When Assessing Intra-Familial Relationships, Are Sociologists, Psychoanalysts and Psychiatrists Really Considering Different Constructs? An Empirical Study

Abstract

This paper aimed to look for the existence of a common core when envisaging intra-familial interactions as perceived by adolescents, which could be shared by sociology, psychoanalysis and child and adolescent psychiatry. An empirical study based on a mixed-method design collected the responses of 194 adolescents to the instruction “In the next half hour, would you please write as freely as you wish about your relationships in your family, explaining how things are”. All answers were then analyzed and 18 dimensions related to 3 different theoretical frameworks were rated blind using numerical scores by two independent raters from each discipline. Inter-rater reliability was good. A parallel analysis evidenced a strong underlying factor explaining a large amount of variance (>50%). This factor is bipolar, it reflects the level of positivity/negativity in the adolescent’s point of view concerning his/her intra-familial relationships. A second factor can marginally be considered (10% of the variance). The 2-factor analysis found one factor related to positive feelings and the other to negative feelings. This finding of unidimensionality supports family study as an intervention science.

Using a self-attention architecture to automate valence categorization of French teenagers’ free descriptions of their family relationships. A proof of concept

“…From this brief overview, we draw several objectives in the present study. We will challenge as a proof-of-concept the use of state-of-the-art NLP learning techniques on the corpus of adolescent texts described in Falissard et al [2]. As described above, this corpus exhibits maximal variance along the positive vs. negative relationships dimension.…”

“…We wish to warmly thank Christine Hassler, Muriel Letrait, Guillaume Macher, François Marty, Elsa Ramos, Anne Revah-Lévy, Philippe Robert, and françois de Singly, for conceiving the multidisciplinary protocol to gather the adolescents' texts dataset used in this study and published in [2].…”

“…Despite the wealth of literature in family research, no consensus has been established on the variables and constructs to describe quantitatively family relationships [2]. The most relevant theoretical frameworks focus on family histories (family development theory), systemic relationships between family members, family relationships with the environment, attachment relationships, social learning by children, etc.…”

“…These different approaches are reflected in a wide variety of experimental methods, but the depth of the epistemological differences constitutes a serious pitfall for the comparison of research results. In order to overcome the differences, Falissard and colleagues have developed a common tool between sociologists, psychoanalysts and adolescent psychiatrists to be applied to the free discourse that adolescents may hold about their own family relationships [2]. 194 French adolescents (mean age : 14.7( s.d .…”

“…Thus, it appears through this interdisciplinary research that relational valence constitutes a key element of the family descriptions produced by adolescents and thus undoubtedly of their mental representations of them. [2] also argue in favor of using this dimension as a primary endpoint in future interventional research. We add to their conclusion, that if it is accepted that the valence of relationships between individuals cannot be ignored in any family assessment, this provides a strong rationale for sentiment analysis studies (see [3] for an overview of this NLP application domain) aimed at automating the analysis of adolescent free speech or writings.…”