An interesting trend that emerged was that in reflecting upon one’s own language use, many heritage bilinguals tended to overestimate via self-report how frequently they used their HL, in comparison to how frequently they actually used their HL as measured by the EAR.
2
—
—
The sentences
The model was significant, F (3, 56) = 9.72, p < 0.001, R 2 = 0.34, with only self-reported HL use serving as a significant predictor of EAR-based HL use, as shown in Table 3 (although English AoA was marginally significant).