Bilingual children show more variation in their language development than monolingual children, a fact that has been linked to their experience with their languages. Bilingual language experience also varies more than monolingual children's, both in terms of how much they hear the language spoken around them (exposure) and how much they speak the language themselves (production). This dissertation investigates the following aspects of the relationship between bilinguals’ language experience and development which are not well-understood: how children’s language production relates to their proficiency in that language, how children’s language exposure relates to receptive versus expressive and lexical versus grammatical skill, and how factors such as social context, cognates, working memory and indirect exposure contribute to bilingual proficiency. I investigate language experience and English proficiency in young school-aged bilinguals acquiring French and English in France. I use data from parental and child interviews to estimate English exposure – how much children regularly hear English – and two facets of English production – output, or how regularly children speak in English, and inter-speaker code-switching, which refers to how regularly children respond in French when spoken to in English. Those measures are then related to English proficiency scores from a picture-identification task, a picture-naming task, and a sentence repetition task targeting grammatical structures ranging in difficulty. The first objective of this study is to better understand bilingual children’s language production as it relates to their language proficiency. I find that how much children switch to speaking in French when addressed in English (inter-speaker code-switching) is closely related to all concurrent English proficiency scores and that this relationship is independent of and stronger than proficiency’s relationship with exposure. The more children switch to French when spoken to in English, the lower they score on all proficiency measures, receptive and expressive vocabulary, and sentence repetition, even when holding their level of English exposure constant. The second objective of this study is to investigate possible limits to the general pattern found in a large body of research on bilingual exposure, which is that lesser exposure leads to lesser skill in that language. First, language exposure may affect receptive skills less than expressive skills. Second, grammatical knowledge may also be less closely related to exposure than lexical knowledge. There are conflicting findings in the literature. My findings are consistent with a weak relationship between receptive skills and language exposure in bilingual children. Despite having lesser exposure to English (34% of their total language exposure), children in this study did not show a relation between variation in exposure and their English receptive vocabulary scores. In these children, the relationship between exposure and grammatical proficiency was similar to that with lexical proficiency. The third objective is to investigate additional contributors to bilingual proficiency. Previous research suggests that children’s socioeconomic status (SES), the status of the languages they speak, and the existence of cognates in their languages make contributions to bilingual children’s proficiency, and may in turn modulate the effect of diminished language exposure (e.g. Cobo-Lewis, Pearson, Eilers, & Umbel, 2002a; 2002b; Thordardottir, 2011). My results suggest that SES and high prestige of the languages being acquired may partially mitigate – though not eliminate – the effect of diminished exposure on bilinguals’ home language proficiency. Similar to findings for other bilingual children from mid- to high-SES backgrounds, these children showed age-related growth in English proficiency, and their English receptive skill differed minimally from monolingual norms. However, the effect of lesser exposure to English can be seen more clearly in their expressive skills, which were lower than monolingual norms and were predicted by variation in their English exposure. The effect of cognates in French and English was also investigated in terms of the advantage they conferred on my measures of lexical proficiency. This effect was significant in both receptive and expressive measures; thus, I conclude that the presence of cognates may also mitigate the effect of bilingual exposure. Finally, this investigation also examines additional individual factors that can influence language proficiency, but which have rarely been taken into account in studies of bilingual proficiency and both its relationship to exposure and production. Specifically, variation in children’s working memory and their exposure to language through overhearing adult conversation have both been linked to language learning in monolingual contexts but are not well understood in the context of bilingual development. In this study, verbal and visuospatial working memory were positively related to English proficiency scores. Indirect exposure from overheard English spoken between parents was not related to proficiency scores when holding direct English exposure from parents constant. However, indirect exposure was related to how much children produce English themselves to their parents, even while holding direct exposure constant, indicating that language use between parents may influence children’s language production with parents. This study contributes to our understanding of how bilingual language exposure and production relate to bilingual language proficiency in the following ways: first and most importantly, it adds to the small but growing literature that shows a strong link between bilingual children’s own production of a language and their lexical and grammatical skill in that language. It is also the first to my knowledge to find that a measure of children’s language production, inter-speaker code-switching, is negatively related not only to expressive but also to receptive lexical skill in the language that children switch from. Secondly, the finding that children’s English exposure is unrelated to their English receptive skill (but related to age, indicating continuing growth in these children) affirms exposure’s differential relationship with receptive versus expressive skills. It also documents a limited role for exposure in a new population (French-English bilinguals in France), supporting the role of cognates, socioeconomic status of children, and high social prestige of languages being acquired in mitigating the effect of bilingual exposure. Finally, in finding an independent contribution of working memory to lexical and grammatical skill in bilinguals, it highlights that these measures should be considered when investigating variation in bilingual proficiency.