Running head: MARITAL CONFLICT AND DIVORCE FOR CHILD DEVELOPMENT 1 Consequences of Marital Conflict and Divorce for Child Development in South Korea

Literature on effects of marital conflict and divorce on child development has long wondered 1) whether divorce followed by parents’ marital conflict exerts distinguishable impacts on children and 2) whether marital conflict is more detrimental to involved children than divorce per se. We address these questions using the Korea Youth Panel Study-Elementary School Students that traced students from 4 grade in 2003 to 8 grade in 2007. Our results indicate 1) that children suffered from parental divorce as well as marital conflict, 2) that the adverse effect seems larger for parental divorce than marital conflict, 3) that there was domain specificity for differential effect of parental divorce by preceding marital conflict, and 4) that children who were under parents’ marital conflict were further disadvantaged if their parents decided to end their marriage. MARITAL CONFLICT AND DIVORCE FOR CHILD DEVELOPMENT 3 Consequences of Marital Conflict and Divorce for Child Development in South Korea Numerous studies have repeatedly reported that parental divorce is detrimental for life chances of those children involved in the process (Amato & Keith, 1991; Hetherington, 1979). Children of divorce, for instance, were found to be more likely to lag in cognitive skills indexed by math and reading test scores than children of intact families from the in-divorce stage onward (Kim, 2011; Sun and Li, 2001). Parental divorce also tended to enhance the risk of high school dropout and lower the likelihood of progress to college (Aston & McLanahan, 1994). Children of divorce appeared to go through a spell of emotional disruption more frequently than their counterparts as well. Using the National Child Development Study, for instance, Cherlin, ChaseLandsdale, and McRae (1998) showed that children of divorce were in a worse mental health compared to children of intact families even after taking various selection factors into account. This quite consistent finding of adverse effect of parental divorce is not restricted geographically to the United States and European Hemisphere. Strong body of literature in South Korea (hereafter, Korea) has established theoretical mechanisms as well as empirical evidence associating observable setbacks in child development with parental divorce (for instance, Jeong, 2011; Ji & Lee, 2012). However, a dominant portion of studies in Korea as well as in Western countries suffer from methodological problems especially because they used crosssectional data obtained from convenient samples (see, for instance, Cherlin et al., 1991; Kim, 2011 for a few notable exceptions). More importantly, a couple of theoretical points are not well understood yet regarding distinct impacts of marital conflict and parental divorce. First, we would like to know whether there is a difference in negative impacts of parental divorce depending on whether children go MARITAL CONFLICT AND DIVORCE FOR CHILD DEVELOPMENT 4 through preceding marital conflict between their parents or not . An answer from this question would illuminate whether preceding marital conflict is a moderator of the divorce effect. More specifically, we are interested in finding distinguishable effect of parental divorce on children who undergo preceding marital conflict. On the other hand, many scholars have wondered a relative strength of marital conflict and divorce per se. For instance, Mechanic and Hansell (1989) found that marital conflict was likely to enhance adolescents’ depressed mood, anxiety, and physical symptoms while divorce per se was not related to changes in health outcomes. In an attempt to close these gaps in the literature, we first tackle the problem of whether there is noticeable effect of parental divorce on diverse dimensions of child development such as cognitive skills and non-cognitive traits in Korea. Next we attempt to unveil whether effect of parental divorce would be heterogeneous by parents’ marital discord that children observed before marital dissolution. If we are to find a similar effect size between two groups of conflictridden marriage and decent marriage, we may infer that marital conflict is not the only one mediator connecting divorce to adverse consequences and there is something unique in the divorce process that gives parental divorce its negative power. Otherwise, we may conclude that a large part of the divorce effect would flow through marital conflict. However, these analytical plan does not illuminate if marital conflict between two parents, whether it leads to divorce or not, inflict negative impacts on child development. This is just because children of intact families include children who suffer from marital conflict. Based on this recognition, we attempt to evaluate effects of parental marital conflict on child development using marital conflict measures between two adjacent survey waves. Namely, we make a variable with four categories depending on existence of marital conflict at time 1 and time 2: 1) no conflict-no conflict, 2) no conflict-in conflict, 3) in conflict no conflict, 4) in conflict in MARITAL CONFLICT AND DIVORCE FOR CHILD DEVELOPMENT 5 conflict. By comparing children in the other three categories with those in the first category, we can appraise impacts of marital conflict on child development. Notice, though, that divorce occurred between time 1 and time 2 can be included any category in the variable such that we are not allowed to examine impacts of parental divorce in this study design. To figure out complex dimensions of effects of parental divorce and marital conflict, in the third study, we propose a study design in which we classify children into six mutually exclusive and exhaustive categories according to existence of marital conflict at time 1 and marital outcome at time 2: 1) no conflict-no conflict, 2) no conflictin conflict, 3) in conflict-no conflict, 4) in conflict-in conflict, 5) no conflict-divorce, 6) in conflict-divorce. By comparing children in the other five categories with those in the first categories, we are in a better position to assess a specific effect of parental divorce and marital conflict on child development. To implement abovementioned study designs, we will analyze the Korea Youth Panel SurveyElementary School Students (KYPS). The panel study has traced students from 4 grade in elementary school in 2004 until 8 grade in 2008. But we investigate data from the first five waves due to data disclosure policy from the data collectors. From a methodological point of view, we employ the framework of the traditional ordinary least squares complemented by the counterfactual framework utilizing a matching estimator in R (Sekhon, 2011). Data and Measurement