The assessment of attachment security in infancy and adulthood is well-studied, but middle childhood has been relatively neglected in the literature. The Child Attachment Interview (CAI) represents a promising assessment of attachment in middle childhood; recent evidence supports its reliability and validity in restricted samples. We assessed the construct validity of the CAI among a diverse sample of 104 8–12 year-old children (M age = 9.80 years; 50 boys; 49 % family income <$40,000; 38 % Latino/a). We evaluated associations among CAI variables and self-report measures of attachment (convergent validity). Further, we evaluated whether CAI variables uniquely relate to theoretically-relevant constructs (internalizing symptoms) above and beyond self-reported attachment (incremental validity) and a theoretically-distinct construct, child temperament (discriminant validity). Our findings support the CAI’s validity, and have important implications for assessing and understanding attachment processes in middle childhood, as well as their development across the lifespan.