Friday, 14 January 2005 - 12:00 PM

This presentation is part of: Poster Session I

Fusion or Familialism: A Construct Problem in Studies of Mexican-American Adolescents

This study was an investigation of intergenerational relationships related to the individuation process as reported by Mexican (N = 2388) and Euro American (N = 2907) adolescents. The primary aim was to examine the construct within theories of adolescent development that emotional separation in parent adolescent relationships is an inherent process that occurs universally (Chin & Kameoka, 2002; Gamst et al., 2002). While this theoretical perspective attempts to balance the need for attachment and separation in family relationships, there are strong underlying assumptions about the benefits of emotional independence. It is unclear whether American ideas about separation, emotional fusion, and feelings of being burdened by parents apply to Mexicans, who may be generally very attached to and involved with their family of origin based on a cultural norm of family obligation coupled with a firm family hierarchy. Method: We conducted a series of confirmatory factor analyses on a subscale (Intergenerational Fusion/Individuation or INFUS) of an instrument developed to measure family process, the Personal Authority in the Family System Questionnaire (PAFS-Q) developed by Bray, Williamson, and Malone (1984). The instrument was based on transgenerational family theory and was structured along a continuum between personal authority and intergenerational intimidation. Transgeneration family theory proposes that a psychological “holding-the-self” in relationship with others is central to healthy functioning. Items on the INFUS subscale ask respondents to quantify the extent to which they agree with statements such as “I am usually able to disagree with my parents without losing my temper.” The subscale has been used in a number of published studies on ethnic differences in adolescent developmental outcomes. Results: We followed the original one-factor measurement model. Using M-Plus Software (Muthen & Muthen, 2001) a configural invariance model was analyzed first. Configural invariance tests the hypothesis that the items have the same configuration of factor loadings across ethnic groups (Steenkamp & Baumgartner, 1998). The results did not show invariance: Χ2 = 2218.85, 69 df, p <.000; CFI = .75; TLI = .80, RMSEA = .10. Horn (1991) and Steenkamp and Baumgartner (1998) state that total metric invariance is “a reasonable ideal” that is scientifically unrealistic. Partial measurement invariance is proposed as a compromise. Partial method invariance can be established if at least one item (other than the one fixed at unity to define the scale of the latent construct) is metrically invariant. Modification indices and expected parameter changes are useful in deciding which loadings should not be constrained across groups. Since there was a lack of invariance in item 4 (MI = 1170.44) the loadings for this item were set free. Results indicated that fit indices were worse Χ2 (60) = 3466.95, CFI = .61, TLI = .64, RMSEA = .14 (.142-151). Implications: Since the scale properties could not be confirmed for Mexican adolescents, there is a need to revisit the concept of autonomy as a universal task during adolescent develpment and the clinical use of enmeshment. Our results also underscore the importance of ethnicity as a context for development.

See more of Poster Session I
See more of Oral and Poster

See more of Celebrating a Decade of SSWR (January 13 - 16, 2005)