Attachment Stories in Middle Childhood: Reliability and Validity of Clinical and Nonclinical Children's Narratives in a Structured Setting.

Abstract

Middle childhood is one of the most understudied periods of development and lacks a gold standard for measuring attachment representations. We investigated the reliability and validity of a Dutch version of the Story-Stem Battery coded using the Little Piggy Narrative (LPN) Coding System in a clinical (N = 162) and a nonclinical group (N = 98) of 4–10-year-old children. Their attachment stories were furthermore coded using the coherence scale. Factor analyses showed that the items of the LPN system formed four attachment scales and a separate scale reflecting distress/anxiety, with sufficient internal consistency for the scales and high interrater reliability (n = 20). Furthermore, we studied construct and discriminatory validity. The attachment scores correlated with coherence and child behavioral problems in the expected direction. Results showed age and gender differences, indicating that separate norm groups are necessary. In particular, disorganized attachment, coherence and distress/anxiety differ between clinical and nonclinical children across age and gender. Results for the other three organized attachment scales were more complex. For instance, older boys from the nonclinical group had higher scores on secure attachment than their clinical peers, while girls from the clinical and nonclinical groups did not differ, even though girls in the nonclinical group had higher secure attachment scores than boys. Results are discussed in light of attachment theory and developmental pathways in middle childhood, as well as their clinical implications.