Background: Socially withdrawn children tend to perform poorer academically than their peers. What remains unknown, is the temporal ordering of the two phenomena. Also, substantial gender differences exist in both social withdrawal and academic achievement; thus, it is conceivable that the strength of the relation between them is gendered as well. Aims: To investigate cross-sectional correlates and test directional effects of social withdrawal and academic achievement from primary to upper secondary school, and to examine potential gendered effects. Methods: Prospective associations were analysed from age 6 to age 14 using biannual teacher ratings of children's social withdrawal and academic achievement in a representative community sample (n = 845), by means of random intercept cross-lagged panel modelling. Results: In boys, increased academic achievement at ages 8 and 12 forecasted decreased social withdrawal 2 years later, whereas increased social withdrawal at age 10 predicted reduced academic achievement at age 12. No such effects were seen in girls. Conclusions: Social withdrawal and academic achievement are bidirectionally related among boys, but not girls. Results are discussed in light of need-to-belong theory, and practical implications for schools and teachers are illuminated.