Peer-reviewed article

Longitudinal associations between memory and emergent literacy from 3 to 6 years of age: a four-wave random-intercept cross-lagged panel study.

Liang JL, Cui XL · 12 Sept 2026

Abstract

Background Memory and emergent literacy are foundational to children's early learning. However, it remains unclear whether their longitudinal association reflects stable differences between children, developmental fluctuations within children, or both. The present study examined their bidirectional associations from 36 to 72 months while distinguishing between-child and within-child variation. Methods This secondary analysis used four waves of caregiver-reported data from the KIT longitudinal study, collected when children were approximately 36, 48, 60, and 72 months old. The analytic sample comprised 1,694 children. A traditional cross-lagged panel model (CLPM) was used to estimate overall longitudinal associations, and a random-intercept cross-lagged panel model (RI-CLPM) was used to separate stable between-child differences from time-specific within-child deviations. Results Mean scores for both constructs increased with age, and memory and emergent literacy were positively correlated within and across waves. In the CLPM, memory was prospectively associated with subsequent emergent literacy across all three intervals, whereas emergent literacy was prospectively associated with subsequent memory in two of the three intervals. In the RI-CLPM, the correlation between the stable between-child components of memory and emergent literacy was 0.81. At the within-child level, deviations in each construct were positively associated with subsequent deviations in the other across all three intervals. The largest estimated emergent-literacy-to-memory coefficient was observed from 60 to 72 months. Conclusions Memory and emergent literacy were bidirectionally associated at the within-child level from 36 to 72 months, while the prospective association from memory to emergent literacy was more consistent across the CLPM and RI-CLPM. Their longitudinal relation involved both stable between-child differences and time-specific within-child variation.