1 Introduction
Moving homes in childhood is a common, but potentially stressful life event (Humke & Schaefer, 1995; Raviv et al., 1990). In most cases, the decision to move to a new home is made by the parent(s). However, children in the family also have to deal with the move and adapt to a new house, a new neighborhood, build new friendships, and potentially find their way in a new school. Moving may hamper children’s feeling of security and stability (Giddens, 1991; Oishi & Talhelm, 2012; Tønnessen et al., 2016) and might disrupt social ties both within the family and with friends and peers outside the family (Coleman, 1988). As such, moving might be detrimental to child and adolescent development, but the perspective of children and adolescents in residential mobility is still underrepresented in the literature. Previous research, mostly using retrospective self-reported mobility data, has shown that residential mobility in childhood (before age 18) was linked to later negative behavioral and socio-emotional outcomes of children or adolescents (for a review, see Jelleyman & Spencer, 2008), such as school drop-out (Haelermans & De Witte, 2015; Tønnessen et al., 2016), risk behavior (Haynie & South, 2005; Wolff et al., 2017) or mental health problems (Mok et al., 2016; Tseliou et al., 2016). A recent meta-analysis (Simsek et al., 2021) showed that residential mobility was overall small to moderately associated with mental health problems, but these relations were stronger for moving in adolescence compared to childhood and for frequent moves (three or more moves) compared to single moves. This highlights that “mobility is not simply an event with specific outcomes, but a set of social and psychological experiences that together result in successful or unsuccessful adjustment to a new environment” (Scanlon & Devine, 2001, p. 122). Thus, the effects of residential mobility may depend on the circumstances of the move, in particular the timing, frequency, or distance of the move(s). In the existing research on residential childhood mobility, the distinction between moves in childhood and adolescence is limited, moving frequency is not always taken into account, and the role of moving distance is still largely understudied (only 3 out of 64 studies in Simsek and colleagues’ (2021) meta-analysis measured moving distance). Moreover, it is known that residential mobility is linked to worse behavioral or educational outcomes (Jelleyman & Spencer, 2008), but it is not known whether mobile adolescents are also less satisfied with their lives. It is relevant to understand how and to what extent moving in adolescence affects life satisfaction in the transition to adulthood, because life satisfaction in adolescence plays an important role in positive development into adulthood (Hawkins et al., 2009).
The current study adds to the existing literature in multiple ways. First, this study incorporates different dimensions of residential mobility: We focus on the timing (i.e., moves in adolescence age 11–17), frequency (i.e., single move versus two or more moves), and the distance of the move (i.e., within the same postal code area, short-distance move < 50 km, or long-distance move > 50 km). Second, we analyze the relation between moving and an often overlooked but key indicator of well-being, general life satisfaction in young adulthood (age 19–22). To examine the relation between residential mobility in adolescence and life satisfaction in young adulthood, a longitudinal German adolescent sample (Children of Immigrant Longitudinal study; CILS4EU-DE v.6.0.0; Kalter et al., 2021) with detailed residential mobility data was used.
Advertisement
2 Background
2.1 Moving in Adolescence and Well-Being
Developmental literature highlights that stressful events can be more detrimental to well-being when they happen in key developmental stages, such as in adolescence (Fowler et al., 2014). Adolescence is a turbulent period in which physical, socio-emotional, and autonomy development take place (Steinberg & Silk, 2002). In such a period of change, one’s home situation, friendships, neighborhood, and school can be stable factors for the adolescent to rely on, an “arena of comfort” (Simmons et al., 1987, p. 1231). Residential mobility may undermine this stability and affect personal development and as such may result in lower well-being (Tønnessen et al., 2016). Additionally, moving is likely to disrupt social ties outside the family, such as friendships or relations with extended family (Coleman, 1988; Rhodes, 2018). In adolescence, peer networks outside the family become increasingly important for socialization, identity formation, and well-being (Brown & Larson, 2009). Given the paramount position of friendships in adolescence, losing these social ties due to a move can be especially detrimental to adolescents’ well-being. In addition, residential may to bring stress to the parents (Boddapati, 2017), that might spill over to their adolescent children, in that way affecting adolescent well-being (Kaufman et al., 2020). Finally, residential mobility may co-occur with other stressful life events, such as divorce (Desmond & Perkins, 2015). The accumulation of stressful life events may affect well-being. In all, there are multiple ways in which residential mobility can, directly or indirectly, affect adolescent life satisfaction. The existing literature supports the idea that moving in adolescence (from age 11 to 17) is more strongly associated with well-being than moving earlier in childhood (Fowler et al., 2014; Li et al., 2019; Simsek et al., 2021). For instance, negative relations were found between adolescent residential mobility and school achievement (Tønnessen et al., 2016), problem behaviors, substance abuse (Brown et al., 2012; Mok et al., 2016; Webb et al., 2016), and delinquent behavior (Fowler et al., 2014; Webb et al., 2016).
Although mobile adolescents experience more mental health problems and problem behaviors, it is still largely unknown whether those who moved in adolescence are also less satisfied with their lives when entering adulthood. Different from objective or affective measures of socio-emotional and behavioral problems (Headey et al., 1993) studied in previous research, life satisfaction is a cognitive, subjective measure of well-being (Dew & Huebner, 1994). As such, it “centers on the person’s own judgements, not upon some criterion which is judged to be important by the researcher” (Diener et al., 1985). Life satisfaction is not a static state but develops over time, especially in the transition into adulthood (Henkens et al., 2022). It can be influenced by (stressful) life events (Fujita & Diener, 2005; McKnight et al., 2002), such as a move, which might put mobile adolescents in a disadvantaged position compared to those who do not move in adolescence. Low life satisfaction is associated with adverse outcomes in multiple domains, such as in school and health (Huebner et al., 2006; Proctor et al., 2008). In all, life satisfaction is important for positive development from adolescence into adulthood (Hawkins et al., 2009). Studies that specifically investigated the relation between adolescent residential mobility and general life satisfaction are lacking. Some studies did examine the relation between moves during the period of childhood and life satisfaction in either adolescence or adulthood. For example, in American samples, negative associations were found between recent childhood moves and frequency of childhood moves and life satisfaction at age 13–14 (Brown & Orthner, 1990) as well as between childhood moves and adult life satisfaction (Oishi & Shimmack, 2010). In all, we expect adolescent residential mobility to be negatively related to life satisfaction in young adulthood.
2.2 Moving Frequency
After a move, the family needs time to settle down, recover from the stress of the move, and build new social networks. When moving frequently, there is no time to rebalance sufficiently and stress may accumulate. In this way, moving repeatedly can result in more negative outcomes than a single move (Mollborn et al., 2018). The existing studies that distinguish between single and frequent moves consistently found that moving more often is negatively related to health and well-being (see Simsek et al., 2021 for a meta-analysis), especially when one move follows shortly after another (Bernard & Vidal, 2020; Vogel et al., 2017). In a longitudinal study, frequent moving between two waves in adolescence predicted a within-person increase in delinquent behaviors for adolescents with no delinquent behavior prior to the moves (Vogel et al., 2017). Studies also found links between frequent moving and lower psychological well-being (Anderson & Leventhal, 2016; Brown et al., 2012; Mok et al., 2016) or lower educational achievement (Haugan & Myhr, 2019; Scanlon & Devine, 2001) in adolescence. Therefore, we expect that those who moved twice or more times in adolescence have lower life satisfaction in young adulthood than those who experienced a single move or no moves.
2.3 Moving Distance
Starting from the social capital perspective we can expect that long-distance moves are more likely to be disruptive for social ties outside the family than short-distance moves. This may relate to the fact that long-distance moves often include school changes (Pribesh & Downey, 1999). Especially for adolescents, for whom social ties outside the family are particularly salient, long-distance moves could be more harmful than short-distance moves as it disrupts social ties in the neighborhood and school (Gillespie, 2013; Haynie et al., 2006). Empirical studies on the role of distance in the relation between residential childhood mobility and later well-being is still limited and inconclusive, depending on the study outcome, design, or sample. Studies conducted in the US found long-distance moves (i.e., crossing county borders) to be related to more delinquency (cross-sectionally: Gillespie, 2017) or behavioral problems (longitudinally: Gillespie, 2013). Other studies from the US and Australia found long-distance moves to be related to higher academic achievement (cross-sectionally: Tucker and colleagues, 1998; longitudinally: Vidal & Baxter, 2018) or to a decrease in delinquency for those with high initial levels of delinquency (Vogel et al., 2017). The latter result suggests that the change of context following a long-distance move can break with negative social ties and behaviors and thus have positive effects.
Advertisement
Based on this social capital perspective, we expect that long-distance moves are more disruptive and thus would be more strongly related to lower life satisfaction in young adulthood compared to moves over shorter distances.
2.4 Alternative Relations Between Residential Mobility and Well-Being
Although the majority of research studied the potential negative effects of residential mobility, (neo) classical migration theory suggests that migration is motivated by pull factors and that people move to improve their quality of life, for themselves or their children (Geist & McManus, 2008; Massey et al., 1993). Moves motivated by positive triggers such as improvements in parental employment, better housing, or neighborhood conditions may thus improve children’s living conditions. The upgrade in living conditions after the move may compensate for the immediate stress of the move and disruption of social capital, leading to no or even positive effects on well-being over time.
In addition, when studying the relation between residential mobility and life satisfaction, it is important to note that research often found pre-existing differences in family background between mobile and non-mobile families. These pre-existing differences, and not the move itself, may account for the differences in life satisfaction (Gasper et al., 2010; Porter & Vogel., 2014). Specifically, previous research found that immigrant families, families with lower SES, or single-parent families are more likely to move (Gasper et al., 2010; Fomby & Sennott, 2013; Kuyvenhoven et al., 20212021; Murphey et al., 2012; Vidal & Baxter, 2018). Children with these backgrounds are also found to report lower life satisfaction (Borraccino et al., 2018; Chappel et al., 2014). Supporting this, a longitudinal study comparing school achievement between movers and non-movers and within movers before and after the move, found that movers had lower achievement than non-movers, but achievement did not decrease within persons after a move (Vidal & Baxter, 2018). This points towards pre-existing differences between mobile and non-mobile adolescents as an explanation for the relations that are found, which highlights the importance of controlling for relevant socio-demographic factors.
2.5 Present Study
The present study contributes to the existing literature by studying moves in adolescence, with a special focus on the role of moving frequency and distance in the relation between adolescent residential mobility and life satisfaction in young adulthood. Furthermore, so far studies on mobility in childhood were mainly focused on the US and UK (Simsek et al., 2021). We study adolescent residential mobility in Germany, one of the most populous countries in Europe, where mobility is moderately low compared to other European countries (Bernard & Vidal, 2020) and the US. It is relevant to examine whether and how the relation between mobility and well-being exists in a context where moving is less normalized. To account for possible confounding factors (i.e., selection effects) the current study takes immigrant background, family SES, and family structure into account. In addition, although exploratory, we examine potential heterogeneity in the relation between adolescent moves and life satisfaction by interacting adolescent moves with gender, immigrant background, family SES, and family structure.
Based on developmental literature and a social capital perspective, we hypothesize that residential mobility in adolescence is related to lower life satisfaction in young adulthood. In addition, we expect frequent moves and long-distance moves to be stronger negatively related to life satisfaction than single or short-distance moves. We analyze unique longitudinal data from a national sample of German adolescents and young adults that allow for investigating the relation between the multidimensional aspects of moving in adolescence and life satisfaction in young adulthood.
3 Methods
3.1 Participants and Procedure
For this study, German data from wave 6 (February 2016 – March 2017) and wave 8 (March – December 2020) of Children of Immigrant Longitudinal Study (CILS4EU-DE v.6.0.0; Kalter et al., 2021) were used. This dataset contains detailed information about residential mobility in adolescence. Participants stem from a longitudinal panel that started in 2010. Using a school-based, stratified three-stage sample design, 100 schools were sampled from a comprehensive national list of all eligible schools. Schools with high proportions of immigrant students were oversampled. In each school, two classes of the 9th grade were randomly selected and all students within these classes were sampled. Data was collected in schools during the first three waves (2010–2013). After that, participants were followed with self-report questionnaires1 up to 2020, resulting in eight waves of data. A random refreshment sample (n = 3513) was added in wave 6 to account for panel attrition. The study sample was built starting from participants who completed the residential history calendar (RHC) in wave 8 (n = 3961). Observations of those younger than 19 or older than 22 at wave 6 (n = 21) or of those with missing data on the dependent variable (n = 4) were excluded. Since we were interested in moves taking place within the family (i.e., not the moves to an independent living situation or out-of-home placements), only participants living in their parental home at wave 6 were included. Germany’s average age of leaving the parental home is 24 years (Eurostat, 2022). Thus, those who had lived separately from their parents for at least 3 months before age 18 (n = 278) or lived independently at wave 6 (n = 649)2 were not included. Finally, we include both first-generation immigrant adolescents and children of immigrants, except for first-generation immigrants who arrived in Germany after age 17 (n = 11). The total sample consisted of 2,998 participants. These were young adults aged 19 to 22 in wave 6 (Mage = 20.18, SD = 0.84), 45.8% boys, 6.8% first-generation immigrants (40.2% Polish background), and 31.0% children of immigrants (36.2% Turkish background). Socio-economic status (ISEI-scores) ranged from 13.24 to 88.51 (Mses = 43.14, SD = 17.6). See Table 1 for subgroup sizes.
3.2 Measures
3.2.1 Residential History Calendar (RHC)
Participants’ residential mobility history was measured retrospectively with the RHC in wave 8 (2020). In this calendar, participants listed for all their residences the age of moving in, the age of moving out, and the postal code of the residence. Adolescent moving frequency was measured by counting all residential changes between ages 11 and 17. Two separate dummy variables were generated: single move (1 if the participant experienced one adolescent move, 0 otherwise) and frequent moves (1 if participant experienced two or more adolescent moves, 0 otherwise). Moving distance was calculated by the difference between two subsequent residences. Three dummy variables were generated that measured moves within the same postal code area, other short-distance (< 50 km) moves, and long-distance moves (> 50 km). These dummies were not mutually exclusive, which means that when participants experienced multiple moves over more than one distance, they were coded 1 on each relevant distance dummy. The number of moves per distance was dichotomized into ever experienced a move on that specific distance (1) or never experienced such a move (0). Finally, to control for the history of childhood moves, moves before age 11 were counted and used as a linear control variable.
3.2.2 Life Satisfaction
The dependent variable life satisfaction was assessed in wave 6 (2016) by the single-item Cantril ladder: “On a scale from 1 to 10, how satisfied are you with your life in general?” The Cantril ladder is repeatedly validated in samples of young people (Jovanović, 2016; Jovanović & Lazić, 2018; Levin & Currie, 2014). The overall mean life satisfaction was 7.8 with a standard deviation of 1.48.
3.2.3 Socio-Economic Status (SES)
Because the refreshment sample entered in wave 6, no complete data of control variables prior to wave 6 were available. Therefore, control variables measured in wave 6 were used. SES was measured using the standard international socio-economic index of current occupational status (or past occupational status if no occupation at the time of measurement; ISEI; for an explanation of this measure, see Ganzeboom et al., 1992). Information from both parents was used to calculate a mean score for family SES. In the ISEI, adolescents report on parental occupation. Engzell and Jonsson (2015) showed that adolescent reports on parental occupation do not differ substantially from that of parents themselves. In the regressions, SES was included as a continuous variable, with higher values indicating higher SES. For the frequency tables, SES was categorized into high (one standard deviation above the mean), middle, and low (one standard deviation below the mean). Since we used adolescent reports, there was no measure available for family income. Hence we used the father’s unemployment as an extra measure of family SES, measured by the question “Does your father currently have a job?” Answers were dichotomized into no (1) and yes (0).
Immigrant background was constructed based on the participant’s country of birth and country of birth of their parents (Dollmann et al., 2014). Participants born in Germany, with both parents born in Germany were coded as of non-immigrant origin (0). Participants born outside of Germany were considered first-generation immigrants (1). Participants born in Germany with at least one parent born outside Germany were considered children of immigrants (2).
Gender was coded as 0 (boy) and 1 (girl).
Family structure was measured at wave 6 with a household box. Adolescents indicated the household members they lived with at that moment as a cumulative measure of household instability over the entire childhood. Two dummy variables were generated: one dummy measured single-parent families in which those living with one parental figure were coded as 1 and 0 otherwise. Another dummy measured stepparent families in which those living with a stepparent were coded 1 and 0 otherwise. Participants living with both of their biological parents at wave 6 were the reference group with 0 on both dummy variables. In addition, a separate control variable captures whether parents were still alive (0) or whether one or both parents were deceased (1).
3.3 Data Analyses
Data were prepared and analyzed in Stata version 17.0 (StataCorp, 2021). From the total sample, 521 respondents (17%) received a short version of the questionnaire, resulting in missings on family structure, father’s employment, and parent deceased. For our main analyses, all missings on independent variables were imputed by multiple imputations with chained equations (10 imputations) in Stata. All analyses were also performed with the unimputed dataset as a robustness check.
The relation between adolescent residential mobility and life satisfaction was examined by ordinary least squares regression models. Separate analyses were performed for moving frequency (model 1a-c) and moving distance (model 2a-c). In model 1a, life satisfaction was predicted by a single move and two or more moves in adolescence. To account for possible selectivity into moving, model 1b included the individual control variables gender and age. In model 1c, the family background characteristics immigrant background, SES, fathers’ employment status, family structure, deceased parent, and moves in childhood (age 0–10) were included. When significant relations between moving and life satisfaction disappear after including family control variables, the relation is probably confounded by family characteristics. In model 2a, life satisfaction was predicted by moves within the postal code area, short-distance moves, and long-distance moves. Model 2b added the individual control variables and model 2c the family background characteristics. Finally, moving frequency and moving distance were interacted with gender, immigrant background, family structure, and SES in separate models to explore potential heterogeneity in the relation between adolescent residential mobility and life satisfaction. Due to the small subgroup sizes of adolescent movers, these analyses should be interpreted as exploratory. Effect sizes are shown as Cohen’s d, in which 0.2 indicates a small effect, 0.5 a medium effect, and 0.8 indicates a large effect (Cohen, 1988).
4 Results
4.1 Descriptive Statistics
Descriptive statistics of adolescent moves are presented in Table 1. The total number of moves experienced in adolescence ranged from 0 to 4 moves. Of the total sample, 18% moved once and 4% moved twice or more in adolescence. Adolescent moves were less common than childhood moves (0–10). In childhood, 28.4% moved once and 14.5% moved twice or more (not presented in Table 1). Bivariate multinomial logistic regressions examined whether moving behavior differed significantly between subgroups (see percentages with asterisks in Table 1)3. Moving behavior did not differ significantly between boys and girls. However, compared to non-immigrants, first-generation adolescents experienced more single and frequent moves. Children of immigrants (second generation) more often experienced a single move compared to non-immigrants. Adolescents from low-SES families experienced more often a single move than those from middle- or high-SES families. Adolescents living in single-parent or stepparent families reported significantly more moves (both single and frequent moves) than those living in two-parent families. Moving frequency did not differ between families with employed or unemployed fathers or between families in which both parents were alive or where one or both parents were deceased. Finally, the number of childhood moves (before age 11) positively predicted single and frequent adolescent moves (OR = 1.71, CI = [1.54–1.89]; OR = 2.22, CI = [1.88–2.62], respectively).
When looking at the moving distances more closely, we found that children of immigrants moved significantly less often over long distances compared to movers without immigrant backgrounds. Compared to adolescent movers from two-parent families, those from single-parent families moved more often short distance. Those from stepparent families moved less often long distance, but note that this was a very small group (only four adolescents).
Table 1
Descriptive statistics of the total sample and those who moved in adolescence
% and (n) of total sample (n = 2998) | % and (n) of movers only (n = 666) | |||||||
|---|---|---|---|---|---|---|---|---|
Never moved | Single move | Two or more moves | Total | Move within postal code | Short-distance move | Long-distance move | Missing data on distance | |
Total | 77.8 | 18.2 | 4.1 | 100 | 42.0 | 46.4 | 8.9 | 9.3 |
(2332) | (544) | (122) | (2998) | (280) | (309) | (59) | (62) | |
Gender | ||||||||
Boy (ref) | 79.4 | 16.9 | 3.7 | 100 | 42.1 | 45.2 | 9.2 | 8.8 |
(1089) | (232) | (51) | (1372) | (119) | (128) | (26) | (25) | |
Girl | 76.5 | 19.2 | 4.4 | 100 | 42.0 | 47.3 | 8.6 | 9.7 |
(1243) | (312) | (71) | (1626) | (161) | (181) | (33) | (37) | |
Immigrant background | ||||||||
Non-migrant (ref) | 82.4 | 13.9 | 3.7 | 100 | 44.6 | 48.0 | 10.8 | 6.2 |
(1515) | (255) | (68) | (1838) | (144) | (155) | (35) | (20) | |
First generation | 54.4 | 35.8* | 9.8* | 100 | 33.3 | 36.6 | 9.7 | 25.8 |
(111) | (73) | (20) | (204) | (31) | (34) | (9) | (24) | |
Children of immigrants | 73.7 | 22.8* | 3.6 | 100 | 42.5 | 47.4 | 6.1* | 7.4 |
(685) | (212) | (33) | (930) | (104) | (116) | (15) | (18) | |
Missing | 80.8 | 15.4 | 3.9 | 100 | 20.0 | 80.0 | 0.0 | 0.0 |
(21) | (4) | (1) | (26) | (1) | (4) | (0) | (0) | |
SES | ||||||||
SES low (ref) | 73.1 | 22.4 | 4.5 | 100 | 46.0 | 42.0 | 8.6 | 9.2 |
(473) | (145) | (29) | (647) | (80) | (73) | (15) | (16) | |
SES middle | 79.1 | 16.8* | 4.1 | 100 | 42.7 | 47.7 | 9.1 | 8.3 |
(1454) | (309) | (75) | (1838) | (164) | (183) | (35) | (32) | |
SES high | 79.0 | 17.5* | 3.5 | 100 | 33.3 | 49.7 | 8.3 | 13.0 |
(405) | (90) | (18) | (513) | (36) | (53) | (9) | (14) | |
Family structure | ||||||||
Two-parent family (ref) | 85.8 | 12.6 | 1.6 | 100 | 40.9 | 41.4 | 10.1 | 9.7 |
(1432) | (211) | (26) | (1669) | (97) | (98) | (24) | (23) | |
Single-parent family | 67.4 | 24.7* | 7.9* | 100 | 40.7 | 52.3* | 8.0 | 10.0 |
(412) | (151) | (48) | (611) | (81) | (104) | (16) | (20) | |
Stepparent family | 48.6 | 40.8* | 10.6* | 100 | 43.5 | 51.1 | 4.4* | 10.9 |
(87) | (73) | (19) | (179) | (40) | (47) | (4) | (10) | |
Missing | 74.4 | 20.2 | 5.4 | 100 | 44.9 | 43.8 | 10.9 | 6.5 |
(401) | (109) | (29) | (539) | (62) | (60) | (15) | (9) | |
Fathers employment | ||||||||
Employed (ref) | 79.6 | 16.9 | 3.5 | 100 | 42.5 | 46.9 | 7.7 | 9.6 |
(1623) | (344) | (72) | (2039) | (177) | (195) | (32) | (40) | |
Unemployed | 76.5 | 19.1 | 4.5 | 100 | 40.5 | 45.6 | 8.9 | 12.7 |
(257) | (64) | (15) | (336) | (32) | (36) | (7) | (10) | |
Missing | 72.6 | 21.8 | 5.6 | 100 | 41.5 | 45.6 | 11.7 | 7.0 |
(452) | (136) | (35) | (623) | (71) | (78) | (20) | (12) | |
Parent | ||||||||
Parents alive (ref) | 79.6 | 16.8 | 3.6 | 100 | 42.1 | 46.4 | 7.9 | 10.2 |
(1831) | (387) | (83) | (2301) | (198) | (218) | (37) | (48) | |
Parent deceased | 67.7 | 27.9* | 4.4 | 100 | 40.9 | 50.0 | 9.1 | 9.1 |
(46) | (19) | (3) | (68) | (9) | (11) | (2) | (2) | |
Missing | 72.3 | 21.9 | 5.7 | 100 | 42.0 | 46.0 | 11.5 | 6.9 |
(455) | (138) | (36) | (229) | (73) | (80) | (20) | (12) | |
4.2 Life Satisfaction and Adolescent Mobility
4.2.1 What type of moves?
In Model 14, the relation between moving frequency in adolescence and life satisfaction in young adulthood was examined (Table 2). Results showed that young adults who experienced only one move in adolescence had comparable levels of life satisfaction as those who never moved. However, those who moved twice or more in adolescence had lower life satisfaction in young adulthood compared to those who did not move in adolescence (Cohen’s d = 0.36). Controlling for individual characteristics in Model 1b hardly changed this effect size (Cohen’s d = 0.35). Adding family background characteristics in Model 1c decreased the effect of frequent moves, but it remained significant (Cohen’s d = 0.24). This substantial decrease in effect size means that part of the relation could be explained by family background characteristics. However, since a small but significant effect of frequent moving remains, it means that frequent moving in adolescence was related to lower life satisfaction above and beyond family background characteristics.
Model 2 examined the relation between ever experiencing at least one adolescent move within the same postal code area, over a short distance (< 50 km), or over a long distance (> 50 km) (compared to those who have never moved that specific distance) and life satisfaction (Table 3)5. Moves within the same postal code area and short-distance moves were weakly but negatively related to life satisfaction in Model 2a (Cohen’s d = 0.14 and 0.18, respectively). Controlling for individual characteristics did not change the effects of moving within the same postal code area and over a short distance (Cohen’s d = 0.14 and 0.18, respectively). The already weak negative relations were no longer significant after controlling for family background characteristics in Model 2c. There was no significant relation between the experience of a long-distance move and life satisfaction. Note that in this model, the moving distance was not interacted with frequency due to the small number of frequent movers.
4.2.2 For Whom?
Interactions between moving frequency in adolescence and immigrant background were tested to examine whether relations differed between those with and without immigrant background (Table 4). The interaction between single moves and children of immigrants was positive and significant (b = 0.35, SE = 0.15, p = .020) and shows that the effect of a single move was weaker for children of immigrants. Indeed, the main effect of a single move became significant, meaning that a single move was weakly related to lower life satisfaction for non-immigrants (Cohen’s d = 0.14), but not for children of immigrants. The interaction between frequent moves and children of immigrants was negative and significant (b = -0.65, SE = 0.32, p = .041) and shows that the negative effect of frequent moving was substantially stronger for children of immigrants than for non-immigrants (Cohen’s d = 0.44 for children of immigrants versus 0.13 for non-immigrants). This suggests that moving frequently in adolescence could be more harmful to children of immigrants than to non-immigrants. SES, family structure, gender, or childhood moves did not interact significantly with moving frequency. Finally, no significant interactions between moving distances and immigrant background, SES, family structure, gender, or childhood moves were found. Again, it should be noted that these interactions were exploratory due to the small numbers of frequent adolescent movers (see Table 1) and should be interpreted as such.
Table 2
Life satisfaction in young adulthood predicted by single and frequent moves in adolescence
Model 1a | Model 1b | Model 1c | |
|---|---|---|---|
Single move | − 0.133 (0.058) | − 0.128 (0.068) | − 0.033 (0.649) |
Frequent moves | − 0.536* (0.000) | − 0.524* (0.000) | − 0.406* (0.004) |
Girl | − 0.093 (0.086) | − 0.075 (0.161) | |
Age | − 0.110* (0.001) | − 0.105* (0.001) | |
First generation | 0.054 (0.630) | ||
Children of immigrants | − 0.158* (0.014) | ||
SES | 0.004* (0.008) | ||
Unemployed father | − 0.347* (0.000) | ||
Single-parent family | − 0.198* (0.005) | ||
Stepparent family | − 0.104 (0.376) | ||
Deceased parent | − 0.324 (0.136) | ||
Childhood moves between 0–10 | − 0.066* (0.042) | ||
Constant | 7.839* (0.000) | 10.113* (0.000) | 9.970* (0.000) |
N | 2998 | 2998 | 2998 |
Table 3
Life satisfaction in young adulthood predicted by moves at different distances in adolescence
Model 2a | Model 2b | Model 2c | |
|---|---|---|---|
Move within postal code area | − 0.212* (0.023) | − 0.205* (0.028) | − 0.109 (0.248) |
Move short distance (< 50) | − 0.265* (0.003) | − 0.262* (0.003) | − 0.160 (0.079) |
Move long distance (> 50) | − 0.150 (0.442) | − 0.137 (0.481) | − 0.071 (0.714) |
Girl | − 0.097 (0.078) | − 0.079 (0.146) | |
Age | − 0.115* (0.000) | − 0.110* (0.001) | |
First generation | 0.034 (0.778) | ||
Children of immigrants | − 0.161* (0.013) | ||
SES | 0.004* (0.021) | ||
Unemployed father | − 0.374* (0.000) | ||
Single-parent family | − 0.201* (0.005) | ||
Stepparent family | − 0.093 (0.442) | ||
Deceased parent | − 0.336 (0.117) | ||
Childhood moves between 0–10 | − 0.069* (0.037) | ||
Constant | 7.844* (0.000) | 10.227* (0.000) | 10.118* (0.000) |
N | 2936 | 2936 | 2936 |
Table 4
Life satisfaction in young adulthood predicted by single and frequent moves in adolescence moderated by immigrant background, SES, single-parent family, gender and childhood moves
Moderation variables | |||||
|---|---|---|---|---|---|
Immigrant background | SES | Single-parent family | Gender | Childhood moves | |
Single move | − 0.211* (0.037) | 0.202 (0.256) | − 0.098 (0.247) | − 0.039 (0.716) | 0.081 (0.406) |
Frequent moves | − 0.195 (0.290) | − 0.611 (0.097) | − 0.200 (0.259) | − 0.360 (0.091) | − 0.365 (0.062) |
Girl | − 0.080 (0.137) | − 0.074 (0.169) | − 0.075 (0.166) | − 0.074 (0.223) | − 0.076 (0.158) |
Age | − 0.102* (0.002) | − 0.104* (0.001) | − 0.106* (0.001) | − 0.105* (0.001) | − 0.104* (0.001) |
First generation | − 0.032 (0.827) | 0.054 (0.630) | 0.057 (0.613) | 0.053 (0.640) | 0.047 (0.677) |
Children of immigrants | − 0.199* (0.006) | − 0.160* (0.013) | − 0.158* (0.014) | − 0.159* (0.013) | − 0.164* (0.010) |
SES | 0.005* (0.007) | 0.005* (0.005) | 0.004* (0.008) | 0.004* (0.008) | 0.004* (0.008) |
Unemployed father | − 0.344* (0.000) | − 0.347* (0.000) | − 0.342* (0.000) | − 0.346* (0.000) | − 0.345* (0.000) |
Single-parent family | − 0.193* (0.007) | − 0.194* (0.006) | − 0.222* (0.008) | − 0.198* (0.005) | − 0.199* (0.005) |
Stepparent family | − 0.096 (0.412) | − 0.104 (0.375) | − 0.103 (0.382) | − 0.103 (0.379) | − 0.109 (0.355) |
Deceased parent | − 0.333 (0.128) | − 0.328 (0.134) | − 0.328 (0.130) | − 0.324 (0.137) | − 0.325 (0.135) |
Childhood moves between 0–10 | − 0.063 (0.052) | − 0.065* (0.046) | − 0.067* (0.037) | − 0.066* (0.043) | − 0.028 (0.477) |
Single move * First generation | 0.362 (0.131) | ||||
Single move * Children of immigrants | 0.353* (0.020) | ||||
Frequent move * First generation | − 0.182 (0.651) | ||||
Frequent move * Children of immigrants | − 0.647* (0.041) | ||||
Single move * SES | − 0.006 (0.147) | ||||
Frequent move * SES | 0.005 (0.545) | ||||
Single move * Single-parent family | 0.242 (0.134) | ||||
Frequent move * Single-parent family | − 0.507 (0.074) | ||||
Single move * Girl | 0.011 (0.939) | ||||
Frequent move * Girl | − 0.080 (0.773) | ||||
Single move * Childhood moves | − 0.136 (0.074) | ||||
Frequent move * Childhood moves | − 0.053 (0.642) | ||||
Constant | 9.931* (0.000) | 9.906* (0.000) | 10.009* (0.000) | 9.973* (0.000) | 9.945* (0.000) |
N | 2998 | 2998 | 2998 | 2998 | 2998 |
5 Discussion
Moving in adolescence could be a stressful experience, adding to the developmental turbulence in this life phase and potentially disturbing the peer networks and friendships that are particularly salient in adolescence. This study examined the relation between adolescent residential mobility and general life satisfaction in young adulthood, taking the multidimensional nature of moving frequency and distance into account. Moreover, we explored whether there might be specific groups of adolescents particularly at risk of lower life satisfaction after moving. We found that, after controlling for individual and family characteristics, it was not so much moving distance, but moving frequency that mattered: Those who moved twice or more in adolescence had lower life satisfaction in young adulthood compared to residentially stable adolescents, even after controlling for family background characteristics. Those who moved within the same postal code area or short distance had lower life satisfaction compared to non-movers or long-distance movers. However, this relation was explained by family SES and family structure. Exploratory analyses pointed towards differences between non-immigrants and children of immigrants: A single move was related to lower life satisfaction only for non-immigrants. Children of immigrants seem a vulnerable group regarding frequent moving, for whom moving frequently in adolescence was even more strongly related to lower life satisfaction in young adulthood than for non-immigrants.
5.1 Frequency Matters
In line with our expectations, young adults who moved twice or more in adolescence reported substantially lower life satisfaction than those who did not move or experienced only a single move in adolescence. This means that frequent moving adolescents are not only more likely to have worse school outcomes, mental health, or physical health (for a meta-analysis, see Simsek et al., 2021; Vidal & Baxter, 2018), they are also more likely to be less satisfied with their lives in general. This is an important insight, since low life satisfaction in adolescence is related to disadvantaged outcomes in later life (Proctor et al., 2008). By moving twice or more often in a period of 7 years (age 11 to 17), adolescents might not have the time to fully recover from the stress of a move before they move again (Bernard & Vidal, 2020; Vogel et al., 2017). A first possible mechanism that may explain this finding comes from the ontological security theory, which argues that people need a sense of stability, security, and continuity for psychological well-being (Giddens, 1991). From this perspective, housing stability, and the ‘sense of home’ can be essential for a person’s satisfaction with life. Frequent moving in adolescence can disrupt this feeling of security and stability, which may lower life satisfaction. Second, from a social capital perspective (Coleman, 1988), frequent moving disrupts social ties time after time (if the adolescent had time to (re-)build a social network in the first place). Social network studies highlighted that mobile adolescents tend to have smaller networks (South & Haynie, 2004), lower-quality friendships, and less social status than non-mobile adolescents (Haynie & South, 2005). Moreover, mobile adolescents were more likely to be involved in deviant peer networks than non-mobile adolescents (Haynie et al., 2006). Given the crucial role of peers and friends in adolescence, the disruption or lack of social ties could negatively affect life satisfaction. Third, frequent moving might increase stress levels in parents (Boddapati, 2017), that spill over to their adolescent children (Anderson et al., 2014; Kaufman et al., 2020). This means that when parents are distressed, either as a result of the move or because of reasons that might have led to the move (e.g., divorce, loss of employment), this stress can be picked up by their children, which can lead to lower life satisfaction. Finally, selectivity into frequent moving partly explained differences in later life satisfaction. That is, we found that first-generation adolescents and adolescents from single- or stepparent families were more likely to move frequently in adolescence than others, in line with recent studies in Europe (Kuyvenhoven et al., 2021; Tønnessen et al., 2016). These adolescents experienced additional stressors on top of the residential mobility (i.e., the international migration for first-generation immigrants, changes in family structure for those in single-parent families) and are more likely to have lower SES (Kogan et al., 2011; McLanahan & Percheski, 2008), which could explain their lower levels of life satisfaction in young adulthood.
Note that Germany is a context in which residential mobility is relatively limited compared to other countries (Bernard & Vidal, 2020). Moving behavior, such as moving frequency and moving distance, is context-dependent (Thomas et al., 2019). In our sample, only 4% of the adolescents moved twice or more between age 11 and 17. In their international comparative study based on retrospective data of older adults, Bernard and Vidal (2020) showed that German respondents moved less in childhood compared to other North European countries such as Denmark, Sweden, and The Netherlands. Moving is thus a rarer experience than in other countries. The impact for adolescents may be larger than in a context in which moving frequently is the norm. In the current study, we found a negative relation between adolescent moves and life satisfaction with two or more moves, but the threshold of ‘frequent moves’ could thus be different in more mobile countries, like the U.S. or Scandinavian countries. These results should therefore be interpreted within the German context and not be generalized into contexts with different moving patterns.
Besides the negative link between frequent adolescent moves and life satisfaction, it is important to emphasize that in general a single move in adolescence was not related to lower life satisfaction in young adulthood. We can thus conclude that although adolescents may experience some stress during or directly after the move, they may also be capable of adapting to their new lives rather rapidly.
5.2 Moving Distance
Contrary to our expectations, life satisfaction was not lower for those who moved over long distances. In line with previous research, we found that long-distance moves were more likely in high-SES and non-immigrant families. These long distance-moves seem thus to be selective and may relate to even further improvement of the family conditions. That is, long-distance moves are found to be more often pro-active moves, motivated by positive triggers such as employment, education, or improvement of quality of life (Geist & McManus, 2008; Morris et al., 2018; Thomas et al., 2019). Since we examined the relation between moves in adolescence and life satisfaction in young adulthood, it could be that these ‘upgrades’ in living conditions have compensated for the immediate loss of social capital. In addition, family SES and positive family relations may buffer the negative effects of long-distance moves in adolescence (Li et al., 2019; Riina et al., 2016).
From a social capital perspective, we expected that long-distance moves would disrupt social ties more than short-distance moves and therefore that especially long-distance moves would be related to lower life satisfaction (Gillespie, 2013). Our findings underline that adolescents may have a smaller geographical scope of life than adults have. That is, their lives mainly take place within a relatively small geographical area in which they go to school, do sports, and meet friends. This may imply that for adolescents, even a short-distance move (i.e., < 50 km) can disrupt their social capital, explaining why short-distance moves were negatively related to life satisfaction before controlling for family characteristics. In addition, Gillespie (2013) hypothesized that long-distance moves may be less related to behavioral problems and lower academic achievement for adolescents than for younger children, because adolescents may be better able to seek social support. Finally, adolescents can use online alternatives like social network sites and social media to stay in contact with their friends also after the move (Reich et al., 2012; Spies Shapiro & Margolin, 2014). These online ways of connection may cushion the physical distance between friends after a long-distance move and give students the time to build a new social life in their new environment.
5.3 Vulnerable Groups?
Although still exploratory, our results hint towards children of immigrants as a vulnerable group when it comes to frequent residential mobility in adolescence. In line with previous studies (Andersson, 2012; De Valk, 2010; Kuyvenhoven et al., 2021), we found that children of immigrants moved more frequently in adolescence than non-immigrants. In addition, children of immigrants who moved frequently (n = 33) had substantially lower life satisfaction than non-immigrant frequent movers. Immigrant families in general are found to have less social and financial capital (Kogan et al., 2011; Nesterko et al., 2013). This means they have fewer resources to rely on during or after the move, which might increase stress and strengthen the adverse effects of the move. Moreover, families with lower socio-economic status may have less control and are more likely to move in reaction to their situation (DeLuca & Jang-Trettien, 2020), which also implies that moves in these families are more often accompanied by other stressful life events (Dong et al., 2005). Thus, especially for immigrant families with lower socio-economic status, there could be an accumulation of childhood stress, resulting in lower life satisfaction in young adulthood. Due to the limited number of frequently moving children of immigrants in our sample, these interaction effects need to be studied in future work with larger samples before solid conclusions can be drawn.
5.4 Strengths and Limitations
The residential history calendar used in this study allowed us to examine the relation between moves in adolescence and life satisfaction in young adulthood. The detailed residential mobility data, with information about the frequency and distance of moves in adolescence, is a strength of this study. However, some limitations should be addressed. First, we studied the relation between moves in adolescence (age 11–17) and life satisfaction in young adulthood (19–22), but we could not control for life satisfaction before the move(s). This means that we cannot draw firm causal conclusions. In order to study within-person changes in life satisfaction after a move, longitudinal data with large samples, or linked population register data as in some of the Nordic European countries exist, are needed, which are still rare. To our knowledge, only a few studies, using indirect well-being outcomes, were able to study residential mobility effects in this way (Vogel et al., 2017 with delinquency and Vidal & Baxter, 2018 with educational achievement).
Second, to reduce the risk of finding spurious relationships, we did control for relevant family background characteristics. However, there were some limitations to the control variables. Residential mobility is likely to co-occur with divorce (Desmond & Perkins, 2015), a major life event that is also likely to affect adolescent life satisfaction (Brown, 2006). Unfortunately, we could not link data about family structure to the moves, which makes it impossible to tell whether changes in family structure co-occurred with a move. When stressful events co-occur in childhood, stress might accumulate and affect life satisfaction even stronger. However, because we controlled for family structure at the sixth wave, the moves that co-occurred with divorce were taken into account. In addition, due to the necessary refreshment sample that entered in wave 6, only the SES measures from wave 6 could be used. Thus, our SES measure was a snapshot, not capturing sudden changes in parental unemployment. Because parental job loss in adolescence can affect life satisfaction in early adulthood (Nikolova & Nikoleav, 2021), the role of SES in our analyses may be over- or underestimated. Finally, there are unobserved factors that could confound the relation between residential mobility and life satisfaction (e.g., remarriage of parents; Fomby & Sennott, 2013). More detailed information was not available in our data, but this should be studied in future work in order to understand which families and adolescents are most affected by moving.
Third, the sample was selective to young adults living with their parent(s). Those who were excluded because they had lived separately from their parents for at least three months before age 18 or lived independently at wave 6 had different characteristics. Additional analyses suggest that we excluded adolescents from non-normative living situations (e.g., problematic family situations, clinical cases, or adolescents who attend boarding schools). Future research should aim to include also these groups with data that cover the full population for example based on population registers.
5.5 Future Research
Several suggestions for future research arise from our study. First, some factors could confound the relation between residential mobility and life satisfaction that could be studied in future research. For example, remarriage of divorced parents, changes in parental employment status or income, and bullying experiences could all influence the mobility behavior of the family and adolescents’ life satisfaction. Second, to understand how and why residential mobility affects life satisfaction, future research should study mediating mechanisms. For example, moving can put pressure on parents, which can negatively affect their parenting (Boddapati, 2017) and in turn negatively affects adolescent well-being and development (Anderson et al., 2014). Third, future research could study the regions adolescents moved from and to. For instance, a move from a metropolitan city to a rural area might have very different effects than the reverse. Fourth, the reasons for moving were not the focus of our study, while it is likely that these also affect the moving experience and its consequences. Proactive moves, motivated by positive triggers and leading to an upgrade in living conditions could have very different outcomes than stressful reactive moves, motivated by triggers such as job loss, divorce, or death of a parent. Fifth, it would be interesting for future studies to compare our findings to the (international) moves among expat families. The available economic and social capital in highly mobile Europeans was found to compensate for the potential negative effects of moving (Favell, 2011), which could also apply to expat families. Finally, our exploratory moderation analyses, revealing children of immigrants as a potentially vulnerable group, should be replicated with larger immigrant samples. Related, we emphasize here that the group of adolescents with immigrant backgrounds is a heterogeneous group, hence future research should distinguish between immigrants from different origins and immigration backgrounds.
6 Conclusion
In conclusion, the novel contribution of this study lies in a detailed measure of residential mobility in adolescence and its relation to life satisfaction in young adulthood. Life satisfaction is a cognitive and evaluative measure of subjective well-being (Pavot & Diener, 2008). Although related, it is different from distress measures such as depression and anxiety (Headey et al., 1993) or objective measures (such as school achievement) that are studied in relation to adolescent residential mobility before. Our study showed that moving frequently in adolescence was associated with lower life satisfaction in young adulthood. Moving only once in adolescence was unrelated to later life satisfaction. Given the importance of life satisfaction for healthy development into adulthood (Hawkins et al., 2009), adolescents who moved frequently may be more vulnerable in their development into adulthood. Moreover, since frequent mobility is more common in disadvantaged and single-parent families (Pribesh & Downey, 1999), frequent moving can accentuate already existing social inequality. Further longitudinal analyses of mobility and life satisfaction in different contexts and for children of diverse origins are needed to understand who is at risk.
Declarations
Ethical Approval
CILS4EU Germany data were collected under the coordination of the University of Mannheim and falls under the German ethical approval guidelines. See Kalter, F., Kogan, I., & Dollmann, J. (2019). Studying integration from adolescence to early adulthood: Design, content, and research potential of the CILS4EU-DE data. European Sociological Review, 35(2), 280–297. Only reduced versions of the data were shared with researchers to make the identification of individuals impossible. The complete data can only be accessed at the Secure Data Center (SDC) of the GESIS Data Archive for the Social Sciences in Köln (Cologne).
Informed Consent
Informed active consent was collected in Germany, which means that signatures of parents were needed for adolescents to participate. All targeted adolescents and their parents were allowed to refuse participation.
Conflict of Interests
The authors have no competing interests to declare that are relevant to the content of this article.
Uniqueness of Publication
To our knowledge, no articles have been published using the elaborated Mobility History Calendar data as part of the CILS 4 EU dataset.
Authorship and Ordering
All authors agree to the content of the manuscript and to the order of authorship.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.