View in EDS HTML Full Text PDF Full Text

Mutual Influence of Marital Conflict and Children's Behavior Problems: Shared and Nonshared Family Risks

Saved in:
Bibliographic Details
Title: Mutual Influence of Marital Conflict and Children's Behavior Problems: Shared and Nonshared Family Risks
Language: English
Authors: Jenkins, Jennifer, Simpson, Anna, Dunn, Judy
Source: Child Development. Jan 2005 76(1):24-39.
Availability: Journal Customer Services, Blackwell Publishing, 350 Main Street, Malden, MA 02148. Tel: 800-835-6770 (Toll Free); Fax: 781-388-8232; e-mail: subscrip@bos.blackwellpublishing.com.
Peer Reviewed: Y
Page Count: 16
Publication Date: 2005
Document Type: Journal Articles
Reports - Evaluative
Education Level: Elementary Secondary Education
Descriptors: Child Behavior, Behavior Problems, Siblings, Conflict, Parent Child Relationship, Longitudinal Studies, Family (Sociological Unit), Marital Satisfaction, Children, Adolescents
DOI: 10.1111/j.1467-8624.2005.00827.x
ISSN: 0009-3920
Abstract: This within-family, longitudinal study including biological and stepfamilies investigated mutual influences between marital conflict and children's behavior problems. Children (4 to 17 years; N296) residing in 127 families drawn from a general population study were investigated at Time 1 and again 2 years later. These nested data were analyzed using multilevel modeling, controlling for previous child behavior or marital conflict. Marital conflict about children predicted change in children's behavior. Children's behavior also predicted an increase in marital conflict, particularly in stepfamilies. Differences between siblings in exposure to conflict and the extent to which siblings were a source of argument increased more in stepfamilies than in biological families. Boys were exposed to more conflict over time than were girls.
Abstractor: Author
Number of References: 61
Entry Date: 2005
Accession Number: EJ685656
Database: ERIC
Full text is not displayed to guests.
FullText Links:
  – Type: pdflink
    Url: https://content.ebscohost.com/cds/retrieve?content=AQICAHj0k_4E0hTGH8RJwT4gCJyBsGNe_WN95AvKlDbXJGqwxwFOfH5rCItrloQk9EmEu0vnAAAA4jCB3wYJKoZIhvcNAQcGoIHRMIHOAgEAMIHIBgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDGJrQm2f-BrxgHqhLAIBEICBmrqUa5GPdR1nJNTYmcRa33MTGAtvm64wXw9uS-Nz8ScYAPC4JQFidiH1Ha5tQvDbYzsIsYIJVMzqhhbiykDAnS0Gj3fGmUx5BnHegKCjpJgU7oMPaSL7UZsHp8sHEwhLpz_UlcrZnYR0QiaWrsFjffsUy-MrGp5O1NWEZEhKRDh23rrc9YBGiwhB7ucggbX2pg64c2Wyd6dscN0=
Text:
  Availability: 1
  Value: <anid>AN0015916620;cdv01jan.05;2019May30.12:36;v2.2.500</anid> <title id="AN0015916620-1">Mutual Influence of Marital Conflict and Children's Behavior Problems: Shared and Nonshared Family Risks. </title> <p>This within‐family, longitudinal study including biological and stepfamilies investigated mutual influences between marital conflict and children's behavior problems. Children (4 to 17 years; N=296) residing in 127 families drawn from a general population study were investigated at Time 1 and again 2 years later. These nested data were analyzed using multilevel modeling, controlling for previous child behavior or marital conflict. Marital conflict about children predicted change in children's behavior. Children's behavior also predicted an increase in marital conflict, particularly in stepfamilies. Differences between siblings in exposure to conflict and the extent to which siblings were a source of argument increased more in stepfamilies than in biological families. Boys were exposed to more conflict over time than were girls.</p> <p>The relationship between marital conflict and child adjustment is well established ([<reflink idref="bib13" id="ref1">13</reflink>]). Marital conflict has been found to be associated with a wide range of both internalizing (e.g., depression), and externalizing (e.g., aggression) outcomes in children, with the association between the latter and marital conflict being particularly robust ([<reflink idref="bib22" id="ref2">22</reflink>], [<reflink idref="bib36" id="ref3">36</reflink>]). The direction of effect between marital conflict and children's behavior problems is usually assumed to be that marital conflict affects children's behavior. Much of this research, however, has been based on cross‐sectional studies from which conclusions about direction of effect are unwarranted. Our aim in this study was to examine both directions of effect by modeling change in both child behavior and marital conflict.</p> <p>Four previous studies provide more convincing longitudinal evidence that marital conflict influences child behavior. These studies used a design whereby marital conflict at Time 1 predicted child behavior at Time 2, controlling for child behavior at Time 1 ([<reflink idref="bib14" id="ref4">14</reflink>], [<reflink idref="bib25" id="ref5">25</reflink>], [<reflink idref="bib27" id="ref6">27</reflink>], [<reflink idref="bib44" id="ref7">44</reflink>]). Only one of these studies, however, has also tested the opposite direction of effect—that difficult children predict increased conflict in the marital relationship. [<reflink idref="bib44" id="ref8">44</reflink>], in a sample of biological, simple (i.e., siblings all biologically related), and complex stepfamilies (i.e., siblings not biologically related) with adolescent children, followed up families for 3 years and tested two directions of effect. First, they demonstrated that marital conflict at Time 1 predicted an increase in adolescent behavior problems from Time 1 to Time 2. Second, they showed that adolescent's externalizing behavior at Time 1 predicted a change in fathers' ratings of negativity and positivity in the marital relationship between Time 1 and Time 2. This relationship was more evident in stepfamilies than in biological families. Although this study is the only one that has examined both directions of effect at the same time, several other findings lead one to think that difficult children may have an adverse effect on marital conflict. First, couples without children have reported greater satisfaction in their marriages than those with children ([<reflink idref="bib55" id="ref9">55</reflink>]). Second, the birth of a baby has been found to be associated with a decline in marital quality and an increase in marital distress ([<reflink idref="bib11" id="ref10">11</reflink>]). Third, the parents of challenging children either because of children's poor temperament ([<reflink idref="bib38" id="ref11">38</reflink>]) or physical health ([<reflink idref="bib18" id="ref12">18</reflink>]) show less relationship satisfaction and more marital distress. Fourth, in families that experienced divorce, the children exhibited behavioral difficulties before the break‐up ([<reflink idref="bib6" id="ref13">6</reflink>], [<reflink idref="bib8" id="ref14">8</reflink>]), a finding that could indicate that having a problem child contributes to the dissolution of the marriage.</p> <p>Further support for the argument that children and parents may both play a role in the observed relationship between marital conflict and child adjustment is obtained from the type of conflict that has been shown to be most strongly associated with child behavior problems. Specifically, marital conflict that is about the child is more strongly related to child adjustment ([<reflink idref="bib23" id="ref15">23</reflink>], [<reflink idref="bib57" id="ref16">57</reflink>]), and child‐related conflict is associated with increased experience of fear, distress, and a greater likelihood of intervention in children ([<reflink idref="bib23" id="ref17">23</reflink>]). Thus, it may be that child reactions are more negative to child‐related conflict and that these reactions then precipitate more conflict about the child. In the present study we distinguished arguments about the child from two other types of marital conflict. General marital conflict is defined here as conflict that is not child related and included topics such as finances, recreation, friends, and so on. We also included a measure of children's presence in the room during marital conflict as the witnessing of conflict between adults has been found to increase aggression in children in experimental studies ([<reflink idref="bib12" id="ref18">12</reflink>]).</p> <p>Why might we expect that both directions of effect will operate? Within a family systems framework, family members are seen as mutually interdependent ([<reflink idref="bib26" id="ref19">26</reflink>], [<reflink idref="bib42" id="ref20">42</reflink>]). Directions of influence are considered to be reciprocal, with children influencing parental behavior and parents influencing child behavior. One account of this model of reciprocal effects has involved spillover of affect from one part of the family system to another part. [<reflink idref="bib40" id="ref21">40</reflink>] interviewed one parent every day for 2 weeks recording conflict within different subsystems of the family: the marital, parent–child, and sibling subsystems, in distressed and nondistressed families. Sequential analyses revealed that when a conflict occurred between a couple, the likelihood that a conflict would occur between siblings in the next 24 hr was increased among the distressed families. Thus, negative affect becomes pervasive in some families as tension from one subsystem spills over into another subsystem and increases the probability of conflict in that subsystem.</p> <p>The parent–child relationship has also been found to be more problematic in families in which marital conflict is high, and this too has been explained by the spillover concept ([<reflink idref="bib17" id="ref22">17</reflink>]). Indeed, some have argued that marital conflict affects the quality of the parent–child relationship, and it is the negative parent–child relationship rather than the marital conflict that has a negative effect on children's behavior. Evidence for both direct and indirect effects have been found ([<reflink idref="bib28" id="ref23">28</reflink>]). In the present study our aim was not to test the difference between direct and indirect effects but rather to ensure that effects of marital conflict were not explained by the quality of the parent–child relationship. To achieve this, maternal negativity was included in our analysis.</p> <p>The following questions about directions of effect between marital conflict and child behavior were examined. Which aspect of marital conflict predicts increased disturbance in children: argument about children, exposure to conflict, or general marital conflict, after accounting for covariates (Question 1b) and maternal negativity (Question 1c)? Does child internalizing or externalizing behavior predict an increase in marital conflict over time (Questions 2b and 3c)? Question numbers are used in the Analysis Description and Plan section.</p> <hd id="AN0015916620-2">Shared and Nonshared Processes in Families</hd> <p>One of the intriguing findings from research in behavioral genetics has been that siblings are more dissimilar than similar once genetic effects have been controlled ([<reflink idref="bib49" id="ref24">49</reflink>]). This finding informed two issues that were examined in this study. Is marital conflict a risk for one sibling but not another? How different are the experiences of siblings with respect to parental marital conflict? [<reflink idref="bib59" id="ref25">59</reflink>] noted an important distinction that emerges from the literature on behavioral genetics: that between the observed environment and the effective environment.</p> <p>The observed environment refers to characteristics of the environment in which children live, such as the frequency of marital conflict and the family's socioeconomic status. It is directly observable and measurable. The observed environment can be divided further into two components: (a) aspects of the environment shared by all siblings and (b) unique experiences of siblings within families. The frequency with which parents argue is the same for all children, but the frequency at which they argue about specific children varies across siblings. Using measures of the observed environment we can model the mean on children's outcomes (examining the mean of child behavior as a function of the environment).</p> <p>The effective environment, in contrast, is not directly measured. The concept comes from behavioral genetics. In this field, inferences about the operation of the environment are made on the basis of the similarity and differences between siblings on behavior and personality (after accounting for genetic effects), given that siblings are raised in the same family environment. For example, if monozygotic twins are raised in the same environment yet behave very differently, arguably an inference can be made about the effect of that environment. Sibling similarity is based on comparisons of within‐ and between‐family variance.</p> <p>If most siblings are dissimilar to one another on behavior after accounting for genetic effects, this suggests two possibilities for how an environmental stressor such as marital conflict operates on siblings. One process may be that siblings are differentially exposed to the stressor, with such differential exposure helping to account for adjustment differences between children ([<reflink idref="bib49" id="ref26">49</reflink>]). In the present study this issue was examined through the differentiation of family‐level and child‐specific processes. Another process may be that shared or family‐level environmental stressors are important in explaining children's development but that these stressors operate to make children more dissimilar rather than similar because shared risks exacerbate individual differences in children ([<reflink idref="bib59" id="ref27">59</reflink>]). This issue is examined in Question 1d. By examining both the effective and the objective environment simultaneously we can achieve a better understanding of the ways in which children are adversely affected by marital conflict. As the observed environment relates to modeling means and the effective environment relates to modeling variance, a statistical technique that can do both was used in the present study.</p> <p>We were also interested in examining the extent to which siblings are differentially the focus of argument between their parents and differentially exposed to such argument. Several studies have examined factors explaining differential parenting and as age, gender, and socioeconomic status ([<reflink idref="bib35" id="ref28">35</reflink>]) have been found to be important predictors of differential parenting, these were included as predictors in this study (Question 3a). Only one other study has examined factors explaining differential sibling exposure to marital conflict. [<reflink idref="bib51" id="ref29">51</reflink>], in a cross‐sectional study, found that more disturbed siblings were exposed to more partner conflict than less disturbed siblings.</p> <hd id="AN0015916620-3">Stepfamilies, Marriage, and Child Behavior</hd> <p>Living in a stepfamily provides a unique context in which to investigate the reciprocal relationships between marital conflict and child behavior. Certain differential qualities of stepfamilies as compared with first marriages may amplify and exacerbate the elements of marriage and child behavior that contribute to conflict dynamics. Remarriages have been found to dissolve more rapidly and frequently than first marriages ([<reflink idref="bib45" id="ref30">45</reflink>]; a finding that led us to include length of time in relationship as a covariate in analyses). This is particularly true in stepfamilies that involve children from previous marriages ([<reflink idref="bib60" id="ref31">60</reflink>]). Moreover, in second marriages, marital quality appears to decline far more quickly than in first marriages ([<reflink idref="bib37" id="ref32">37</reflink>]). The affective environment of stepfamilies involves more tension and open, overt conflict ([<reflink idref="bib31" id="ref33">31</reflink>]). Children represent a frequent source of conflict in stepfamilies ([<reflink idref="bib10" id="ref34">10</reflink>], [<reflink idref="bib32" id="ref35">32</reflink>]).</p> <p>These findings suggest that stepfamily relationships are harder to negotiate than biological family relationships. Furthermore, not only is the stepfamily context associated with more marital conflict and marital instability, the children within stepfamilies are more likely to have behavioral difficulties ([<reflink idref="bib30" id="ref36">30</reflink>]). Being in a stepfamily may therefore amplify problematic family processes. In the present study we examined whether child disturbance had a more negative effect on marital conflict in stepfamilies than in biological families (Question 2c). We also hypothesized, as part of amplification, whether siblings in stepfamilies are more differentiated over time in the extent to which they were a focus of parental argument and the extent to which they were exposed to argument (Questions 2d and 3d).</p> <p>This study adds to the few studies that have examined the effect of marital conflict on children longitudinally and to the one previous study that has examined the effect of children's behavior on marital conflict ([<reflink idref="bib44" id="ref37">44</reflink>]). We extended the latter study by distinguishing between different types of marital conflict. Furthermore, no previous study has simultaneously examined observed and effective processes of marital conflict, differentiated between family‐ and child‐level processes in their effects on children's behavior, or modeled differential sibling experiences of marital conflict longitudinally.</p> <hd id="AN0015916620-4">Method</hd> <p></p> <hd id="AN0015916620-5">Sample</hd> <p>The Avon Longitudinal Study of Parents and Children (ALSPAC) is a longitudinal, prospective study of women, their partners, and an index child. The study design included all pregnant women living in the health district of Avon, England, who were to deliver their baby between April 1991 and December 1992 ([<reflink idref="bib19" id="ref38">19</reflink>]). It was estimated that 85% to 90% of the eligible population took part (<emph>N</emph>=14,000). The current study, the Avon Brothers and Sisters Study (ABSS), is a substudy within ALSPAC that capitalized on the community nature of the ALSPAC sample and used it as a sampling frame to select a representative group of nonstepfamilies, single‐parent families (not included in this report), and stepfamilies ([<reflink idref="bib16" id="ref39">16</reflink>]). In this study we have differentiated between stepfather families where all the siblings are biologically related (stepfather families) and families that we have called complex stepfamilies in which either the siblings are not all related to one another or the stepparent is a stepmother (complex/stepmother families). The intention of the ABSS sampling design was to select at random 50 families of each of four types from the larger community sample for intensive longitudinal study. We included families with two or more children in the sampling procedure because we were especially interested in within‐family variation in children's adjustment and family process. The representativeness of families within each of the four family types in ABSS was supported by the finding that ABSS families did not differ from families of the same type in the larger community sample on maternal education, income, and child adjustment. Overall, 83% of eligible families who were approached (in the ALSPAC sample, based on the preceding criteria) agreed to participate in ABSS. In the ABSS sample, the average age at Time 1 of the youngest child (the ALSPAC target child) was 4.8 years (<emph>SD</emph>=.38) and the age range of the siblings was 6 to 17 years (<emph>M</emph>=10.2 years, <emph>SD</emph>=2.9). The sample was generally middle class but included a range of socioeconomic levels.</p> <p>There have been two waves of data collection, and data for the current study are based on both waves. Wave 1 included 50 intact or biological families (<emph>n</emph>=113 children), 49 stepfather families (<emph>n</emph>=122 children) and 45 complex/stepmother families (<emph>n</emph>=123 children). Single parents (<emph>n=</emph>106) were also included in the sample but, apart from those in couple relationships described later, these families were not included in the present study. Criteria for inclusion in the present study involved stability in the couple relationship between Time 1 and Time 2, participation in both waves of data collection, and primary residence with the ABSS family. Although most couples that participated in data collection were sharing the same residence at Time 1, five maintained separate residences. Four of these five had become cohabiting stepfather families by Time 2 and were treated as such in the analysis, and the single anomalous family was dropped from analysis. (All analyses were completed with and without these families and no inclusion or exclusion altered substantive results.) The follow‐up sample included 45 biological families (<emph>n</emph>=101 children), 44 stepfather families (<emph>n</emph>=109 children), and 38 complex families (<emph>n</emph>=86 children). These families had 3 participating children (<emph>n</emph>=44 families), 2 participating children (<emph>n</emph>=81) or 1 participating child (<emph>n</emph>=2), as only one child in each family met participation criteria). Seventy‐five percent of families were married and the remainder were cohabiting. For this reason we refer to the conflict between parents as partner conflict.</p> <hd id="AN0015916620-6">Measures</hd> <p> <bold>Child externalizing and internalizing behavior. </bold> At both Time 1 and Time 2, teachers completed the Teacher's Report Form (TRF). This contains 118 problem behaviors that are scored on 3‐point scales ([<reflink idref="bib1" id="ref40">1</reflink>], [<reflink idref="bib2" id="ref41">2</reflink>]). Teachers indicated whether behavioral descriptions were not at all true (0), somewhat true (<reflink idref="bib1" id="ref42">1</reflink>), or very true (<reflink idref="bib2" id="ref43">2</reflink>) of their child. The child externalizing score is based on items from the aggression and delinquency indexes whereas the internalizing score is based on items from the Withdrawn, Somatic Complaints, and Anxious/Depressed scales. Scores for both externalizing and internalizing were skewed, and a square root transformation was applied to the data. For multilevel modeling these variables were centered. Although behavioral problems in children are assessed at the individual child level, a multilevel model involves the simultaneous estimation of family‐level (based on the family mean) and child‐level (based on the child‐specific score) effects. For this reason the family average on childhood internalizing and externalizing symptoms was derived from the centered scores.</p> <p> <bold>General partner conflict—family level. </bold> This measure is based on the Disagreement Scale from the Locke–Wallace Marital Adjustment Test ([<reflink idref="bib39" id="ref44">39</reflink>]). Mothers are asked to rate the frequency of conflict on a 6‐point scale ranging from 1 (<emph>always agree</emph>) to 6 (<emph>always disagree</emph>) pertaining to eight areas of family life: handling finances, recreation, demonstration of affection, friends, sex, conventional behavior, philosophy of life, and dealing with in‐laws. This is a family‐level variable as the score is the same for all children in the family. Higher scores represented more conflict. Internal consistency was α=.85.</p> <p> <bold>Argument about children. </bold> Mothers completed the Child Rearing Inventory ([<reflink idref="bib31" id="ref45">31</reflink>]) at Time 1 and Time 2 about each child in the family. The format of the questionnaire was as follows: "Sometimes parents have disagreements about bringing up children, over issues like how to handle quarrels between brothers and sisters, or bedtimes etc. How much have you and your partner disagreed in relation to (child's name) about the following: your child's behavior towards brothers and sisters, how to handle tantrums, defiance and disobedience towards your partner, bedtimes, meal times, TV, table manners/politeness etc., making too much noise, behavior in school, playing outside (where, when, with whom)?" Ratings were made on a 7‐point scale ranging from 1 (<emph>not at all in the last month</emph>) to 7 (<emph>more than once per day</emph>). The internal consistency (α) of the scale was α=.87. As well as a score that is unique to each child in the family (child specific), the family average on argument about children (called argument about children—family level) was also derived.</p> <p> <bold>Exposure to conflict. </bold> The item used to assess children's exposure to conflict was drawn from the Family Conflict Inventory ([<reflink idref="bib32" id="ref46">32</reflink>]). Mothers were asked, "How often is (name of child) present when you argue?" and they rated their responses on a 5‐point‐scale ranging from 1 (<emph>never</emph>) to 5 (<emph>almost all the time</emph>). As well as the child‐specific score, the family average was also derived (called exposure to conflict—family level).</p> <p> <bold>Stepfather families, complex/stepmother families, and biological families. </bold> Three family types were defined: biological, stepfather (father was not biologically related to at least one child but the mother was biologically related to all children), and complex/stepmother families (mother was not biologically related to at least one child in the family, or both the mother and father were not biologically related to at least one child in the family, but not the same child). Stepfather families have been found to differ in important ways from more complex stepfamilies ([<reflink idref="bib9" id="ref47">9</reflink>], [<reflink idref="bib32" id="ref48">32</reflink>]). We combined stepmother and complex stepfamilies because there were too few stepmother families to examine separately, even in the large community sample from which the ABSS sample was selected. Biological families were the control condition.</p> <p> <bold>Occupational status. </bold> This was assessed using the Registrar General's classification of occupational status ([<reflink idref="bib46" id="ref49">46</reflink>]). There are six categories of occupation from professional (<reflink idref="bib1" id="ref50">1</reflink>) to unskilled (<reflink idref="bib6" id="ref51">6</reflink>). The occupation of both parents was coded, and the highest code for the family was used in the analysis.</p> <p> <bold>Time in current relationship. </bold> The duration of the partner relationship was coded in years from the time the couple began to cohabit.</p> <p> <bold>Children's age and gender. </bold> Children's age was obtained from mothers and coded in years. Girls were coded 1 and boys were the reference category (coded 0).</p> <p> <bold>Mother negativity. </bold> Maternal negativity toward the child was constructed as a composite of four scales using standardized scores. The internal consistency was α=.76. The four scales were as follows: First, the frequency of parent–child conflict was assessed with a modified version of the Daily Routines Scale ([<reflink idref="bib31" id="ref52">31</reflink>]). It included 4 items assessed on a 7 point scale (α=.86). Second, parental criticism and nagging of the child was assessed using the negative factor from the Parent–Child Relationship Scale ([<reflink idref="bib31" id="ref53">31</reflink>]). This included 11 items rated on a 7‐point scale (α=.82). Third, frequency of different types of punitive discipline techniques, such as yelling at or ridiculing the child, was measured using a modified version of the Negative Sanctions Scale ([<reflink idref="bib31" id="ref54">31</reflink>]). This included 4 items rated on a 5‐point scale (α=.66). Fourth, hostility was assessed using an investigator‐based interview measure of four questions, coded on 4‐ to 5‐point scales, concerning the frequency and level of hostility (α=.57; [<reflink idref="bib16" id="ref55">16</reflink>]).</p> <hd id="AN0015916620-7">Results</hd> <p></p> <hd id="AN0015916620-8">Missing Data</hd> <p>The range of data points missing on single variables across models was between 0% and 22%. In order not to introduce bias into the sample through listwise deletion of missing data, we used the multiple imputation technique described by [<reflink idref="bib53" id="ref56">53</reflink>] and [<reflink idref="bib56" id="ref57">56</reflink>]. This has been identified as one of the best techniques for handling missing data ([<reflink idref="bib3" id="ref58">3</reflink>]). The basic idea of multiple imputation is that an imputation model can be used to generate a series of data sets. In the imputation model the joint distribution of all variables in the model are generated. The data are not balanced, as some individuals will have missing data on some of the responses. An estimation technique is required that can fit a multivariate response model that includes missing data. We used Markov Chain Monte Carlo (MCMC) estimation (see [<reflink idref="bib7" id="ref59">7</reflink>]). This model includes all missing data values as parameters to be estimated. Where an individual had missing observations we estimated a mean and variance for each missing observation based on the overall covariance structure of the variables and the nonmissing values present for the individual. The distribution of every missing value in the data set was estimated. The MCMC method provided samples from these distributions. The imputation data sets contained original values where data were not missing and a draw from the missing value's distribution where data were missing. Having sampled a series of complete data sets, it was then possible to fit the model of interest to six data sets estimating the mean and standard error for these data sets. In our results tables we quote parameters and standard errors. The parameter estimates are simply the mean of the parameter across the series of complete data sets. The standard errors are calculated from a more elaborate formula that combines the within‐data‐set and the between‐data‐set variance for that parameter (see [<reflink idref="bib7" id="ref60">7</reflink>]). Conventional single imputation techniques ignore the fact that there is uncertainty around the imputed values. Multiple imputation techniques adjust the final estimates for this uncertainty and are therefore widely regarded as a preferable method. As a check, the parameter estimates derived from multiple imputation were compared with those from listwise deletion, and substantive results were the same. Imputation data are presented in tables.</p> <hd id="AN0015916620-9">Analysis Description and Plan</hd> <p>First, bivariate relationships between variables were examined. Second, hypothesis testing was carried out with the use of multilevel modeling using MLwiN, version 1.10.0006 (Rasbashet al., 2000). This is a regression analysis that takes into account the nested structure of the data set: children within families. Variance is partitioned into levels corresponding to this nested structure, in this case, between and within family variance. The technique allows for the modeling of means (why some children show higher scores than other children), as in standard regression, as well as the modeling of variance. The within‐family variance estimate represents the extent to which children within the family differ from one another, with higher values indicating more dissimilarity between siblings. The between‐family variance estimate indicates the extent to which families differ from one another on the family mean for the response variable. Fixed effects (predictor variables that enable us to model changes in the response variable, i.e. modeling the mean) are presented in the top part of Tables 3, 4 and 5. These estimates and standard errors are interpreted as in a regression model. An estimate that is approximately twice the size of its standard error is significant (<emph>p</emph><.05). The difference to consider between a standard regression and a multilevel model with respect to fixed effects is that in these multilevel models the predictor variables differentiate between child‐specific processes (those that are unique to an individual child in the family) and family‐level processes (those that are shared between siblings) and are so divided in the tables. The variance estimates given in the bottom half of the tables are unique to a multilevel model. They are divided into within‐ and between‐family variances. It is possible within a multilevel model to model the within‐ and between‐family variances. That is, the variance is allowed to vary as a function of a variable of interest (see [<reflink idref="bib20" id="ref61">20</reflink>], pp. 47–58). For instance, siblings in stepfamilies may be more varied on the extent to which parents argue about children than they are in biological families.</p> <p>3 
 Fixed‐Effects Estimates and Variance–Covariance Estimates for Models of the Predictors of Externalizing Behavior in Children, Time 2 (N=296)</p> <p> <ephtml> <table><thead valign="bottom"><tr><th /><th align="center">Model 1</th><th align="center">Model 2</th><th align="center">Model 3</th><th align="center">Model 4</th><th align="center">Model 5</th><th align="center">Model 6</th></tr></thead><tbody valign="top"><tr><td /><td align="center">Fixed effects</td></tr><tr><td>Intercept</td><td><bold>.72 (.10)</bold></td><td><bold>.72 (.10)</bold></td><td><bold>.70(.10)</bold></td><td><bold>.70(.10)</bold></td><td><bold>.70 (.10)</bold></td><td><bold>.70(.09)</bold></td></tr><tr><td>Child specific</td></tr><tr><td> Children's externalizing T1</td><td /><td><bold>.39 (.08)</bold></td><td><bold>.39(.08)</bold></td><td><bold>.39(.08)</bold></td><td><bold>.40(.08)</bold></td><td><bold>.41(.08)</bold></td></tr><tr><td> Mother negativity</td><td /><td /><td /><td /><td>−.01(.02)</td><td>.01(.02)</td></tr><tr><td>Family</td></tr><tr><td> FA children's externalizing T1</td><td /><td>.17(.10)</td><td>.17(.10)</td><td>.11 (.10)</td><td>.12(.10)</td><td>.08(.10)</td></tr><tr><td> Stepfather</td><td /><td /><td>.03 (.02)</td><td>.03(.02)</td><td>.03 (.02)</td><td>.03(.02)</td></tr><tr><td> Complex</td><td /><td /><td>.01(.02)</td><td>.02(.02)</td><td>.02(.02)</td><td>.02(.02)</td></tr><tr><td> FA argument about children T1</td><td /><td /><td /><td><bold>.05(.01)</bold></td><td><bold>.05(.01)</bold></td><td><bold>.05(.02)</bold></td></tr><tr><td /><td align="center">Random parameters</td></tr><tr><td>Between families</td><td>.003(.002)</td><td>.001(.001)</td><td>.001(.001)</td><td>.000(.001)</td><td>.000(.00)</td><td>.000(.001)</td></tr><tr><td>Within families</td><td><bold>.016(.003)</bold></td><td><bold>.013(.002)</bold></td><td><bold>.013(.003)</bold></td><td><bold>.013(.002)</bold></td><td><bold>.013(.002)</bold></td><td /></tr><tr><td> Within‐family variance for FA argument about children</td><td /><td /><td /><td /><td /><td><bold>.012(.005)</bold></td></tr><tr><td> Within‐family variance for a one‐Unit increase in FA argument about children</td><td /><td /><td /><td /><td /><td><bold>.018 (.004)</bold></td></tr><tr><td>−2* log likelihood</td><td>−336.32</td><td>−435.65</td><td>−439.68</td><td>−453.80</td><td>−455.88</td><td>−461.69</td></tr><tr><td>Change in model fit (χ<sup>2</sup>) from prior model to present; <italic>dfs</italic></td><td /><td>99.33***</td><td>4.03</td><td>14.12***</td><td>2.08</td><td>5.81*</td></tr><tr><td /><td /><td>2 <italic>df</italic></td><td>2 <italic>df</italic></td><td>1 <italic>df</italic></td><td>1 <italic>df</italic></td><td>1 <italic>df</italic></td></tr></tbody></table> </ephtml> </p> <p>1 <emph>Note</emph>. Variance–covariance estimates are also referred to as random parameters. T1=Time 1; FA=family average.</p> <ulist> <item>2 * <emph>p</emph> <.05.</item> <item>3 *** <emph>p</emph> <.001.</item> <item>4 
 Fixed‐Effects Estimates and Variance–Covariance Estimates for Models of the Predictors of Argument About Children, Time 2 (N=296)</item> </ulist> <p> <ephtml> <table><thead valign="bottom"><tr><th /><th align="center">Model 1</th><th align="center">Model 2</th><th align="center">Model 3</th><th align="center">Model 4</th><th align="center">Model 5</th><th align="center">Model 6</th></tr></thead><tbody valign="top"><tr><td /><td align="center">Fixed effects</td></tr><tr><td>Intercept</td><td><bold>1.50(.05)</bold></td><td><bold>1.49 (.04)</bold></td><td><bold>1.45 (.06)</bold></td><td><bold>1.44 (.06)</bold></td><td><bold>1.43(.05)</bold></td><td><bold>1.42(.05)</bold></td></tr><tr><td>Child specific</td></tr><tr><td> Parental argument T1</td><td /><td><bold>.38(.09)</bold></td><td><bold>.38(.09)</bold></td><td><bold>.38(.09)</bold></td><td><bold>.38(.09)</bold></td><td><bold>.37(08)</bold></td></tr><tr><td> Children's externalizing T1</td><td /><td /><td /><td>.01(.30)</td><td>.01(.29)</td><td>.02(.30)</td></tr><tr><td>Family level</td></tr><tr><td> FA parental argument T1</td><td /><td><bold>.35(.10)</bold></td><td><bold>.35(.10)</bold></td><td><bold>.28(.10)</bold></td><td><bold>.28(.10)</bold></td><td><bold>.30(.10)</bold></td></tr><tr><td> FA externalizing T1</td><td /><td /><td /><td><bold>.89(.43)</bold></td><td>−.09(.60)</td><td>−.19(.55)</td></tr><tr><td>FA Externalizing T1 × Children's Externalizing TI</td><td /><td /><td /><td><bold>2.82(1.41)</bold></td><td><bold>3.13(1.49)</bold></td><td><bold>3.16(1.52)</bold></td></tr><tr><td> Stepfather</td><td /><td /><td>.09(.08)</td><td>.08(.08)</td><td>.09(.07)</td><td>.09(.07)</td></tr><tr><td> Complex</td><td /><td /><td>.04(.09)</td><td>−.01(.08)</td><td>.01(.08)</td><td>−.01(.08)</td></tr><tr><td> Step × FA Externalizing Tl</td><td /><td /><td /><td /><td><bold>1.88(.83)</bold></td><td><bold>1.91(.84)</bold></td></tr><tr><td> Complex × FA Externalizing T1</td><td /><td /><td /><td /><td>1.17(.72)</td><td>1.20(.66)</td></tr><tr><td /><td align="center">Random parameters</td></tr><tr><td>Between‐family variance</td><td><bold>.21(.04)</bold></td><td><bold>.06(.02)</bold></td><td><bold>.06(.02)</bold></td><td><bold>.04(.02)</bold></td><td>.03(.02)</td><td>.03(.02)</td></tr><tr><td>Within‐family variance</td><td><bold>.22(.03)</bold></td><td><bold>.19(.02)</bold></td><td><bold>.19(.02)</bold></td><td><bold>.19 (.02)</bold></td><td>.19(.02)</td><td /></tr><tr><td> Biological</td><td /><td /><td /><td /><td /><td>.14(.03)</td></tr><tr><td> Stepfather</td><td /><td /><td /><td /><td /><td><bold>.24(.04)</bold></td></tr><tr><td> Complex</td><td /><td /><td /><td /><td /><td>.18(.04)</td></tr><tr><td>−2* log likelihood</td><td>537.59</td><td>419.45</td><td>418.02</td><td>397.45</td><td>390.81</td><td>385.14</td></tr><tr><td>Change in model fit (χ<sup>2</sup>) from prior model to present; <italic>dfs</italic></td><td /><td>118.14***</td><td>1.43</td><td>20.57***</td><td>6.64*</td><td>5.67†</td></tr><tr><td /><td>2 <italic>df</italic></td><td>2 <italic>df</italic></td><td>3 <italic>df</italic></td><td>2 <italic>df</italic></td><td>2 <italic>df</italic></td></tr></tbody></table> </ephtml> </p> <ulist> <item>4 <emph>Note</emph>. Variance–covariance estimates are also referred to as random parameters. T1=Time 1; FA=family average.</item> <item>5 † <emph>p</emph> <.10.</item> <item>6 * <emph>p</emph> <.05.</item> <item>7 *** <emph>p</emph> <.001.</item> <item>5 
 Fixed‐Effects Estimates and Variance–Covariance Estimates for Models of the Predictors of Child‐Specific Exposure to Partner Conflict, Time 2 (N=296)</item> </ulist> <p> <ephtml> <table><thead valign="bottom"><tr><th /><th align="center">Model 1</th><th align="center">Model 2</th><th align="center">Model 3</th><th align="center">Model 4</th><th align="center">Model 5</th><th align="center">Model 6</th></tr></thead><tbody valign="top"><tr><td /><td align="center">Fixed effects</td></tr><tr><td>Intercept</td><td><bold>3.71(.07)</bold></td><td><bold>3.72(.05)</bold></td><td><bold>3.63(.06)</bold></td><td><bold>3.70(.09)</bold></td><td><bold>3.69 (.09)</bold></td><td><bold>3.69 (.08)</bold></td></tr><tr><td>Child specific</td></tr><tr><td> Exposure T1</td><td /><td><bold>.32(.12)</bold></td><td><bold>.29(.12)</bold></td><td><bold>.29(.12)</bold></td><td><bold>.31(.12)</bold></td><td><bold>.27(.12)</bold></td></tr><tr><td> Girl</td><td /><td /><td><bold>−.19(.07)</bold></td><td><bold>−.19(.08)</bold></td><td><bold>−.21(.07)</bold></td><td><bold>−.21(.06)</bold></td></tr><tr><td> Children's externalizing T1</td><td /><td /><td /><td /><td>.51(.33)</td><td>.46(.35)</td></tr><tr><td>Family level</td></tr><tr><td> FA exposure T1</td><td /><td><bold>.32(.15)</bold></td><td><bold>.34(.14)</bold></td><td><bold>.33(.15)</bold></td><td><bold>.32(.15)</bold></td><td><bold>.36(.15)</bold></td></tr><tr><td> Stepfather</td><td /><td /><td /><td>.04(.12)</td><td>.05(.11)</td><td>.05(.12)</td></tr><tr><td> Complex</td><td /><td /><td /><td>.18(.13)</td><td>.19(.13)</td><td>.19(.13)</td></tr><tr><td> FA externalizing</td><td /><td /><td /><td /><td>.17(.62)</td><td>.10(.65)</td></tr><tr><td /><td align="center">Random parameters</td></tr><tr><td>Between families</td><td><bold>.44(.08)</bold></td><td><bold>.19(.04)</bold></td><td><bold>.19(.04)</bold></td><td><bold>.18(.04)</bold></td><td><bold>.18(.04)</bold></td><td><bold>.20(.04)</bold></td></tr><tr><td>Within families</td><td><bold>.30(.04)</bold></td><td><bold>.28(.03)</bold></td><td><bold>.26(.03)</bold></td><td><bold>.26(.03)</bold></td><td><bold>.26(.03)</bold></td><td /></tr><tr><td> Biological</td><td /><td /><td /><td /><td /><td>.09(.02)</td></tr><tr><td> Stepfather</td><td /><td /><td /><td /><td /><td><bold>.39(.08)</bold></td></tr><tr><td> Complex</td><td /><td /><td /><td /><td /><td><bold>.29(.06)</bold></td></tr><tr><td>−2* log likelihood</td><td>669.01</td><td>577.63</td><td>570.43</td><td>567.91</td><td>563.76</td><td>532.26</td></tr><tr><td>Change in model fit (χ<sup>2</sup>) from prior model to present; dfs</td><td /><td>91.38***</td><td>7.2*</td><td>2.52</td><td>4.15</td><td>31.5***</td></tr><tr><td /><td>2 <italic>df</italic></td><td>1 <italic>df</italic></td><td>2 <italic>df</italic></td><td>2 <italic>df</italic></td><td>2 <italic>df</italic></td></tr></tbody></table> </ephtml> </p> <ulist> <item>8 <emph>Note</emph>. Variance–covariance estimates are also referred to as random parameters. T1=Time 1; FA=family average.</item> <item>9 * p<.05.</item> <item>10 *** <sups>***</sups> p<.001.</item> </ulist> <p>Separate analyses were carried out for child behavior and each of the partner conflict measures. Multilevel models within each response variable (or dependent variable) were estimated sequentially. Each model is compared with the previous model (using change in the log likelihood) to determine whether the addition of the new parameter(s) improved the fit of the model. Table 1 shows, for each response variable, the relationship among the study questions, the parameters that are entered into the regression equation to answer each question (column labeled "parameters in equation"), and the model sequence number.</p> <p>1 
 Links Among Study Questions, Parameters in Equation, and Model Sequence for Response Variables: Externalizing Behavior, Argument About Children, and Exposure to Conflict</p> <p> <ephtml> <table><thead valign="bottom"><tr><th>Response variable and study questions</th><th align="center">Parameters in equation</th><th>Model number</th></tr></thead><tbody valign="top"><tr><td>1. Predicting change in children's externalizing behavior</td><td /><td>Table 3</td></tr><tr><td> a. Is this predicted by being in a stepfamily?</td><td>Entry of stepfamily status into equation.</td><td>3</td></tr><tr><td> b. Is this predicted by argument about children, exposure to conflict, or general partner conflict, and at which level: child specific or family average?</td><td>Entry of partner conflict predictors into equation: child specific and family average.</td><td>4</td></tr><tr><td> c. Does partner conflict remain a significant predictor of child externalizing after controlling for maternal negativity?</td><td>Entry of maternal negativity into equation. Does partner conflict remain significant?</td><td>5</td></tr><tr><td> d. Does sibling dissimilarity on externalizing behavior increase at higher levels of partner conflict?</td><td>Allowing the within‐family variance to vary across levels of partner conflict.</td><td>6</td></tr><tr><td>2. Predicting change in argument about children</td><td /><td>Table 4</td></tr><tr><td> a. Is this predicted by being in a stepfamily?</td><td>Entry of stepfamily status into equation.</td><td>3</td></tr><tr><td> b. Is this predicted by children's externalizing behavior?</td><td>Entry of externalizing into equation (child specific, family average, and their interaction).</td><td>4</td></tr><tr><td> c. Does externalizing behavior predict change in argument about children more in stepfamilies than in biological families?</td><td>Entry of interaction term: Stepfamily Status × Children's Externalizing Behavior.</td><td>5</td></tr><tr><td> d. Does argument about children become more differential over time in stepfamilies than in intact families?</td><td>Estimating separate within‐family variances for intact families and stepfamilies.</td><td>6</td></tr><tr><td>3. Predicting change in exposure to conflict</td><td /><td>Table 5</td></tr><tr><td> a. Is this predicted by gender, age, socioeconomic status, or time in current relationship?</td><td>Entry of gender, age, socioeconomic status, and time in current relationship.</td><td>3</td></tr><tr><td> b. Is this predicted by being in a stepfamily?</td><td>Entry of stepfamily status.</td><td>4</td></tr><tr><td> c. Is this predicted by externalizing behavior?</td><td>Entry of externalizing into equation (child specific and family average).</td><td>5</td></tr><tr><td> d. Does exposure to conflict become more differential between siblings over time in stepfamilies than in intact families?</td><td>Estimating separate within‐family variances for intact families and stepfamilies.</td><td>6</td></tr></tbody></table> </ephtml> </p> <p>All analyses started with the null model without predictor variables. The null model showed the partitioning of variance into within‐ and between‐family variance. Next, the Time 1 version of the response variable was entered into the model (Model 2), and this showed the degree of stability in the construct over time. Next, effects of child age, gender, time in current relationship, and socioeconomic status were entered for all response variables. In fact, only one of these variables (gender) was found to be significant for one response variable (exposure to conflict) and all others were dropped from the models. Family type was entered next, as this was part of the sampling frame for the study. After this step the entry of variables across models differed according to study questions as outlined in Table 1. Parameters that are significant at <emph>p</emph><.05 are bolded in the text and in the last line of the table the change in the model fit from the previous model to the present model is given.</p> <p>Children's internalizing behavior was examined as a predictor and as a response variable. In neither case were hypothesized effects seen. As a consequence, only results for externalizing child behavior are described in multilevel models.</p> <hd id="AN0015916620-10">Correlations</hd> <p>The intercorrelations of variables are in Table 2. Several findings are worth mentioning. First, significant continuity over 2 years within partner conflict measures and within childhood adjustment measures was evident. Second, externalizing and internalizing disturbances were more consistently related to child‐directed conflict measures than general partner conflict, although differences between correlations were not significant. Third, all partner conflict measures were significantly correlated, making for partial redundancy across measures. Fourth, lower socioeconomic status was associated with greater argument about children, and general partner conflict (Time 1). It was not associated with increased exposure to argument or three of four of the internalizing and externalizing measures. Older children were exposed to more partner conflict, and they showed more externalizing disturbance. Being a girl was associated with less argument about children, exposure to conflict, and externalizing symptomatology.</p> <p>2 
 Correlation Matrix of Children's Behavior and Partner Conflict at Time 1 and Time 2 and Covariates (N=296)</p> <p> <ephtml> <table><thead valign="bottom"><tr><th /><th align="center">1</th><th align="center">2</th><th align="center">3</th><th align="center">4</th><th align="center">5</th><th align="center">6</th><th align="center">7</th><th align="center">8</th><th align="center">9</th><th align="center">10</th><th align="center">11</th><th align="center">12</th><th align="center">13</th><th align="center">14</th><th align="center">15</th></tr></thead><tbody valign="top"><tr><td>1. Child age</td></tr><tr><td>2. Girl</td><td>−.10</td><td /><td /><td /><td /><td /><td /><td /><td /><td /><td /><td /><td /><td /><td /></tr><tr><td>3. Time in current relationship</td><td>−.00</td><td>−.07</td><td /><td /><td /><td /><td /><td /><td /><td /><td /><td /><td /><td /><td /></tr><tr><td>4. Occupational status</td><td>−.06</td><td>−.08</td><td>−.12</td><td /><td /><td /><td /><td /><td /><td /><td /><td /><td /><td /><td /></tr><tr><td>5. Mother–child negativity T1</td><td>−.07</td><td>−.04</td><td>.01</td><td>.00</td><td /><td /><td /><td /><td /><td /><td /><td /><td /><td /><td /></tr><tr><td>6. Externalizing T1</td><td>.17**</td><td>−.14*</td><td>−.08</td><td>.09</td><td>.24***</td><td /><td /><td /><td /><td /><td /><td /><td /><td /><td /></tr><tr><td>7. Externalizing T2</td><td>.18**</td><td>−.14*</td><td>−.12*</td><td>.12*</td><td>.14*</td><td>.52***</td><td /><td /><td /><td /><td /><td /><td /><td /><td /></tr><tr><td>8. Internalizing T1</td><td>.05</td><td>−.03</td><td>−.14*</td><td>.01</td><td>.11</td><td>.34***</td><td>.06</td><td /><td /><td /><td /><td /><td /><td /><td /></tr><tr><td>9. Internalizing T2</td><td>−.03</td><td>−.04</td><td>−.06</td><td>−.02</td><td>.02</td><td>.14*</td><td>.25***</td><td>.30***</td><td /><td /><td /><td /><td /><td /><td /></tr><tr><td>10. General partner conflict T1: family level</td><td>.08</td><td>−.05</td><td>.10</td><td>.16**</td><td>.23***</td><td>.00</td><td>.09</td><td>.01</td><td>.01</td><td /><td /><td /><td /><td /><td /></tr><tr><td>11. General partner conflict T2: family level</td><td>.02</td><td>−.08</td><td>.13*</td><td>.10</td><td>.20***</td><td>.05</td><td>.07</td><td>.11</td><td>.16**</td><td>.68***</td><td /><td /><td /><td /><td /></tr><tr><td>12. Argument about children T1</td><td>.00</td><td>−.10</td><td>.15*</td><td>.15*</td><td>.43***</td><td>.17**</td><td>.26***</td><td>.11</td><td>.09</td><td>.31***</td><td>.40***</td><td /><td /><td /><td /></tr><tr><td>13. Argument about children T2</td><td>−.08</td><td>−.12*</td><td>.03</td><td>.18**</td><td>.39***</td><td>.27***</td><td>.28***</td><td>.19**</td><td>.09</td><td>.22***</td><td>.32***</td><td>.62***</td><td /><td /><td /></tr><tr><td>14. Exposure to argument T1</td><td>.16**</td><td>−.15*</td><td>.14*</td><td>−.05</td><td>.16**</td><td>.11</td><td>.12*</td><td>.06</td><td>.01</td><td>.49***</td><td>.44***</td><td>.31***</td><td>.15*</td><td /><td /></tr><tr><td>15. Exposure to argument T2</td><td>.08</td><td>−.18**</td><td>.09</td><td>−.05</td><td>.11</td><td>.00</td><td>.08</td><td>−.07</td><td>.06</td><td>.36***</td><td>.40***</td><td>.16**</td><td>.18**</td><td>.59***</td><td /></tr><tr><td>16. Step</td><td>.08</td><td>.02</td><td>−.31***</td><td>.15</td><td>−.03</td><td>.00</td><td>.09</td><td>.04</td><td>−.08</td><td>.07</td><td>−.05</td><td>−.02</td><td>.04</td><td>.05</td><td>−.01</td></tr><tr><td>17. Complex</td><td>.10</td><td>.01</td><td>−.32***</td><td>−.09</td><td>.00</td><td>.06</td><td>.03</td><td>.04</td><td>.08</td><td>.01</td><td>.08</td><td>−.04</td><td>.04</td><td>.02</td><td>.10</td></tr></tbody></table> </ephtml> </p> <ulist> <item>11 <emph>Note</emph>. T1=Time 1; T2=Time 2.</item> <item>12 * <emph>p</emph><.05.</item> <item>13 ** <emph>p</emph><.01.</item> <item>14 *** <emph>p</emph><.001.</item> </ulist> <hd id="AN0015916620-11">Partner Conflict Affects the Development of Children's Behavior Problems</hd> <p>Table 3 gives results for children's externalizing behavior. The null model allowed us to examine the similarity between siblings on children's externalizing behavior. The intraclass correlation, calculated on the proportion of family‐level variance over the total variance (.003/(.003+.016)=.16, demonstrated that there was only a small degree of similarity among siblings on externalizing behavior. The sample mean for children's externalizing behavior can also be derived from Model 1 (see fixed effects, intercept). Model 2 showed that children's externalizing behavior at Time 1 was related to their externalizing behavior at Time 2, but the family average for children's externalizing at Time 1 did not add to the prediction. For models that follow (Models 3 through 6), all predictor variables can be interpreted as predicting a change in child behavior, as children's previous externalizing behavior remains in the model. All covariates (age, gender, time in current relationship, and socioeconomic status) were entered into the model, but as none was significant they were not retained.</p> <p>In Model 3 dummy variables for family type were entered as fixed effects (see fixed effects, family level, step/complex). Complex and stepfamily status were not found to predict a significant change in children's externalizing behavior. In Model 4, partner conflict measures were entered. The only significant predictor of change in children's externalizing behavior was argument about children at Time 1. When the two measures of argument about children (family average and child specific) were in the model, they split the variance between them and neither coefficient was significant. In contrast, both measures were significant when in the model on their own and both reduced between‐family variance. As model fit was better when argument about children—family average was retained rather than the child‐specific measure, the latter was dropped from the equation. Argument about children explained differences between families on children's externalizing behavior rather than differences within families. This suggested a family style on argument about children such that families differed more from other families than siblings differed from one another. Results for Model 5 showed that argument about children—family level remained significant even when maternal negativity was in the model. Thus, effects of argument about children on a change in externalizing were not explained by maternal negativity toward children. In Model 6 the within‐family variance was allowed to vary as a function of argument about children—family level, and this significantly improved model fit. The more parents argued about children, the more dissimilar siblings were to one another on change in externalizing behavior. The within‐family variance increased from.12 to.18 for a one‐unit increase in the family average for argument about children.</p> <p>In summary, first the family average rather than child‐specific argument about children predicted an increase in children's externalizing behavior. Second, at higher levels of the family average for argument about children, the within‐family variance for children's externalizing behavior increased over time.</p> <hd id="AN0015916620-12">Children Affect the Development of Partner Conflict: Within‐Family Dynamics of Conflict</hd> <p> <bold>Parental argument about children. </bold> Results can be seen in Table 4. Two pieces of information were derived from the null model. The intercept (<emph>SE</emph>) value in Model 1 (see fixed effects, intercept) indicated that the mean for argument about children at Time 2 was 1.5. The intraclass correlation derived from Model 1 was.49. This demonstrated that children in the same family were relatively similar to one another in the amount of argument between parents that parents have about them. This clustering within families on argument about children helps make sense of the finding presented in Table 3 that family average for argument about children was a stronger predictor of child externalizing behavior than child‐specific scores. These findings suggest there is a family style related to children being the subject of conflict, operating to differentiate one family from another (rather than only one child from another). Model 2 shows that both argument about children at Time 1 and the family average for argument about children at Time 1 were significant predictors of outcome. Thus, both the extent to which the parents argued about a specific child and the extent to which all children were a source of argument in the family explained variance in individual children's Time 2 scores.</p> <p>Next, covariates (age, gender, time in current relationship, and socioeconomic status) were entered into the model, but as none approached significance these were not retained. For Model 3 we entered the two dummy variables for family type but these were not associated with an increase in argument about children. For Model 4, child‐specific externalizing behavior, family average for children's externalizing behavior, and the interaction between these were entered. This interaction was significant and is shown in Figure 1. We truncated lines based on the real value of child‐specific scores for low, medium, and high family average externalizing scores. It can be seen from Figure 1 that in families with a low family average externalizing score there was no increase in argument about children as a function of the child‐specific externalizing score. This was not true, however, in the other two groups of families. When family average externalizing was medium or high, the higher the child‐specific externalizing score was, the greater was the increase in the parental argument about that child. Thus, for families in which the children were generally not problematic, having a child who was a little bit more of a problem than the others was not associated with an increase in argument about that child. In families with more externalizing problems, argument about children was found to increase over time as a function of the child's externalizing behavior at Time 1. In Model 5 the interactions for children's externalizing behavior at Time 1 (both family and child specific) were entered into the model. The interactions between children's externalizing—family average, stepfather, and complex family status were significant. (As the child‐specific interactions were not significant they were dropped from the models.) The positive interactions between complex/stepfamilies and family average externalizing shed further light on the pattern shown in Figure 1. The effect of the family average for children's externalizing behavior operates more strongly in stepfamilies and complex families than in biological families. Thus, when children in the family show more externalizing problems, this results in a greater increase in arguments about children in stepfamilies and complex families than it does in biological families.</p> <p>Graph: 1 Change in argument about children as a function of the family average for children's externalizing behavior (low, middle, high) and child‐specific externalizing.</p> <p>In Model 6 we allowed separate within‐family variances to be estimated as a function of stepfather and complex families. Parents were found to argue more differentially over time about siblings in stepfather families, but not about siblings in complex families, when they were compared with biological families (labeled "within‐family variance, stepfather"). Change in the model fit was of borderline significance when separate within‐family variances were allowed for both but was significant when only allowed for stepfamilies. Thus, stepfather status was associated with some children in the family being the focus of a lot of argument and others being the focus of only a little, with such differences increasing over time, whereas siblings in biological families were the focus of similar amounts of argument over time.</p> <p> <bold>Exposure to conflict. </bold> Results for exposure to conflict are in Table 5. Model 1 shows that most of the variance in children's exposure to conflict was at the family level, with an intraclass correlation of.57. This suggests a family style related to exposing children to parental conflict, which operated to differentiate one family from another more than one child from another. Results for Model 2 showed that continuity was high at the individual child level. It was also notable that the family average for exposure at Time 1 predicted an increase in exposure over time over and above the child‐specific effect. Covariates of child gender, age, socioeconomic status, and time parents have been living together were entered into Model 3, but only gender was retained as the other three variables were not significant. The change in children's exposure to conflict was found to be greater for boys than for girls. Boys experienced.19 more change in exposure to conflict than did girls. In Model 4 stepfamilies and complex/stepmother families were added to the model, but neither was associated with an increase in exposure to conflict. In Model 5 children's externalizing behaviors, both child specific and family level, were entered into the model but neither was a significant predictor of change in children's exposure. In Model 6 interactions between stepfamily status and child behavior were entered into the model, but as none was significant these are not displayed. Next, separate within‐family variances were allowed for stepfamilies and complex/stepmother families. These variances were significant for both, showing that in both types of stepfamilies, children became more differentiated over time in the extent to which they were exposed to partner conflict as compared with biological families.</p> <p> <bold>General partner conflict—family level. </bold> Neither children's externalizing behavior nor their internalizing behavior was found to be a significant predictor of general partner conflict.</p> <p> <bold>Multivariate response model. </bold> As we had found a direction of effect from argument about children to externalizing behavior in children and another from externalizing behavior in children to argument about children, it was important to ensure that both directions of effect would be evident if all predictive relationships were estimated within the same model. This can be achieved with a multivariate response model ([<reflink idref="bib21" id="ref62">21</reflink>], [<reflink idref="bib50" id="ref63">50</reflink>]). This model was estimated with two outcomes: argument about children and externalizing behavior in children. Substantive findings were the same. Argument about children at Time 1 was a significant predictor of change over time in children's externalizing behavior, and children's externalizing behavior at Time 1 was a significant predictor of change in argument about children when both directions of effect were estimated simultaneously. The details of this model are available from the authors.</p> <hd id="AN0015916620-13">Discussion</hd> <p></p> <hd id="AN0015916620-14">Partner Conflict Increases the Risk of Children's Behavior Problems</hd> <p>One goal of the study was to explore the direction of effect between partner conflict and children's emotional and behavioral problems. Partner conflict at Time 1 was found to predict a change in child externalizing behavior between Time 1 and Time 2. As this effect was evident over and above the effect of maternal negativity on children, we can conclude that parental conflict does have a direct effect on change in children's externalizing behaviors. Four other studies have used a design that involved examining change in children's behavior as a function of earlier partner interaction. Two of these reported direct effects and two reported indirect effects (through the parent–child relationship) of partner conflict on children's behavior ([<reflink idref="bib14" id="ref64">14</reflink>], [<reflink idref="bib25" id="ref65">25</reflink>], [<reflink idref="bib27" id="ref66">27</reflink>], [<reflink idref="bib44" id="ref67">44</reflink>]).</p> <p>Three types of partner conflict, assessed at the family or the child‐specific level, or both, were examined in this study: general partner conflict, arguments about children, and exposure to conflict. Argument about children was the only aspect of partner conflict found to predict an increase in children's externalizing behavior, and none of the partner conflict measures predicted an increase in internalizing behavior in children. Others have also found that parental arguments about children compared with arguments about non‐child‐related topics are associated with increased behavioral disturbance in children ([<reflink idref="bib23" id="ref68">23</reflink>], [<reflink idref="bib57" id="ref69">57</reflink>]). It may be that children blame themselves ([<reflink idref="bib24" id="ref70">24</reflink>]) or it may be that argument about children is much more salient to children than other types of argument, and thus it is this type of interaction that clues children in to the fact that their parents are in distress.</p> <p>This study is the first within‐family study of the effects of partner conflict on children. The within‐family design allowed us to differentiate between measures of partner conflict shared by all children in the family and measures of partner conflict that were specific to individual children. This issue is of central importance to understanding how environments affect children. Argument about children clustered at the family level, suggesting a family style in which children are the topic of partner conflict. Second, the family average of argument about children predicted change in children's externalizing more strongly than the child‐specific measure, suggesting that a shared family risk factor was more potent than a child‐specific risk factor in predicting change in child behavior. The distinction between the observed and the effective environment raised in the Introduction is important to consider here. The observed environment shows that the most potent stress on children is a family‐level or shared‐risk factor. We also found that sibling differences on externalizing behavior increased over time when parents argued more about children. This finding pertains to the effective environment. Thus, although shared family risk increases externalizing disturbance in children, it also increases differences between children. This suggests that one child within the family may increase a lot on externalizing and the other only a little when exposed to high levels of parental argument. A family risk shared by siblings has a nonshared effect in that the risk increases siblings' externalizing behaviors differentially. What processes involving individual vulnerability may be operating to increase differentiation between children? A range of factors have been found to moderate the effects of marital conflict and other family stressors on children's outcomes: temperament ([<reflink idref="bib54" id="ref71">54</reflink>]) including emotional reactivity, attributional processes ([<reflink idref="bib23" id="ref72">23</reflink>]), perceptions of threat ([<reflink idref="bib14" id="ref73">14</reflink>]), and compensatory social relationships ([<reflink idref="bib36" id="ref74">36</reflink>]). It would be valuable for future studies to use a within‐family design and to include measurement on such individual variables to explain within‐family differences in children's reactions to shared stressors.</p> <p>Another issue to consider is the mechanism involved in children's increased risk of externalizing behavior as a function of partner conflict. The spillover process described in the Introduction gives an account of the way in which affect from one family member or dyad influences another. Spillover may operate through several mechanisms. For instance, affect contagion described by [<reflink idref="bib29" id="ref75">29</reflink>], as below conscious awareness and involving mimicry and synchronization, may play a role in the way affect spreads from one person to another. Alternatively, more sophisticated mechanisms involving enculturation ([<reflink idref="bib41" id="ref76">41</reflink>]) may operate. Enculturation involves the communication of values, including social goals, through language and behavior. Parents may influence their children's appraisal systems through family talk such that an interpretation of an event leading to anger is more likely than one leading to another emotion ([<reflink idref="bib52" id="ref77">52</reflink>]).</p> <hd id="AN0015916620-15">Children's Behavior Problems Increase the Risk of Partner Conflict</hd> <p>The only aspect of child behavior found to predict increased marital conflict was children's externalizing symptomatology. Furthermore, the only aspect of marital conflict predicted by child behavior was parental argument about children. Thus, we found that the higher the children's externalizing scores at Time 1, the more the parental argument about that child increased over time. This family context effect was most marked in families with middle to high family average externalizing. In families with low family average externalizing (in which by definition none of the children is showing major externalizing problems), the externalizing behavior of children did not result in increased argument between parents. This is likely to be because even though the child showed more problems than his or her siblings, the level of behavior was not sufficient to increase parental conflict. The effect of externalizing behavior on parental argument about children was stronger in stepfamilies than in biological families.</p> <p>The first issue to consider is why externalizing behavior in children elicits more argument between couples. Externalizing problems in children may elicit strong feelings of frustration, disappointment, and humiliation in parents ([<reflink idref="bib48" id="ref78">48</reflink>]). With potential consequences being serious, parents may feel more highly invested in their own perception of the problem. This mixture of serious consequences and heightened affective arousal may make conflict more likely. It may also be that affect contagion and enculturation, described earlier, operate when we consider the influence of children's affect (i.e., anger and aggression seen in externalizing problems) on parental conflict. Interacting with an angry child through processes of affect matching may increase parental anger, which is then more readily activated when the couple are interacting with one another. With respect to enculturation, as externalizing children, through talk and behavior, show that they blame others for negative events ([<reflink idref="bib15" id="ref79">15</reflink>]) and place a high value on dominance in interaction ([<reflink idref="bib34" id="ref80">34</reflink>]), parents' own goals in relationship may be influenced, both in their parenting and in their marriage.</p> <p>The second issue to consider is why externalizing behavior has a more negative influence on partner conflict in stepfamilies than conflict in biological families. Previous research has shown stepfamily relationships to be more complex and potentially more problematic than relationships in biological families ([<reflink idref="bib32" id="ref81">32</reflink>]). In stepfamilies the nonbiological parent is less involved in discipline and decision making than the biological parent. When stepparents are drawn into the arguments, however, these are more intense ([<reflink idref="bib31" id="ref82">31</reflink>]). It may be, therefore, that the increased effect of children's externalizing behavior on partner conflict in stepfamilies may be mediated by the nonbiological parents' involvement in discipline, an issue to be investigated in future studies.</p> <p>Family systems theorists have long argued that the effect of the family system on the behavior of individual family members is more than the sum of its parts ([<reflink idref="bib43" id="ref83">43</reflink>]). These higher order family processes, such as the family average for parental argument about children and the effect of child‐specific externalizing behavior being contingent on the family average for children's externalizing behavior, are examples of the value of conceptualizing the family system as more than the sum of individual's behavior. Multilevel modeling has just begun to be used for the analysis of family data ([<reflink idref="bib35" id="ref84">35</reflink>]), and one of its advantages is its ability to examine simultaneously the effect of individual‐ and family‐level processes.</p> <hd id="AN0015916620-16">Differential Sibling Experiences of Conflict in Stepfamilies and Biological Families</hd> <p>Although we certainly found some evidence for a family style when it comes to arguing about children and exposure to conflict, differential sibling experience was also evident. Siblings in stepfather families and complex stepfamilies experienced more differential argument and more differential exposure than did siblings living in biological families. We also found that the level of externalizing behavior among children had a more negative effect on argument about children in stepfather and complex families than in biological families. Thus, the stepfamily configuration is amplifying the effect of children's externalizing behavior on the parental marriage, supporting the findings of others that negative family processes can be exacerbated in stepfamilies ([<reflink idref="bib30" id="ref85">30</reflink>], [<reflink idref="bib44" id="ref86">44</reflink>]).</p> <p>Differential experiences of siblings were also evident for children's exposure to parental conflict, although again we also found evidence for a family style of children being exposed to argument. Sibling experiences in stepfather and complex families on their exposure to conflict were found to be more differential than siblings in intact families. Boys were exposed to more parental conflict over time than were girls, as others have found ([<reflink idref="bib33" id="ref87">33</reflink>]). This was not explained by the increased level of behavioral problems in boys. Also, it was not that boys were exposed to conflict more than were girls because they elicited more parental conflict, as boys were not found to elicit more conflict over time than were girls. There has been much literature to suggest that girls are more relationally attuned than boys (e.g., [<reflink idref="bib58" id="ref88">58</reflink>]). It may therefore be that parents try to protect their girls from exposure to their conflict as they fear it will have a greater impact on them as compared with boys.</p> <p>Some limitations in this study must be considered. First, although the autoregressive design we used enabled us to make stronger statements about causal processes than designs that do not examine change in the phenomenon of interest, it remains possible that variables that were not in our model, that covaried with our predictor variables, were the variables responsible for the outcome. It is important, therefore, for future studies to test the hypothesized mechanisms experimentally. For instance, in a design similar to that used by [<reflink idref="bib4" id="ref89">4</reflink>] it would be possible to pair externalizing and nonexternalizing children with married couples and set the couple the task of getting the child to do something that he or she was reluctant to do (e.g., homework, a clean‐up task). Is conflict between the couple more evident in couples dealing with children with externalizing problems? If siblings are observed pre‐ and post‐parental conflict, with conflict elicited by the experimenters, is sibling behavior more differentiated during postconflict than during preconflict?</p> <p>Limitations in the measures should also be considered. First, the measurement of argument about children involved multiple items and was therefore better than the single‐item measurement for exposure to conflict. This may have resulted in stronger findings for argument about children. Second, the same person (the mother) reported on all siblings' exposure to conflict and argument about children. If mothers are not very sensitive to the arguments they have with their partners about their different children, the measure may be functioning more as a between‐family rather than a within‐family measure even though it was designed as the latter. It may also be that ability to differentiate accurately between children either on their own behavior or on the behavior they evoke in others varies as a function of environmental stress. Wahler and Dumas (1989) have suggested that mothers' accuracy of perceptions about their children's behavior may be a function of the effect of stress on the attentional system. They argued that as attention becomes narrowed under conditions of stress, parents have been shown to misperceive child behavior. This may lead parents to see siblings as behaving more similarly, as results pertaining to argument about children from this study suggest. There are also data to suggest the opposite effect, however. [<reflink idref="bib35" id="ref90">35</reflink>] found that mothers reported treating their children more dissimilarly (with respect to positive engagement) at high levels of socioeconomic stress. As within‐family research proceeds, it will be important to find ways to distinguish between these different sources of the shared family effect. Incorporation of experience‐sampling methodology ([<reflink idref="bib5" id="ref91">5</reflink>]) or diaries (Papp, Cummings, & Goeke‐Morey, 2002) for events such as argument about children that are unlikely to be observed might prove useful tools in the interpretation of shared and child‐specific effects. Similarly, the incorporation of child‐report data will also be useful in this regard, although this will constrain the age of the sample given concerns about reliability of report in young children.</p> <ref id="AN0015916620-17"> <title> Footnotes </title> <blist> <bibl id="bib1" idref="ref40" type="bt">1</bibl> <bibtext> Jennifer Jenkins, Department of Human Development and Applied Psychology, University of Toronto; Judy Dunn, Institute of Psychiatry, University of London; Thomas G. O'Connor, University of Rochester Medical Center; Jon Rasbash, Institute of Education, University of London; Anna Simpson, Department of Human Development and Applied Psychology, University of Toronto.</bibtext> </blist> <blist> <bibl id="bib2" idref="ref41" type="bt">2</bibl> <bibtext> We would like to thank the Medical Research Council, UK, for funding this study and the Avon Brothers and Sisters Study families for taking part. We would also like to thank the Social Sciences and Humanities Research Council, Canada, and the UK Economic and Social Research Council's Research Methods Program for support to individuals. Thanks, too, to Tracey Curwen, Tom Hollenstein, Claudia Megna, Paulo Pires, and Kevin Runions for comments on an earlier draft.</bibtext> </blist> </ref> <ref id="AN0015916620-18"> <title> References </title> <blist> <bibtext> Achenbach, T. M. (1991a). Manual for the Child Behavior Checklist/4–18 and 1991 profiles. Burlington, VT: University of Vermont, Department of Psychiatry.</bibtext> </blist> <blist> <bibtext> Achenbach, T. M. (1991b). Manual for the Teacher's Report Form and 1991 profiles. Burlington, VT: University of Vermont, Department of Psychiatry.</bibtext> </blist> <blist> <bibl id="bib3" idref="ref58" type="bt">3</bibl> <bibtext> Allison, P. D. (2002). Missing data. Thousand Oaks, CA: Sage.</bibtext> </blist> <blist> <bibl id="bib4" idref="ref89" type="bt">4</bibl> <bibtext> Anderson, K., Lytton, H., & Romney, D. M. (1986). Mothers' interactions with normal and conduct‐disordered boys. Developmental Psychology, 22, 604 – 609.</bibtext> </blist> <blist> <bibl id="bib5" idref="ref91" type="bt">5</bibl> <bibtext> Barrett, L. F., & Barrett, D. J. (2001). An introduction to computerized experience sampling in psychology. Social Science Computer Review, 19, 175 – 185.</bibtext> </blist> <blist> <bibl id="bib6" idref="ref13" type="bt">6</bibl> <bibtext> Block, J., Block, J. H., & Gjerde, P. F. (1989). Parental functioning and the home environment of divorce : Prospective and concurrent analyses. Annual Progress in Child Psychiatry and Child Development, 192 – 207.</bibtext> </blist> <blist> <bibl id="bib7" idref="ref59" type="bt">7</bibl> <bibtext> Browne, W. (2003). Multivariate normal response models and missing data. In MCMC estimation in MLwiN (pp. 207 – 227). London: Institute of Education.</bibtext> </blist> <blist> <bibl id="bib8" idref="ref14" type="bt">8</bibl> <bibtext> Cherlin, A. J., Furstenberg, F. F., Chase‐Lansdale, P. L., Kiernan, K. E., Robins, P. K., Morrison, D. R., et al. (1991). Longitudinal studies of effects of divorce in children in Great Britain and the United States. Science, 252, 1386 – 1389.</bibtext> </blist> <blist> <bibl id="bib9" idref="ref47" type="bt">9</bibl> <bibtext> Clingempeel, W. G., Ievoli, R., & Brand, E. (1984). Structural complexity and the quality of stepfather–stepchild relationships. Family Process, 23, 547 – 560.</bibtext> </blist> <blist> <bibtext> Coleman, M., Fine, M. A., Ganong, L. H., Downs, K. J. M., & Puak, N. (2001), Personal Relationship, 8, 55 – 73.</bibtext> </blist> <blist> <bibtext> Cowan, C. P., & Cowan, P. (2000). When partners become parents: The big life change for couples. Mahwah, NJ: Erlbaum.</bibtext> </blist> <blist> <bibtext> Cummings, E. M. (1987). Coping with background anger in early childhood. Child Development, 58, 976 – 984.</bibtext> </blist> <blist> <bibtext> Cummings, E. M., & Davies, D. (1994). Children and marital conflict: The impact of family dispute and resolution. New York: Guilford.</bibtext> </blist> <blist> <bibtext> Davies, P. T., Harold, G. T., Goeke‐Morey, M. C., & Cummings, E. M. (2002). Child emotional security and interparental conflict. Monographs of the Society for Research in Child Development, 67 Serial number 270(3).</bibtext> </blist> <blist> <bibtext> Dodge, K. A., Bates, J. E., & Pettit, G. S. (1990). Mechanisms in the cycle of violence. Science, 250, 1678 – 1683.</bibtext> </blist> <blist> <bibtext> Dunn, J., Davies, L. C., O'Connor, T. G., & Sturgess, W. (2000). Parents' and partners' life course and family experiences : Links with parent–child relationships in different family settings. Journal of Child Psychology and Psychiatry and Allied Disciplines, 41, 955 – 968.</bibtext> </blist> <blist> <bibtext> Erel, O., & Burman, B. (1995). Interrelatedness of marital relations and parent–child relations : A meta‐analytic review. Psychological Bulletin, 118, 108 – 132.</bibtext> </blist> <blist> <bibtext> Gaither, R., Bingen, K., & Hopkins, J. (2000). When the bough breaks : The relationship between chronic illness in children and couple functioning. In K. B. Schmaling & T. G. Sher (Eds.), The psychology of couples and illness: Theory research, & practice (pp. 337 – 365). Washington, D.C.: American Psychological Association.</bibtext> </blist> <blist> <bibtext> Golding, J. (1996). Children of the nineties : A resource for assessing the magnitude of long‐term effects of prenatal and perinatal events. Contemporary Reviews in Obstetrics and Gynaecology, 8, 89 – 92.</bibtext> </blist> <blist> <bibtext> Goldstein, H. M. (1995). Multilevel statistical models. London: Edward Arnold.</bibtext> </blist> <blist> <bibtext> Goldstein, H. M. (2003). Multilevel statistical models (3rd ed.), London: Hodder Arnold.</bibtext> </blist> <blist> <bibtext> Grych, J. H., & Fincham, F. D. (1990). Marital conflict and children's adjustment : A cognitive‐contextual framework. Psychological Bulletin, 108, 267 – 290.</bibtext> </blist> <blist> <bibtext> Grych, J. H., & Fincham, F. D. (1993). Children's appraisals of marital conflict : Initial investigations of the cognitive‐contextual framework. Child Development, 64, 215 – 230.</bibtext> </blist> <blist> <bibtext> Grych, J. H., & Fincham, F. D. (2001). Child development and interparental conflict. New York: Cambridge University Press.</bibtext> </blist> <blist> <bibtext> Grych, J. H., Harold, G. T., & Miles, C. J. (2003). A prospective investigation of appraisals as mediators of the link between interparental conflict and child adjustment. Child Development., 74, 1176 – 1193.</bibtext> </blist> <blist> <bibtext> Haley, J. (1976). Problem‐solving therapy: New strategies for effective family therapy. San Francisco: Jossey‐Bass.</bibtext> </blist> <blist> <bibtext> Harold, G. T., & Conger, R. D. (1997). Marital conflict and adolescent distress : The role of adolescent awareness. Child Development, 68, 333 – 350.</bibtext> </blist> <blist> <bibtext> Harold, G. T., Fincham, F. D., Osborne, L. N., & Conger, R. D. (1997). Mom and dad are at it again : Adolescent perceptions of marital conflict and adolescent psychological distress. Developmental Psychology, 33, 333 – 350.</bibtext> </blist> <blist> <bibtext> Hatfield, E., Cacioppo, J. T., & Rapson, R. L. (1994). Emotional contagion. Cambridge, England: Cambridge University Press.</bibtext> </blist> <blist> <bibtext> Hetherington, E. M., Bridges, M., & Insabella, G. M. (1998). What matters? What does not? Five perspectives on the association between marital transition and children's adjustment. American Psychologist, 53, 167 – 184.</bibtext> </blist> <blist> <bibtext> Hetherington, E. M., & Clingempeel, W. G. (1992). Coping with marital transitions. Monographs of the Society for Research in Child Development, 57 (2‐3, Serial No. 227).</bibtext> </blist> <blist> <bibtext> Hetherington, E. M., Henderson, S., & Reiss, D. (1999). Adolescent sibling in stepfamilies : Family functioning and adolescent adjustment. Monographs of the Society for Research in Child Development, 64, (4, Serial No. 259).</bibtext> </blist> <blist> <bibtext> Hetherington, E. M., Stanley‐Hagan, M., & Anderson, E. R. (1989). Marital transitions : A child's perspective. American Psychologist, 44, 303 – 312.</bibtext> </blist> <blist> <bibtext> Jenkins, J., & Greenbaum, R. (1999). Intention and emotion in child psychopathology : Building cooperative plans. In P. Zelazo, J. W. Astington, & D. Olson, (Eds.), Theories of mind in action: Development and evolution of social understanding and self control (pp. 269 – 291). Mahwah, NJ: Erlbaum.</bibtext> </blist> <blist> <bibtext> Jenkins, J. M., Rasbash, J., & O'Connor, T. G. (2003). The role of shared family context in differential parenting. Developmental Psychology, 39, 99 – 113.</bibtext> </blist> <blist> <bibtext> Jenkins, J. M., & Smith, M. A. (1990). Factors protecting children living in disharmonious homes : Maternal reports. Journal of the American Academy of Child and Adolescent Psychiatry, 29, 60 – 69.</bibtext> </blist> <blist> <bibtext> Kurdek, L. A. (1991). Marital stability and changes in marital quality in newly wed couples : A test of the contextual model. Journal of Social and Personal Relationships, 8, 27 – 48.</bibtext> </blist> <blist> <bibtext> Leve, L. D., Scaramella, L. V., & Fagot, B. I. (2001). Infant temperament, pleasure in parenting, and marital happiness in adoptive families. Infant Mental Health Journal, 22, 545 – 558.</bibtext> </blist> <blist> <bibtext> Locke, H. J., & Wallace, K. M. (1987). Marital Adjustment Test. In N. Freedman & R. Sherman (Eds.), Handbook of measurements for marriage and family therapy (pp. 46 – 50). New York: Bruner/Mazel.</bibtext> </blist> <blist> <bibtext> Margolin, G., Christensen, A., & John, R. S. (1996). The continuance and spillover of everyday tensions in distressed and nondistressed families. Journal of Family Psychology, 10, 304 – 321.</bibtext> </blist> <blist> <bibtext> Mesquita, B. (2003). Emotions as dynamic cultural phenomena. In R. J. Davidson, K. Scherer, & H. H. Goldsmith (Eds.), Handbook of affective sciences (pp. 871 – 890). Oxford, England: Oxford University Press.</bibtext> </blist> <blist> <bibtext> Minuchin, S. (1974). Families and family therapy. Cambridge, MA: Harvard University Press.</bibtext> </blist> <blist> <bibtext> Nichols, M. P., & Schwartz, R. C. (2001). Family therapy: Concepts and methods. Boston: Allyn & Bacon.</bibtext> </blist> <blist> <bibtext> O'Connor, T. G., & Insabella, G. (1999). Marital satisfaction, relationships and roles. In E. M. Hetherington, S. Henderson, & D. Reiss (Eds.), Adolescent sibling in stepfamilies: Family functioning and adolescent adjustment. Monographs of the Society for Research in Child Development, 64, (4, Serial No. 259)</bibtext> </blist> <blist> <bibtext> O'Connor, T. G., Pickering, K., Dunn, J., & Golding, J. (1999). Frequency and predictors of relationship dissolution in a community sample in England. Journal of Family Psychology, 13, 436 – 449.</bibtext> </blist> <blist> <bibtext> Office of Population Censuses and Surveys and Employment Department Group. (1980). Standard classification of occupations. London: Her Majesty's Stationery Office.</bibtext> </blist> <blist> <bibtext> Papp, L. M., Cummings, E. M., & Goeke‐Morey, M. C. (2002). Marital conflicts in the home when children are present versus absent. Developmental Psychology, 38, 774 – 783.</bibtext> </blist> <blist> <bibtext> Patterson, G. R. (1982). Coercive family process. Eugene, OR: Castalia.</bibtext> </blist> <blist> <bibtext> Plomin, R., & Daniels, D. (1987). Why are children in the same family so different from one another? Behavioral & Brain Sciences, 10, 1 – 16.</bibtext> </blist> <blist> <bibtext> Rasbash, J., Browne, W., Goldstein, H., Yang, M., Plewis, I., Healy, M., et al. (2000). A user's guide to MLwiN, Version 2.1. London: University of London, Institute of Education.</bibtext> </blist> <blist> <bibtext> Richmond, M. K., & Stocker, C. M. (2003). Siblings' differential experiences of marital conflict and differences in psychological adjustment. Journal of Family Psychology, 17, 339 – 350.</bibtext> </blist> <blist> <bibtext> Root, C. A., & Jenkins, J. M. (in press). Maternal appraisal styles, family risk status and anger biases of children. Journal of Abnormal Child Psychology.</bibtext> </blist> <blist> <bibtext> Rubin, D. B. (1987). Multiple imputation for non‐response in surveys. New York: Wiley.</bibtext> </blist> <blist> <bibtext> Rubin, K. H., Burgess, K. B., Dwyer, K. M., & Hastings, P. D. (2003). Predicting preschoolers' externalizing behaviors from toddler temperament, conflict, and maternal negativity. Developmental Psychology, 39, 164 – 176.</bibtext> </blist> <blist> <bibtext> Ryder, R. G. (1973). Longitudinal data relating marriage satisfaction to having a child. Journal of Marriage and the Family, 35, 604 – 606.</bibtext> </blist> <blist> <bibtext> Schafer, J. L. (1996). Analysis of incomplete multivariate data. London: Chapman & Hall.</bibtext> </blist> <blist> <bibtext> Snyder, D. K., Klein, M. A., Gdowski, C. L., Faulstich, C., & LaCombe, J. (1988). Generalized dysfunction in clinic and nonclinic families : A comparative analysis. Journal of Abnormal Child Psychology, 16, 97 – 109.</bibtext> </blist> <blist> <bibtext> Strough, J., & Berg, C. A. (2000). Goals as a mediator of gender differences in high‐affiliation dyadic conversations. Developmental Psychology, 36, 117 – 125.</bibtext> </blist> <blist> <bibtext> Turkeimer, E., & Waldron, M. (2000). Nonshared environment : A theoretical, methodological and quantitative review. Psychological Bulletin, 126, 78 – 108.</bibtext> </blist> <blist> <bibtext> Tzeng, J. M., & Mare, R. D. (1995). Labor market and socioeconomic effects on marital stability. Social Science Research, 24, 329 – 351.</bibtext> </blist> <blist> <bibtext> Wahler, R. G., & Dumas, J. E. (1989). Attentional problems in dysfunctional mother‐child interactions: An interbehavioral model. Psychological Bulletin, 105, 116 – 130.</bibtext> </blist> </ref> <aug> <p>By Jennifer Jenkins; Anna Simpson; Judy Dunn; Jon Rasbash and Thomas G. O'Connor</p> <p>Reported by Author; Author; Author; Author; Author</p> </aug> <nolink nlid="nl1" bibid="bib13" firstref="ref1"></nolink> <nolink nlid="nl2" bibid="bib22" firstref="ref2"></nolink> <nolink nlid="nl3" bibid="bib36" firstref="ref3"></nolink> <nolink nlid="nl4" bibid="bib14" firstref="ref4"></nolink> <nolink nlid="nl5" bibid="bib25" firstref="ref5"></nolink> <nolink nlid="nl6" bibid="bib27" firstref="ref6"></nolink> <nolink nlid="nl7" bibid="bib44" firstref="ref7"></nolink> <nolink nlid="nl8" bibid="bib55" firstref="ref9"></nolink> <nolink nlid="nl9" bibid="bib11" firstref="ref10"></nolink> <nolink nlid="nl10" bibid="bib38" firstref="ref11"></nolink> <nolink nlid="nl11" bibid="bib18" firstref="ref12"></nolink> <nolink nlid="nl12" bibid="bib23" firstref="ref15"></nolink> <nolink nlid="nl13" bibid="bib57" firstref="ref16"></nolink> <nolink nlid="nl14" bibid="bib12" firstref="ref18"></nolink> <nolink nlid="nl15" bibid="bib26" firstref="ref19"></nolink> <nolink nlid="nl16" bibid="bib42" firstref="ref20"></nolink> <nolink nlid="nl17" bibid="bib40" firstref="ref21"></nolink> <nolink nlid="nl18" bibid="bib17" firstref="ref22"></nolink> <nolink nlid="nl19" bibid="bib28" firstref="ref23"></nolink> <nolink nlid="nl20" bibid="bib49" firstref="ref24"></nolink> <nolink nlid="nl21" bibid="bib59" firstref="ref25"></nolink> <nolink nlid="nl22" bibid="bib35" firstref="ref28"></nolink> <nolink nlid="nl23" bibid="bib51" firstref="ref29"></nolink> <nolink nlid="nl24" bibid="bib45" firstref="ref30"></nolink> <nolink nlid="nl25" bibid="bib60" firstref="ref31"></nolink> <nolink nlid="nl26" bibid="bib37" firstref="ref32"></nolink> <nolink nlid="nl27" bibid="bib31" firstref="ref33"></nolink> <nolink nlid="nl28" bibid="bib10" firstref="ref34"></nolink> <nolink nlid="nl29" bibid="bib32" firstref="ref35"></nolink> <nolink nlid="nl30" bibid="bib30" firstref="ref36"></nolink> <nolink nlid="nl31" bibid="bib19" firstref="ref38"></nolink> <nolink nlid="nl32" bibid="bib16" firstref="ref39"></nolink> <nolink nlid="nl33" bibid="bib39" firstref="ref44"></nolink> <nolink nlid="nl34" bibid="bib46" firstref="ref49"></nolink> <nolink nlid="nl35" bibid="bib53" firstref="ref56"></nolink> <nolink nlid="nl36" bibid="bib56" firstref="ref57"></nolink> <nolink nlid="nl37" bibid="bib20" firstref="ref61"></nolink> <nolink nlid="nl38" bibid="bib21" firstref="ref62"></nolink> <nolink nlid="nl39" bibid="bib50" firstref="ref63"></nolink> <nolink nlid="nl40" bibid="bib24" firstref="ref70"></nolink> <nolink nlid="nl41" bibid="bib54" firstref="ref71"></nolink> <nolink nlid="nl42" bibid="bib29" firstref="ref75"></nolink> <nolink nlid="nl43" bibid="bib41" firstref="ref76"></nolink> <nolink nlid="nl44" bibid="bib52" firstref="ref77"></nolink> <nolink nlid="nl45" bibid="bib48" firstref="ref78"></nolink> <nolink nlid="nl46" bibid="bib15" firstref="ref79"></nolink> <nolink nlid="nl47" bibid="bib34" firstref="ref80"></nolink> <nolink nlid="nl48" bibid="bib43" firstref="ref83"></nolink> <nolink nlid="nl49" bibid="bib33" firstref="ref87"></nolink> <nolink nlid="nl50" bibid="bib58" firstref="ref88"></nolink>
Header DbId: eric
DbLabel: ERIC
An: EJ685656
AccessLevel: 3
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Mutual Influence of Marital Conflict and Children's Behavior Problems: Shared and Nonshared Family Risks
– Name: Language
  Label: Language
  Group: Lang
  Data: English
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Jenkins%2C+Jennifer%22">Jenkins, Jennifer</searchLink><br /><searchLink fieldCode="AR" term="%22Simpson%2C+Anna%22">Simpson, Anna</searchLink><br /><searchLink fieldCode="AR" term="%22Dunn%2C+Judy%22">Dunn, Judy</searchLink>
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="SO" term="%22Child+Development%22"><i>Child Development</i></searchLink>. Jan 2005 76(1):24-39.
– Name: Avail
  Label: Availability
  Group: Avail
  Data: Journal Customer Services, Blackwell Publishing, 350 Main Street, Malden, MA 02148. Tel: 800-835-6770 (Toll Free); Fax: 781-388-8232; e-mail: subscrip@bos.blackwellpublishing.com.
– Name: PeerReviewed
  Label: Peer Reviewed
  Group: SrcInfo
  Data: Y
– Name: Pages
  Label: Page Count
  Group: Src
  Data: 16
– Name: DatePubCY
  Label: Publication Date
  Group: Date
  Data: 2005
– Name: TypeDocument
  Label: Document Type
  Group: TypDoc
  Data: Journal Articles<br />Reports - Evaluative
– Name: Audience
  Label: Education Level
  Group: Audnce
  Data: <searchLink fieldCode="EL" term="%22Elementary+Secondary+Education%22">Elementary Secondary Education</searchLink>
– Name: Subject
  Label: Descriptors
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Child+Behavior%22">Child Behavior</searchLink><br /><searchLink fieldCode="DE" term="%22Behavior+Problems%22">Behavior Problems</searchLink><br /><searchLink fieldCode="DE" term="%22Siblings%22">Siblings</searchLink><br /><searchLink fieldCode="DE" term="%22Conflict%22">Conflict</searchLink><br /><searchLink fieldCode="DE" term="%22Parent+Child+Relationship%22">Parent Child Relationship</searchLink><br /><searchLink fieldCode="DE" term="%22Longitudinal+Studies%22">Longitudinal Studies</searchLink><br /><searchLink fieldCode="DE" term="%22Family+%28Sociological+Unit%29%22">Family (Sociological Unit)</searchLink><br /><searchLink fieldCode="DE" term="%22Marital+Satisfaction%22">Marital Satisfaction</searchLink><br /><searchLink fieldCode="DE" term="%22Children%22">Children</searchLink><br /><searchLink fieldCode="DE" term="%22Adolescents%22">Adolescents</searchLink>
– Name: DOI
  Label: DOI
  Group: ID
  Data: 10.1111/j.1467-8624.2005.00827.x
– Name: ISSN
  Label: ISSN
  Group: ISSN
  Data: 0009-3920
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: This within-family, longitudinal study including biological and stepfamilies investigated mutual influences between marital conflict and children's behavior problems. Children (4 to 17 years; N296) residing in 127 families drawn from a general population study were investigated at Time 1 and again 2 years later. These nested data were analyzed using multilevel modeling, controlling for previous child behavior or marital conflict. Marital conflict about children predicted change in children's behavior. Children's behavior also predicted an increase in marital conflict, particularly in stepfamilies. Differences between siblings in exposure to conflict and the extent to which siblings were a source of argument increased more in stepfamilies than in biological families. Boys were exposed to more conflict over time than were girls.
– Name: AbstractInfo
  Label: Abstractor
  Group: Ab
  Data: Author
– Name: Ref
  Label: Number of References
  Group: RefInfo
  Data: 61
– Name: DateEntry
  Label: Entry Date
  Group: Date
  Data: 2005
– Name: AN
  Label: Accession Number
  Group: ID
  Data: EJ685656
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=eric&AN=EJ685656
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1111/j.1467-8624.2005.00827.x
    Languages:
      – Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 16
        StartPage: 24
    Subjects:
      – SubjectFull: Child Behavior
        Type: general
      – SubjectFull: Behavior Problems
        Type: general
      – SubjectFull: Siblings
        Type: general
      – SubjectFull: Conflict
        Type: general
      – SubjectFull: Parent Child Relationship
        Type: general
      – SubjectFull: Longitudinal Studies
        Type: general
      – SubjectFull: Family (Sociological Unit)
        Type: general
      – SubjectFull: Marital Satisfaction
        Type: general
      – SubjectFull: Children
        Type: general
      – SubjectFull: Adolescents
        Type: general
    Titles:
      – TitleFull: Mutual Influence of Marital Conflict and Children's Behavior Problems: Shared and Nonshared Family Risks
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Jenkins, Jennifer
      – PersonEntity:
          Name:
            NameFull: Simpson, Anna
      – PersonEntity:
          Name:
            NameFull: Dunn, Judy
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 01
              Type: published
              Y: 2005
          Identifiers:
            – Type: issn-print
              Value: 0009-3920
          Numbering:
            – Type: volume
              Value: 76
            – Type: issue
              Value: 1
          Titles:
            – TitleFull: Child Development
              Type: main
ResultId 1