 77
Get help with problems in school
or studies
yes (no) 83
Problems preparing for tests no (yes) 65
Plans for future education snr. high school and university (other) 43
Teachers expect too much no (yes) 68
Teachers encourage pupils to express
their views
yes (no) 54
Problems following teaching no (yes) 83
Problems doing tasks that require
writing
no (yes) 83
Problems doing tasks that require
reading
no (yes) 81
Problems doing homework no (yes) 73
Personally chosen courses 0±4 (more) 43
Symptoms weekly no (one or more) 34
Common colds during past half year no more than one (more often) 49
Chronic disease no (yes) 90
Asthma no (yes) 94
Means for self-ful®lmentHealth
Status
MULTILEVEL STUDY OF PUPILS' WELL-BEING 193
Fig. 1. The school well-being model (Konu & Rimpela
È, in press).
Fig. 2. Distribution of subjective well-being indicator (N87 341 pupils).
194 ANNE I. KONU ET AL.
1993; Young & Reynolds, 1996). This study concentrated on pupils' general
subjective well-being and found that in Finnish schools the variation of
subjective well-being occurred mainly at the individual level. One percent
of the variation appeared at the school level. This ®nding is in line with
Opdenakker's and Van Damme's study (2000) showing that the relative
in¯uence of classes and schools was much higher on achievements than on
well-being.
Any time individuals form natural groups like classes and schools, it is
crucial to take into account the natural hierarchical structure of the data. This
is possible using the multilevel approach (Goldstein, 1995; Rice & Leyland,
1996) to analyse a continuous dependent variable. The studied variable,
general subjective well-being, was continuous and normally distributed
(Fig. 2). The variation was studied at school and individual levels. The class
level was omitted, because in Finnish schools pupils form groups according to
their personally chosen subjects. The data from the SHPS offered an excellent
possibility for multilevel analysis with 458 schools and no less than 70 pupils
in each school.
Measuring non-cognitive outcomes in school is quite problematic (Knuver
& Brandsma, 1993). Well-being measures among school children are still not
Fig. 3. Distribution of school averages of pupils' subjective well-being indicator (N458
schools).
MULTILEVEL STUDY OF PUPILS' WELL-BEING 195
Table 2. Final Model Estimates and Their Standard Errors.
Level Estimate Standard Error
Random part
School general subjective well-being 0.030 0.006
Individual general subjective well-being 7.188 0.042
Fixed part: variable
Cons ÿ0.286 0.084
Individual Gender 0.558 0.025
Individual Grade 0.187 0.022
Individual Parents' unemployment 0.153 0.024
Individual Family structure 0.105 0.027
Individual Guardians' education 0.129 0.024
Individual Locker where to keep things 0.106 0.029
Individual School lunch is a relaxing break 0.199 0.029
Individual The amount of school work is right 0.157 0.024
Individual Temperature disturbing school work ÿ0.018 0.025
Individual Dirtiness disturbing school work ÿ0.181 0.026
Individual Inappropriate desks complicating school work ÿ0.013 0.025
Individual Restlessness disturbing school work ÿ0.097 0.028
Individual Haste disturbing school work 0.128 0.025
Individual Risk of accidents disturbing school work ÿ0.116 0.032
Individual Getting to meet school nurse 0.081 0.035
Individual Good work atmosphere in class 0.081 0.024
Individual Being bullied at school 0.708 0.025
Individual Teachers are interested in how pupil is doing 0.418 0.026
Individual Problems working in teams 0.280 0.050
Individual Problems getting along with school friends 0.171 0.056
Individual Having bullied at school 0.153 0.025
Individual Pupils enjoy being together 0.271 0.026
Individual Problems getting along with teachers ÿ0.100 0.033
Individual Teachers treat pupils fairly 0.043 0.025
Individual Pupils' views are respected in school development 0.070 0.024
Individual Con®dentiality in school health service 0.190 0.029
Individual Problems ®nding a personal way to study 0.433 0.030
Individual Problems with tasks requiring personal activity 0.533 0.030
Individual Get help with problems in school or studies 0.624 0.032
Individual Problems preparing for tests 0.352 0.030
Individual Plans for future education 0.427 0.025
Individual Teachers expect too much 0.136 0.026
Individual Teachers encourage pupils to express their views 0.141 0.024
Individual Problems following teaching 0.129 0.035
Individual Problems doing tasks that require writing 0.125 0.034
Individual Problems doing tasks that require reading 0.074 0.032
Individual Problems doing homework 0.087 0.032
196 ANNE I. KONU ET AL.
well established and further methodological research is needed, for example
on reliability and validity issues. This present study was cross-sectional in
design and needs to be complemented with longitudinal work. Further, all
school-level variables were aggregated from the individual level. Variables
measured at the school level would be an important information source.
A surprisingly small share of the variation was observed to occur between
schools. This may be due to quite homogeneous conditions in Finnish schools.
However, even though the share of the between-school variation appeared to
be small, we have to keep in mind that the variation was notable as can be seen
in Fig. 3. The variation between pupils within schools was so much greater
that the share of school variation appeared small. The differences in pupils'
average well-being between schools in the opposite ends of the school
distribution were, however, notable with more than a 2-point difference on
a scale from 0 to 13. Furthermore, the dependent variable in this study was
general subjective well-being. To ®nd out more speci®c school effects on
well-being, the dependent indicator should concentrate more on well-being
areas speci®c to schools.
In Finnish schools, the variation of pupils' general subjective well-being
occurred mostly at the individual level. This indicates that a more detailed
analysis of this speci®c indicator, the general subjective well-being, can be
Table 2. (Continued).
Individual Symptoms weekly 0.929 0.026
Individual Common colds during past half year 0.288 0.023
Individual Chronic disease 0.110 0.037
School Parents' unemployment ÿ0.003 0.002
School Family structure 0.003 0.002
School Guardians' education ÿ0.006 0.002
School School lunch is a relaxing break ÿ0.005 0.002
School Ventilation disturbing school work 0.001 0.001
School Being bullied at school ÿ0.005 0.002
School Problems getting along with teachers 0.006 0.004
School Teachers treat pupils fairly ÿ0.004 0.002
School Get help with problems in school or studies 0.005 0.004
School Problems preparing for tests 0.004 0.004
School Plans for future education 0.013 0.002
School Problems doing homework ÿ0.005 0.004
School Common colds during past half year 0.008 0.002
MULTILEVEL STUDY OF PUPILS' WELL-BEING 197
done at the individual level. However, at the individual level, a considerable
part of the variation was connected with items concerning studying and other
aspects of school life (see also Konu et al., in press). It seems that school does
have an effect on pupils' general subjective well-being, but Finnish schools
are relatively homogeneous in this respect.
Notable between-school variation has been found when studying achieve-
ments (Goldstein et al., 1993; Opdenakker & Van Damme, 2000). This is quite
natural since the school's focus is on teaching and measuring achievements.
Should the schools consider well-being aspects as important as achievements,
an impact might be seen on well-being at the school level. It is crucial to ®nd
the best practices to raise the well-being of pupils in school. Educators in
schools would do well if they considered aspects affecting pupils' well-being
more.
ACKNOWLEDGEMENTS
This study was ®nancially supported by The Finnish Cultural Foundation and
The Foundation of Juho Vainio.
REFERENCES
Allardt, E. (1989). An updated indicator system: Having, loving, being (Working papers 48).
Helsinki, Finland: Department of Sociology, University of Helsinki.
Argyle, M., Martin, M., & Crossland, J. (1989). Happiness as a function of personality and
social encounters. In J.P. Forgas & J.M. Innes (Eds.), Recent advances in social
psychology. An international perspective (pp. 189±203). Amsterdam: Elsevier Science.
Arnold, C. (1992). An introduction to hierarchical linear models. Measurement and Evaluation
in Counseling and Development,25, 58±89.
Beck, A.T., & Beck, R.W. (1972, December). Screening depressed patients in family practice.
A rapid technic. Postgraduate Medicine, 81±85.
Beck, A.T., Steer, R.A., & Garbin, M.G. (1988). Psychometric properties of the Beck
Depression Inventory: Twenty-®ve years of evaluation. Clinical Psychology Review,8,
77±100.
Black, T.R. (1999). Doing quantitative research in the social sciences. An integrated approach
to research design, measurement and statistics. London: SAGE.
Bowling, A. (1991). Measuring health: A review of quality of life measurement scales.
Buckingham, UK: Open University Press.
Dupuy, H.J. (1984). The Psychological General Well-Being (PGWB) Index. In N.K. Wenger,
M.E. Mattson, C.D. Furberg, & J. Elinson (Eds.), Assessment of quality of life in clinical
trials of cardiovascular therapies (pp. 170 ±183). New York: LeJacq.
198 ANNE I. KONU ET AL.
Goldberg, D. (1978). Manual of the General Health Questionnaire. Windsor, UK: NFER.
Goldstein, H. (1995). Multilevel statistical models. London: Edward Arnold.
Goldstein, H., Rasbash, J., Yang, M., Woodhouse, G., Pan, H., Nuttal, D., & Thomas, S. (1993).
A multilevel analysis of school examination results. Oxford Review of Education,19,
425 ± 433.
Hawe, P. (1998). Making sense of context-level in¯uences on health [Editorial]. Health
Education Research,13, i±iii.
Huebner, E.S. (1991). Further validation of the Students' Life Satisfaction Scale: The
independence of satisfaction and affect ratings. Journal of Psychoeducational
Assessment,9(4), 363±368.
Huebner, E.S., Gilman, R., & Laughlin, J.E. (1999). A multimethod investigation of the
multidimensionality of children's well-being reports: Discriminant validity of life
satisfaction and self-esteem. Social Indicators Research,46, 1±22.
Knuver, A.W.M., & Brandsma, H.P. (1993). Cognitive and affective outcomes in school
effectiveness research. School Effectiveness and School Improvement,4, 189±204.
Konu, A., & Rimpela
È, M. (2002) Well-being in schools ± a conceptual model. Health
Promotion International,17, 79±87.
Konu, A., Lintonen, T., & Rimpela
È, M. (in press). Factors associated with schoolchildren's
general subjective well-being. Health Education Research.
McDowell, I., & Newell, C. (1996). Measuring health. A guide to rating scales and
questionnaires. New York: Oxford University Press.
McCullough, G., Huebner, E.S., & Laughlin, J.E. (2000). Life events, self-concept, and
adolescents' positive subjective well-being. Psychology in the Schools,37(3), 281±290.
Navarrate, L.A. (1999). Melancholy in the millennium: A study of depression among
adolescents with and without learning disabilities. High School Journal,82(3), 137±149.
Neugarten, B.L., Havighurst, R.J., & Tobin, S.S. (1961). The measurement of life satisfaction.
Journal of Gerontology,16, 134±143.
Opdenakker, M.-C., & Van Damme, J. (2000). Effects of schools, teaching staff and classes on
achievement and well-being in secondary education: Similarities and differences
between school outcomes. School Effectiveness and School Improvement,11, 165±196.
Raitasalo, R. (1995). Ela
Èma
Ènhallinta sosiaalipolitiikan tavoitteena [Coping as the target of
social policy]. Helsinki, Finland: Kansanela
Èkelaitos, Sosiaali-ja terveysturvan tutki-
muksia 1.
Rice, N., & Leyland, A. (1996). Multilevel models: Applications to health data. Journal of
Health Services Research and Policy,1, 154±164.
Samdal, O. (1998). The school environment as a risk or resource for students' health-related
behaviours and subjective well-being. Research Centre for Health Promotion, University
of Bergen, Norway.
Statistics Finland. (1999). Pupils in Comprehensive School 1999 [On-line]. Available: http://
www.tilastokeskus.®/ April 19, 2000.
Stoel, W.G.R. (1980). De beleving van de school door leerlingen in het voortgezet onderwijs.
De ontwikkeling van een schoolbelevingsschaal [Pupils' well-being in secondary
schools. The development of an instrument for measuring pupils' well-being]. Haren,
The Netherlands: RION.
Terry, T., & Huebner, E.S. (1995). The relationship between self-concept and life satisfaction in
children. Social Indicators Research,35(7), 39±52.
MULTILEVEL STUDY OF PUPILS' WELL-BEING 199
Veenhoven, R. (1991). Questions on happiness: Classical topics, modern answers, blind spots.
In F. Strack, M. Argyle, & N. Schwarz (Eds.), Subjective well-being. An interdisciplinary
perspective (pp. 7±26). Oxford: Pergamon Press.
Young, D.J., & Reynolds, A.J. (1996). Science achievement and educational productivity: A
hierarchical linear model. Journal of Educational Research,89, 272±278.
200 ANNE I. KONU ET AL.
