Child Sexual Abuse & Family Risk: Pathways to Psychosocial Adjustment (2008) PDF

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/5442921 Child Sexual Abuse, Early Family Risk, and Childhood Parentification: Pathways to Current Psychosocial Adjustment Article in Journal of Family Psychology · April 2008 DOI: 10.1037/0...

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/5442921 Child Sexual Abuse, Early Family Risk, and Childhood Parentification: Pathways to Current Psychosocial Adjustment Article in Journal of Family Psychology · April 2008 DOI: 10.1037/0893-3200.22.2.320 · Source: PubMed CITATIONS READS 67 2,319 6 authors, including: Heidi M Zinzow Joan Lynne Jackson Clemson University University of Georgia 56 PUBLICATIONS 3,413 CITATIONS 40 PUBLICATIONS 2,160 CITATIONS SEE PROFILE SEE PROFILE All content following this page was uploaded by Heidi M Zinzow on 05 June 2014. The user has requested enhancement of the downloaded file. Journal of Family Psychology Copyright 2008 by the American Psychological Association 2008, Vol. 22, No. 2, 320 –324 0893-3200/08/$12.00 DOI: 10.1037/0893-3200.22.2.320 Child Sexual Abuse, Early Family Risk, and Childhood Parentification: Pathways to Current Psychosocial Adjustment Monica M. Fitzgerald Renee A. Schneider Medical University of South Carolina Center for Health Care Evaluation, Veterans Affairs Palo Alto Health Care System and Stanford University School of Medicine Seoka Salstrom Heidi M. Zinzow Rosalind Franklin University of Medicine and Science Medical University of South Carolina Joan Jackson Rebecca V. Fossel University of Georgia Longmont, Colorado In this study, the authors examined the role of parentification (children assuming adult-like roles in the family) as it relates to family risk (parental psychopathology, parental illness, and domestic violence), child sexual abuse (CSA), and psychosocial adjustment in 499 college women. Structural equation modeling was used to test a model of direct, indirect, and mediational pathways through which CSA, family risk, and parentification contributed to later psychosocial maladjustment. Results indicate that CSA and family risk independently and directly predicted higher levels of maladjustment, but only family risk positively predicted parentification in childhood. Parentification was unexpectedly related to less mal- adjustment. Parentification failed to mediate the relation between early family risk and maladjustment. Findings suggest that family risk factors may contribute to parentification and that parentification is not always related to poorer psychosocial outcomes. Future research should examine the impact of parentification on other aspects of functioning and should assess how individual, familial, and cultural variables (e.g., age, gender, duration, perceived fairness, ethnicity, and family support) moderate the impact of parentification on long-term adjustment. Keywords: parentification, child sexual abuse, family risk, psychosocial adjustment Mika, Bergner, and Baum (1987) defined parentification providing comfort, advice, and protection to family mem- as “a family interactional pattern in which children and bers (Jurkovic, 1997; Mika et al., 1987). Although parenti- adolescents are assigned roles and responsibilities normally fication may promote the development of desirable at- the province of adults in a given culture, but which parents tributes, such as responsible behavior and resourcefulness in a particular family have abdicated” (p. 229). Parentified (Barnett & Parker, 1998; Chase, 2001), a majority of re- children often fulfill instrumental caretaking roles within the search documents the adverse consequences of childhood family system, such as paying the bills and caring for parentification, including internalized emotional distress, younger siblings, and emotional caregiving roles, such as externalizing behavior problems, and interpersonal difficul- ties (see Earley & Cushway, 2002). Preliminary research suggests that child sexual abuse Monica M. Fitzgerald and Heidi M. Zinzow, National Crime (CSA), which involves children fulfilling the sexual needs Victims Research and Treatment Center, Medical University of of adults, may relate to the premature assumption of other South Carolina; Renee A. Schneider, Center for Health Care types of parentified roles and caretaking behaviors within Evaluation, Veterans Affairs Palo Alto Health Care System and family and peer relationships (e.g., Alexander, Teti, & Stanford University School of Medicine, Palo Alto, California; Anderson, 2000; Green & Jurkovic, 2002). There is also Seoka Salstrom, Department of Psychology; Rosalind Franklin growing evidence that parentification often occurs within University of Medicine and Science; Joan Jackson, Department of families that experience major stressors or family dysfunc- Psychology, University of Georgia; Rebecca V. Fossel, indepen- tion, such as parental illness and psychopathology, divorce, dent practice, Longmont, Colorado. Correspondence concerning this article should be addressed to and marital conflict (Barnett & Parker, 1998; Earley & Monica M. Fitzgerald, Medical University of South Carolina, Cushway, 2002; Mayseless, Bartholomew, Henderson, & National Crime Victims Research and Treatment Center, 165 Trinke, 2004). Additionally, the experience of CSA tends to Cannon Street, P.O. Box 250852, Charleston, SC 29425. E-mail: co-occur with familial dysfunction, such as parental psy- [email protected] chopathology, substance use, and spousal violence (e.g., 320  BRIEF REPORTS 321 Long & Jackson, 1994). These family dysfunction variables 66 participants who reported CSA by a peer were excluded are considered family risk factors because they consistently from analyses. The final sample included 499 women, rang- predict subsequent adverse psychosocial outcomes, such as ing from 17 to 42 years of age (M ⫽ 19.29, SD ⫽ 1.95). internalizing problems, decreased self-esteem, and interper- Ethnicity was 82% White, 11% African American, 4% sonal problems (e.g., El-Sheikh & Flanagan, 2001). Asian American, 1% Hispanic, and 2% other. Average From a developmental psychopathology perspective (Cic- socioeconomic status was upper middle class. The Institu- chetti & Toth, 1995), an important area of research is identi- tional Review Board of the University of Georgia granted fying familial processes, such as parentification, that may in- approval to conduct this study. terfere with or promote psychosocial adjustment in children exposed to adverse life experiences. Assuming adult-like roles Measures within certain contexts (e.g., parental divorce, illness) may serve an adaptive, protective function—such as ensuring fam- Family risk and child abuse history. The Life Experi- ily cohesion and personal well-being— but possibly with a cost ences Questionnaire (Jackson, Calhoun, Amick, Maddever, to the youth’s socioemotional functioning (Hetherington, & Habif, 1990) assessed family risk (i.e., witnessing do- 1999). In the present study, we tested a model examining mestic violence, parental impairment) and CSA. Witnessing relations among CSA, family risk, parentification, and long- domestic violence was assessed with the following item term adjustment. We hypothesized that family risk and CSA rated on a 5-point scale (ranging from 1 ⫽ never to 5 ⫽ very would be directly related to increased psychosocial maladjust- frequently): “Did you ever witness any physical violence in ment and higher levels of parentification. It was also expected your home between your parents, stepparents, or between that parentification would, in turn, be related to higher levels of your parent and a dating partner? (i.e., pushing, shoving, psychosocial maladjustment and that parentification might par- slapping, hitting, punching, etc.).” Mother and father im- tially mediate the relations between CSA and psychosocial pairment were measured by summing four items with outcomes and between family risk and psychosocial outcomes 5-point scale ratings (1 ⫽ never to 5 ⫽ almost always) (see Figure 1). assessing parental substance use, anxiety, depression, and illness (total scores ranged from 4 to 20; mother impairment Method ␣ ⫽.67, father impairment ␣ ⫽.64). CSA history was assessed with a behaviorally specific item (i.e., “As a child Participants or adolescent [under age 18 years] did anyone who was at Participants were 565 undergraduate students enrolled in least 5 years or older than you involve you in any kind of psychology classes at a southeastern university. Data from sexual contact with him/her?”). The Life Experiences Ques- Family Risk.39*.44.76.62.35*.22* Witness Mother Father Domestic Impairment Impairment Violence Psychosocial 42.61 Parentification Maladjustment.81 -.13*.85.77.57.84 -.70.84 Child Sexual Spousal Parental Sibling Symptom Self- Social Abuse Adjustment Severity Esteem.29.29.28 1.0.41.50.68 Abuse.14* 0.0 Figure 1. Final model for family risk, child sexual abuse, parentification, and psychosocial maladjustment. Latent constructs are shown in ellipses, and observed variables are shown in rectangles: ␹2(30, N ⫽ 499) ⫽ 61.15, p ⬍.05, root-mean-square error of approximation ⫽.05, nonnormed fit index ⫽.96, standardized root-mean-square residual ⫽.04. Standardized path coefficients are shown for paths that are statistically significant; dashed lines indicate hypothesized paths that are not significant. *p ⬍.05. 322 BRIEF REPORTS tionnaire includes examples of sexual contact and follow-up directions (see Figure 1). When the confirmed factor questions about the victim–perpetrator relationship, the on- structures were used in the structural model analysis, the set of abuse, and the type of contact and frequency of abuse, overall fit of the full model was good, ␹2(30, N ⫽ 499) ⫽ regarding up to three separate incidences. 61.15, p ⬍.05, RMSEA ⫽.05, NNFI ⫽.96, SRMR ⫽.04. Parentification. Three subscales from the Parentifica- As hypothesized, both family risk and CSA had a significant tion Scale (Mika et al., 1987) measured acting as a (a) direct effect on psychosocial maladjustment in the positive parent to her parent(s) (␣ ⫽.76), (b) spouse to her parent(s) direction (␥ ⫽.39, p ⬍.05; ␥ ⫽.14, p ⬍.05, respectively); (␣ ⫽.78), and (c) parent to her sibling(s) (␣ ⫽.86). Items however, only family risk had a significant, positive direct were rated on a 5-point scale (ranging from very often to effect on parentification (␥ ⫽.35, p ⬍.05). Contrary to the never or doesn’t apply) according to the frequency of the hypotheses, CSA did not have a significant direct effect on behavior before the age of 14 years and between the ages of parentification (␥ ⫽.01, ns), and parentification related to 14 and 16 years. Items were weighted according to content psychosocial maladjustment in the negative direction (␤ ⫽ and age and summed within each subscale. ⫺.13, p ⬍.05). Family risk failed to have a significant, Psychosocial adjustment. The global severity index of indirect effect on psychosocial maladjustment through pa- the Symptom Checklist–90 –Revised (Derogatis, 1983) rentification (indirect effect ⫽ ⫺.04, ns). As expected, served as a measure of current psychological distress (␣ ⫽ family risk and CSA significantly covaried (␸ ⫽.22)..97). The Social and Leisure subscale (11 items excluding 2 To test for mediation, we first tested the direct effect dating behavior items) of the Social Adjustment Scale— model by removing the parentification factor to determine Self-Report version (Weissman & Bothwell, 1976) mea- whether significant paths existed between family risk and sured women’s social adjustment (␣ ⫽.62). The 10-item psychosocial maladjustment and between CSA and psycho- Rosenberg Self-Esteem Scale (Rosenberg, 1979) measured social maladjustment (Holmbeck, 1997). The resulting global self-esteem (␣ ⫽.89). model fit the data well, ␹2(12, N ⫽ 499) ⫽ 27.90, p ⫽.00, RMSEA ⫽.05, NNFI ⫽.95, SRMR ⫽.03, and yielded Data Analysis significant paths from family risk to psychosocial malad- justment (␤ ⫽.35, p ⬍.05) and from CSA to psychosocial We used structural equation modeling with Mplus soft- maladjustment (␤ ⫽.13, p ⬍.05). Next, these model results ware (Muthén & Muthén, 1998 –2004) to evaluate the pro- posed measurement and structural models. The direct max- were compared with the full model results. An examination imum likelihood method of estimation was used to account of path coefficients showed that when the potential mediator for missing data (Allison, 2003). The following fit indices of parentification was excluded from the full model, the were used in conjunction with the chi-square statistic to relationship between family risk and psychosocial malad- assess model fit: (a) the root-mean-square error of approx- justment remained strong (␤ ⫽.35 vs. ␤ ⫽.39), and the imation (RMSEA), (b) the nonnormed fit index (NNFI; relationship between CSA and psychosocial maladjustment Tucker–Lewis Index), and (c) the standardized root-mean- did not change (␤ ⫽.14 vs. ␤ ⫽.14). The final step in square residual (SRMR; cf. Vandenberg & Lance, 2000). assessing a mediational effect was to assess the fit of the model (a) when the direct paths from family risk to psycho- social maladjustment and from CSA to psychosocial mal- Results adjustment were constrained to zero and (b) when the direct Of the sample, 22% reported CSA, with an average age of paths were not constrained (i.e., full model; Holmbeck, onset of 11.57 years (SD ⫽ 4.56). With regard to the first or 1997). As noted in Table 1, when the paths were fixed to only victimization experience, 23% of perpetrators were zero, the model was adequate, and when these results were family members/relatives, 44% were family friends, 29% compared with the full model, no mediation was found. Post were other known adults, and 4% were strangers. hoc analyses were conducted to further examine relations Confirmatory factor analysis indicated that the measure- between the parentification indices (i.e., spousal, parental, ment model fit the data well according to all criteria, ␹2(24, sibling) and psychosocial maladjustment. Structural equa- N ⫽ 499) ⫽ 57.82, p ⬍.05, RMSEA ⫽.05, NNFI ⫽.96, tion modeling analyses revealed that sibling parentification SRMR ⫽.04. Furthermore, all indicators loaded signifi- was negatively associated with psychosocial maladjustment cantly on their latent constructs (p ⬍.05) in the expected (␤ ⫽ ⫺.11, p ⬍.05), whereas relations between psycho- Table 1 Goodness of Fit Indices of Models Model ␹2 df ⌬␹2 ⌬df RMSEA NNFI (TLI) SRMR * Reduced/restricted model 105.93 32.07.92.07 Full model 61.15* 30.05.96.04 Comparison Models 1 and 2 44.78* 2 Note. N ⫽ 499. RMSEA ⫽ root-mean-square error of approximation; NNFI ⫽ nonnormed fit index; TLI ⫽ Tucker–Lewis Index; SRMR ⫽ standardized root-mean-square residual. * p ⬍.001.  BRIEF REPORTS 323 social maladjustment and the spousal (␤ ⫽.07, ns) and port for a mediational model. Despite limitations, however, parental (␤ ⫽.10, ns) indices were nonsignificant. the present study represents an important step in the attempt to understand the underlying mechanisms linking child Discussion abuse and family risk to later psychosocial difficulties. CSA and family risk independently predicted psychoso- cial maladjustment; however, contrary to expectations, CSA References was not directly related to parentification. Relations be- tween parentification and CSA may be specific to father– Alexander, P. C., Teti, L., & Anderson, C. L. (2000). Childhood sexual abuse history and role reversal in parenting. Child Abuse daughter incest or intrafamilial abuse. It is also likely that and Neglect, 24(6), 829 – 838. parentification is more common when abuse is chronic, and Allison, P. D. (2003). Missing data techniques for structural equa- future research should examine how the frequency, severity, tion modeling. Journal of Abnormal Psychology, 112(4), 545– and duration of CSA and the nature of the victim– 557. perpetrator relationship influence parentification. Family Anderson, L. P. (1999). Parentification in the context of the Afri- risk was positively associated with parentification, but can American family. In N. D. Chase (Ed.), Burdened children: higher levels of parentification related to less maladjust- Theory, research and treatment of parentification (pp. 154 –170). ment. These findings were unexpected given that the ma- London: Sage. jority of prior studies with college students have docu- Barnett, B., & Parker, G. (1998). The parentified child: Early mented negative consequences of parentification (e.g., competence or childhood deprivation? Child Psychology & Psy- Jacobvitz & Bush, 1996). However, at least two factors may chiatry Review, 3, 146 –155. Chase, N. D. (2001). Parentified children grow-up: Dual patterns help to explain this relationship. First, because family risk of high and low functioning. In B. E. Robinson & N. D. Chase and maladjustment were correlated and because the direc- (Eds.), High-performing families: Causes, consequences, and tion and magnitude of the relationship between parentifica- clinical solutions (pp. 157–189). Alexandria, VA: American tion and maladjustment changed with family risk in the Counseling Association. model, it appears that family risk is a negative suppressor Cicchetti, D., & Toth, S. (1995). A developmental psychopathol- variable for parentification. That is, family risk accounted ogy perspective on child abuse and neglect. American Academy for variance in maladjustment not accounted for by paren- of Child and Adolescent Psychiatry, 34, 541–564. tification, allowing a stronger relationship between parenti- Derogatis, L. R. (1983). SCL-90-R: Administration, scoring, and fication and maladjustment to emerge. Second, results of procedures manual. Towson, MD: Clinical Psychometric Re- post hoc analyses also suggest that different types of paren- search. tification are associated with differential outcomes. In par- Earley, L., & Cushway, D. (2002). The parentified child. Clinical Child Psychology and Psychiatry, 7(2), 163–178. ticular, findings suggest that the negative association be- El-Sheikh, M., & Flanagan, E. (2001). Parental problem drinking tween parentification and maladjustment was driven by the and children’s adjustment: Family conflict and parental depres- sibling parentification index. In particular, caring for sib- sion as mediators and moderators of risk. Journal of Abnormal lings may confer skills and a sense of self-efficacy that are Child Psychology, 29, 419 – 432. important to psychosocial adjustment, at least among Green, M. L., & Jurkovic, G. J. (2002, November). The impact of young, female college students. family functioning on trauma symptomatology in child and ad- Further, the hypothesis that parentification would mediate olescent victims of intrafamilial sexual abuse. Poster presented at relations between CSA or family risk and psychosocial the annual meeting of the International Society for Traumatic adjustment was not supported. Future research should ex- Stress Studies, Boston, MA. amine how age, gender, ethnicity, race, type and duration of Hetherington, E. M. (1999). Should we stay together for the sake of the children? In E. M. Hetherington (Ed.), Coping with di- parentification, and children’s perceptions of fairness influ- vorce, single parenting, and remarriage: A risk and resiliency ence relationships between parentification and adjustment perspective (pp. 93–116). Mahwah, NJ: Erlbaum. (Anderson, 1999; Jurkovic, 1997; Jurkovic et al., 2004). A Holmbeck, G. N. (1997). Toward terminological, conceptual, and cultural– ecological approach (see Jurkovic, 1997) to study- statistical clarity in the study of mediators and moderators: ing parentification and the use of multimethod assessment Examples from the child-clinical and pediatric psychology liter- measures are also encouraged. atures. Journal of Consulting and Clinical Psychology, 65(4), Without replication, it is premature to conclude that 599 – 610. parentification facilitates long-term adjustment or that Jackson, J., Calhoun, K., Amick, A., Maddever, H., & Habif, V. parentification fails to mediate psychosocial adjustment (1990). Young adult women who report childhood intrafamilial in the context of CSA or family risk. Methodological sexual abuse: Subsequent adjustment. Archives of Sexual Behav- weaknesses—including a cross-sectional design and an in- ior, 19, 211–221. Jacobvitz, D. B., & Bush, N. F. (1996). Reconstructions of family ability to draw causal conclusions, an exclusive reliance on relationships: Parent– child alliances, personal distress, and self- retrospective self-report, problems related to shared method esteem. Developmental Psychology, 32(4), 732–743. variance, a narrow definition of psychosocial maladjust- Jurkovic, G. J. (1997). Lost childhoods: The plight of the paren- ment, and the use of a predominantly White, upper middle- tified child. Philadelphia: Brunner/Mazel. class college student sample—limit the generalizability of Jurkovic, G. J., Kuperminc, G., Perilla, J., Murphy, A., Ibañez, G., findings. Research with diverse populations and a more & Casey, S. (2004). Ecological and ethical perspectives on filial sophisticated definition of parentification could yield sup- responsibility: Implications for primary prevention with immi- 324 BRIEF REPORTS grant Latino adolescents. Journal of Primary Prevention, 25, Rosenberg, M. (1979). Conceiving the self. New York: Basic 81–104. Books. Long, P. J., & Jackson, J. L. (1994). Childhood sexual abuse: An Vandenberg, R. J., & Lance, C. E. (2000). A review of synthesis examination of family functioning. Journal of Interpersonal Vi- of the measurement invariance literature: Suggestions, practices, olence, 9(2), 270 –277. and recommendations for organizational research. Organiza- Mayseless, O., Bartholomew, K., Henderson, A., & Trinke, S. tional Research Methods, 3(1), 4 – 69. (2004). “I was more her mom than she was mine”: Role reversal Weissman, M. M., & Bothwell, S. (1976). Assessment of social in a community sample. Family Relations, 53, 78 – 86. adjustment by patient self-report. Archives of General Psychia- Mika, P., Bergner, R. M., & Baum, M. C. (1987). The develop- try, 33, 1111–1115. ment of a scale for the assessment of parentification. Family Therapy, 14(3), 229 –235. Received September 14, 2006 Muthén, L. K., & Muthén, B. O. (1998 –2004). Mplus user’s guide Revision received August 15, 2007 (3rd ed.). Los Angeles: Author. Accepted August 16, 2007 䡲 View publication stats