The Family Interaction and Alliance Assessment Scales (FIAAS): A validity and reliability study
The aim of this study is to design and validate an assessment tool for father-mother-infant interactions: the FIAAS (Family Interaction and Alliance Assessment Scale). Family-level dynamics are more and more acknowledged as one of the determining factor of infant social adjustment and psychopathology. But assessing family-level interactions in a reliable way is a methodological challenge, and efforts in this direction are scarce. So there is now a clear need for a validated and clinically relevant tool. Moreover, the FIAAS will be one of the essential tool of a larger project study on parentality troubles which will be applied later. The FIAAS will also fulfil the request of teams who collaborate with us (UR-CEF, "Unité de Recherche du Centre d'Etude de la Famille", part of the "Institut Universitaire de Psychothérapie" in Lausanne), at the national and international level. Some of them will already contribute by sharing samples. Indeed the validation study will be done on two samples: one non referred (Lausanne sample; N=51) and one clinically referred (Montpellier sample; N=9). The main steps of this study are: (i) item selection, based upon the UR-CEF expertise in nonverbal microanalysis of triadic interactions, (ii) coding of a first sub-sample and test for the reliability of the FIAAS, (iii) coding of a second sub-sample and test for construct validity and for criterion-related validity, (iv) definitive selection of items and final elaboration of the FIAAS.
Main results (May, 2006):
- The FAAS scales are highly correlated with the results of the former micro-analytic coding, but with more dimensions evaluated and much less time needed for the coding (Rbp between.39* and.70***, only one non significant correlation with the scale "Readiness to interact").
- The scales have no link with self-reported marital satisfaction by the parents (DAS scales); this absence of link between observational measures and self-reported seems to be a matter of methodology rather than a validity problem of the observational tool.
- Correlations with the CFRS scales show that the family scales of the FAAS are correlated with the conceptually relevant co-parental CFRS scales. All FAAS scales are significantly correlated with at least one relevant CFRS scale (the least significant correlation being.38*, the most.67***). Only one FAAS scale, "parental scaffolding", does not correlate at all with the CFRS.
- A matched pairs comparison with the small referred sample from the Montpellier study show that one scale of the FAAS is discriminating between the samples, the "coparental coordination", for which the normative Lausanne sample got higher results (Kolmogorov-Smirnov Z = 1.41*). This is clinically relevant.
- Exploratory factor analysis show that there is two underlying dimensions in family interactions (eigenvalues>1, Varimax rotation), one related to the non verbal structure of the play (the "family partnership", accounting for 34.2% of the variance), the other related to the cognitive and affective exchanges (the "emotional benevolence", accounting for 29.9% of the variance).
Validation of an observational grid for assessing father-mother-infant/child interactions. Main method is systematic observation. First step will be item selection by using relevant information in the scientific litterature, as well as knowledge coming from our previous researches in Lausanne. Second step will be to evaluate the content validity and the face validity of the grid by submitting it to 5 experts of the field - family therapists, researchers in family interactions, etc. Reliability and construct validity + criterion-based validity will then be evaluated by coding family interactions in a non referred samples (volunteers families taking part in a larger research project of the UR-CEF; N=51) and in a referred sample (families in therapy for infant's symptomatology; N=36). The observation situation is the Lausanne Trilogue Play, a semi-standardized laboratory situation. At least two coders will code each of the LTP to ensure interraters reliability.
Démarche méthodologique: Expériences de laboratoire, Observation non participante
|Ethical approval||No||Study type||
|Start - End date||01.04.2005 - 28.04.2006|