This study explores the extent to which a bilingual advantage can be observed for executive function tasks in children of varying levels of language dominance, and examines the contributions of general cognitive knowledge, linguistic abilities, language use and socio-economic level to performance. Welsh-English bilingual and English monolingual primary school age and teenage children were tested on two executive function tasks, a tapping task and a Stroop task. Bilingual children came from homes in which only Welsh, Welsh and English, or only English was spoken. Results differed by task. On the tapping task, bilingual children from only Welsh homes showed overall superior performance, monolinguals inferior, with the other two bilingual groups between them. Performance correlated with general cognitive abilities with number and pattern discrimination, as well as, at the older age, with balanced use of the two languages. On the Stroop task, language tested mattered; there were no differences across groups in Welsh, but complex patterns in English. The only strictly bilingual advantage on the Stroop was at the younger age for children from Welsh and English homes when tested in English, and performance was also influenced by SES level, cognitive abilities, vocabulary levels and use of the two languages. (Contains 10 tables and 10 figures.)