Marital Conflict and Support Seeking by Parents in Adolescence: Empirical Support for the Parentification Construct

Abstract

Parentification, a parent–child dynamic wherein children come to provide ongoing emotional support for their parents, has been documented extensively in the clinical literature; however, it rarely has been studied systematically. Using a community sample of 83 couples and their adolescent children (mean age = 15.26 years; 52% male, 48% female), the authors linked adolescent self-report of parentification to specific youth and adult behaviors using multiple methods and examined its associations with youth adjustment problems. The parentification measure demonstrated strong internal consistency and 1-year stability. Parentification was associated with marital conflict, youth perceptions of threat, low warmth in the parent–child relationship, and the tendency for youths to intervene in marital conflict. Links were also found with youth reports of internalizing and externalizing behavior and poorer competency in close friendships. These findings thus support the parentification construct and provide evidence that parentification may contribute to poor youth outcomes by burdening children with developmentally inappropriate responsibilities.

Exposure to marital conflict places children at increased risk for a host of adjustment problems (Cummings & Davies, 2002), and considerable research suggests that declines in parenting and parent–child relationships may mediate this risk (Cummings & Davies, 1994). Typically, work in this arena has emphasized specific parenting problems likely to emerge in the context of marital discord, such as inconsistent discipline or disengaged or authoritarian parenting (Kelly, 2000). Although the effects of these practices are well established, parenting behaviors that may directly convey the strains of marital discord have received less attention. In particular, there has been limited empirical examination of the tendency for parents to rely on their children for emotional support during times of stress. This situation is commonly referred to as parentification because of the parental role it requires children to assume. Despite frequent reference in the marital conflict and divorce literatures, parentification has remained a theoretically compelling construct with little empirical support (Peris & Emery, 2005).

In the present investigation, we address this gap in the literature, examining the reliability and validity of a measure of parentification and exploring its association with adolescent adjustment. Specifically, we assessed (a) the stability of a particular dimension of the construct (emotional parentification) over time, as well as its links to (b) specific adult and child behaviors with which it is theoretically associated, and (c) adolescent adjustment problems.

Parenting Behavior as a Mediator of the Effects of Marital Conflict

Marital conflict is linked to a host of ineffective parenting practices, ranging from the use of inconsistent discipline to diminished displays of warmth and responsivity in the parent–child relationship (Krishnakumar & Buehler, 2000). Although these practices are documented as mediators of the link between marital conflict and poor youth outcomes (Cummings & Davies, 1994; O’Leary & Vidair, 2005; Schoppe-Sullivan, Schermerhorn, & Cummings, 2007), parent–child interactions that reflect more subtle affective family dynamics have received less attention. Parentification holds utility in this regard, as it is a dynamic that on the surface suggests closeness in the parent–child relationship but conceptually may be a source of child distress. Specifically, parentified children are doubly burdened because they not only witness parental conflict as a third party to marital discord but are also called upon to comfort parents concerning adult distress rather than their own. Viewed through the lens of attachment theory (Bowlby, 1980), the distress created by such parental behavior may foster an approach–avoidance conflict in which the parent is both the desired source of support as well as a simultaneous threat to security (Byng-Hall, 2002). Although appealing from various theoretical perspectives, parentification has remained an under-researched aspect of parenting, and its function as a risk factor for poor adolescent adjustment has yet to be established empirically.

Defining Emotional Parentification

To date, parentification has been described with a variety of labels (e.g., “boundary dissolution,” “role reversal,” “spousification,” and “seductive parenting”), some of which are conceptually distinct, and some of which share overlapping meanings (see Kerig, 2005, for a review). Broadly construed, parentification is a form of role reversal wherein children come to care for their parents. However, caring can encompass a range of responsibilities, and its consequences can vary depending on the developmental status of the child. As emotional caregiving is a particularly taxing responsibility, and the effects of marital conflict on parent–child relationships are especially salient as youths transition from early to midadolescence (Hetherington, 2006), in the present investigation we focused specifically on emotional caregiving during middle adolescence. Emotional parentification (hereafter referred to as parentification) is operationally defined herein as the child’s perception that he or she is expected to meet a parent’s emotional need for support or companionship (e.g., by serving as confidante, dispute mediator, or decision maker). Parentification may originate from either the parent or the child. For the parent, it may involve inappropriate support seeking from the child or efforts to elevate the child to peer status. From the child’s perspective, it may include active efforts to comfort a parent or relate as a peer, or otherwise assume responsibility for the parent’s well being.

Empirical Investigation of Parentification

Most extant research has focused on the link between childhood parentification and difficulty in later adult relationships. Adults who retrospectively report experiencing this dynamic during childhood are likely to report feeling that they were treated unfairly and to express discontent with the quality of their homelife (Jurkovic, Thirkfield, & Morrell, 2001). Research has also linked parentification to shame and guilt (Wells & Jones, 2000), dependent behavior in adult relationships (Wells, Glickauf-Hughes, & Jones, 1999), poorer academic performance among college students (Chase, Deming, & Wells, 1998), higher levels of depression and anxiety in early adulthood (Hetherington, 1999), and difficulty in peer relationships (M. T. Martin, 1996). Notably, these findings consistently suggest that girls are at greater risk for parentification than boys. However, gender may moderate risk such that paternal parentification is more detrimental to boys, whereas maternal role reversal carries more adverse consequences for girls (Macfie, Houts, McElwain, & Cox, 2005).

Studies of parentification and concurrent childhood adjustment also suggest this dynamic has harmful effects. Specifically, parentification has been linked to both internalizing and externalizing behavior problems (Carlson, Jacobvitz, & Sroufe, 1995; Jacobvitz, Hazen, Curran, & Hitchens, 2004; Johnston, 1990; Stein, Riedel, & Rotheram-Borus, 1999) and to poorer social adjustment (Macfie et al., 2005). Notably, research with youth populations has focused primarily on toddlers and children; examination of the parentification dynamic in adolescent samples has been limited (e.g., Hetherington, 1999). This is striking given that emotional caregiving is likely to be especially problematic during adolescence, when the increasingly salient drive for independence competes with needs for closeness (Davies, 2002). Indeed, Cummings, Ballard, El-Sheikh, and Lake (1991) found that children’s tendencies to mediate in parents’ conflicts peaked in middle adolescence (i.e., 14 years of age). Given that this is a prime time for such behavior, one might also expect parentification to be at a high point in this period.

Several other aspects of the parentification construct are in need of empirical scrutiny. First, although theory posits that the parent–child dynamic is a stable aspect of the family system (Macfie et al., 2005), its continuity during adolescence has yet to be established. Second, the mechanisms by which parentification works to undermine child well-being require elucidation, particularly with regard to the specific child or adult behaviors that may account for its harmful effects. In particular, it has yet to be established empirically that the parentification process draws children into marital conflict to the extent that they take steps to mediate or resolve it. In addition, differences in how parents and children view the dynamic must be examined, particularly in light of theories stipulating that parents may view the relationship as warm and supportive, whereas children may view it as lacking in these qualities (Boszormenyi-Nagy & Spark, 1984). Understanding disparities in how parents and children view the parentification dynamic is central to understanding the process by which it may influence child outcomes. Finally, distinctions between parentification and related constructs have yet to be examined empirically. This is particularly important given that parentification is one of myriad terms used to describe situations in which boundaries are violated and child autonomy is undermined. The concept of psychological control (Barber, Stolz, & Olsen, 2005), for example, addresses the extent to which parents use guilt induction and other control strategies to maintain closeness to their children. Distinguishing parentification from this concept is an important step toward validating the construct.

Applying a multi-method, multi-informant approach that measures parentification in both mothers and fathers, in the present study we attempted to addresses these issues. We hypothesized that parentification, as measured via youth self-report, will be (a) stable over time, (b) more common among mothers and their daughters, (c) associated with higher levels of marital conflict and youth appraisals of threat, (d) linked to increased likelihood of adolescent attempts to intervene in marital conflict, (e) associated with adolescent report of lower levels of parental support but with increased parental perceptions of closeness and support, (f) associated with higher rates of internalizing and externalizing behaviors in adolescents, and (g) predictive of adolescent adjustment problems above and beyond existing measures of parental psychological control.

Method

Overview

These data were collected as part of a larger, longitudinal investigation of marital conflict and child development. In brief, this larger study employed a longitudinal design to follow 300 children (M = 11.1 years, SD = 2.3) and their maritally intact parents across three annual waves of data collection. Attrition from Wave 1 to Wave 3 was minimal (16%), and families who discontinued study involvement did not differ from the remaining sample with regard to marital satisfaction or parent report of child behavior problems on the Child Behavior Checklist/4–18 (CBCL; Achenbach, 1991a). Sample characteristics and procedures relevant to the current purposes of examining parentification are presented below.

Participants

Eighty-three midadolescents between the ages of 14 and 18 years (mean age = 15.26 years) and their maritally intact or cohabiting parents participated in this study (mean length of marriage = 17.73 years). The present sample was roughly evenly divided between male (51%) and female (49%) adolescents who were African American (11%), Latino (10%), and European American (79%). Of the sample, 42% (29% wives, 25% husbands) reported that their marriages were disharmonious, as indicated by a score of less than 100 on at least one of the spouse’s Marital Satisfaction Inventory questionnaires (Locke & Wallace, 1959). Moreover, 21% of couples (13% of wives, 10% of husbands) reported that their marriages were extremely disharmonious, as indicated by scores of less than 85 on at least one of the spouse’s Marital Satisfaction Inventory questionnaires. Family income was also diverse (median range = $40,000–$65,000; full range = less than $10,000, more than $80,000).

The parentification measure was developed for use with youths 14 years of age and older and thus was administered to a subset of the sample. Given the limited number of youths meeting this age criterion at Waves 1 and 2, in the present study we centered on Wave 3 of the study (gathered between 2001 and 2003), when the largest number of children were in middle adolescence and thus eligible to complete the parentification measure. For analyses examining the stability of parentification over time, we used a subset of the sample with available data at both Waves 2 and 3 (n = 49). These youths did not differ from the larger Wave 3 sample with regard to race, age, family income, or years of marriage; however, they were significantly different with respect to gender, t(81) = −2.37, p < .05, with the stability subsample being more heavily female (60%), and the full sample being equally divided between male and female youth participants. As noted below, there were no differences in the stability of male and female Emotional Parentification Questionnaire (EPQ; M. T. Martin, 1996) reports.

Procedure

We recruited families through letters sent home with children from local schools, postcards sent to community residents, flyers distributed at churches and community events, and newspaper and media advertisements. To be included in the study, parents in each family had to be married or cohabitating for at least 2 years. During each laboratory visit, the purpose of the study was reviewed, informed consent and assent were obtained (from mother, father, and child), and confidentiality guidelines were discussed; the project was approved by the university’s committee for protection of human participants. Parents began by completing self-report measures, and then they participated in a videotaped conflict resolution task. Children began by participating in the analogue tasks, which were interspersed with self-report measures so as to avoid fatigue.

Measures

EPQ (M. T. Martin, 1996)

The EPQ is an 18-item self-report measure that evaluates the extent to which adolescents 14 years of age and older perceive their parents to be emotionally reliant upon them. Adolescents complete separate forms for each parent, responding to items using a 5-point Likert scale ranging from 1 (strongly agree) to 5 (strongly disagree). Items reflect specific behaviors indicative of a role reversal (e.g., “I give my mother a lot of advice about how to handle the problems in her life”) as well as the emotions they may elicit (e.g., “I feel I am responsible for how things are going in my family”). The EPQ demonstrates sound psychometric properties with young adult samples (Mika, Bergner, & Baum, 1987). At Wave 2, the EPQ form for mothers produced an α = .89, and the form for fathers produced an α = .90. At Wave 3, alpha values were .91 and .90 for mothers and fathers, respectively.

Analogue procedures

Youth participants viewed 15 counterbalanced videotaped scenarios in which male and female actors portrayed a range of conflictual marital interactions. They then answered questions regarding their thoughts and feelings and how they might respond if the interaction was to occur at home (Goeke-Morey, Cummings, Harold, & Shelton, 2003). The tendency to intervene in marital conflict was coded on the basis of youth responses to the question, “What would you do if you were in the room with your parents?” Responses to each vignette were rated on a 5-point scale ranging from 0 (no intervention) to 4 (extreme intervention), with youths attempting to call the police, discipline parents, or pleading with them to stop. Two raters coded each vignette response, with high overall interrater reliability among coders (r = .95). The average of the two ratings was used to create an intervention score for each response. These scores were then used to produce an intervention composite score that tabulated the sum of all vignette responses for each child.

Children’s Perceptions of Interparental Conflict Scale (CPIC; Grych, Seid, & Fincham, 1992)

The 12-item Perceived Threat subscale of the CPIC was used to measure how the degree to which youths feel threatened and unable to cope in the face of interparental discord. The CPIC is a psychometrically sound instrument (Cummings & Davies, 1994), with Cronbach’s alpha for this sample = .77.

Children’s Report of Parenting Behavior Inventory (CRPBI; Margolies & Weintraub, 1977)

The 56-item form of the CRPBI is a widely used and well-validated Likert scale measuring the quality of the parent–child relationship. The scale yields separate measures for acceptance versus rejection (the extent to which parents are warm and supportive), lax control versus firm control (the extent to which disciplinary strategies are permissive vs. firm), and psychological control versus autonomy (the degree to which parents promote independence). The measure was completed by target youths separately regarding their mother and father. Parents each completed a self-report version of the Parenting Behavior Inventory regarding their own parenting. Cronbach’s alphas ranged from .78 to .95.

CBCL (Achenbach, 1991a)

To obtain multiple perspectives on adolescent functioning, we had both mothers and fathers complete this widely used, psychometrically sound measure. The measure is composed of 113 items that assess behavior problems in both the internalizing and externalizing spectra. All scales of the CBCL are well-normed and validated (Achenbach, 1991a).

Youth Self-Report (YSR; Achenbach, 1991b)

The youth-report version of the CBCL/4-18 was administered to adolescents as an additional measure of youth adjustment. Children with a fifth-grade reading level can easily complete the measure, which has extensive norms and validation (Achenbach, 1991b).

Self-Perception Profile for Children (SPPC; Harter, 1985)

This youth self-report measure provides an index of positive mental health in five different domains (scholastic competence, social acceptance, athletic competence, physical appearance, and behavioral conduct) as well as an index of global self-worth. The SPPC is a psychometrically sound measure that has demonstrated good internal consistency, reliability, and validity (see Winters, Meyers, & Proud, 2002). Within the present sample, alpha values ranged from .69 to .86.

O’Leary Porter Scale (OPS; Porter & O’Leary, 1980)

The OPS is a 10-item measure on which parents report the frequency of overt marital conflict witnessed by children using a 5-point Likert scale ranging from never to very often. Items are summed to produce a total score for inter-parental conflict. The OPS has well-established reliability and validity (Kerig, 2005). Cronbach’s alphas were .74 and .77 for mothers and fathers, respectively.

Results

Overview

Data analysis primarily focused on identifying concurrent associations between emotional parentification and theoretically related constructs (i.e., marital conflict, youth intervention in marital conflict, and parenting behavior). On the basis of the extant literature (e.g., M. T. Martin, 1996), we anticipated small to medium size effects (Cohen, 1988), with 86% power to detect effects in the present analyses. With the exception of the analog task (6% missing), study measures had no more than 1% missing data. For all measures, missing values were random in their distribution and were imputed using the sample mean for continuous variables and the mode for categorical variables. Study measures were plotted and assessed for distribution normality and for the presence of outliers; these data were rare and thus were not removed.

Descriptive Analyses

The means, standard deviations, and ranges for the EPQ and other key study measures are presented in Table 1. Youth reports on both mothers and fathers revealed generally low levels of emotional parentification. Further, youth reports of maternal and paternal emotional parentification were highly correlated at both study assessment points, with Wave 2 reports producing an r = .81, p < .01, and Wave 3 reports producing an r = .82, p < .01. These correlations did not differ significantly by child gender.

Table 1.

Descriptive Statistics for the Emotional Parentification Questionnaire (EPQ) and Other Key Study Measures

Measure (Respondent)	Wave 3
	M	SD	Range
EPQ—Youth for mother	38.30	14.35	18–81
EPQ—Youth for father	35.48	12.65	18–72
CPIC—Youth	5.87	1.22	0–21
Analogue task—Youth	16.29	11.89	0–42.50
CRPBI—Youth for mother
Accept/reject	60.98	7.86	35–72
Firm/lax	38.66	4.75	22–47
Psychological control/autonomy	36.59	7.13	20–48
CRPBI—Youth for father
Accept/reject	56.84	10.49	26–71
Firm/lax	37.63	5.56	19–48
Psychological control/autonomy	38.33	6.26	19–48
OPS—Maternal report	18.12	5.16	9–34
OPS—Paternal report	18.30	4.69	9–35
CBCL—Mother internalizing T-score	46.23	10.45	31–76
CBCL—Mother externalizing T-score	48.08	9.64	32–76
CBCL—Mother total problems T-score	46.35	11.10	23–75
CBCL—Father internalizing T-score	45.25	10.38	31–77
CBCL—Father externalizing T-score	47.94	9.17	32–71
CBCL—Father total problems T-score	45.69	10.27	23–68
YSR—Youth internalizing T-score	46.12	10.58	26–73
YSR—Youth externalizing T-score	51.45	9.49	30–79
YSR—Youth total problems T-score	48.77	9.48	27–72
SPPC—Youth behavioral conduct	15.86	2.76	7–20
SPPC—Youth close friendship	16.88	3.45	7–20
SPPC—Youth global self-worth	16.87	2.69	9–20

Note. N = 83. CPIC = Children’s Perceptions of Interparental Conflict Scale (Perceived Threat subscale); CRPBI = Children’s Report of Parenting Behavior Inventory; OPS = O’Leary Porter Scale; CBCL = Child Behavior Checklist/4–18; YSR = Youth Self-Report; SPPC = Self-Perception Profile for Children.

Does Parent or Child Gender Influence Parentification?

A repeated measures analysis of variance was conducted to examine main effects for parent and child gender as well as an interaction between these two variables. This analysis indicated a main effect for parent gender, F(1, 82) = 8.76, p < .01, with youths reporting higher levels of parentification from their mothers versus their fathers. No significant differences emerged for youth gender, with boys and girls making very similar reports on the EPQ. However, a significant interaction emerged between youth and parent gender, F(1, 82) = 5.97, p < .05, with girls more likely to identify mothers as higher in parentification (M = 37.66, SD = 14.56) than fathers (M = 33.21, SD = 13.42), but with boys making no such distinction between mothers (M = 38.95, SD = 14.27) and fathers (M = 37.79, SD = 11.50). In addition, fathers were seen as higher in parentification of boys than of girls, F(1, 82) = 4.14, p < .05, but child gender did not seem to be a factor in mothers’ parentification, F(1, 82) = 0.26, ns.

Is Parentification Stable Over Time?

To address the hypothesis that parentification would be stable over time, we conducted bivariate correlation analyses using Wave 2 and Wave 3 data. The EPQ demonstrated high levels of stability over the 1-year period between study waves, with youth reports of maternal parentification producing an r(49) = .70, p < .01, and reports of paternal parentification producing an r(49) = .71, p < .01.

Is Parentification Associated With Marital Conflict and Youth Appraisals of Threat?

Maternal, but not paternal, OPS ratings of marital conflict were associated with youth report of parentification from both mothers (r = .33, p < .01) and fathers (r = .27, p < .05). Moreover, as expected, youths who reported higher levels of perceived threat (CPIC) in response to marital conflict also reported higher levels of maternal parentification (r = .28, p < .05). Youth reports of perceived threat were not associated with paternal parentification.

Are Parentified Youths More Likely to Intervene in Marital Conflict?

Consistent with expectations, bivariate correlation analyses revealed that youth reports of maternal parentification were associated with increased tendency to intervene in conflict as assessed by the analog task (r = .24, p < .05). Youth reports of paternal parentification were not linked to any measure of youth intervention in marital conflict, and gender did not moderate results.

Is Parentification Linked to Different Parent and Youth Views of the Parent–Child Relationship?

These analyses began by assessing the degree to which parents and children agreed in their overall reports of parenting behavior. With regard to firm/lax discipline, there was small to moderate agreement between youths and their mothers (r = .34, p < .01) and fathers (r = .51, p < .001). Likewise, on measures of acceptance/rejection, adolescents produced similar reports to their mothers (r = .41, p < .001) and fathers (r = .24, p < .05). Finally, on measures of psychological control/autonomy, there was small to moderate agreement between youths and their mothers (r = .53, p < .001) and fathers (r = .36, p < .001).

Correlation analyses were conducted to test the hypothesis that higher levels of parentification would be associated with lower levels of youth-perceived parental support on the acceptance/rejection scale of the CRPBI but higher levels of parent-perceived support. As expected, these analyses indicated that increases in youth report of maternal parentification were associated with decreases in youth perceptions of warmth and support from both mothers (r = −.22, p < .05) and fathers (r = −.28, p < .01). Similarly, high levels of maternal and paternal parentification were associated with youth perceptions of lower levels of autonomy granting behavior from mothers (r = −.24, p < .05) and fathers (r = −.24, p < .05). By contrast, parents’ CRPBI reports were not linked to youth report of parentification.

Is Parentification Associated With Poor Adjustment in Adolescence?

Preliminary t tests revealed no significant differences between reports for boys and girls on any outcome measure (e.g., CBCL, SPPC). Similarly, mother and father reports did not differ significantly on any subscale of the CBCL, and moderate to large correlations emerged between maternal and paternal report of CBCL externalizing (r = .63), internalizing (r = .39), and total behavior problems (r = .61).

Links between the EPQ and youth functioning

Correlation analyses examined associations between youth ratings of maternal and paternal parentification and youth adjustment (see Table 2). These analyses began with maternal and paternal ratings of adolescent adjustment, in an effort to reduce shared method variance stemming from use of youth EPQ reports and youth adjustment measures alone. However, youth adjustment measures were also examined, in light of evidence suggesting that youths may be better respondents on adjustment difficulties during adolescence (Yeh & Weisz, 2001).

Table 2.

Bivariate Correlations Between Wave 3 Maternal Emotional Parentification and Youth Adjustment

Variable	1	2	3	4	5	6	7	8
1. Mom EPQ	1
2. Dad EPQ	.82**	1
3. Internalizing (YSR)	.27*	.12	1
4. Externalizing (YSR)	.32**	.22	.52**	1
5. Total behavior problems (YSR)	.31**	.19	.84**	.84**	1
6. SPPC—Behavior conduct	−.13	−.11	−.25*	−.48**	−.43**	1
7. SPPC—Close friendships	−.24*	−.27*	−.33**	−.13	−.25*	.14	1
8. SPPC—Global self-worth	−.18	−.15	−.41**	−.34**	−.48**	.31**	.31**	1

Note. N = 83. EPQ = Emotional Parentification Questionnaire (completed by youths); YSR = Youth Self-Report; SPPC = Self-Perception Profile for Children.

^*p < .05.

^**p < .01.

Maternal parentification was linked to higher levels of youth reports of internalizing, externalizing, and total behavior problems (see Table 2). Similarly, higher scores on the EPQ were linked to lower levels of perceived competency in close friendships. With regard to parent CBCLs, no significant links were found between youth reports of maternal or paternal parentification and parent report of youth internalizing, externalizing, or total behavior problems.

Does Parentification Predict Youth Adjustment Above and Beyond Psychological Control?

Hierarchical regression analyses were conducted to examine whether parentification predicted youth behavior problems above and beyond measures of psychological control derived from the CRPBI. To reduce common method variance, and because parent and child CRPBI reports of maternal psychological control were highly correlated (r = .53, p < .01), we created a new variable using the average of these two reports (i.e., mother and child). The average of the two raters (rather than simply maternal report) was used on the basis of convention with this measure as well as prior evidence suggesting that children may be better raters of parental psychological control compared with parents (Cummings, Keller, & Davies, 2005). Regression analyses indicated that maternal parentification predicted internalizing, externalizing, and total problems above and beyond a measure of psychological control (see Table 3).

Table 3.

Hierarchical Regression To Predict Wave 3 Youth Behavior Problems (YSR, CBCL) From Maternal Psychological Control (CRPBI) and Maternal Emotional Parentification (EPQ)

Predictor	βa	βb	R2	ΔR2
1. Internalizing T-score (YSR)
Maternal psychological control	−.10	−0.13	.02	.02
Maternal emotional parentification	.25	7.07	.08	.06*
2. Externalizing T-score (YSR)
Maternal psychological control	−.37	−0.44	.16	.16**
Maternal emotional parentification	.27	6.86	.23	.07**
3. Total behavior problems T-score (YSR)
Maternal psychological control	−.27	−0.32	.09	.09**
Maternal emotional parentification	.28	6.96	.17	.08**
4. Youth somatic complaints T-score (CBCL)
Maternal psychological control—Youth report	.03	0.03	.00	.00
Maternal emotional parentification	.24	3.47	.06	.05*

Note. YSR = Youth Self-Report; CBCL = Child Behavior Checklist/4–18; CRPBI = Children’s Report of Parenting Behavior Inventory; EPQ = Emotional Parentification Questionnaire.

^aBeta weights indicate simultaneous control of all other predictors, following entry of the last variable.

^bUnstandardized beta weights.

^*p < .05.

^**p < .01.

Discussion

In this study, we employed a multi-informant, multi-method approach to link adolescent self-report of parentification with specific youth and adult behaviors. Consistent with expectations, parentification ratings were stable over the 1-year period between study waves and were associated with marital conflict, higher levels of perceived threat for adolescents, and an increased tendency for youths to intervene in marital conflict. In addition, parentification was associated with youth perceptions of low warmth and support in the parent–child relationship and with youth reports of internalizing and externalizing symptomatology.

In considering these findings, it is important to note that youths reported relatively low levels of parentification at both time points. Thus, stable ratings of parentification tended to indicate a consistent lack of enmeshed dynamics rather than a continuous presence of them. It may be the case that the subtle nature of this complex family dynamic was not adequately captured by the EPQ measure. However, it seems equally plausible that in a sample of primarily maritally intact families, parentification was a low base-rate occurrence. That is, in families in which both parents are present, there may be less of a need for parents to seek support from children or for children to assume peer status among adults. Regardless of its rate of occurrence, the stability of parentification ratings provides valuable insight into the continuity of affective family dynamics. Moreover, these findings shed light on parentification in community samples and suggest that the relevance of the construct may extend beyond high-risk or divorce samples.

From a methodological standpoint, these findings extend M. T. Martin’s (1996) work on the EPQ measure and provide further evidence for its reliability. From a conceptual perspective, they suggest that the family dynamic captured by the EPQ may reflect a stable pattern of interaction rather than a transient response to stressful circumstances or life events. Certainly, this is in keeping with earlier work in the area (Sroufe, Jacobvitz, Magelsdorf, DeAngelo, & Ward, 1985) as well as a substantial body of literature on the continuity of dysfunctional parenting strategies (see Holden & Miller, 1999).

The present findings also shed light on the issue of who seeks support and who is called upon to provide it. Implicit in most conceptualizations of parentification is the idea that mothers may be more likely than fathers to seek emotional support from their children (Jurkovic, 1997). However, the present investigation is one of only a handful of studies to obtain parentification data on both mothers and fathers (Jacobvitz et al., 2004; Kerig, 2005; Rowa, Kerig, & Geller, 2001). Consistent with earlier findings, youths in this study perceived higher levels of emotional support seeking from their mothers versus fathers, and girls were more likely than boys to report parentification from their mothers (Jacobvitz et al., 2004; Macfie et al., 2005).

This study also attempted to link parentification to key parenting behaviors with which it is associated in the literature theoretically. Specifically, theories of parentification stipulate that emotional support seeking from parents is linked to a lack of sensitivity to child needs (Boszormenyi-Nagy & Spark, 1984). Moreover, parents and children are hypothesized to view parentification differently, with parents perceiving closeness and warmth and children viewing these qualities as lacking (Jurkovic, 1997). The present investigation provides empirical support for these ideas, demonstrating that mothers who were rated high in parentification were more likely to employ parenting strategies that their children viewed as low in warmth and support. Rather than viewing parental pulls for support as a sign of relationship closeness, youths may experience the dynamic as a detriment to their own support and nurturance. By contrast, parents rated high in parentification by their children did not tend to view the relationship as less warm or supportive.

Interestingly, compromises in parenting may occur regardless of which parent does the support seeking. Indeed, maternal parentification was linked to youth perceptions of low warmth in mothers and fathers. Family systems theory suggests that the collective experience of the family unit has implications for individual dyadic relationships. Thus, in this regard, it makes sense that when mothers turn to their children for support, it may have consequences for how children view their fathers. Youths who are drawn into alliances with their mothers may naturally be inclined to view their fathers as less warm or supportive. Alternatively, the difficulties in the interparental subsystem that elicit maternal support seeking may be the very same stressors that cause both parents to be compromised in their parenting abilities.

Critically, these findings suggest that parents and children do not view the parentification dynamic the same way. Although the hypothesis that parentification would be associated with parental perceptions of warmth in the parent–child relationship was not supported, it is clear that youths differ from their parents in their views of the parentification dynamic. Although parents may not view support seeking as a marker of closeness in the parent–child relationship, they may also fail to recognize it as an impediment to support and nurture in the same way that youths do.

As a step toward understanding the process underlying the parentification dynamic, in this study we also examined its links to marital conflict, youth perceptions of threat, and youth intervention in marital disputes. Theories of parentification suggest that parentification emerges in response to family stress (e.g., marital conflict) and that children are at risk because they feel responsible for their parents’ problems and make efforts to manage them. However, research on whether children feel threatened by parental disclosures is limited (Jurkovic et al., 2001), and the specific issue of whether parentification is tied to active efforts to intervene in marital conflict has yet to be addressed. The present findings indicate that for parents, parentification is linked to higher levels of marital conflict, and for youths, it is linked to greater perceptions of threat when conflict occurs. Moreover, they indicate that parentification is associated with an increased tendency for youths to intervene in interparental conflict, supporting the notion that the dynamic effectively engages children in adult problems. Viewed through the lens of the emotional security hypothesis (Cummings & Davies, 2002; Davies & Cummings, 1994; Davies, Cummings, & Winter, 2004), which suggests that youths act strategically to maintain a sense of safety, it is understandable that in situations in which parents share their distress directly with their children, youths may feel threatened and, in turn, all the more compelled to take action.

These findings also provide support for the idea that youths who experience parentification from their mothers may be likely to report behavior problems in both the internalizing and externalizing spectra. Extending earlier work in this area (Jacobvitz et al., 2004; Jurkovic et al., 2001) to adolescent populations, they provide the first evidence that parentification makes a unique contribution to behavior problems, above and beyond ratings of psychological control. Notably, the EPQ was not associated with most parent reports of youth adjustment problems, and it is important to recognize that the link between parentification and youth outcomes is modest. However, in some respects, this finding is not surprising; research suggests that parents and children often do not agree on such problems (Hawley & Weisz, 2003). Moreover, youth internalizing symptoms in particular are likely to be overlooked by adults, especially during adolescence (J. L. Martin, Ford, Dyer-Friedman, Tang, & Huffman, 2004; Yeh & Weisz, 2001). Finally, a number of studies suggest that stress in the family system, and maternal stress in particular, may be linked to parents’ failure to identify youth behavior problems (Briggs-Gowan, Carter, & Schwab-Stone, 1996; Youngstrom, Loeber, & Stouthamer-Loeber, 2000). To the extent that parentification reflects a particular response to maternal stress (as well as, potentially, a degree of self-centeredness), it seems plausible that mothers in this sample may have overlooked certain youth behavior problems. Alternatively, it may be the case that parents failed to endorse child adjustment problems because parentified youths appear outwardly competent and resilient. Indeed, Hetherington and Kelly (2002) reported that although parentified youths typically grew into high achieving and socially skilled adults, they tended to feel unsure of themselves and to be vulnerable to persistent symptoms of dysthymia.

In interpreting these findings, it is important to note several study limitations. First, the majority of aims in this investigation were addressed via cross-sectional, correlation analyses with a low-risk sample. As a result, it is difficult to parse apart the direction of effects or to make inferences about causality. In addition, it is difficult to ascertain the degree to which these findings hold for the high-risk samples cited in much of the theoretical literature on parentification. Second, the homogenous, high-functioning nature of this community sample of intact families may limit generalization of study findings. Reliance on a single self-report inventory to measure the central construct of interest in this study also constrains interpretation of findings. Finally, it is important to note that although efforts were made to reduce method variance by relying on multiple informants for key outcome measures, many central findings are likely inflated by youth report on both the predictor and criterion measures.

Despite these weaknesses, the present investigation marks a step forward in empirical work with the parentification construct. This study provides evidence for the reliability and construct validity of the parentification dynamic, and it makes valuable strides in distinguishing it from related concepts such as psychological control. Prior work has demonstrated the potential harm that parentification may foster but has not addressed the process by which this damage occurs. Taken together, the results of this study take a small but critical step toward addressing this gap. They elucidate core aspects of the parentification construct and provide a much-needed empirical backbone for existing theories of this dynamic.