Session: Sibling Research Reimagined: Measurement and Methodology Considerations (Society for Social Work and Research 28th Annual Conference - Recentering & Democratizing Knowledge: The Next 30 Years of Social Work Science)

Background: Siblings are one of the most foundational relationships in a child's life, particularly during middle childhood, and endure well beyond a child's relationship with their parents. Despite the majority of children (77-82%) in the U.S. growing up with at least one sibling, researchers have commonly overlooked the role of sibling relationships in child or family research. However, research suggests siblings influence human development and behavior (e.g., social, emotional, and behavioral skill development) through everyday moments of play and family activities. Underdeveloped measurement tools may have hindered researchers and clinicians from fully understanding the sibling bond. This symposium presents results from three studies: (1) a systematic review of sibling relationship quality measures examining their robust instrument development and validation standards as defined by the NIH PROMIS; (2) a case study to reproduce the measurement structure of the most widely-used sibling relationship quality measure; and (3) an exploratory methodological study using different forms of artificial intelligence (AI) technology, in combination with more traditional approaches, to identify applications of AI in measurement development. Methods: Study 1 is a systematic review of general population sibling relationship quality among children aged 2 to 17 in the United States and each netted measure was examined for robust instrument development and validation standards as defined by the NIH PROMIS®. Study 2 estimated exploratory factor analysis (EFA) models using maximum likelihood factor extraction to evaluate the degree to which the factor configuration described in the initial development of Furman & Buhurmester's Sibling Relationship Questionnaire (SRQ) could be empirically reproduced and thus evidence its construct validity. Study 3 took a multi-pronged approach that included (1) thematic analysis; (2) AI concept analysis combined with text-search analysis; and (3) AI coding analysis on items (n=340) pooled across 10 sibling relationship quality measures. Results: Study 1 search yielded 947 citations after deduplication with a final total of 87 articles on general population sibling relationship quality that utilized 13 measures of sibling relationship quality. None of the 13 measures identified through the systematic review met all criteria specified by PROMIS® for developing reliable and valid measures of child and adult health and well-being. Study 2 EFA results did not evidence a clear factor structure of the SRQ, even when specifying up to four factors to mirror the commonly described four core domains of the measure. Study 3 found that the codes generated by the concept mapping/text search strictly identified the most common words and could be useful in generating open codes for subsequent analysis. The codes generated by AI coding offer potential "bins" of items (i.e., groups of items according to meaning and/or latent constructs) and may be useful for informing higher-level themes for thematic analysis methods. Conclusions and Implications: Underdeveloped measurement tools have hindered researchers and clinicians from fully understanding the sibling bond, and potentially developing related and effective interventions that harness the potential of this critical relationship. Methodological challenges of measurement development, dyadic research, and harnessing the power of AI will be discussed.529 modified by 12.245.207.222 on 4-15-2023-->