Creativity is only partly recognised in education. A recent meta-analysis estimated a correlation of r = 0.22 between creativity and educational achievement across many international student samples of all educational levels. In the meta-analysis, creativity was measured with a variety of measures, including divergent thinking and remote association tasks. The differences in the measures influenced the strength of the relationship between creativity and educational achievement. More research is needed to establish reliable measures of creativity, especially in primary school children, whose creativity remains poorly evaluated. The present study measured creativity in written stories in children at age 9 using the Consensual Assessment Technique (CAT). The study employed a longitudinal design, using CAT creativity scores as a predictor of educational achievement at age 16. Each of the stories from 59 children were coded by 6 different judges for 10 dimensions, including creativity. The inter-rater reliabilities between the judges for the 10 dimensions were high (α = .76 -.95). Among the dimensions, a factor analysis revealed two factors: Creative Expressiveness and Logic. The Creative Expressiveness factor explained an additional 7 % of variance in English grades, but not in Maths, beyond intelligence, previous achievement and personality traits associated with creativity. Overall, the study showed that CAT is a robust and reliable measure to detect verbal creativity in childhood. The results also suggest that early creativity predicts later academic achievement, calling for more attention to early creativity assessment and development.