As researchers have dedicated more attention to the study of romantic involvement across the life course, there is evidence of continuity across development, especially when considering the impact of family context. The family context represents one of the most important contexts in which individuals develop (Gray & Steinberg, 1999; Collins & Laursen, 2004; Buhrmester & Furman, 1987). The family context is consistently associated with romantic relationships (e.g., Conger, Cui, Bryant, & Elder, 2000; Feldman, Gowen, & Fisher, 1998; Benson, Larson, Wilson & Demo, 1993). Even the quality of parent-child relationships early in life have been shown to impact individuals' capacity for close relationships as they mature (e.g. Sroufe & Waters, 1977; Furman, Simon, Shaffer, & Bouchey, 2002).
Characteristics of the family, particularly marital instability, have been highlighted as consequential for individuals' romantic involvement. As children spend increasing time growing up outside of two-biological married parents (Bumpass & Lu, 2000; McLanahan & Sandefur, 1994), researchers have noted both the long and short term consequences of family instability. Long term implications of marital instability are evidenced in the intergenerational transmission of divorce where exposure to parental divorce is associated with an greater likelihood of children's divorce (Amato & DeBoer, 2001). This elevated risk of divorce may be due to entry into marriage at younger ages and with less education (Bumpass, Martin, & Sweet, 1991; Mueller & Pope, 1977; Thornton, 1991). Children whose parents divorced also report poorer quality relationships in their own marriages (Amato & Booth, 1991; Amato & Booth, 2001).
Others have focused on more proximal effects of marital instability by exploring links between family instability and sexual activity during adolescence. Biological married parents provide a buffer against precocious sexual activity as adolescents from divorced or single parent families face a significantly greater risk of initiating sexual activity earlier than their peers (e.g., Santelli, Lowry, Brener, & Robin, 2000; White & DeBlassie, 1992; Lammers, Ireland, Resnick, & Blum, 2000; Davis & Friel, 2001; Johnson & Tyler, 2007; Quinlan, 2003). Father absence is particularly risky for females, though family disruption predicts early sexual debut for males (Newcomber & Udry 1987; Ellis, et al., 2003; Campa & Eckenrode, 2006). The presence of a stepfather is associated with a greater risk of early sexual debut for adolescents though adolescents with higher rates of stepfather involvement report more restrictive attitudes regarding sexual activity (Quinlan, 2003; Day, 1992; Menning, Holtzman, & Kapinus, 2007). These effects have been shown to be independent of other family processes such as parental involvement (Pearson, Muller, & Frisco, 2006; South, Haynie, & Bose, 2005).
New research has examined the effects of family instability with a focus on adolescent romantic involvement. Adolescents from two-biological parent families were less likely to report romantic involvement than those from step or single parent families (Cavanagh, Crissey, & Raley, in press) The cumulative effects of multiple marital transitions also increased the odds that adolescents would be romantically involved. Marital transitions during middle childhood and early adolescence had the strongest impact on adolescents' romantic involvement. Adolescents who were exposed to marital disruption also begin dating earlier than those whose parents were continuously married (Sullivan, 2008).
The literature on the short and long term effects of divorce show consistent effects of family instability on romantic involvement across the life course. Exposure to divorce is associated with poorer marital outcomes during adulthood and accelerated romantic and sexual activity during adolescence. Yet it remains unclear how exposure to divorce impacts the emergence and development of romantic relationships earlier in life during childhood and preadolescence. The primary goal of this chapter is to explore links between family instability and children and preadolescents' romantic interest and involvement, with a particular focus on the effects of parental repartnering following divorce. This chapter takes advantage of the Texas Families Project (TFP) design to explore a more nuanced definition of parental repartnering. Unlike previous literature which generally only examines divorce and marriage, there are additional transitions that children are exposed to as mothers and fathers recover from divorce and reenter the romantic domain that warrant study (Heard, 2007; Amato, 2005; Davis & Friel, 2001). For example, children may witness parental repartnering such as cohabitation (e.g., Brown, 2006) and dating, break-ups, sleep-overs (e.g., Anderson & Greene, 2005; Anderson, et al., 2004). After a review of the family processes that may be impacted by maternal repartnering and divorce, hypotheses regarding the impact of maternal repartnering on children and preadolescents' romantic interest and involvement will be discussed.
Adapting related literature to fit a younger sample (e.g., Cavanagh, et al., in press; Sullivan, 2008), the effects of family instability may be influential on the developmental trajectory of children and preadolescents' romantic interest and development. This may be particularly true as previous research has shown that divorce that occurs during middle childhood and preadolescence has the strongest effects on children's outcomes (Cavanagh, et al., 2008; Amato, 1996). Researchers have noted three specific aspects of the family context in trying to uncover the processes that account for these effects: family socialization, control, and involvement.
The family is an important source of information about appropriate behavior (Davis & Friel, 2001). The socialization of attitudes, norms, and behaviors with a romantic and sexual context originate within the home environment through parents' behaviors and attitudes (Thornton & Camburn, 1987). For example, the marital relationship, particularly the types of attitudes and behaviors parents exhibit, provides children with a model of forming and maintaining romantic relationships (Crockett, Bingham, Chopak, & Vicary, 1996; Bandura, 1977). The marriage provides children with a model of characteristics and commitment in close relationships (Wolfinger, 2005). Children may internalize different attitudes regarding romantic involvement depending on their exposure to family instability. Children who grow up in families with continuously married parents may be more likely to expect and value long term, stable relationships. These types of intimate relationships have high barriers to entry, making it likely for these individuals to postpone romantic involvement. Children who experience marital disruption may perceive romantic relationships as transient following exposure to the formation and dissolution of their parents' marital and romantic relationships.
However, this chapter focuses on the experiences of children and preadolescents following parental divorce. For these indivdiuals, I am interested in the variation in romantic interest and involvement that can be attributed to mothers and fathers' romantic repartnering in the two years following divorce. It is easy to predict the consequences of models of romantic relationships in instances of very low or very high levels parental repartnering. Parents who are involved in few or no romantic relationships following divorce may present a model that romantic relationships are serious and not to be entered lightly. They may may signal that the breakup of romantic relationships is very painful and children and preadolescents should be hesitant to open themselves to such pain. Family relationships compared to relationships with nonfamily members may be emphasized in these households. The transmission of these messages would be evidenced in low or delayed romantic interest and involvement among children and preadolescents. Parents who report high rates of repartnering may impress an entirely different model of romance on their children. Higher rates of instability and exposure to parental repartnering may increase adolescents' awareness of romantic relationships. Greater awareness of romances may accelerate children and preadolescents' attention to and involvement in opposite sex relationships.
It is less clear what models parents provide their children when they are involved in moderate amounts of involvement in romantic relationships. Parents' romantic involvement may signal recovery from the divorce and an appropriate type of relationship to establish. Moderate involvement may make children and preadolescents more aware of romance but in doses they are better equipped to handle. These children and preadolescents' romantic involvement and interest may fall between those who are exposed to low or high rates of repartnering.
Additionally, it is uncertain if involvement in specific types of relationships will have different effects on children and preadolescents' romantic interest and involvement. If parents provide models of involvement only serious and stable relationships following divorce, these experiences may more closely resemble the experiences children have from continuously married families. When parents limit their romantic involvement to stable and committed relationships, children may report depressed romantic interest and involvement. In instances where parents are involved in more casual relationships, children and preadolescents may perceive romance as more accessible to them given the temporary nature modeled by parents.
Inherent in this discussion of parental repartnering is exposure to breakups. It is unclear what model of romance children and preadolescents may internalize in the face of parents' breakups. Parents' breakups may serve as additional evidence of the transient nature of romantic relationships for some children and preadolescents, thereby reducing their perceived barriers to these relationships. However, breakups may also expose children and preadolescents to the heartache that can accompany romantic involvement. This may parallel the pain they witnessed during divorce and result in reluctance to expose themselves to the pain and hurt of breakups.
Following divorce, there is some pressure for parents to recover and reengage romantic involvement (Hetherington & Kelly, 2002; Anderson & Greene, 2005; Anderson, Greene, Walker, Malerba, Forgatch, & DeGarmo, 2003). Movement back into the romantic domain may be precipitated by changes in parents' attitudes and behaviors regarding romantic involvement. There is evidence that these attitudes may also contribute to the models of relationships that children internalize (Whitbeck, Simons, & Kao, 1994). Parents' repartnering attitudes and behaviors including greater interest in and efforts to date may also increase children and preadolescents' awareness of romantic relationships.
Instability also has implications for parents' control of children and preadolescents due to changes in the makeup of the household (Thompson, McLanahan, & Curtin, 1992; Thornton, 1991). Marital dissolution often means that children primarily live with one parent, thereby reducing the number of adults in the household that can provide oversight (McLanahan & Sandefur, 1994; Hogan & Kitagawa, 1985). Single parents are often less equipped to provide children with guidance, rules, and monitoring in the wake of instability (Lauritsen, 1994; Newcomber & Udry, 1987; Cottrell et al., 2003). Reductions in the amount of monitoring that parents can provide their children has implications for their romantic and sexual behavior (Longmore, Manning, & Giordano, 2001). In light of reduced control, children and preadolescents may be more free to develop romantic interest and involvement. Parents may be able to place fewer restrictions on children's attitudes and behaviors, resulting in accelerated romantic interest and involvement.
Changes in the family environment can also result from the addition of new household members such as a parent's romantic partner. Romantic partners are often not considered a legitimate authority figure in the household, weakening parental control over children (Baxter, Braithwaite, Bryant, & Wagner, 2004). Reduced control over children resulting from changes in household structure leave parents less able to exert rules and restrictions. In households characterized by greater parental repartnering, children and preadolescents may have fewer restrictions on romantic interest and involvement. The ramifications of reduced parental control may be evidenced in reduced supervision and restrictions on children and preadolescents and accelerated rates of romantic interest and involvement. The introduction of romantic partners may aggravate the accelerated development of romantic interest and involvement.
Following instability, parents struggle to adapt to changes in the family (Amato & Booth, 1996; Zill, Morrison, & Coiro, 1993; Hines, 1997). Whether recovering from marital dissolution or adapting to new family members following a marriage, parents often are less able to devote time and attention to their children due to the stress and poorer communication, interactions, and relationship quality that accompanies instability (Lauritsen, 1994). This leaves children without the emotional support and closeness that they require (Sweeney, 2007; Cavanagh & Huston, 2006). Children may react to the loss of parental support and closeness by seeking out alternative relationships. Children and preadolescents who are exposed to higher rates of repartnering may be compelled to form their own romantic relationships as a substitute for parent-child relationships.
Parents' attitudes may also impact children's perceptions of family involvement. Following divorce, many parents feel the added pressure of being a single parent while also balancing their new social life (Anderson & Greene, 2005). The strain of balancing two different aspects of their new life may lead parents to place more focus on one role versus another. The parents who desire to develop their romantic lives may be more likely to have children who seek out romantic relationships to replace the loss of the parent-child relationship. Parents who place more emphasis on their role as a mother may have stronger relationships with their children, thereby reducing children's need to seek out alternative sources of closeness and support.
Data for this chapter will focus on parental repartnering characteristics following divorce. The analyses will examine the effects of both parental repartnering behaviors and maternal attitudes. Data on paternal repartnering was collected at each yearly assessment. Recruitment for fathers' participation in the study required mothers approval, which was rarely given (N = ). In this chapter, mothers' reports of fathers' repartnering will be used. Mothers were asked to provide their best guess about fathers repartnering status. Mothers only reported on the repartnering behaviors of which their children were aware and do not provide information about fathers' attitudes and efforts to repartner.
Data on maternal repartnering behaviors were collected continuously throughout participation in the study. Thus, extensive data management and cleaning procedures must be undertaken before any analyses can be completed. One of the predominant strengths of the TFP was the level of attention given to transitions in maternal repartnering characteristics. Besides collecting information at the six and 12 month assessments, information about romantic involvement was also collected on a monthly basis to track changes in mothers' romantic lives. To fit data collected on a monthly basis to the multilevel model with three time points, I will compile the information collected between the baseline to 12 month and 12 to 24 month assessments to compute several key categorical variables, including involvement in a romantic relationship, sleep over, cohabitation, marriage, or break-up. These variables will reflect whether the target child was exposed to any of these relationships during the previous 12 months. Data for history of romantic involvement at baseline will be compiled from information provided by the mother describing her romances since separation.
To help ensure a model can be fitted to this data, I have attempted to include a concise yet comprehensive set of variables. For example, I have tried to compile relationship information into relatively few categorical variables. For example, cohabitation and sleep over both include the suggestion of a sexual relationship where the mother and romantic partner spend the night together in the same house as the child. Yet I believe that the potentially more formal or serious nature of cohabitations may have different implications for children and preadolescents' romantic interest and involvement than exposure to sleep overs. This example remains an empirical question however and patterns in the data will determine if these are qualitatively different experiences.
Another data management issue revolves around how to best classify target children's exposure to repartnering. If a child is exposed to both a romantic relationship, cohabitation, and breakup within a year, what is the best way to describe his or her experiences? I will explore several different avenues of capturing children's experiences. This may include a hierarchical breakdown of children's experiences where children are grouped based on their mothers' most serious relationship. It may also be better informed after simply looking at the patterns evidenced in maternal repartnering. I also will create a variable that represents the sum of all transitions children have experienced by the 24 month assessment.
Alternative measures of maternal repartnering characteristics will be assessed through maternal attitudes regarding repartnering following divorce. These analyses will determine if maternal attitudes have an independent effect on their children's romantic interest and involvement. Three main attitudes will be assessed: mothers' efforts to repartner, interest in forming a romantic relationship, and role strain. Efforts to repartner are assessed through mothers' reports about the activities she would consider or has done to meet new romantic partners. Thirteen items and three free response items were completed by every mother. Items included activities such as "Gone to bars, clubs, or parties," "Joined a gym, visited parks, or attended sporting events," and "Placed your profile on an internet dating service." The variable is computed through a count of all the activities the mother reported she had done.
Finally role strain is a measurement of mothers' struggle to balance her own needs and the needs of her children. Role strain will be assessed through up to 36 items as the measurement of this construct is still being finalized. Further analysis is required to determine the final factor structure. Items such as "I think of myself as a mother before any other role in my life" and "I would not marry someone my child disliked" may be included.
Multilevel modeling techniques will also be employed in the investigation of the effects of maternal repartnering on children and preadolescents' romantic interest and involvement. Building on the base model from Equation 3.1, three models will be examined in this chapter. The first model will include variables regarding maternal repartnering behaviors. Categorical variables measuring romantic involvement, cohabitational relationships, marriage, and break-ups will be examined. The reference category will be the children who were not exposed to repartnering during study participation. The second model will include a continuous variable measuring the total number of repartnering transitions the child experienced entered into the equation for the intercept.
Level 1
ROMij = π0j + π1jAGE + π2jDAT + π3jCOH + π4jMAR + π5jBRK + εij
Level 2
π0j = γ00 + ζ0j
π1j = γ10 + ζ1j
π2j = γ20 + ζ2j
π3j = γ30 + ζ3j
π4j = γ40 + ζ4j
π5j = γ50 + ζ5j
[Equation 5.1]
The third model will include maternal repartnering attitudes. These analyses will include continuous measures of mothers' efforts to date, interest in forming a romantic relationship, and role strain (Equation 5.2). The fourth model will include both maternal repartnering behaviors and attitudes. At this point it is unclear what additional control variables will be included in these models. Depending on the results from Chapter 4, specifically the analyses modeling the effects of gender, age, and race/ethnicity, these variables may or may not be included here.
Level 1
ROMij = π0j + π1jAGE + π2jEFF + π3jINT + π4jROL + εij
Level 2
π0j = γ00 + ζ0j
π1j = γ10 + ζ1j
π2j = γ20 + ζ2j
π3j = γ30 + ζ3j
π4j = γ40 + ζ4j
[Equation 5.2]