Assessing Children’s Responses to Interparental Conflict: Developing and Validating Concise Scales for a Scandinavian Country

• In Event: 2-012 - Interparental Conflict, Emotional Security and Child Adjustment: Cross-Cultural Studies Exploring Processes, Contexts and Methods

Integrative Statement

Objective: The Security in the Interparental Subsystem (SIS) and The Children’s Perception of the Interparental Conflict Scale (CPIC) are two extant and widely used self-report scales capturing (a) children’s unique reactions to interparental conflict, and (b) children’s own perception of the conflict rate between the parents. The two scales have been validated to varying degrees, but no validation has taken place in Scandinavian countries. Relatedly, there is a need for brief and concise versions of the scales in order to facilitate implementation in clinical practice as well as in research studies. The primary aim of the current study was therefore to validate the Norwegian versions of the SIS and the part of the CPIC measuring the children’s perception of the conflict (CPIC-properties), and to develop concise short versions of the scales. Method: In a sample of N = 392 children and adolescents aged 10-15 years old (M age=11.1, SD=0.5; 52.3% girls) confirmatory factor analyses (CFA) were used to analyze the underlying factor structure of the SIS and the CPIC-properties. Stepwise regression analysis was used in creating the short versions of the scales. CFA was subsequently used to factor analyze the short scales and regression analyses was conducted to investigate the predictive strengths of the short version subscales. Results: Both the full version and the short version of the SIS had excellent fit according to a three-level model (SIS-full version: χ2(658) = 1049.37, p < .05, RMSA= 0.039, CFI = 0.958, TLI = 0.956 and SIS short version: χ2(111) = 258.87, p < .05, RMSEA = 0.059, CFI = 0.972 and TLI = 0.965). The SIS-model was organized as six first-order factors, three second-order factors (Emotional reactivity, Regulation of exposure to affect, and Internal representation) and one third-order factor (‘reactions’). The full and short versions of the CPIC-properties also showed good fit CPIC-Properties-full version: χ2(270) = 417.95, p < .05, RMSEA = 0.038, CFI = 0.957, TLI = 0.952 and CPIC-properties-short: χ2(114) = 218.56, p < .05, RMSEA = 0.050, CFI = 0.959, TLI = 0.951). The CPIC-properties model was organized as five first-order properties factors organized under a second-order latent variable ‘Conflict properties’.), The SIS was reduced from 38 items in the original version to 17 items in the short version and the CPIC-properties was reduced from 25 items to 17 items in the short version. The internal consistencies of both original and short versions were satisfactory. The predictive strengths of the short subscales were comparable of the full subscales’ predictive strengths. Conclusions: The findings support the validity of SIS and CPIC-properties in a Norwegian sample of youth. The development of the concise short version has several advantages and should be tested out in future studies outside Scandinavian countries.

Authors

• Tonje Holt, Norwegian Institute of Public Health
  Presenting Author

• Maren Sand Helland, Norwegian Institute of Public Helath
  Non-Presenting Author

• Anh Ha, University of Notre Dame
  Non-Presenting Author

• E. Mark Cummings, University of Notre Dame
  Non-Presenting Author