Journal of Youth and Adolescence

, Volume 39, Issue 8, pp 870–881

Trajectories of Life Satisfaction Across the Transition to Post-Compulsory Education: Do Adolescents Follow Different Pathways?

Authors

    • Helsinki University Collegium for Advanced Studies, University of Helsinki
  • Lotta Tynkkynen
    • University of Jyväskylä
Empirical Research

DOI: 10.1007/s10964-009-9464-2

Cite this article as:
Salmela-Aro, K. & Tynkkynen, L. J Youth Adolescence (2010) 39: 870. doi:10.1007/s10964-009-9464-2
  • 268 Views

Abstract

To examine the developmental trajectories of life satisfaction shown by adolescents during a major educational transition, 687 (327 girls, 360 boys) Finnish adolescents responded to measures of life satisfaction during the beginning of their last year in comprehensive school and three times annually thereafter during the transition to post-comprehensive education. Growth Mixture Modeling identified three latent groups based on life satisfaction: high-decreasing (18%), low-increasing (16%), and high-stable (66%). The results showed that boys and those with high academic achievement were overrepresented in the high-stable life satisfaction group. The results further showed that those with high school satisfaction at the last measurement time were more likely to belong to the high-stable or low-increasing life satisfaction group. Finally, adolescents in the high-stable life satisfaction group were more often on an academic track at the last measurement time.

Keywords

AdolescenceLife satisfactionTrajectoriesLongitudinalEducational transitionGrowth mixture modelingGender

Introduction

The majority of young people manage adolescence without severe problems (Graber and Brooks-Gunn 1996; Powers et al. 1989). However, some have difficulties in adapting to the transitions and changes that characterize this age period (Rutter 1990). A gender difference also has been noted, boys showing higher life satisfaction than girls (e.g., Neto 1993). Although much research has been conducted on life satisfaction during adulthood (e.g., Steger and Kashdan 2007), it is only recently that attention has been paid to changes in life satisfaction during the years of adolescence. The previous results are mixed as some studies have found a decrease in life satisfaction during adolescence (e.g., Goldbeck et al. 2007), while others have found an increase (e.g., Salmela-Aro and Tuominen-Soini in press). However, a single growth trajectory runs the risk of oversimplifying the heterogeneity of the changes between adolescents in life satisfaction. Moreover, it has only recently been suggested that educational transitions may play a destabilizing role in life satisfaction (Salmela-Aro et al. 2008a; Tram and Cole 2006). However, to our knowledge, no previous study has focused on heterogeneity in the developmental trajectories of life satisfaction across an educational transition. Moreover, few studies have examined the role played by school-related factors, such as academic achievement and school satisfaction, in life satisfaction during educational transitions. To find answers to these questions, the present 4-wave longitudinal study examined whether adolescents’ life satisfaction changes and what kinds of life satisfaction trajectories adolescents show across the transition to post-compulsory education. In addition, the role of gender, academic achievement, educational track and satisfaction with studies in relation to membership of differing trajectories of life satisfaction were examined.

Life Satisfaction During Adolescence

Life satisfaction can be defined as an individual’s overall appraisal of the quality of his or her life (Diener et al. 1985), including the perception that one is progressing towards important life goals (Diener et al. 1999). Shin and Johnson (1978) defined life satisfaction as a global assessment of a person’s quality of life according to his or her chosen criteria, and Frisch (2000) defined it as a person’s subjective evaluation of the degree to which his or her most important needs, goals, and wishes have been fulfilled (Compas et al. 1995; Gilman and Huebner 2003; Huebner 2004; Lerner and Steinberg 2004). Most previous studies on life satisfaction have examined adults or emerging adults, while research on adolescents’ life satisfaction has lagged behind (see Funk et al. 2006).

The studies conducted on life satisfaction among adolescents have shown that most adolescents view their overall lives positively (Gilman and Huebner 2003). In a US study, 73% of the adolescents surveyed rated their global life satisfaction in a positive manner and 11% rated it negatively (Huebner et al. 2000). Some studies have examined changes in life satisfaction during adolescence and the results have been mixed: some have found life satisfaction to decrease (Goldbeck et al. 2007), while others have found it to increase during adolescence (Salmela-Aro and Tuominen-Soini in press). Previous studies have, however, mainly examined the stability of life satisfaction. Over moderate lengths of time, scores have been very stable, with test–retest correlations ranging from 0.79 to 0.89 over periods ranging from 2 weeks to 2 months (Lyubomirsky 2001; Pavot and Diener 1993). Life satisfaction also appears to be stable over longer periods of time (Ehrhardt et al. 2000; Fujita and Diener 2005; Schimmack and Oishi 2005, for review). For example, a meta-analysis of test–retest studies found 1-year test–retest coefficients to range from 0.55 to 0.75 when multi-item life satisfaction measures were used (Schimmack and Oishi 2005). In the long term, life satisfaction appears moderately stable, yet with considerable unexplained variance, suggesting that it might be sensitive to life events (Pavot and Diener 1993). However, a recent study (Steger and Kashdan 2007) conducted among university students found less stability in life satisfaction scores than previously reported. One possible explanation for this is that there appears to be an increase in the stability of life satisfaction scores with age (Ehrhardt et al. 2000). The low stability estimates of Steger and Kashdan (2007) might thus be due to the relatively young age of the sample.

One further limitation of the previous studies on life satisfaction among adolescents is that most have examined the developmental course of life satisfaction during adolescence only at the mean level (see also Costello et al. 2008) and have assumed that a single trajectory and individual variation around it can adequately approximate an entire population (Jung and Wickrama 2008; Muthén and Muthén 2000). The problem with this approach is that it does not investigate the possibility that individual trajectories are different. Recently, it has been suggested that person-oriented (Bergman et al. 2003) or idiographic methods, such as trajectory analysis (Nagin 1999; Nagin and Land 1993; Nagin and Tremblay 1999), or growth mixture modeling (Muthén 2004; Muthén and Muthén 2000; Muthén and Shedden 1999) may be more fruitful methods for examining development than simple analysis of the mean level change. The aim of growth mixture modeling is to search for the optimal number of latent classes formed on the basis of the growth factor means of level and slope so that each class defines a different trajectory over time and the variances of these growth factors and the covariance between them are allowed to vary between latent classes (Muthén 2001). However, to the best of our knowledge, none of the previous studies have focused on investigating the developmental trajectories of life satisfaction during adolescence, which is the main aim of this study.

Latent Trajectories of Life Satisfaction Across Educational Transitions During Adolescence

Formal education provides an important developmental context for adolescents’ psychological functioning (Eccles 2004). The transition to post-compulsory education is a key educational transition in many European educational systems and a challenge for school adjustment. It determines the quality and the kinds of learning opportunities available to students (Oakes et al. 1992), while exposing them to a wide range of peers and teachers. For some students, educational transitions may act as a turning point in life satisfaction: for some adolescents such transitions will be successful while for others they may be disruptive (Crockett et al. 1989; Eccles and Midgley 1989; Graber and Brooks-Gunn 1996; Hirsch and Rapkin 1987). However, studies investigating the development of life satisfaction across an educational transition are lacking (Salmela-Aro and Tuominen-Soini in press; Tram and Cole 2006). Consequently, the present study examined what kinds of developmental trajectories adolescents show in their life satisfaction across the transition to post-compulsory education.

Previous research has shown that adolescents’ perceptions and experiences of school are associated with various adjustment outcomes. Poor academic performance and academic failures are related to psychological stress and negative affect (Cole et al. 1999; Crystal et al. 1994), while high academic achievement is related to high emotional well-being and protects against maladjustment (Gerard and Buehler 2004). In the present study we examine the extent to which academic achievement plays a role in adolescents’ life satisfaction trajectories. Our expectation is that high achievement will be related to a high life satisfaction trajectory.

A successful school life plays an important role in adolescents’ lives and is an important developmental task of adolescence (Schulenberg et al. 2005). The positive resolution of school-related challenges is thus important for a successful transition to adulthood, while failure to benefit from school carries the risk for subsequent developmental disadvantage and problems (Schulenberg et al. 2004). Life satisfaction during adolescence can be expected to lead to success in dealing with the major school-related challenges: a successful transition to and satisfaction with school life could be seen as a potentially important consequence of life satisfaction (Gilman and Huebner 2005; Salmela-Aro and Tuominen-Soini in press). Consequently, we expected those on the high life satisfaction trajectory to experience high school satisfaction later on.

Gender and Life Satisfaction

Gender might also play a role in determining different trajectories of life satisfaction during adolescence (Ge et al. 2001; Nolen-Hoeksema and Girgus 1994). While some studies have found no gender differences among adolescents in life satisfaction (Adelman et al. 1989; Dew and Huebner 1994), others have found a gender difference (Neto 1993; Verkuyton 1996), with boys experiencing higher life satisfaction than girls. Previous research also has shown gender differences in school adjustment. For example, girls experience higher levels of stress (e.g., Ge et al. 1994; Jose and Ratcliffe 2004; Matud 2004), school burnout (Salmela-Aro et al. 2008a), and more internalized symptoms (e.g., Hoffmann et al. 2004; Nolen-Hoeksema and Girgus 1994; Pomerantz et al. 2002) than boys. There is some evidence to suggest that girls respond more negatively to competitive learning conditions. In line with this, research shows that girls not only are more exposed to stressful life events, but also are more vulnerable to their negative effects (Ge et al. 1994; Kessler and McLeod 1984; Salmela-Aro et al. 2008a, b). Consequently, we expected boys to be found more often in a high life satisfaction trajectory and girls in trajectory characterized by a low level of life satisfaction.

Aims and Hypotheses

Our first aim was to examine to what extent adolescents’ life satisfaction changes during the transition to post-comprehensive education. Our second aim was to examine to what extent heterogeneity exits in adolescents’ life satisfaction during the transition from comprehensive school to post-compulsory education and the kinds of developmental trajectories adolescents show in their life satisfaction during the transition to post-compulsory education. We expected, first, that most adolescents would be found in the trajectory of high and stable life satisfaction. However, we further assumed that a single growth trajectory would oversimplify the heterogeneity of the changes in adolescents’ life satisfaction during the transition to post-compulsory education, some experiencing an increase and some a decrease in life satisfaction during the transition, but most passing through the same stage with a stable and high level of life satisfaction (Hypothesis 1). We expected that during the educational transition changes showing either an increase or a decrease in life satisfaction would take place in the life satisfaction trajectories (Hypothesis 2).

Our third aim was to examine the extent to which gender and academic achievement would be related to certain life satisfaction trajectory groups during the transition to post-compulsory education. It was expected that being a boy (Neto 1993) and having high academic achievement would increase the likelihood of membership of a high life satisfaction trajectory (Hypothesis 3). Our final aim was to examine the extent to which belonging to certain latent trajectories of life satisfaction would be related to school satisfaction and educational tracks later on. It was assumed (Hypothesis 4) that those in the high life satisfaction groups would have high school satisfaction and be on an academic track later on.

Schooling in Finland

Finnish children start their education at kindergarten during the year of their sixth birthday. One year later, at age seven, they move to compulsory comprehensive school, where they continue for the next 9 years. Up until age sixteen, all Finnish adolescents receive a similar basic education. After comprehensive school, adolescents’ educational trajectories begin to diverge. About 55% of all adolescents enter senior high schools and 37% vocational schools, 2% stay on for a voluntary tenth grade, and 6% exit formal education (School Statistics, Central Statistical Office of Finland, 2003). Average academic achievement in the ninth grade is the minimum requirement for admission to senior high school. Senior high school education, in turn, is a bridge to further education and, most likely, higher education. Secondary vocational education serves as a route to working life in academically less demanding occupations, and also to tertiary-level education, most likely vocational education. Thus, educational choices at the end of comprehensive school channel young Finns onto either an academic or a vocational track (OECD 1998). Finnish girls graduate from senior high schools and enter universities more often than boys (Education in Finland 1999). Education in Finland is state-provided and tuition is free.

Method

Participants

The present study is part of the ongoing FinEdu study. The aim of FinEdu is to examine adolescents’ life-planning and well-being during educational transitions in middle and late adolescence. At the beginning of the present study, the participants were ninth-graders (median age = 15) facing the transition to post-comprehensive schooling. All the ninth-grade students in a medium-sized town (population = 88,000) in Central Finland were recruited for the study (N = 954). The participants were asked to participate four times. Two measurements were carried out before the transition to senior secondary (academic track) or vocational (vocational track) education: one at the beginning of the 9th grade, which is the final term of comprehensive school (Time 1), and the other at the end of the 9th grade (Time 2). Two measurements were also carried out after the transition to post-comprehensive schooling: the first was half a year after the transition (Time 3) and the second was 1 year after Time 3 (Time 4).

At Time 1,687 (327 girls, 360 boys) adolescents out of the 954 students attending the nine comprehensive schools participated in the study (retention rate 72%); at Time 2, the number of participants was 642 (317 girls, 325 boys; retention rate 67%); at Time 3, the number was 818 (396 girls, 422 boys; retention rate 86%); and at Time 4, the number was 749 (368 girls, 381 boys; retention rate 79%).

The majority of the participants (99%) were Finnish-speaking, 1% of them having some other mother tongue. This ratio agrees well with the figures for ethnic minorities at the national level. The questionnaires were group-administered to the students in their classrooms during regular school hours or sent to their postal address. The adolescents answered questions concerning their life satisfaction at all four measurement points. Gender and GPA were measured at Time 1, and educational track and school satisfaction were measured at Time 4.

Attrition Analyses

Attrition analyses were carried out to examine attrition between the measurements by comparing the adolescents who participated in the study at each measurement point (N = 469) with those who had missing data at some measurement point (N = 389) with regard to the variables life satisfaction, academic achievement, and attained educational track. The results showed that there was a small selection effect with respect to the life satisfaction variables. Those who did not answer at all four measurement points had slightly lower scores on life satisfaction at Time 1 (M = 4.38, SD = 1.55), Time 2 (M = 4.58, SD = 1.46), Time 3 (M = 4.56, SD = 1.40), and Time 4 (M = 4.74, SD = 1.30) than those who did (Time 1: M = 4.70, SD = 1.28, t(335) = −2.61, p < 0.05; Time 2: M = 4.90, SD = 1.25, t(265) = −2.57, p < 0.05; Time 3: M = 5.00, SD = 1.13, t(542) = −4.46, p < 0.001; Time 4: M = 4.95, SD = 1.21, t(633) = −2.11, p < 0.05). Also, those who did not answer at all four measurement points reported lower school achievement (M = 7.66, SD = 0.90) than those who did (M = 8.01, SD = 0.79, t(227) = −4.20, p < 0.001). However, at Time 4, there was no selection effect for attained educational track between those who did not answer at all the measurement points (M = 1.37, SD = 0.49) and those who did (M = 1.33, SD = 0.47, t(599) = 1.08, p = 0.280). By using the missing data procedure (for details see the description below of the analytical strategy used), we were able to supply data for all the participants in the analyses.

Measures

Life Satisfaction

Life satisfaction was assessed at all four measurement points with the five-item Satisfaction with Life Scale (Diener et al. 1985). The items (e.g., “I am satisfied with my life”, “In most ways my life is close to my ideal”, “The conditions of my life are excellent.”, “So far I have gotten the important things I want in life”, “If I could live my life over, I would change almost nothing”) were rated on a 7-point Likert-type scale ranging from 1 (I totally disagree) to 7 (I totally agree). A sum score was calculated from all five items. Cronbach’s alpha reliabilities for all the measures were .88, .90, .89 and .89, respectively. The results showed that variables measuring continuous life satisfaction were skewed and consequently the parameters of the analyses were estimated using the MLR estimator (Muthén and Muthén 1998–2007). However, no ceiling effect for life satisfaction was found. Life satisfaction had good concurrent validity and it correlated significantly and positively with self-esteem r = .60, p < .001 (Rosenberg 1965) and negatively with depressive symptoms r = −.52, p < .001 (Salokangas et al. 1995).

Gender

Gender was measured by asking the adolescents to circle their gender (1 = girl, 2 = boy). This variable was dummy-coded further so that 1 indicated a girl and 0 indicated a boy.

Academic Achievement

Academic achievement was measured at Time 1 by asking the participants to report their Grade Point Average (GPA) from the preceding spring term (i.e., Time 0). GPA ranged from 4 (lowest) to 10 (highest). Self-reported GPA has shown a correlation of .96 with actual GPA (Holopainen and Savolainen 2005).

Educational Track

The participants’ educational track after comprehensive school (Time 4) was measured by the following questions: (1) “Are you in education at the moment?” (1 = yes, 0 = no); (2) “If you are in education, what is the name of your school?” Next, an educational track variable was created by contrasting academic track with vocational track. Both adolescents who were in vocational school and those who were in vocational school and also attending senior secondary school courses were considered vocational school students. Adolescents in senior secondary school (academic track; n = 426) were coded 1 and adolescents in vocational school (vocational track; n = 232) were coded 0. The number of participants who had dropped out of the educational system 1 year after comprehensive school education was 21 (3%). They were also included in the data.

School Satisfaction

Adolescents’ school satisfaction after the transition to post-comprehensive schooling was measured by the Satisfaction with Educational Track scale (Nurmi et al. 2003). The scale consisted of four questions (1 = “How satisfied are you with your educational track?” 2 = “How interested are you in the subjects you are being taught?” 3 = “Do you enjoy going to school?” 4 = “Do you feel that your choice of post-comprehensive education was successful?”) all of which were rated on a 5-point scale (1 = not at all; 5 = very much). A mean sum score measuring school satisfaction was calculated from all four items. The Cronbach’s alpha for the scale was .87.

Analytic Strategy

The primary aim of the present study was to examine what developmental trajectories adolescents show in their life satisfaction during the transition to post-compulsatory education. Moreover, we examined whether gender and academic achievement would be related to membership in the latent trajectories of life satisfaction and whether belonging to certain latent trajectories would be related to school satisfaction and being on a specific educational track later on. These research questions were examined by using growth mixture modeling (GMM; Muthén 2001, 2004; Muthén and Muthén 2000). GMM is an excellent tool for examining heterogeneity over time by identifying homogeneous subgroups (i.e., latent classes) of participants that differ with respect to their developmental trajectories.

The GMM analyses were carried out in three stages: (1) a Latent Growth Model (LGM) was first calculated for the whole sample. (2) GMM analyses for different latent groups were carried out and their fit indices and class frequencies were compared. The solution that best fitted the data based on the indicators (see indicators below) and was also deemed reasonable from the standpoint of the interpretation was chosen as the final growth mixture model. Next (3) gender and GPA and (4) educational track and school satisfaction were added into the final GMM model. The growth mixture analyses aimed at determining whether distinct latent trajectory classes (groups of homogeneous subjects) could be identified on the basis of the different growth factor means (i.e., means of the intercept and linear slope). The overall latent growth model was used as a basic model for searching for the optimal number of latent groups.

The analyses were performed by using the Mplus statistical package (Version 5, Muthén and Muthén 1998–2007), assuming the standard MAR approach (missing at random) to missingness, which is not as strict as CMAR approach. This missing data method uses all of the available data in order to estimate the model without imputing data. By using the missing data method, we were able to utilize all the available observations in the data set when estimating the parameters of the models. Because the variables measuring continuous life satisfaction were skewed, the parameters of the LGM and GGM analyses were estimated using the MLR estimator (Muthén and Muthén 1998–2007). The goodness-of-fit of the estimated latent growth models was evaluated according to the following four indicators: (a) chi-square test, (b) Comparative Fit Index (CFI), (c) Root Mean Square Error of Approximation (RMSEA), and (d) Standardized Root Mean Square Residual (SRMR). In turn, we used the following four indices to select the number of latent trajectory classes in the growth mixture models: (a) the Bayesian information criterion (BIC), the adjusted Bayesian information criterion (aBIC), the Akaike’s information criterion (AIC), (b) the Lo-Mendell Rubin adjusted likelihood ratio test (aLRT), and Vuong-Mendel-Mendell-Rubin Likelihood ratio test, (c) entropy values (entropy values range from 0 to 1, values close to 1 indicating a clear classification), and (d) the practical usefulness, theoretical justification, and interpretability of the latent groups solution (see also Bauer and Curran 2003; Marsh et al. 2009; Muthén 2003).

Results

Latent Growth Model of Life Satisfaction for the Whole Sample

The sample correlation matrix, means, and standard deviations of the observed continuous variables are shown in Table 1. The correlations between the variables measuring life satisfaction range from .48 to .70.
Table 1

Correlations and descriptive statistics for the observed variables

 

1.

2.

3.

4.

5.

6.

7.

1. Life satisfaction T1

.48**

.49**

.54**

.23**

−.24**

.13

2. Life satisfaction T2

.70**

.50**

.51**

.19**

−.05

.12

3. Life satisfaction T3

.52**

.56**

.63**

.23**

−.13*

.26**

4. Life satisfaction T4

.57**

.52**

.62**

.23**

−.13*

.30**

5. GPA T1

.33**

.30**

.30**

.30**

−.69**

.15*

6. Educational track T4

−.10

−.16**

−.17**

−.19**

−.60**

.11

7. School satisfaction T4

.23**

.18**

.32**

.37**

.13

.05

Girls

     

 M

4.41

4.59

4.73

4.75

8.04

1.30

3.60

 SD

1.39

1.34

1.31

1.30

.81

.46

1.30

Boys

 M

4.77

5.03

4.91

5.01

7.81

1.41

3.38

 SD

1.34

1.25

1.22

1.16

.84

.49

1.31

Whole sample

 Range

1–7

1–7

1–7

1–7

4–10

1–2

0–6

Correlations for girls are below the diagonal and for boys above the diagonal

T1 = Time 1; T2 = Time 2; T3 = Time 3; T4 = Time 4

0 = vocational track, 1 = academic track

p < 0.05; ** p < 0.01

The results of the LGM for life satisfaction in which only statistically significant parameters were included showed that the two growth components (i.e., level and linear change) described the shape of the change well (χ2(854, 5) = 20.97, p < 0.001; CFI = .97; RMSEA = .06; SRMR = .06). The results at the mean level showed that the level of life satisfaction was significant, indicating that the level of life satisfaction was significantly different from zero (4.640(0.046), p < .001). The linear change of life satisfaction was also statistically significant (0.046(0.010), p < .001), indicating that life satisfaction increased during the transition. The results for the variances of the growth components showed significant individual variation in life satisfaction both in level (1.127(0.100), p < .001) and in linear change (0.021(0.006), p < .001), indicating that there were significant individual differences in these two growth components. Moreover, the covariance between the intercept and linear trend was negative (r = −33, p < .01), suggesting that the lower the initial level of life satisfaction, the greater the increase in it across the transition. However, it is possible that there is regression towards the mean.

Growth Mixture Models for Life Satisfaction

The main aim of our study was to examine what kinds of developmental trajectories adolescents show in life satisfaction. To examine this, Growth Mixture Models (GMM) were applied to the longitudinal data on life satisfaction. The mean structure of the latent growth factors described previously was used to search for the optimal number of latent groups (covariance structure was held equal across the latent groups).

Table 2 shows the fit indices and class frequencies for the compared growth mixture solutions. Because the three-group solution was theoretically meaningful and the goodness-of fit indices suggested that the third latent group was needed, the three-latent-group solution was selected as the best model. Table 2 shows the results for the other solutions as well.
Table 2

Fit indices and class frequencies for growth mixture models for life satisfaction with different numbers of latent trajectory groups

No. of groups

BIC

aBIC

AIC

Entropy

aLRT, p

VLMR, p

No. of est. param.

1

10,071.97

10,052.92

10,045.23

   

6

2 (n1 = 85%, n2 = 15%)

8,084.93

8,053.17

8,037.45

.67

44.45, p < .001

4.16, p < .001

10

3 (n1 = 16%, n2 = 66%, n3 = 18%)

8,065.47

8,024.19

8,003.76

.68

37.83, p < 0.01

7.63, p < .001

13

4 (n1 = 11%, n2 = 12%, n3 = 12%, n4 = 65%)

8,073.77

8,022.96

7,997.81

.66

11.38, p < .08

0.38, p < .07

16

BIC bayesian information criterion; aLRT lo-mendel rubin adjusted likelihood ratio test; VLMR vuong-lo-mendell-rubin likelihood ratio test

The results of the final growth mixture model for life satisfaction are presented in Table 3. Figure 1 displays the estimated growth curves for the different latent trajectories of life satisfaction. The results (Table 3) showed that the first latent group (66%) was characterized by a very high initial level of life satisfaction, which stayed stable over time. The results for the adolescents in the second group (18%) showed that the initial level of life satisfaction was high, but decreased over time. The third latent group (16%) was characterized by a low initial level of life satisfaction, which increased over time. The latent groups were labeled as follows: high-stable, high-decreasing and low-increasing. These results showed that most—two-thirds—of the adolescents belonged to the stable trajectory of life satisfaction, while the minority belonged to a trajectory characterized by change.
Table 3

Estimation results of the final growth mixture model with three latent groups (unstandardized estimates; standard errors in parenthesis)

Growth components

High-stable (66%)

Low-increasing (18%)

High-decreasing (16%)

Mean structure

 Level

5.18 (0.09)a

2.91 (0.16)a

4.19 (0.26)a

 Linear change

0.04 (0.02)

0.62 (0.08)a

−0.31 (0.10)b

Covariance structure

 Variance of the level

0.54 (0.07)a

0.54 (0.07)a

0.54 (0.07)a

 Variance of the linear slope

0*

0*

0*

 Covariance of level and linear slope

0*

0*

0*

Residual variances

 Life satisfaction (T1)

0.50 (0.07)a

0.50 (0.07)a

0.50 (0.07)a

 Life satisfaction (T2)

0.75 (0.07)a

0.75 (0.07)a

0.75 (0.07)a

 Life satisfaction (T3)

0.70 (0.07)a

0.70 (0.07)a

0.70 (0.07)a

 Life satisfaction (T4)

0.44 (0.05)a

0.44 (0.05)a

0.44 (0.05)a

Variances are held equal across the different latent groups

* Fixed to zero

ap < .001; p < .01

https://static-content.springer.com/image/art%3A10.1007%2Fs10964-009-9464-2/MediaObjects/10964_2009_9464_Fig1_HTML.gif
Fig. 1

Estimated growth curves for different latent trajectories of life satisfaction across the transition to post-comprehensive education

The next aim was to examine whether gender and academic achievement would be related to membership in the latent trajectories. Consequently, gender and GPA were added to the previous model. When they were added to the model, the results remained very similar with respect to the size and interpretation of the latent groups. The results showed that adolescents with high grades (GPA) were more likely to belong to the high-stable than low-increasing or high-decreasing latent groups (see Table 4). Furthermore, a significant main effect of gender was found: compared to girls, boys were overrepresented in the high-stable latent group than in the low-increasing group, whereas girls were overrepresented in the low-increasing latent group compared to high-stable group.
Table 4

Estimated odds ratios for gender and GPA at Time 1, and education track and school satisfaction at Time 4 of latent trajectory group membership of life satisfaction

Comparison

Gender OR

GPA OR

Track OR

School satisfaction OR

High-decreasing

0.82

0.96

0.17

−2.75**

High-stable vs. low-increasinga

3.12**

3.28***

2.39**

1.57

High-stable vs. high-decreasinga

1.27

3.42***

2.31**

4.17***

*** p < .001; ** p < .01

aReference class in italic

The final aim was to examine whether trajectory membership would be related to educational track and school satisfaction later on. Consequently, educational track and school satisfaction were added to the previous model. When educational track and school satisfaction were added into the model, the results remained closely similar with respect to the size and interpretation of the latent groups. The results showed that those with high school satisfaction at Time 4 were more likely to belong to the high-stable latent group than high-decreasing group and less likely to belong to high-decreasing than low-increasing latent group life satisfaction trajectory group (see Table 4). Finally, the results showed that those who were on an academic track at Time 4 were more likely to belong to the high-stable latent group compared to high-decreasing or low-increasing latent groups (see Table 4).

Discussion

The present study indicated that the development of life satisfaction during the transition to post-compulsory education showed heterogeneity. Three trajectories of life satisfaction were identified: high-stable, high-decreasing, and low-increasing. Most adolescents experienced high and stable life satisfaction throughout the follow-up. However, change in life satisfaction during the transition was characteristic of two trajectory classes representing one-third of the adolescents. These results are significant as to our knowledge no previous study has examined heterogeneity in adolescents’ trajectories of life satisfaction. Most earlier studies have been cross-sectional (Gilman and Huebner 2003) and heterogeneity in life satisfaction has been wholly left out of accounts.

Developmental Trajectories of Life Satisfaction Across the Transition to Post-Compulsory Education

The main aim of the present study was to use a person-oriented approach and to examine what kinds of developmental trajectories in life satisfaction adolescents show during the transition to post-compulsory education. Growth Mixture Modeling yielded a three latent group solution for life satisfaction. In accordance with Hypothesis 1, the results showed that the majority of the adolescents (66%) were members of a latent trajectory group characterized by a very high and stable level of life satisfaction across the entire follow up. This supports the earlier studies showing that most adolescents view their overall lives positively (Gilman and Huebner 2003). However, our study adds to this by showing that among most adolescents this high level of life satisfaction remains stable during the years of middle adolescence.

While the majority of adolescents belonged to the latent trajectory group characterized by a high and stable level of life satisfaction, two other smaller latent trajectory groups were also identified in the sample. In line with our expectation (Hypothesis 1), a single growth trajectory would have oversimplified the heterogeneity of the change in adolescents’ life satisfaction during the transition to post-compulsory education. Indeed, both of the smaller latent trajectory groups showed, in line with our Hypothesis 2, change in life satisfaction during the school transition period. Another group of adolescents (18%) showed an initially low but increasing level of life satisfaction, and another small group of adolescents (16%) was characterized by a high initial but then decreasing level of life satisfaction. These results broaden our knowledge of the trajectories of life satisfaction during the transition to post-compulsory education by showing that both positive and negative development in life satisfaction occurs among some adolescents during this period. It was recently concluded by Lucas and Donnellan (2007) that, although life satisfaction is moderately stable over long periods of time, it also shows an appreciable degree of instability that might depend on contextual circumstances. It is likely that some of the adolescents’ difficulties in adapting to the transitions and changes that accompany adolescence are only temporary or limited to the stages of adolescence. However, it might be that over a longer time span life satisfaction may vary around an individual set point (Fujita and Diener 2005).

The stage-environment theory (Eccles 2004; Eccles and Midgley 1989) offers one explanation for this result: depending on the ability of the new school environment to meet the individual needs of the students, life satisfaction among young people may either improve or decline following a major educational transition (see also Ash and Huebner 2001). However, one has to bear in mind that, besides a major educational transition, adolescents face several other key transitions during this age period, such as neurobiological, social and autonomy transitions (Caspi 2002; Erikson 1968; Lefkowitz 2005; Masten et al. 1999; Schulenberg et al. 2003; Shanahan 2000; Shiner and Masten 2002) which might have contributed to the changes found in life satisfaction in this study.

The Role of Gender and Academic Achievements in the Membership of Latent Trajectories of Life Satisfaction

The next aim of the present study was to examine whether gender and academic achievement would be related to belonging to specific latent trajectories of life satisfaction. In support of Hypothesis 3, the results showed that boys and those with high levels of academic achievements were more likely to belong to the high-stable life satisfaction group than to the other latent trajectory groups. A happiness gap seems to exist between girls and boys: adolescence seems to be harder on females than males. Previous research has explained girls’ higher vulnerability to low well-being by changes in hormonal levels (Angold et al. 1999), entering puberty early (Ge et al. 2001), being more reactive to stress and negative life events (Crick and Zahn-Waxler 2003), having a higher tendency towards ruminative coping styles (Broderick 1998; Burwell and Shirk 2007; Nolen-Hoeksema 1994), and lower body satisfaction (Barker and Galambor 2003). Previous research has shown also that, compared to boys, girls experience more academic achievement-related stress (Murberg and Bru 2004) and school-related burnout (Salmela-Aro et al. 2008a, b). Consequently, our results support the earlier studies showing that boys rather than girls were more likely to be a member of the high-stable life satisfaction trajectory. Boys and girls seem also to differ in the ways in which they adjust their self-perceptions according to external feedback on their academic ability (Crosnoe et al. 2007). Girls feel particularly bad when they do not measure up to the standards set by their peers and family, whereas no special sensitivity to their more intimate contexts appears to exist among boys. Girls and boys differ in their attributions for failing a class (Crosnoe et al. 2007). The results further showed that adolescent girls were more likely to belong to the low-increasing latent group than to the high-stable latent group (cf. Hypothesis 3). One explanation for this result is that girls find new academic and social challenges positive after the transition to post-compulsory education (Cole et al. 1999). However, more research is needed on this topic.

Life Satisfaction Trajectories, Educational Track and Later School Satisfaction

The last aim of the present study was to examine whether belonging to certain latent trajectories would be related to a subsequent level of school satisfaction and educational track. The results showed that adolescents from the high-stable and low-increasing latent trajectory groups showed higher subsequent school satisfaction compared to adolescents from the high-decreasing group, thereby supporting Hypothesis 4. Moreover, adolescents who belonged to the high-stable trajectory of life satisfaction were more often on an academic track at Time 4. There are two possible explanations for these results. Firstly, a high level of life satisfaction may decrease stress and anxiety in achievement situations, leading to high effort, high academic performance, and, consequently, high school satisfaction (see also Bandura 1982; Rosenberg 1965). Secondly, life satisfaction can also increase social relations and, thus, increase a sense of school belonging. To conclude, our results support earlier findings according to which life satisfaction during adolescence is related to success in dealing with major school-related challenges (Gilman and Huebner 2005; Salmela-Aro and Tuominen-Soini in press).

Limitations

At least seven limitations should be taken into account in any effort to generalize the results of the present study. First, all the measures included in the present study were based on self-report measures, which are not always the most valid or reliable method of data collection (Shaffer 2002). The fact that the study was wholly based on adolescents’ self-reports means that the data are subject to common-method variance. This is a particular danger where all the data are self-report measures of satisfaction. Accordingly, it would have been important to have had register data on achievement as well as on the neurobiological and hormonal changes taking place during this age period. Second, we did not have information about possible natural oscillations in life satisfaction, that is, the tendency of adolescents to cycle. Consequently, the changes in life satisfaction that took place concurrently with the present educational transition may in part be explained by natural oscillations in life satisfaction. In addition, we did not have information about other transitions, such as puberty, which might have taken place during the age period of interest. Third, the educational trajectories under investigation spanned only the transition from comprehensive school to upper secondary or vocational education. Future studies are needed to examine the role of life satisfaction trajectories during other educational transitions and educational and occupational trajectories across a more extended period of time. Fourth, the attrition analyses showed that those who provided data at each measurement point had to some extent higher life satisfaction than those who did not. Fifth, the present study was carried out in Finland and thus one has to be cautious in generalizing the results to other school contexts. However, many European countries have a similar educational system, in which students attend comprehensive school and then go on to an academic or a vocational track. However, it is very likely that those who dropped out of the educational system are the same adolescents who dropped out of this longitudinal data collection, and this bias should be taken into account as a possible limitation of this study. Moreover, third variables, such as family problems, may affect both academic progress and life satisfaction. This should be taken into account in future studies. Finally, our results do not allow us to propose any causal explanations and thus more research is needed to examine possible reverse paths from school performance and satisfaction to life satisfaction.

Implications for Future Research and Practice

As most of the longitudinal research conducted among adolescents has focused on the negative side of adolescent development, there is evident need for longitudinal studies focusing on adolescents’ strengths and well-being (Rich 2003). In addition, there is need to take into account the different changes adolescents experience, such as hormonal changes and their impact on life satisfaction. Moreover, future studies should be conducted also during stable time periods to be able to examine if changes in trajectories take place in stable as well as transitional periods. For future studies, more intensive data also need to be collected, for example by means of diary methods, reveals in greater detail the development of adolescents’ life satisfaction during educational transitions in greater detail (Huebner et al. 2000). The results also have some practical implications. On the basis of our study, adolescence seems to be a period of possibility, during which changes in life satisfaction can take place. Early interventions should, in particular, target adolescents who exhibit decreasing life satisfaction in order to prevent the accumulation of problems.

Acknowledgments

This study forms part of a larger project under the title Finnish Educational Transitions (FinEdu) and was funded by grants from the Academy of Finland (121 0319) and the Jacobs Foundation.

Copyright information

© Springer Science+Business Media, LLC 2009