Current evidence suggests that conflicted student-teacher relationships may increase behavior problems in children and vice-versa, but this may be due to confounding. We therefore analyzed their relation applying a within-person approach that adjusts for all time-invariant confounding effects, involving samples from Norway (n = 964, 50.9% females) and the USA (n = 1,150, 48.3% females) followed from age 4-12 years with similar measures. Increased parent-reported behavior problems forecasted increased student-teacher conflict to a similar extent in both countries (β = 0.07, p = .010), whereas teacher-reported behavior problems predicted increased student-teacher conflict more strongly in Norway (β = 0.14, p = .001) than in the US (β = 0.08, p = .050). Increased teacher-child conflict also predicted increased parent-reported (β = 0.07, p = .010), but not teacher-reported, behavior problems in both countries. Findings underscore the reciprocal relation between behavior problems and a conflictual student-teacher relationship.