Effects of Family Structure and the Experience of Parental Separation: A Study on Adolescents’ Well-Being Effects of Family Structure and the Experience of Parental Separation: A Study on Adolescents’ Well-Being Sabine Walper, Carolin Thönnissen, Philipp Alt Abstract: Large numbers of studies, mostly from the U.S., have addressed the effects of parental separation and divorce, pointing to disadvantages of children and adolescents growing up in separated families. However, evidence on this topic varies across countries and is limited for Germany. Using longitudinal data from waves 1 and 3 of the German Family Panel pairfam, we investigated differences in adolescents’ well-being by comparing stable nuclear families (n = 1968), single mother families (n = 360), and stepfather families (n = 214), as well as an additional smaller group of adolescents whose parents separated between waves 1 and 3 (“prospective separators”; n = 76). Adolescents’ satisfaction with different domains of life (family, education/work, and their general life satisfaction) as well as their self-esteem were used as indicators of well-being. A series of multiple regression analyses tested the effects of family structure on well-being at T1 and changes in well-being over time, controlling for various background factors. Furthermore, likely mediation effects of infrequent contact to the non-resident father and economic strain were tested. The fi ndings show (relatively minor) effects of parental separa- tion, namely lower well-being among youth1 in single mother families compared to nuclear families. Disadvantages of youth in single mother families could only be partly explained by the higher fi nancial strain generally experienced in these fami- lies. Youth in stepfather families reported a similar overall well-being as adolescents in nuclear families, but indicated a greater decrease in family satisfaction over time. Pre-separation disadvantages among prospective separators were limited to great- er dissatisfaction with school. Infrequent contact with the non-resident father did not affect adolescents’ well-being. Effects of family structure did not differ between boys and girls, but maternal education moderated the effects of family structure on adolescents’ life satisfaction. Overall, the fi ndings are in line with other evidence from Germany, which points towards only limited disadvantages of youth in sepa- rated or divorced families. Keywords: Parental separation · Divorce · Stepfamily · Adolescent well-being · Economic deprivation Comparative Population Studies Vol. 40, 3 (2015): 335-364 (Date of release: 27.11.2015) © Federal Institute for Population Research 2015 URL: www.comparativepopulationstudies.de DOI: 10.12765/CPoS-2015-12en URN: urn:nbn:de:bib-cpos-2015-12en3 1 The terms “youth” and “adolescent” are used interchangeably. • Sabine Walper, Carolin Thönnissen, Philipp Alt336 1 Introduction Only few recent changes in family life have received as much public and scientifi c attention as the increasing instability of marriage. Ever since the “Golden Age of Marriage” of the 1960s started to wane, divorce became a prominent issue not only among demographers and family sociologists, but also among developmental and clinical psychologists seeking to determine the impact of parental divorce for the children involved. Meanwhile, marital instability has increased considerably across many countries – as refl ected in the crude divorce rate for Europe (EU-27), which doubled from 1.0 divorces per year per 1 000 inhabitants in 1970 to 2.0 divorces by 2010 (Eurostat 2013). At the same time, marriage rates declined, cohabitation increased, and a rising share of children is born to unmarried parents. In the U.S., the percentage of births to unmarried women increased relatively steadily since the 1960s, peaking in 2009 and levelling off since, with 40.6 percent in 2013 (Curtin et al. 2014). Other countries show similar trends, albeit at different levels. Given that un- married childbearing increasingly occurs in cohabiting unions and that such unmar- ried unions have a higher risk of breakup (e.g. Liefbroer/Dourleijn 2006), even when joint children are involved (Bastin et al. 2012; Schnor 2012), it is no longer only the instability of parents’ marriages, but also the instability of unmarried unions which has become an issue for child development. Not surprisingly, concerns about the impact of these changes on the next gen- eration are high. However, the focus and theoretical framing have changed over the years. While early studies from the 1950s, ‘60s and ‘70s (Hetherington/Stanley- Hagan 1999) on the effects of parental divorce were largely guided by a common defi cit model assuming that children from divorced homes are per se worse off than children from nuclear families, these generalising assumptions proved to be inadequate. The initial methodological approach, which relied on simple compari- sons of children’s development in nuclear and separated (single parent) families, also turned out to be insuffi cient. As pointed out by Amato (2010: 661), “focusing on the average effects of divorce masks the substantial degree of variability that ex- ists in people’s adjustment”. Over the past decades, large numbers of studies have contributed to a refi ned understanding of relevant conditions that may – or may not – occur in the context of divorce and may contribute to or ameliorate the many chal- lenges in children’s and adolescents’ coping with parental breakup (for an overview see Amato 2010). Accordingly, understanding the effects of parental breakup has become a complex enterprise. This paper starts by providing an overview of relevant perspectives, models, and fi ndings focusing on several key issues: the role of economic deprivation in linking family structure to child development, the infl uence of children’s contact with the non-resident parent on children’s well-being, effects of interparental confl ict, the infl uence of pre-separation stress in families which eventually separate, and costs and benefi ts of stepfamily formation. The majority of such fi ndings come from the U.S., but we will also look into evidence from Europe, particularly from Germany. The empirical analyses presented here investigate adolescents’ well-being in differ- ent family types and address the mediating role of fi nancial hardship and reduced Effects of Parental Separation • 337 contact with the non-resident father. We focus particularly on the experience of parental separation in mid- to late adolescence using prospective data from the German family panel pairfam, through which we can address possible strains prior to and following parental separation. 2 Perspectives on the Effects of Divorce Research on the well-being of children and adolescents in divorced or separated homes has examined a wide variety of consequences, including emotional well-be- ing and mental health (Chase-Lansdale et al. 1995; Strohschein 2005), behavioural problems and delinquency (Burt et al. 2008; Fergusson et al. 1992), cognitive com- petencies (Sanz-De-Galdeano/Vuri 2007), academic achievement and educational attainment (Francesconi et al. 2010; Hilmert 2002), as well as life-course trajectories with respect to home-leaving, employment and earnings, partnership stability, and early childbearing (Cherlin et al. 1995; Ross/Mirowsky 1999). Given the many empir- ical studies which address differences in children’s well-being by family structure, several meta-analyses have sought to integrate available empirical evidence. In an update of his earlier meta-analysis (Amato/Keith 1991), Amato reviewed empirical evidence from studies mainly conducted in the U.S. during the 1990s (Amato 2001) and found consistent evidence for overall disadvantages among children from di- vorced families when compared to children from nuclear families, even though the effect sizes were rather small. Elaborating on this approach and looking at evidence from Europe, a recent meta-analysis covered 17 studies from Europe (Amato 2014), encompassing inves- tigations on behavioural and emotional problems (from Bulgaria, the Netherlands, Greece, Germany, and Norway), research on educational achievement (from Italy and Sweden), studies that have looked at health problems, substance use (alcohol and tobacco), and risky sexual behaviour (from Germany, France, Greece, the Neth- erlands, and Slovakia), two studies on delinquency (from Denmark) and one study on attachment security (from Germany). Almost all of these studies reported poorer outcomes for children with divorced parents than for children with continuously married parents. The weighted effect (accounting for differences in sample size) revealed a mean value of -.17 across all studies – a value which proved small and very similar to the overall effect sizes from American studies. Accordingly, Amato (2014: 15) concluded that “irrespective of national and cultural characteristics, the gap between children with divorced and continuously married parents is about the same on both sides of the Atlantic.” At the same time, there is considerable variability in fi ndings across studies and countries (Amato/James 2010). While some authors stress the dramatic consequenc- es which parental breakup may have for the offspring’s well-being (Wallerstein et al. 1988), others highlight the coping potential in divorced families (Hetherington/Kelly 2002). Such divergent views also fi nd support in evidence from Germany: several studies based on large samples did not fi nd higher depressiveness, impaired self- esteem, more problems in peer relations, or increased behaviour problems among • Sabine Walper, Carolin Thönnissen, Philipp Alt338 children and adolescents from separated single parent families compared to nucle- ar families (Walper 2002; Walper/Wendt 2005; Wendt/Walper 2007). Furthermore, a recent study on the effects of parental separation on young children whose parents cohabited, but were not married when the children were born, found no evidence for increased emotional or behavioural problems among those who experienced parental separation (Walper/Langmeyer 2014). Similarly, the large-scale PISA as- sessment of academic competencies among 15-year-old adolescents did not sug- gest any differences between youth from single parent and two-parent families in Germany, whereas marked differences were found in other countries, such as the U.S. (Ehmke et al. 2004). However, other research conducted in Germany reported increased emotional and behavioural problems for up to two years after parental separation (Schmidt-Denter 2000), an increased likelihood of insecure attachment representation for boys in single parent families (Gloger-Tippelt/König 2007), and – in a clinical sample – more instances of conduct disorder among youth raised in separated single parent families (Steinhausen et al. 1987). In seeking to account for the variability of fi ndings across as well as within sam- ples and to better understand the origin of differences between children from sepa- rated and nuclear families, researchers have long pointed out that divorce is not a uniform single event, but typically comes along with other stressors triggered by union dissolution (Amato 2000; Hetherington/Stanley-Hagan 1999). This divorce- stress-adjustment perspective views disadvantages among offspring from sepa- rated families as caused by divorce-related stressors and differences in coping re- sources. In contrast, the selection perspective cautions against overestimating the causal role of separation and disadvantages and points towards common causes for parental separation and offspring strain (Amato 2000). Both perspectives shall be shortly reviewed. 2.1 Parental Separation as Stressor The divorce-stress-adjustment perspective (Amato 2000) points out that parental breakup is neither a uniform nor a single event, but instead is likely to be embedded in a longer series of events and stressors which partly precede but mostly follow separation or divorce. Such stressors differ in prevalence and strength, and they occur in the context of varying coping resources. Both factors – divorce-triggered stressors and the available coping resources – jointly shape the individual experi- ence and outcomes of parental separation. Prominent divorce-related stressors are (1) economic problems which arise when available fi nancial resources have to cover the expenses of two households, (2) reduced contact to the non-resident parent, and (3) confl ict between parents (e.g. about child custody and child-support payments). Findings from many countries have shown that single parenthood is linked to an increased risk of income poverty and downward mobility (Amato 2010), even though effects of family structure on in- come vary by family policy and welfare programmes, with a weaker effect in Scan- dinavian countries than in the U.K. and the U.S. (Garriga/Härkönen 2009). Poverty, in turn, has been shown to increase the risk for compromised child development in Effects of Parental Separation • 339 terms of health, education, emotional well-being, social relations, and behavioural problems (Conger et al. 2000; Duncan/Brooks-Gunn 1997; Walper 2009). In line with the economic strain hypothesis, several fi ndings suggest that a large share of the disadvantages found for children in single parent families can be explained by the increased risk of inadequate fi nancial resources (see also Amato 2010; McLanahan 1999), at least in countries where divorce typically leads to fi nancial hardship (Gar- riga/Härkönen 2009). Support has also been provided for the interparental confl ict hypothesis, as chil- dren’s increased emotional and behavioural problems in divorced families could be largely accounted for by the higher prevalence of intense confl ict between separat- ed parents (e.g. Schick 2002). Similarly, looking at subgroups of children with differ- ent developmental trajectories after parental separation, more negative trajectories were linked to more severe problems in the post-divorce relationship (Schmidt- Denter 2001). In general, for nuclear as well as for separated families, interparental confl ict has been pointed out as a major stressor for children, because it is likely to spill over into parenting (Buehler/Gerard 2002; Erel/Burman 1995), contributes to loyalty confl icts, and undermines children’s emotional security in the family system (Davies et al. 2002; Walper/Beckh 2006). However, since post-divorce confl ict may have its roots in pre-separation marital problems, it is less clear whether increased interparental tension and confl ict in the post-separation period can be interpreted as resulting from parental breakup, or whether it actually preceded the parents’ decision to separate. In general, it is quite likely that confl ict was already present in the pre-divorce period. Accordingly, interparental confl ict is also a factor addressed by the selection perspective. In contrast, parental separation is very likely to affect a child’s relation to the non-resident parent, at least in terms of frequency of contact. Many studies fi nd that contact with the non-resident parent – usually the father – diminishes as time goes by, especially if the father lacks the economic resources to provide for the children and if the interparental relation is strained (Amendt 2004). A decrease in contact with the non-resident parent can also be observed if the children were very young at the time of parental breakup and once they reach adolescence and invest more time in peer relations (see Walper/Krey 2009). In legal decision-making, counsel- ling, and public debate, considerable concern relates to such effects, because it is commonly held that children thrive on a continuous relation to both parents despite parental breakup. Interestingly, however, available evidence does not suggest that the frequency of contact is a crucial variable for well-being (Amato 1993, 2010). In a meta-analysis of links between child well-being and several aspect of non-resident fathers’ involvement, almost no effects of contact frequency were found, whereas the quality of fathers’ parenting had the strongest effect (Amato/Gilbreth 1999). Fur- thermore, a study from Germany on adolescents who never had contact to their separated father evidenced no differences in well-being compared to adolescents in nuclear families (Walper/Wendt 2011). A longitudinal study from the U.S. involv- ing a large sample of adolescents from separated families even found that the link between fathers’ engagement and adolescent behaviour had to be interpreted in the opposite direction: Fathers’ engagement did not affect their child’s subsequent • Sabine Walper, Carolin Thönnissen, Philipp Alt340 well-being, but an increase in adolescents behavioural problems led to a decrease in paternal involvement in the subsequent year (Hawkins et al. 2007). While the ef- fects of contact frequency on children’s well-being seem to be less reliable, contact does affect the chances of establishing a close relationship with the non-resident parent (Walper/Krey 2009; Whiteside/Becker 2000). Finally, union dissolution opens the options for parents to repartner and remarry, thus potentially triggering additional changes in children’s family composition and parenting experiences. In Germany, about eleven percent of all minors are estimat- ed to live with a stepparent in their primary household (Steinbach 2008). Further- more, in about 5 percent of all nuclear and 17 percent of stepfamilies, at least one parent has one or more children outside of the household, suggesting a secondary or “weekend” stepfamily constellation. Although parents usually benefi t from the support of a new partner (Schmidt-Denter 2001), the advantages of such changes are less clear for children (Coleman et al. 2000; Sweeney 2010). Typically, they not only have to adapt to a new person in the household, who is likely to claim infl u- ence on decisions and may be inexperienced in child rearing, but they also lose some of the time and attention of their biological parent. Adolescence has spe- cifi cally been pointed out as a diffi cult phase in stepfamily life, and some studies suggest more negative fi ndings for girl’s adjustment in stepfamilies (Coleman et al. 2000; Hetherington/Jodl 1994). Integrating fi ndings from 61 studies, a meta-analysis found that stepchildren generally fare worse than children in nuclear families (with both biological parents) regarding academic achievement and psychological well- being (Jeynes 2006). When compared to children in single parent families, however, stepchildren did not show any advantages. They did not differ in psychological well- being and showed lower scholastic achievement than youth from divorced single parent families. 2.2 The selection perspective While the divorce-stress-adjustment perspective highlights the causal role of di- vorce-related stressors in explaining disadvantages of children whose parents separated, the selection perspective addresses possible differences between sta- ble and disrupted families preceding parental breakup, which may affect not only the likelihood for separation or divorce but also for children’s compromised de- velopment. Likely factors are parents’ personality problems, low socio-economic resources, problems in the parents’ relationship, but also genetic factors that play a role in social behaviour (Amato 2000; Garriga/Härkönen 2009). Several approaches have been developed to address issues of selectivity (see Amato 2010; Garriga/Harkönen 2009). Taking into account pre-disruption differ- ences in family characteristics and child well-being is one of the main methods for estimating selection effects. In fact, much support for the selection perspec- tive came from prospective longitudinal studies which were able to trace family characteristics and child development across time prior to parental separation (e.g. Block et al. 1986; Cherlin et al. 1991). Many of these fi ndings suggested that disad- vantages in child development can be observed well ahead of their parents’ separa- Effects of Parental Separation • 341 tion. Interestingly, boys seemed to be more strongly affected in the pre-separation period than girls. In the study by Cherlin et al. (1991), effects of parental breakup were considerably reduced once such pre-separation differences were taken into account. In some cases, the effects of parental separation disappear when control- ling for children’s pre-divorce well-being and competencies. For example, a study based on data from the National Education Longitudinal Study observed teenagers from divorced families before and after the divorce and compared them to youth from intact homes (Sanz-De-Galdeano/Vuri 2007). The authors did not fi nd that pa- rental divorce negatively affected adolescents’ cognitive skills when controlling for pre-divorce competencies. However, selection effects such as these are not consistently found in all stud- ies on the topic (e.g. Allison/Furstenberg 1989; Forehand et al. 1997). In a study on urban youth in Germany, only limited pre-separation differences were found when comparing adolescents from stable and instable families one year prior to parental separation, in the year of parental breakup, and one year later (Schwarz 1999). Pre- separation disadvantages were restricted to adolescents’ relationships to parents and peers, but were not found for their transgression proneness (positive attitudes to delinquency) or their self-derogation. Other data suggest a combination of selec- tion effects as well as additional stresses caused by parental separation (Cherlin et al. 1998; Cherlin et al. 1995; Sigle-Rushton et al. 2005). For example, a 14-year longitudinal study which compared children of unmarried or divorced mothers to children with married mothers drew from information about the mothers’ devel- opment during their adolescent years, long before the children were born (Emery et al. 1999). As expected, children from single mother families showed increased problematic behaviour, which could partly be explained by the higher delinquency of these mothers during their adolescent years. It is quite likely that mothers‘ ear- lier problem behaviour prompts risky choices in mate selection and contributes to problems with the child’s father, thus triggering partnership instability at a later stage. Furthermore, problems in self-regulation or anti-social traits may undermine mothers’ parenting competencies, thus reducing the chances for positive child de- velopment. At the same time, parental separation seems to have caused additional problems in child development. Given such direct similarities in parents’ and children’s behaviour, addressing genetic issues seems to be a logical next step, even more so since some evidence points to genetic infl uences on the likelihood to divorce (see Amato 2010; Garriga/ Harkonen 2009). However, when comparing the effects of parental divorce in bio- logical and adoptive families, child outcomes in terms of problem behaviour and substance use were similar for biological and adopted children (e.g. Amato/Cheadle 2008; O’Connor et al. 2000). Accordingly, it is unlikely that these effects of parental separation can simply be attributed to genetic selectivity. Overall, there is evidence for selection effects which often, but not consistently, lead to increased problems among children even prior to parental separation. In general, these fi ndings are not at odds with the divorce-stress-adjustment perspec- tive, even more so since increased strain among youth shortly before their parents’ separation could still be seen as part of the separation process as a whole. This • Sabine Walper, Carolin Thönnissen, Philipp Alt342 phase of decision making is often particularly stressful for the adults involved. Thus, selection and socialisation may jointly be at work. 2.3 Differential effects across contexts and child characteristics Some attention has been paid to the question of whether the negative consequenc- es of divorce fade as divorce becomes more prevalent and a more common ex- perience in children’s lives. Such an increased “institutionalisation“ of divorce is likely to reduce the risk of stigmatisation and may facilitate successful coping with the demands of parental breakup given that many peers share the same experi- ence. However, comparing the disadvantages of children from divorced families across time, the meta-analysis by Amato (2001) does not provide evidence of a linear trend. Although effects in the 1980s were less marked than in older studies conducted up until the 1970s, more recent fi ndings from the 1990s show an increase in disadvantages for children from divorced families. Since most of these studies were conducted in the U.S., where social benefi ts for single parent families had been cut back during the 1990s, such changes in policy have been suggested as a likely explanation (see Amato 2001). Findings from other countries point towards similar evidence. A large multi-country study based on data from the Generations and Gender Survey provided no support for the „institutionalisation hypothesis“, which would have predicted less disadvantages for divorced families where divorce is highly prevalent. Instead, these data indicated larger effects of parental separa- tion on the offspring’s educational attainment in countries and at times with higher divorce rates (Bernardi/Radl 2014). Such fi ndings also raise issues of self-selection into divorce. When divorces are rare, the threshold to separate is high. Hence, couples with highly dysfunctional relationships are more likely to select into the group of those who opt for divorce. Since many fi ndings show that children’s well-being is undermined by frequent and intensive interparental confl ict (Fincham 1998; Krishnakumar/Buehler 2000), one might expect more negative consequences of parental breakups for children in such highly stressed families. However, prospective studies suggest otherwise (see Amato 2010): when tracing the effects of divorce for families with more or less problems in the parents’ pre-divorce relationship, it rather seems that children from high-confl ict families have less to loose and more to gain if their parents separate. In contrast, offspring from previously unstressed marriages showed considerable disadvantages in their adult life (e.g. Amato et al. 1995). Many studies have addressed the issue of children’s vulnerabilities, examining age and gender as likely moderators. Although some fi ndings are in line with the wide-spread notion that younger children are more dependent on a well-function- ing family and thus more vulnerable to family disruption (Allison/Furstenberg 1989; Emery 1988), other researchers have suggested that it is adolescent children who are particularly vulnerable to parental breakups (Chase-Lansdale et al. 1995). The majority of fi ndings do not suggest differential effects by children’s age (Amato 2001). Similarly, gender differences in children’s coping with parental breakup are not very clear. In line with evidence suggesting that boys are more vulnerable to Effects of Parental Separation • 343 family stress than girls (e.g. Gloger-Tippelt/König 2007), pre-separation distress was found to be higher among boys whose parents break up during childhood (Block et al. 1986; Cherlin et al. 1991). In general, however, divorce does not seem to affect boys and girls differently (e.g. Fergusson et al. 1994). For the adolescent years, the evidence is even less consistent. Since divorce has been found to be more prevalent among families with low so- cio-economic resources, studies usually control for parental education, occupation, or a SES (socioeconomic status) composite when estimating the effects of divorce. However, little attention has been paid to differential effects of parental separation depending on parents’ SES resources. In a recent study using data from the Genera- tions and Gender Survey (covering 14 countries), Bernardi and Radl (2014) investi- gated the long-term consequences of parental separation for children’s educational attainment, focusing on the likelihood of achieving tertiary education. Comparing the effects of parental breakup for families with different levels of parental educa- tion, they found no evidence for the “social origin compensation hypothesis“, which claims that parents with higher SES resources are better able to cope with divorce and protect their children from negative consequences. Rather, their data provided clear support of the „fl oor effect“, i.e. for weaker effects of parental breakup among families with low parental education whose children face substantially lower chanc- es for tertiary education anyway. Interestingly, this „fl oor effect“ was not observed in countries with early tracking, i.e. early performance-based selection into differ- ent school tracks (like in Germany). In these countries, offspring from families with higher as well as lower parental education experienced markedly lower chances for tertiary education when their parents separated. In the following study, we sought to follow up on these issues by looking at various aspects of adolescents’ well- being, including their satisfaction with their domain of education and work. 2.4 Research Questions In the following analyses, we compare adolescents’ well-being in different family structures over a 2-year time period. We employ a broad concept of parental sepa- ration, including all cases with both parents alive but not living in a union, irrespec- tive of whether they have been married in the past or not or whether they are still formally married but separated and no longer live in an intact partnership. When testing possible disadvantages of adolescents in stable single parent households and stepfamilies as opposed to nuclear families, we control for important back- ground factors and explore whether both types of separated families are linked to similar disadvantages (separation disadvantage hypothesis) or whether either of these two types of separated families is less likely to promote adolescent well-being (single mother or stepfamily strain hypothesis). Although single parent families are more likely to suffer from fi nancial hardship and may thus provide a more stressful environment for youth development, available evidence does not suggest advan- tages of stepfamilies compared to single parent families for children’s well-being (Jeynes 2006). Hence, we expect similar disadvantages for youth from both types of separated families when compared to adolescents from nuclear families. Further- • Sabine Walper, Carolin Thönnissen, Philipp Alt344 more, we seek to explore effects of a recent parental separation employing a pro- spective design. Based on previous fi ndings from prospective studies, we expect to fi nd lower well-being among adolescents prior to parental separation than among adolescents from nuclear families (pre-separation strain hypotheses). Secondly, we investigate whether changes in outcomes differ across time by fam- ily type. In this respect, we expect a reduction in adolescents’ well-being after pa- rental separation among those whose parents separated between both waves, while controlling for pre-separation well-being and background factors (post-separation strain hypothesis). These analyses also explore whether other types of separated families show different trends in adolescents’ well-being than nuclear families do. Thirdly, we ask whether infrequent contact to the non-residential parent and/or economic deprivation explain any effects of family type on changes in well-being across time. Such evidence would suggest that contact and/or economic resources function as mediating links between family structure and adolescent outcomes. While public discourse gives major importance to the amount of contact children have to their non-residential parent, research on the topic has not found substantial effects of such contact on children’s and adolescents’ well-being. Hence, we in- clude this factor, but do not expect it to explain differences in adolescents’ well-be- ing when comparing youth from separated families to those from nuclear families. In contrast, economic deprivation has been pointed out as an important mediator for explaining disadvantages of youth in single mother families, who are most likely to be hit by fi nancial strain. We expect to replicate this fi nding (economic strain hypothesis). Finally, we test whether adolescents’ gender and parental education moderate the link between family structure and adolescent well-being. With respect to paren- tal education, we expect to fi nd more negative effects of a recent parental separa- tion, single motherhood, and stepfamily life in the less educated group (resource hypothesis). Although previous fi ndings have not pointed to the signifi cance of a child’s gender as a factor that moderates the effects of parental separation, pro- spective studies have suggested that boys may be more strained prior to parental separation. Thus, we test such moderation effects for T1 differences between family types as well as for T3 changes in adolescents’ well-being in different family types. 3 Method 3.1 Sample The paper’s sample was drawn from the German Family Panel pairfam (Huinink et al. 2011), a three-cohort longitudinal study on family development with annual as- sessments, which started in 2008/2009 for three birth cohorts. Participants were recruited through register data and personal visits of the interviewer, who conduct- ed the interviews in the participants’ home. For our purposes, the sample was re- stricted to the adolescent cohort (born 1991–1993). Adolescents were mostly 15 to 17 years old during wave 1, with only very few younger or older participants. Of the Effects of Parental Separation • 345 initial adolescent cohort in wave 1 (n = 4334), 72.3 percent were re-interviewed two years later for wave 3 (n = 3132). Our analysis are based on the data from the fi rst three waves of pairfam, release 4.0 (Nauck et al. 2013). Data on household composition in waves 1 and 3 was available for 3128 adoles- cents. The family type was identifi ed by the adolescents’ report on whether both parents currently lived in a joint household and the adolescents’ household com- position at their main residence. In a fi rst step, the sample was restricted to youth whose parents were both still alive (n = 3019) and who lived with at least one parent in waves 1 and 3 (n = 2760). This second criterion excluded youth who had left the parental home prior to wave 3 (n = 259). As might be expected, the likelihood of having left the parental household differed for youth from higher- and lower-track schools. Adolescents who had already left the parental home by wave 3 were less likely to be on the higher school track than those who still stayed with their parents (35.1 percent vs. 44.6 percent; Chi2 = 8.69; df = 2; p<.01). In a second step, we se- lected family types of suffi cient group size. Adolescents who lived with a single father (n = 57) or with their father and his new partner/a stepmother (n = 27) in waves 1 and 3 could not be considered due to their low prevalence in the sample. For similar reasons, newly formed or separated stepfather families (n = 27 and 31) were not included in the analyses. In addition, we excluded all cases in which no information about either maternal or paternal education was provided (5.1 percent of all cases; n = 160). Based on data for waves 1 and 3, we categorised all remaining cases as stable nuclear families (with both biological parents in adolescents’ household, n = 1968), stable single mother families (biological mother, without a partner in the household; n = 360), stable stepfather families (biological mother and her partner in the house- hold; n = 214), or prospective separators, i.e. families who started as nuclear fami- lies in wave 1 but experienced parental separation between waves 1 and 3 (n = 76). In total, n = 2618 adolescents who participated in waves 1 and 3 were available, among these up to 2483 had complete data for our analyses. As analyses regarding selective drop-out across time revealed, longitudinal drop-out was not randomly distributed between nuclear and separated families (Chi2 = 13.44; df = 1; p < .01). Adolescents from separated families were more likely to drop out by wave 3 than adolescents from nuclear families were (32.5 percent vs. 26.8 percent). Table 1 provides an overview of the sample, subdivided by family type as well as providing totals. As shown in the last column, adolescents’ average age during wave 1 was 15.92 years (SD = .88), and about half of the sample was female. Almost half of all adolescents were currently enrolled or had already fi nished grammar school (“Gymnasium”), the highest school track which – if successfully fi nished – provides access to university studies. The share of families with high maternal education (university or advanced technical college entrance qualifi cation) was about a third. One-fi fth of the adolescents had a migration background, i.e. they and/or at least one of their parents were not born in Germany. In line with population statistics, only a minority (18.0 percent) of participants lived in eastern Germany. To test for signifi cant differences by gender, maternal education, migration back- ground, and region of residence between the subgroups defi ned by family type, • Sabine Walper, Carolin Thönnissen, Philipp Alt346 Chi2-Tests were carried out. The share of adolescents in grammar schools proved signifi cantly lower in single mother families and stepfather families than in nuclear families (Chi2 = 34.28; df = 3; p < .01). A similar pattern was found for maternal education. The share of high maternal education (graduation from grammar school) proved signifi cantly lower in single mother families and stepfather families than in nuclear families (Chi2 = 8.03; df = 3; p < .05). Signifi cant differences were also found by region (Chi2 = 12.39; df = 3; p < .01): Adolescents in stepfather families were more likely to live in eastern Germany, while regional distribution did not differ for the other family types. Furthermore, migration background differed signifi cantly across family types (Chi2 = 15.42; df = 3; p < .01): Prospective separators and ado- lescents in stepfather families were less likely to have a migration background than youth in single mother families. In order to inspect the fi nancial situation of the family, we decided not to rely on household income, which was only available as reported by adolescents. Instead, we used a presumably more valid indicator: a two-item-scale which was assessed in wave 3 and measured fi nancial strain in the household (see indicators below). Analyses of variance comparing the four family types revealed highly signifi cant effects of family type (F = 85.37; df = 3; p < .001): Economic deprivation was sig- nifi cantly lower in nuclear families than in all types of separated families. According to post-hoc tests, youth from single mother families reported signifi cantly higher economic deprivation than youth from stepfather and prospective separation fami- lies. Stepfather families and prospective separators did not differ signifi cantly. Tab. 1: Descriptives of the Sample by Family Type Nuclear Families Prospective Separators* Single Mother Families Stepfather Families Total N 1968 76 360 214 2618 Age at T1 M 15.90 15.91 15.97 15.95 15.92 SD .87 .80 .89 .94 .88 Gender (% Female) 47.4 53.9 47.2 46.7 47.5 School Track (% Grammar School) 48.2 42.7 33.9 35.5 45.1 High Maternal Education (% yes) 32.9 27.6 27.2 26.6 32.5 Migration Background (% yes) 21.3 11.8 23.3 11.2 20.8 Region (% East) 16.9 22.4 18.3 26.2 18.0 Economic Deprivation T2/3 M 2.10 2.55 2.91 2.32 2.24 SD .87 .92 1.01 .90 .93 % With Contact to Non-Resident Father at T3 -- 97.3 77.7 76.7 79.7 * Adolescents whose parents separated between waves 1 and 3 Source: pairfam data release 4.0 Effects of Parental Separation • 347 Finally, we compared the share of adolescents’ who still had contact to their non-residential father in wave 3 and found signifi cant differences between the three types of separated families (Chi2 = 16.11; df = 2; p < .001). As might be expected, youth in recently separated families (prospective separators) were less likely to have lost contact to their biological fathers at T3 than youth in stable single mother and stable stepfather families. No signifi cant differences regarding age or gender were found. 3.2 Indicators Economic Deprivation. To assess Economic Deprivation in the household, two items from a larger scale were used in wave 3 indicating the inadequacy of the family’s budget for ordinary living expenses (e.g. “My parents often have to forego some- thing because they have to watch their budget”). The indicator was computed as the mean of both items. More specifi c information on this, as well as all other indicators used in our analyses is presented in table 2. Contact to Father. For separated families, information on the frequency of con- tact to the non-residential father was assessed by a single question: “How often are you in contact with your father, adding up all visits, letters, phone calls, etc.?” For the purpose of our analyses, we focused on the effects of infrequent contact and recoded the categories into a dichotomous variable combining the categories “con- Tab. 2: Description of Indicators Variable Number of items Source Response format Cron- bach’s Alpha Mean (standard deviation) Wave 1 Wave 3 Economic Deprivation 2 Schwarz et al. (1997) 1 = not at all correct to 5 = completely correct .89 (W3) 2.24 (.93) Contact to Father 1 pairfam 1 = daily to 7 = never 10 = contact never existed - 1.87 (1.75) Satisfaction with Family 1 pairfam 0 = very dissatisfi ed to 10 = very satisfi ed 8.83 (1.54) 8.55 (1.58) Satisfaction with School, Education, Career 1 pairfam 0 = very dissatisfi ed to 10 = very satisfi ed 7.36 (1.85) 7.42 (1.98) Overall Life Satisfaction 1 pairfam 0 = very dissatisfi ed to 10 = very satisfi ed 7.99 (1.47) 7.90 (1.45) Self-Esteem 3 Adapted from Rosenberg (1965) 1 = not at all to 5 = absolutely .78 (W3) 4.13 (.74) 3.98 (.82) • Sabine Walper, Carolin Thönnissen, Philipp Alt348 tact never existed”, ”never” and “less than several times per year” into the category infrequent (no or rare) contact (n = 186; coded as 1), while the other categories (from “several times per year” up to “daily”) served as reference category indicating more frequent contact (n = 2407; coded as 0). Although this coding implies a loss of information about variations in frequency of contact, it provided a variable which was less confounded with family structure than the original categories. Adolescents’ Well-Being. To assess adolescents’ well-being in waves 1 and 3, we relied on indicators which were similarly measured in both waves. Three single- item indicators gave information about adolescents’ satisfaction with their family, school/work, and life in general. The items were introduced with the question “How satisfi ed are you with the following domains of your life?”. The domains were named “Family” and “School, Education, Career”. Overall life satisfaction was assessed with the question “All in all, how satisfi ed are you with your life at the moment?”. Furthermore, a 3-item-indicator of self-esteem was included in both waves. Since the assessment mode of self-esteem was changed between waves 1 and 2 from computer-assisted personal interviews (wave 1: CAPI) to computer-assisted self-completed interview (waves 2 and later waves: CASI), the scale had to be cor- rected for mean-level differences which emerged between wave 1 and later waves (with higher levels of self-esteem reported in the CAPI than in the CASI interviews). In order to correct for mean level differences without changing the rank order of participants at each wave, the scale was z-standardised for the entire sample of each wave. Due to the loss of sample size for the present analyses (with a higher loss in wave 1 than in wave 3), the distribution of the self-esteem scales does not completely match the original means and standard deviations. More detailed in- formation on all indicators used in the present analyses can be found in the scale manual for the pairfam project report, available online (Thönnissen et al. 2014). 3.3 Analyses Multiple regression analyses were used to test effects of family structure on ado- lescents’ well-being at waves 1 and 3 (T1 and T3), controlling for adolescents’ age, gender, maternal education, migration background, and region of residence. Fam- ily type was dummy-coded using nuclear families as the reference category for comparisons with recent separators, single mother families, and stepfather families (each coded as a dichotomous dummy variable). For each dependent variable, fi ve analyses were conducted: (1) In model 1, data for T1 inform about mean differences between youth from dif- ferent family types, with a special focus on possible pre-separation strain among those whose parents later separated in the following two years (between T1 and T3). These fi rst analyses control for background variables only. (2) In model 2, we added interaction effects between family type and gender or ma- ternal education to test whether pre-separation well-being among prospective separators was more compromised for boys or youth from families with lower maternal education than their respective counterparts. In addition, these analy- ses examined similarities and differences in the effects of single motherhood Effects of Parental Separation • 349 and stepfamily life between boys and girls and between families with higher or lower educational resources. (3) The third model used data for T3 as dependent variables to assess changes in well-being between both waves, controlling for the stability of the respective outcome variable across time. It tested effects of family type, controlling for the background variables as well as for effects of the outcome variable at T1. (4) Model 4 added the effects of infrequent contact to the biological father at T3 and economic deprivation at T3 as likely mediators between family structure and adolescent outcomes. In these latter analyses, we were interested in seeing (a) whether youth from separated family types evidenced less favourable changes in well-being across the two-year time period than youth from nuclear families, (b) whether infrequent contact to the father and/or economic deprivation affect- ed adolescent outcomes, and (c) whether and to what extent the effects of fam- ily structure were diminished when introducing infrequent contact to the father and economic deprivation as additional predictors. Given that infrequent con- tact to the father is only found in separated families and economic deprivation has been shown to be substantially higher in single mother families (see sample description), both are potential mediators which may explain disadvantages in adolescent well-being when comparing separated to nuclear families (for tests on mediator effects see Baron/Kenny 1986). Furthermore, particular attention was paid to effects of a prospective separation between T1 and T3. If a recent parental separation was more stressful than the pre-separation period, adoles- cents’ well-being should be signifi cantly compromised over time for this group of prospective separators (controlling for the previous level of the respective outcome at T1). (5) In a fi nal model, we tested interactions between family type and adolescent gen- der or maternal education, using multiplication terms between the dichotomous family type variables and either gender or education (each variable coded 0/1). All interaction terms were entered simultaneously (controlling for the predictors used in the model 4 for T3). If any interaction term proved signifi cant, the mul- tiple regression analyses were repeated separately for the relevant subgroups (males vs. females or adolescents with higher- vs. lower-educated mother) to allow for more detailed comparisons. 4 Results Satisfaction With Family. The fi ndings for multiple regressions on adolescents’ sat- isfaction with their family are depicted in table 3. As shown in the fi rst column for family satisfaction at T1, youth from single mother families indicated signifi cantly lower satisfaction with family life, whereas youth in stepfather families did not dif- fer from nuclear families. Furthermore, youth in nuclear families whose parents separated within the subsequent two years (prospective separators) also indicated signifi cantly lower satisfaction with their family at T1, even though this effect was rather weak. There were neither age nor gender differences, nor did the region of • Sabine Walper, Carolin Thönnissen, Philipp Alt350 residence or a migration background have any effect. However, adolescents with highly-educated mothers reported signifi cantly lower satisfaction with family life than youth with a less highly-educated mother. As can be seen in model 2, no inter- action effects were found, i.e. the effects of family type did not differ by adolescent gender or educational resources. At T3 (model 3), disadvantages of separated families had slightly increased, as evident in signifi cant negative changes when controlling for family satisfaction at T1. This particularly holds for prospective separators, whose family satisfaction showed a signifi cantly stronger decline than found among youth in nuclear families. Youth in single mother or stepfather families also indicated a small, but signifi cantly stronger decrease in satisfaction with their family than youth in nuclear families, but these effects were weak overall. As can be seen in model 4, these effects were not diminished when controlling for economic deprivation and infrequent contact to the father even though economic deprivation had a signifi cant negative effect on adolescents’ satisfaction with their family. As expected, infrequent contact to the T1 T3 (1) (2) (3) (4) (5) Gender (Male) -.006 -.005 .017 .016 .008 Age -.021 -.022 -.010 -.010 -.010 Migration Background .039 .038 .021 .035 .034 Maternal Education -.096** -.091*** -.012 -.028 -.044* Region (East) .022 .021 -.014 -.001 -.003 Prospective Separators+ -.084** -.042 -.106*** -.099*** -.150*** Single Mother Families -.139** -.152*** -.081*** -.059** -.080* Stepfather Families -.024 -.016 -.054*** -.050** -.061* Family Satisfaction T1 .393*** .387*** .388*** Economic Deprivation -.089*** -.089*** Infrequent Contact to Father .004 .008 Prosp.Sep. X Gender -.053 .032 Single Fam. X Gender .029 .022 Stepfather Fam. X Gender -.010 -.006 Prosp.Sep. X Education -.012 .057 Single Fam. X Education -.013 .007 Stepfather Fam. X Education .000 .027 R2 .031 .031 .186 .192 .193 N 2481 2481 2427 Tab. 3: Effects on Adolescents’ Family Satisfaction at T1 and T3: Standardised Regression Coeffi cients Note: Signifi cance: * p < .05; ** p < .01; *** p < .001 + Adolescents whose parents separated between waves 1 and 3 Source: pairfam data release 4.0 Effects of Parental Separation • 351 father did not affect adolescents’ satisfaction with their family. Model 5 introduced the interaction terms, but none of these proved signifi cant. Satisfaction With School, Education, and Occupation. Similar to the previous analyses, youth from single mother families as well as prospective separators re- ported lower satisfaction with their education or work situation at T1 when com- pared to adolescents from nuclear families (see Table 4). This indicates not only disadvantages of youth with single mothers, but also pre-separation strain in the scholastic or occupational domain among adolescents whose parents later separat- ed. However, both effects were rather small. Adolescents from stepfather families did not differ from youth who lived with both biological parents. Age, gender, ma- ternal education, and migration background showed no effects. There were small effects for region, as youth from eastern Germany tended to report less satisfaction with their education and work than youth from western Germany. In model 2, no interaction effect emerged. Regarding change across time (model 3), prospective separators as well as youth from single mother or stepfather families did not differ Tab. 4: Effects on Adolescents’ Satisfaction with School, Education, and Occupation at T1 and T3: Standardised Regression Coeffi cients T1 T3 (1) (2) (3) (4) (5) Gender (Male) -.008 -.012 .057** .055** .060** Age .038 -.038 -.018 -.017 -.018 Migration Background -.010 -.009 -.047* -.029 -.031 Maternal Education -.026 -.024 -.015 -.035 -.031 Region (East) -.052* -.051* -.043* -.028 -.030 Prospective Separators+ -.052** -.062* -.021 -.012 -.023 Single Mother Families -.090*** -.102** -.032 .001 .016 Stepfather Families -.027 -.013 -.010 .000 .017 Satisfaction with School, Education, and Occupation T1 .208*** .197*** .197*** Economic Deprivation -.106*** -.106*** Infrequent Contact to Father -.014 -.016 Prosp.Sep. X Gender .028 -.009 Single Fam. X Gender .009 .006 Stepfather Fam. X Gender -.011 -.026 Prosp.Sep. X Education -.018 .033 Single Fam. X Education .011 -.035 Stepfather Fam. X Education -.010 .004 R2 .011 .010 .050 .059 .059 N 2481 2481 2426 Note: Signifi cance: * p < .05; ** p < .01; *** p < .001 + Adolescents whose parents separated between waves 1 and 3 Source: pairfam data release 4.0 • Sabine Walper, Carolin Thönnissen, Philipp Alt352 from youth in nuclear families. A small gender effect was found, as male adoles- cents reported a slightly stronger increase in satisfaction than female adolescents. Economic deprivation proved to predict lower satisfaction with this domain but did not explain any effect of family structure (model 4), as there were no such effects in model 3. Interestingly, satisfaction with this domain was substantially less stable than family satisfaction. No interaction effects between family structure and gender or maternal education were found (model 5). Overall Life Satisfaction. Table 5 shows a summary of multiple regressions on adolescent’s life satisfaction. A signifi cant effect of family type was found at T1 and T3 (model 1), with lower overall life satisfaction at T1 and more negative changes (T3) reported by youth from single mother families (see model 1 and 3). The lat- ter effect was reduced to insignifi cance when controlling for economic deprivation (model 4), which proved to be a highly signifi cant predictor of compromised life satisfaction. Infrequent contact to the father did not affect overall life satisfaction. As for the control variables, males and younger adolescents proved more satisfi ed Tab. 5: Effects on Adolescents’ Life Satisfaction at T1 and T3: Standardised Regression Coeffi cients T1 T3 (1) (2) (3) (4) (5) Gender (Male) .074* .063** .057** .055** .062** Age -.067* -.069** -.028 -.029 -.028 Migration Background .018 .017 -.045* -.013 -.014 Maternal Education -.024 -.006** .017 -.019 -.039 Region (East) -.028 -.028 -.070*** -.042* -.043* Prospective Separators+ -.001 .014 -.030 -.014 -.012 Single Mother Families -.076* -.062*** -.076*** -.018 -.052 Stepfather Families -.013 -.040 -.015 .002 .021 Life Satisfaction T1 - .335*** .308*** .311*** Economic Deprivation - -.188*** -.190*** Infrequent Contact to Father - -.020 -.014 Prosp.Sep. X Gender - -.008 -.023 Single Fam. X Gender - .023 .011 Stepfather Fam. X Gender - .032 -.035 Prosp.Sep. X Education - -.019 .025 Single Fam. X Education - -.060* .048* Stepfather Fam. X Education - .010 .008 R2 .014 .015 .134 .163 .164 N 2483 2483 2426 Note: Signifi cance: * p < .05; ** p < .01; *** p < .001 + Adolescents whose parents separated between waves 1 and 3 Source: pairfam data release 4.0 Effects of Parental Separation • 353 than females and older adolescents, while maternal education and migration back- ground did not matter. A small effect of region emerged at T3 (model 3), indicating a slightly stronger decrease in life satisfaction among youth in eastern than in west- ern Germany. In model 4, however, controlling for economic deprivation, this effect was no longer signifi cant. Finally, both tests for interaction effects (model 2 and 5) evidenced differential effects of single motherhood among families with higher or lower educational resources. However, the effect differed for both models, i.e. changed over time. Separate regression analyses (as in model 1) were conducted for both educational groups. Among families with higher maternal education, youth from single mother families reported lower life satisfaction than youth from nuclear families (beta = -.153, p < .001), while no such effect was evident among youth with less highly-educated mother (beta = -.046, n.s.). While this fi nding does not support the resource hypothesis, changes over time (model 3) were more in line with the re- source notion. These changes were less favourable for youth from less resourceful families (beta = -.102, p < .001) than for adolescents with a higher-educated mother (beta = -.004, n.s.). Tab. 6: Effects on Adolescents’ Self-Esteem at T1 and T3: Standardised Regression Coeffi cients T1 T3 (1) (2) (3) (4) (5) Gender (Male) .213*** .197*** .113*** .115*** .107** Age -.009 -.010 .001 .000 .001 Migration Background .049* .049* .016 .026 .026 Maternal Education .016 .026 -.010 -.021 -.027 Region (East) -.028 -.028 -.025 -.016 -.017* Prospective Separators+ -.023 -.029 -.001 .002 -.031 Single Mother Families -.012 -.011 -.052** -.049* -.071* Stepfather Families -.016 -.050 -.017 .010 .013 Self-Esteem T1 - .380*** .369*** .369*** Economic Deprivation - -.075*** -.075*** Infrequent Contact to Father - .040 .043* Prosp.Sep. X Gender - -.020 .023 Single Fam. X Gender - .024 .023 Stepfather Fam. X Gender - -.036 .000 Prosp.Sep. X Education - -.014 .033 Single Fam. X Education - -.036 .009 Stepfather Fam. X Education - .016 -.007 R2 .047 .047 .131 .164 .165 N 2482 2416 Note: Signifi cance: * p < .05; ** p < .01; *** p < .001 + Adolescents whose parents separated between waves 1 and 3 Source: pairfam data release 4.0 • Sabine Walper, Carolin Thönnissen, Philipp Alt354 Self-Esteem. Finally, turning to adolescents’ self-esteem, table 6 shows the fi nd- ings from our regression analyses. As can be seen, no effects of family type were found at T1. Prospective separators did not differ from youth in nuclear families, nei- ther prior to parental separation (T1), nor shortly after (T3). Similarly, we did not fi nd any disadvantage of youth in stepfather families at either measurement point. Only at T3 did adolescents growing up with a single mother indicate slightly lower self- esteem. Although signifi cant, this effect was very weak and could not be accounted for by the negative effects of economic deprivation on adolescents’ self-esteem (model 4). There were no interaction effects (model 5). With respect to the control variables, males consistently reported higher self- esteem than females at T1 and T3, and youth with a migration background had slightly higher self-esteem than youth without a migration background. Age, ma- ternal education, and the region of residence did not have any signifi cant effects at T1 (model 1) or over time (model 3). No interaction effects were found in model 5. 5 Discussion The aim of the current study was to assess the effects of family structure on dif- ferent aspects of adolescent well-being across a period of two years. We not only wanted to explore possible disadvantages for adolescents in stable single mother and stepfather families compared to nuclear families, but also took the opportunity of following a group of adolescents whose parents separated within the two-year interval between the two longitudinal assessments. This group of so-called “pro- spective separators” allowed for addressing issues of selectivity and pre-separation strain that may compromise adolescents’ well-being prior to parental separation. We also examined whether changes in adolescents’ well-being differed by family type and expected to fi nd more negative change among prospective separators. Secondly, we explored the mediating role of infrequent contact to the non-residen- tial parent as well as economic deprivation in linking family structure to adolescent outcomes in well-being. Finally, we tested whether gender and maternal education moderated the effects of family structure at T1 or across time. In the following, we summarise and discuss the overall pattern of results linking our fi ndings to the hy- potheses previously specifi ed. We end with a brief outline of strengths and limita- tions of our study and of challenges for future research. 5.1 Effects of Family Type Effects of family structure on adolescents’ well-being were evident for all outcome variables, but differed with respect to the domain under consideration. Most con- sistent evidence was found for disadvantages of youth from single mother families, who reported lower satisfaction with family life and with the domain of education and occupation as well as a limited overall life satisfaction compared to their peers in nuclear families at T1. These youth with single mothers also evidenced a stronger decrease in family satisfaction, overall life satisfaction, and self-esteem than youth Effects of Parental Separation • 355 in nuclear families. Although these fi ndings are in line with the large body of re- search on parental separation (Amato 2010), it should be noted that the effects are mostly very small. As might be expected, the most pronounced effects were found for family satisfaction, indicating that the problems faced by these youth are indeed rooted in the family system and not in confounded external factors. In contrast, there was almost no evidence for disadvantages of adolescents raised in stepfather families. At T1, they did not differ in any of the outcome vari- ables from youth in nuclear families. When looking at changes over time, only fam- ily satisfaction was found to show a stronger decline among youth living with a stepfather, while the other three indicators did not reveal any less advantageous changes than found for youth in nuclear families. Overall, these fi ndings suggest that the majority of stepfamilies manage to adjust successfully to the challenges of stepfamily life and provide a positive context for adolescents’ well-being. This fi nd- ing is noteworthy, since other evidence has instead focused on the challenges and problems in stepfamily life (Amato 1994; Sweeney 2010). Contrary to the fi ndings from the large body of data available from international studies (Jeynes 2006), our data do not suggest that adolescents raised in a stepfamily are similarly strained or even worse off than youth living in a single mother family. The slightly stronger de- crease in satisfaction with family life may indicate that stepfamily life becomes more stressful in late adolescence when autonomy is more fully established. However, this effect was quite small and should not be over-interpreted. With respect to prospective separators, the expected pre-separation strain at T1 was evidenced by lower family satisfaction than found for youth from stable nuclear families. Furthermore, such pre-separation impairment of well-being was also evi- dent in the domain of education and work, which these adolescents were less satis- fi ed with than those from stable nuclear families. It should be noted that – contrary to youth from single mother and stepfather families – prospective separators were not less likely than their peers from stable nuclear families to be enrolled in or have fi nished the highest track of schooling. This may suggest that such problems with education/work were not yet evident at the end of elementary school but rather emerged more recently, refl ecting distractions emanating from discord at home. While these effects are in line with other evidence suggesting higher levels of strain in the pre-separation phase (e.g. Cherlin et al. 1991), they prove to be rather weak. Moreover, neither overall life satisfaction nor adolescents’ self-esteem suggest any major general pre-separation strain extending into adolescents’ outlook on life. Overall, our fi ndings at least partly resemble other evidence from Germany, which provided only limited evidence for pre-separation effects (Schwarz 1999). Similarly, more negative changes across time are restricted to family satisfaction, indicating that family problems increase when parents separate. This fi nding is no surprise. Rather, it must be emphasized that this effect is quite limited, not only in content, but also in size. Neither did we fi nd any signs for a spillover of post-separation problems into the domain of education or work, nor did the experience of parental separation seem to affect more general features of well-being, such as life satisfac- tion or self-esteem. • Sabine Walper, Carolin Thönnissen, Philipp Alt356 In summary, while the expected effects of parental separation were evident for almost every outcome variable, these effects proved rather weak and fell below the average effect sizes reported in meta-analyses for data, most of which come from the US (Amato 2001). The most pronounced effects were found for adolescents’ family satisfaction, which – along with adolescents’ satisfaction in the domain of education and work – also provided support for the expected pre-separation strain. An additional decrease in well-being shortly after a recent parental separation was only evident for family satisfaction. Finally, stepfamilies seem to provide a rather unobtrusive context for adolescents’ well-being. 5.2 The Mediating Role of Infrequent Contact to the Father and Economic Deprivation Our analyses also tested for possible mediation effects of infrequent contact to the father and economic deprivation. Throughout the analysis, infrequent contact to father had no effect on adolescents’ well-being and thus could not explain any effects of the family type. Overall, these fi ndings are well in line with other studies (Amato/Gilbreth 1999). Economic deprivation, however, proved to mediate at least some of the effects of family structure. Such mediation effects of economic deprivation were restricted to disadvantages of single mother families (who experienced the highest fi nancial strain). The stronger decrease in overall life satisfaction which was found among adolescents in single mother families could be fully accounted for by the higher economic deprivation reported by these youth. With respect to family satisfaction, however, only partial mediation was found, and our results for self-esteem did not indicate any mediation by economic deprivation. Although self-esteem decreased more strongly in single mother families than in nuclear families, this effect remained unchanged when controlling for fi nancial conditions. Overall, these fi nding are at least partly in line with other evidence suggest- ing that a substantial share of disadvantages found for children raised by a single mother can be accounted for by the increased fi nancial problems these families face (e.g. McLanahan 1999). At the same time, parental separation seems to also play a distinct role, as particularly evident in cases of recent separations. Elevated post-separation strain among prospective separators, as evident in reduced family satisfaction, could not be explained by economic deprivation. This suggests that a recent separation does provide a unique stressor which cannot be reduced to fi nancial problems. 5.3 The Moderating Role of Gender and Maternal Education We also addressed possible moderating effects of gender and adolescents’ school type to test for differential effects of family type in the respective subgroups. In line with other fi ndings, gender did not moderate any effects of growing up in a single mother family. While other evidence suggests that girls fare worse than boys in stepfather families (Hetherington/Jodl 1994), we did not fi nd stronger disadvan- Effects of Parental Separation • 357 tages in well-being among girls (or boys) growing up with a stepfather. Similarly, we could not replicate earlier fi ndings indicating stronger pre-separation strain among boys. While the group of prospective separators may have been too small to detect such effects, the overall pattern suggests rather homogeneous effects of family type for males and females. Maternal education was found to moderate the effects of family type in one case only: for adolescents’ overall life satisfaction. At the same time, this effect was com- plex and not easy to interpret. Contrary to the resource hypothesis, but in line with the fl oor effect found by Bernardi and Radl (2014), our data also indicate a “divorce penalty” among more privileged youth only: Negative effects of parental separa- tion at T1 were restricted to youth with a higher-educated mother. In this subgroup, youth from single mother families indicated lower life satisfaction than those from nuclear families, while no such difference was found for adolescents with a less highly educated mother. It may well be that adolescents from families with higher educational resources have more to lose when their parents separate, at least in terms of fi nancial resources. Interestingly, however, support for the resource hy- potheses was found when looking at changes across time. Adolescents’ life satis- faction decreased more strongly in single mother than nuclear families with a lower education, while no such effect was found among adolescents with a higher-edu- cated mother. As expected, the decrease in life satisfaction among youth from less resourceful single mother families could be fully explained by fi nancial strain. This suggests that fi nancial resources linked to parental education play a major role in differential outcomes of single parenthood. 5.4 Strengths and Limitations of the Study This study used a large sample of adolescents in Germany for systematically testing several research questions regarding the effects of parental separation on adoles- cents’ well-being. Taking a broad view on separation and re-partnering, we did not only focus our analyses strictly on parental divorce or on remarried families, but also included youth whose parents were still or never married but separated, and those who lived with their mother’s new partner, no matter whether they were mar- ried or not. Overall, our data provide support for the detrimental effects of increased fi nancial strain among single mother families, while the role of contact to the non- residential father for adolescents’ well-being seems to be much less marked. This latter fi nding is well in line with broad evidence from international studies (Amato/ Gilbreth 1999) as well as studies conducted in Germany (Walper/Wendt 2011; Walp- er/Krey 2009), but not with current legal practice, which places much emphasis on preserving the continuity of adolescents’ relationship to their non-custodial par- ent, even against the children’s will. Furthermore, we could identify increased pre- separation strain in the subgroup of prospective separators, who also seem to face additional post-separation strain, not with a broad impact on general well-being, but more specifi cally in the domain of family life. Finally, we found evidence for the moderating role of socio-economic resources as addressed by maternal education. • Sabine Walper, Carolin Thönnissen, Philipp Alt358 Nevertheless, these fi ndings should be interpreted with caution. First, as pointed out above, the effects of family structure were mostly very weak. Many of them would have remained undetected using a smaller sample. While the large sample may be considered an advantage of this study, the subgroups in focus differ con- siderably in size. Particularly the group of prospective separators was very small in comparison to the large group of nuclear families. Restricting the number of nuclear families by selecting a random subsample or by matching them to the separated families might have solved this problem. However, such procedures carry other risks. We decided to capitalise on the full sample to provide a broader picture of adolescents growing up in different family types. Second, we restricted our analyses to adolescents who still lived with their mother in any case. Although the number of adolescents who had already moved out by age 17 to 19 (the majority age range at T3) was small, our selectivity analyses showed that offspring from separated families is overrepresented in this group. Hence, we may underestimate the effects of family structure if those who were most strained by their parents’ separation moved out earlier. Third, we could only rely on a limited range of outcome variables. Indicators of adolescents’ depression differed at both measurement points and thus could not be considered. Information about adolescents’ problem behaviour was not available at either measurement point. Hence, we may underestimate psychological and behav- ioural problems that have been pointed out to be more sensitive to negative effects of parental separation (Amato 2001). Finally, our analyses did not inform about some key variables previously dis- cussed, such as the role of interparental confl ict and the number of family transi- tions adolescents went through prior to our study. Some of these data are available in the pairfam panel, and can be addressed in future analyses. Acknowledgement This research was supported by a grant from the German Research Foundation (DFG) to Sabine Walper (Wa 949/11-1 through 3) as principle co-investigator of the German Family Panel, pairfam, at Ludwig-Maximilian-University, Munich. We are grateful to all participants of the pairfam panel. References Allison, Paul D.; Furstenberg, Frank F. Jr. 1989: How marital dissolution affects chil- dren: variations by age and sex. In: Developmental Psychology 25,4: 540-549 [doi: 10.1037/0012-1649.25.4.540]. Amato, Paul R. 1993: Children’s adjustment to divorce: Theories, hypotheses, and em- pirical support. In: Journal of Marriage and Family 55: 23-38 [doi: 10.2307/352954]. Amato, Paul R. 1994: The implications of research fi ndings on children in stepfamilies. In: Booth, Alan; Dunn, Judith (Eds.): Stepfamilies: Who benefi ts? Who does not? Hill- sdale, NJ: Erlbaum: 81-87. Effects of Parental Separation • 359 Amato, Paul R. 2000: The consequences of divorce for adults and children. In: Journal of Marriage and Family 62: 1269-1287 [doi: 10.1111/j.1741-3737.2000.01269.x]. Amato, Paul R. 2001: Children of divorce in the 1990s: An update of the Amato and Keith (1991) meta-analysis. In: Journal of Family Psychology 15,3: 355-370 [doi: 10.1037/0893-3200.15.3.355]. Amato, Paul R. 2010: Research on divorce: Continuing trends and new developments. In: Journal of Marriage and Family 72,3: 650-666 [doi: 10.1111/j.1741-3737.2010.00723.x]. Amato, Paul R. 2014: The consequences of divorce for adults and children: an update. In: Drustvena istrazivanja: Journal for General Social Issues 23,1: 5-24 [doi: 10.5559/ di.5523.5551.5501]. Amato, Paul R.; Keith, Bruce 1991: Parental divorce and the well-being of children: a meta-analysis. In: Psychological Bulletin 110: 26-46 [doi: 10.1037/0012-1649.25.4.540]. Amato, Paul R.; Gilbreth, Joan G. 1999: Nonresident fathers and children’s well-being: A meta-analysis. In: Journal of Marriage and Family 61: 557-573 [doi: 10.2307/353560]. Amato, Paul R.; Loomis, Laura S.; Booth, Alan 1995: Parental divorce, marital confl ict, and offspring well-being during early adulthood. In: Social Forces 73: 895-915 [doi: 10.2307/2580551]. Amato, Paul R; Cheadle, Jacob E. 2008: Parental divorce, marital confl ict and children’s behavior problems: A comparison of adopted and biological children. In: Social Forc- es 86,3: 1139-1161 [doi: 10.1353/sof.0.0025]. Amato, Paul R.; James, Spencer 2010: Divorce in Europe and the United States: commonalities and differences across nations. In: Family Science 1,1: 2-13 [doi: 10.1080/19424620903381583]. Amendt, Gerhard 2004: Scheidungsväter. Schriftenreihe des Instituts für Geschlechter- und Generationenforschung Band 6. Bremen: Ikaru-Verlag. Baron, Reuben M.; Kenny, David A. 1986: The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. In: Journal of Personality and Social Psychology 51: 1173-1182 [doi: 10.1037/0022- 3514.51.6.1173]. Bastin, Sonja; Kreyenfeld, Michaela; Schnor, Christine 2012: Diversität von Familienfor- men in Ost- und Westdeutschland. In: Krüger, Dorothea; Herma, Holger; Schierbaum, Anja (Eds.): Familie(n) heute Entwicklungen, Kontroversen, Prognosen. Wiesbaden: Juventa: 126-145. Bernardi, Fabrizio; Radl, Jonas 2014: The long-term consequences of parental divorce for children’s educational attainment. In: Demographic Research 30,61: 1653-1680 [doi: 10.4054/DemRes.2014.30.61]. Block, Jeanne H.; Block, Jack; Gjerde, Per F. 1986: The personality of children prior to di- vorce: A prospective study. In: Child Development 57: 827-840 [doi: 10.2307/1130360]. Buehler, Cheryl; Gerard, Jean M. 2002: Marital confl ict, ineffective parenting and chil- dren’s and adolescents’ maladjustment. In: Journal of Marriage and Family 64,1: 78-92 [doi: 10.1111/j.1741-3737.2002.00078.x]. Burt, S. Alexandra et al. 2008: Parental divorce and adolescent delinquency: ruling out the impact of common genes. In: Developmental Psychology 44,6: 1668-1677 [doi: 10.1037/a0013477]. Chase-Lansdale, P. Lindsay; Cherlin, Andrew J.; Kiernan, Kathleen E. 1995: The long- term effects of parental divorce on the mental health of young adults: A developmen- tal perspective. In: Child Development 66,6: 1614-1634 [doi: 10.1111/j.1467-8624.1995. tb00955.x]. • Sabine Walper, Carolin Thönnissen, Philipp Alt360 Cherlin, Andrew J.; Chase-Landsdale, P. Lindsay; Mcrae, Christine 1998: Effects of pa- rental divorce on mental health throughout the life course. In: American Sociological Review 63,2: 239-249 [doi: 10.2307/2657325]. Cherlin, Andrew J. et al. 1991: Longitudinal studies of effects of divorce on children in Great Britain and the United States. In: Science 252: 1386-1389 [doi: 10.1126/sci- ence.2047851]. Cherlin, Andrew J.; Kiernan, Kathleen E.; Chase-Lansdale, P. Lindsay 1995: Parental di- vorce in childhood and demographic outcomes in young adulthood. In: Demography 32,3: 299-318 [doi: 10.2307/2061682]. Coleman, Marilyn; Ganong, Lawrence; Fine, Mark 2000: Reinvestigating remarriage: Another decade of progress. In: Journal of Marriage and Family 62,4: 1288-1307 [doi: 10.1111/j.1741-3737.2000.01288.x]. Conger, Katherine J.; Rueter, Martha A.; Conger, Rand D. 2000: The role of economic pressure in the lives of parents and their adolescents: The Family Stress Model. In: Crockett, Lisa J.; Silbereisen, Rainer K. (Eds.): Negotiating adolescence in times of social change. Cambridge: Cambridge University Press: 201-233. Curtin, Sally C.; Ventura, Stephanie J.; Martinez, Gladys M. 2014: Recent Declines in Nonmarital Childbearing in the United States. In: NCHS data brief 162: 1-8. Davies, Patrick T. et al. 2002: Child emotional security and interparental confl ict. Boston, MA: Blackwell. Duncan, Greg J.; Brooks-Gunn, Jeanne (Eds.) 1997: The consequences of growing up poor. New York: Russell Sage Foundation. Ehmke, Timo et al. 2004: Familiäre Lebensverhältnisse, Bildungsbeteiligung und Kom- petenzerwerb. In: Deutschland, Pisa-Konsortium (Ed.): PISA 2003. Der Bildungsstand der Jugendlichen in Deutschland Ergebnisse des zweiten internationalen Vergleichs. Münster: Waxmann. Emery, Robert E. 1988: Marriage, divorce, and children’s adjustment. Newbury Park, CA: Sage. Emery, Robert E. et al. 1999: Delinquent behavior, future divorce or nonmarital child- bearing, and externalizing behavior among offspring: A 14-year prospective study. In: Journal of Family Psychology 13,4: 568-579 [doi: 10.1037/0893-3200.13.4.568]. Erel, Osnat; Burman, Bonnie 1995: Interrelatedness of marital relations and parent- child-relations: A meta-analytic review. In: Psychological Bulletin 118,1: 108-132 [doi: 10.1037/0033-2909.118.1.108]. Eurostat 2013: European social statistics. 2013 edition. Luxembourg: European Union. Fergusson, D. David M.; Horwood, L. John; Lynskey, Michael T. 1992: Family change, parental discord, and early offending. In: Journal of Child Psychology and Psychiatry Review 33: 1059-1075 [doi: 10.1111/j.1469-7610.1992.tb00925.x]. Fergusson, David M; Horwood, L. John; Lynskey, Michael T. 1994: Parental separation, adolescent psychopathology, and problem behaviors. In: Journal of the American Academy of Child & Adolescent Psychiatry 33,8: 1122-1133 [doi: 10.1097/00004583- 199410000-00008]. Fincham, Frank D. 1998: Child development and marital relations. In: Child Development 69,2: 543-574 [doi: 10.2307/1132183]. Forehand, Rex; Armistead, Lisa; David, Corinne 1997: Is adolescent adjustment follow- ing parental divorce a function of predivorce adjustment? In: Journal of Abnormal Child Psychology 25,2: 157-164 [doi: 10.1023/A:1025735629914]. Effects of Parental Separation • 361 Francesconi, Marco; Jenkins, Stephen P.; Siedler, Thomas 2010: Childhood family struc- ture and schooling outcomes: Evidence for Germany. In: Journal of Population Eco- nomics 23,3: 1073-1103 [doi: 10.1007/s00148-009-0242-y]. Garriga, Anna; Härkönen, Juho 2009: The effects of marital instability on children’s well- being and intergenerational relations. In: EQUALSOC State-of-the-art report. Pompeu Fabra University and Stockholm University. Gloger-Tippelt, Gabriele; König, Lilith 2007: Attachment representations of 6-year-old children from one and two parent families in Germany. In: School Psychology Interna- tional 28,3: 313-330 [doi: 10.1177/0143034307078540]. Hawkins, Daniel N.; Amato, Paul R.; King, Valerie 2007: Nonresident father involvement and adolescent well-being: Father effects or child effects? In: American Sociological Review 72,6: 990-1010 [doi: 10.1177/000312240707200607]. Hetherington, Eileen Mavis; Jodl, Kathleen M. 1994: Stepfamilies as settings for child development. In: Booth, Alan; Dunn, Judy (Eds.): Stepfamilies. Who benefi ts? Who does not? Hillsdale, N.J.: Lawrence Erlbaum: 55-79. Hetherington, Eileen Mavis; Stanley-Hagan, Margaret 1999: The adjustment of children with divorced parents: a risk and resiliency perspective. In: Journal of Child Psychol- ogy and Psychiatry 40,1: 129-140 [doi: 10.1177/000312240707200607]. Hetherington, E. Mavis; Kelly, John 2002: For better or for worse. Divorce reconsidered. New York: Norton. Hilmert, Steffen 2002: Familiale Resourcen und Bildungschancen: Konsequenzen eines frühzeitigen Elternverlustes (Family Resources and Educational Opportunities: Con- sequences of Early Loss of Parents). In: Zeitschrift für Familienforschung 14: 44-69. Huinink, Johannes et al. 2011: Panel Analysis of Intimate Relationships and Family Dy- namics (pairfam) Conceptual Framework and Design. In: Zeitschrift für Familien- forschung/Journal of Family Research 23,1: 77-100. Jeynes, William H. 2006: The impact of parental remarriage on children: A meta-analy- sis. In: Marriage & family review 40,4: 75-102 [doi: 10.1300/J002v40n04_05]. Krishnakumar, Ambika; Buehler, Cheryl 2000: Interparental confl ict and parenting be- haviors. A meta-analytic review. In: Family Relations 49: 25-44 [doi: 10.1111/j.1741- 3729.2000.00025.x]. Liefbroer, Aart C.; Dourleijn, Edith 2006: Unmarried cohabitation and union stability: Testing the role of diffusion using data from 16 European countries. In: Demography 43,2: 203-221 [doi: 10.1353/dem.2006.0018]. McLanahan, Sara 1999: Father absence and the welfare of children. In: Hetherington, Eileen Mavis (Ed.): Coping with divorce, single parenting, and remarriage. A risk and resilience perspective. Mahwah: Lawrence Erlbaum Associates: 117-146. Nauck, Bernhard et al. 2013: The German Family Panel (pairfam). GESIS Data Archive, Cologne. ZA5678 Data fi le Version 4.0.0, [doi:10.4232/pairfam.5678.4.0.0]. O’Connor, Thomas G. et al. 2000: Are associations between parental divorce and chil- dren’s adjustment genetically mediated? An adoption study. In: Developmental Psy- chology 36,4: 429-437 [doi: 10.1037/0012-1649.36.4.429]. Rosenberg, Morris 1965: Society and the adolescent self-image. Princeton: Princeton University Press. Ross, Catherine E.; Mirowsky, John 1999: Parental divorce, life-course disrup- tion, and adult depression. In: Journal of Marriage and Family 61: 1034-1045 [doi: 10.2307/354022]. • Sabine Walper, Carolin Thönnissen, Philipp Alt362 Sanz-De-Galdeano, Anna; Vuri, Daniela 2007: Parental Divorce and Students’ Perfor- mance: Evidence from Longitudinal Data. In: Oxford Bulletin of Economics and Statis- tics 69,3: 321-338 [doi: 10.1111/j.1468-0084.2006.00199.x]. Schick, Andreas 2002: Behavioral and emotional differences between children of di- vorce and children from intact famlies: Clinical signifi cance and mediating processes. In: Swiss Journal of Psychology 61,1: 5-14 [doi: 10.1024//1421-0185.61.1.5]. Schmidt-Denter, Ulrich 2000: Entwicklung von Trennungs- und Scheidungsfamilien: Die Kölner Längsschnittstudie. In: Schneewind, Klaus A. (Ed.): Familienpsychologie im Aufwind. Brückenschläge zwischen Forschung und Praxis. Göttingen: Hogrefe: 203- 221. Schmidt-Denter, Ulrich 2001: Differentielle Entwicklungsverläufe von Scheidungs- kindern. In: Walper, Sabine; Pekrun, Reinhard (Eds.): Familie und Entwicklung. Ak- tuelle Perspektiven der Familienpsychologie. Göttingen: Hogrefe: 292-313. Schnor, Christine 2012: Trennungsrisiko von Paaren mit Kindern: Der Einfl uss der Re- ligion in West- und Ostdeutschland. In: Huinink, Johannes; Kreyenfeld, Michaela; Trappe, Heike (Eds.): Familie und Partnerschaft in Ost- und Westdeutschland: Ähnlich und doch immer noch anders (Sonderheft 9 der Zeitschrift für Familienforschung). Leverkusen: Verlag Barbara Budrich: 229-255. Schwarz, Beate 1999: Die Entwicklung Jugendlicher in Scheidungsfamilien. Weinheim: Beltz PVU. Schwarz, Beate; Walper, Sabine; Gödde, Mechtild; Jurasic, Stephanie 1997: Dokumen- tation der Erhebungsinstrumente der 1. Haupterhebung (überarb. Version). Berichte aus der Arbeitsgruppe „Familienentwicklung nach der Trennung“ # 14/1997. Sigle-Rushton, Wendy; Hobcraft, John; Kiernan, Kathleen 2005: Parental Divorce and Subsequent Disadvantage: A Cross-Cohort Comparison. In: Demography 42,3: 427- 446 [doi: 10.1353/dem.2005.0026]. Steinbach, Anja 2008: Stieffamilien in Deutschland. In: Zeitschrift für Bevölkerungswis- senschaft 33,2: 153-180 [doi: 10.1007/s12523-009-0009-2]. Steinhausen, Hans-Christoph; von Aster, Sigrid; Goebel, Dietmar 1987: Family compo- sition and child psychiatric disorders. In: Journal of the American Academy of Child & Adolescent Psychiatry 26,2: 242-247 [doi: 10.1097/00004583-198703000-00022]. Strohschein, Lisa 2005: Parental divorce and child mental health trajectories. In: Journal of Marriage and Family 67,5: 1286-1300 [doi: 10.1111/j.1741-3737.2005.00217.x]. Sweeney, Megan M. 2010: Remarriage and Stepfamilies: Strategic sites for scholarship in the 21st century. In: Journal of Marriage and Family 72,3: 667-684 [doi: 10.1111/j.1741- 3737.2010.00724.x]. Thönnissen, Carolin et al. 2014: Scales Manual of the German Family Panel. Waves 1 to 5. Release 5.0 (with contributions from: Petra Buhr, Veronika Salzburger, Nadia Lois and Claudia Schmiedeberg) [http://www.pairfam.de/fi leadmin/user_upload/redak- teur/publis/Dokumentation/Manuals/Scales_Manual_pairfam_5.0.pdf]. Wallerstein, Judith S.; Corbin, Shauna B.; Lewis, Julia M. 1988: Children of divorce: A ten-year study. In: Hetherington, Eileen Mavis; Arasteh, J. D. (Eds.): Impact of divorce, single parenting, and stepparenting on children. Hilldale, N.J.: Erlbaum: 198-214. Walper, Sabine 2002: Einfl üsse von Trennung und neuer Partnerschaft der Eltern. Ein Vergleich von Jungen und Mädchen in Ost- und Westdeutschland. In: Zeitschrift für Soziologie der Erziehung und Sozialisation 22,1: 25-46. Effects of Parental Separation • 363 Walper, Sabine 2009: Links of perceived economic deprivation to adolescents’ well- being six years later. In: Zeitschrift für Familienforschung/Journal of Family Research 21: 107-127. Walper, Sabine; Beckh, Katharina 2006: Adolescents’ development in high-confl ict and separated families. Evidence from a German longitudinal study. In: Clarke-Stewart, Alison; Dunn, Judy (Eds.): Families count: Effects on child and adolescent develop- ment. Cambridge, MA: Cambridge University Press: 238-270. Walper, Sabine; Krey, Mari 2009: Familienbeziehungen nach Trennungen. In: Lenz, Karl; Nestmann, Frank (Eds.): Persönliche Beziehungen. Weinheim: Juventa. Walper, Sabine; Langmeyer, Alexandra 2014: Auswirkungen einer elterlichen Trennung auf die Entwicklung von Kindern in den ersten Lebensjahren. In: Kißgen, Rüdiger; Hei- nen, Norbert (Eds.): Trennung, Tod und Trauer in den ersten Lebensjahren: Begleitung und Beratung von Kindern und Eltern. Stuttgart: Klett-Cotta: 159-176. Walper, Sabine; Wendt, Eva-Verena 2005: Nicht mit beiden Eltern aufwachsen ein Ri- siko? Kinder von Alleinerziehenden und Stieffamilien. In: Alt, Christian (Ed.): Kinderle- ben Aufwachsen zwischen Familie, Freunden und Institutionen. Band 1: Aufwachsen in Familien. Wiesbaden: VS Verlag für Sozialwissenschaften. DJI Deutsches Jugend- institut: 187-216. Walper, Sabine; Wendt, Eva-Verena 2011: Die Bedeutung der Abstammung für die Iden- titätsfi ndung und Persönlichkeitsentwicklung in der Adoleszenz: Adoption, Samen- spende und frühe Vaterabwesenheit nach Trennung der Eltern. In: Vaskovics, Laszlo (Ed.): Pluralisierung der Elternschaft und Kindschaft (Sonderheft der Zeitschrift für Familienforschung): 211-237. Wendt, Eva-Verena; Walper, Sabine 2007: Entwicklungsverläufe von Kindern in Ein- Eltern- und Stieffamilien. In: Alt, Christian (Ed.): Kinderleben Start in die Grundschu- le Band 3: Ergebnisse aus der zweiten Welle. Wiesbaden: VS Verlag für Sozialwissen- schaften: 211-242. Whiteside, Mary F; Becker, Betsy Jane 2000: Parental factors and the young child’s post- divorce adjustment: a meta-analysis with implications for parenting arrangements. In: Journal of Family Psychology 14,1: 5-26 [doi: 10.1037/0893-3200.14.1.5]. Date of submission: 14.11.2014 Date of acceptance: 21.09.2015 Prof. Dr. Sabine Walper (). German Youth Institute. Munich, Germany. E-mail: walper@dji.de URL: http://www.dji.de/~walper Ludwig-Maximilians-University of Munich. Munich, Germany. URL: http://www.dji.de/~walperl Dr. Carolin Thönnissen, Philipp Alt. Ludwig-Maximilians-University of Munich. Munich, Germany. E-mail: c.thoennissen@edu.lmu.de, Philipp.Alt@edu.lmu.de URL: http://www.edu.lmu.de/apb/personen/wiss_ma/thoennissen/index.html http://www.edu.lmu.de/apb/personen/wiss_ma/alt/index.html Published by / Herausgegeben von Prof. Dr. Norbert F. Schneider Federal Institute for Population Research D-65180 Wiesbaden / Germany Managing Editor / Verantwortlicher Redakteur Frank Swiaczny Assistant Managing Editor / Stellvertretende Redakteurin Katrin Schiefer Copy Editor (German) / Lektorat (deutsch) Dr. Evelyn Grünheid Layout / Satz Beatriz Feiler-Fuchs E-mail: cpos@bib.bund.de Scientifi c Advisory Board / Wissenschaftlicher Beirat Paul Gans (Mannheim) Johannes Huinink (Bremen) Michaela Kreyenfeld (Rostock) Marc Luy (Wien) Clara H. Mulder (Groningen) Notburga Ott (Bochum) Peter Preisendörfer (Mainz) Zsolt Spéder (Budapest) Comparative Population Studies www.comparativepopulationstudies.de ISSN: 1869-8980 (Print) – 1869-8999 (Internet) Board of Reviewers / Gutachterbeirat Martin Abraham (Erlangen) Laura Bernardi (Lausanne) Hansjörg Bucher (Bonn) Claudia Diehl (Konstanz) Andreas Diekmann (Zürich) Gabriele Doblhammer-Reiter (Rostock) Jürgen Dorbritz (Wiesbaden) Anette Eva Fasang (Berlin) E.-Jürgen Flöthmann (Bielefeld) Alexia Fürnkranz-Prskawetz (Wien) Beat Fux (Salzburg) Joshua Goldstein (Berkeley) Karsten Hank (Köln) Sonja Haug (Regensburg) Hill Kulu (Liverpool) Aart C. Liefbroer (Den Haag) Kurt Lüscher (Konstanz) Emma Lundholm (Umeå) Nadja Milewski (Rostock) Dimiter Philipov (Wien) Roland Rau (Rostock) Tomáš Sobotka (Wien) Jeroen Spijker (Barcelona) Olivier Thévenon (Paris) Helga de Valk (Brussel) Heike Trappe (Rostock) Michael Wagner (Köln) © Federal Institute for Population Research 2015 – All rights reserved