Politicians in various European countries politicians are debating on methods of giving immigrant children opportunities to learn the national language inside or outside their homes, before entering school. This paper reports on the efficacy of attempts to improve the educational chances of immigrant children in the Netherlands through center-based early intervention. Large- scale studies show that children from low-SES, immigrant families lag behind their native peers in Dutch primary education, particularly with respect to Dutch language proficiency (Dagevos, Gijsberts & van Praag, 2003). The fact that this educational lag can already be observed in kindergarten and is not compensated for in the subsequent eight years of primary school, has lead Dutch national and local governments to invest in early intervention activities. Although there are other approaches to early intervention, policy nowadays seems mainly directed at center-based projects: various intervention programmes have been developed, which are to be carried out in preschool playgroups. One of the assumptions is that these playgroups provide immigrant children, in particular, with a rich second language environment that will enhance their acquisition of Dutch, a claim which seems to be supported by evaluation studies of the American Head Start project (cf. Currie & Thomas, 1999). Within the Dutch context, however, this assumption is much debated: some researchers suggest that these playgroup programmes do contribute to children's language development (Schonewille, Kloprogge & van der Leij, 2000), while others fail to establish a positive effect of playgroup participation (Driessen & Doesborgh, 2003; Leseman, 2000). In the paper presented here we will discuss some findings of two related studies into the effects of center-based intervention on the second language acquisition of low-SES, immigrant children in the Netherlands. The first is a quantitative study, in which two groups of children were compared: a group of immigrant children who took part in a playgroup before entering kindergarten at the age of 4 (N=37) and a comparable group of children who did not (N=32). No significantly better results were found for the playgroup children on school vocabulary and emergent literacy skills in kindergarten, nor on vocabulary, spelling, reading ability and reading comprehension in grades 1 and 2. We conclude, in accordance with the previous findings by Driessen & Doesborgh (2003) and Leseman (2000), that playgroup participation does not have the expected positive effect on immigrant children's early school success. One possible explanation of this finding is, that these children, when entering a playgroup, do not have the necessary linguistic tools to benefit from the 'rich language environment' provided there. The second study examines whether this lack of linguistic tools is reflected in teacher-child interactions in everyday practice in the playgroup. Teacher- initiated interactions were observed between four low-SES, immigrant children (all of Turkish origin) and their Dutch playgroup teachers, during a period of 10 months. We specifically observed 'higher-level' interactions, i.e. interactions aimed at providing the children with decontextualized, school-type language use (or Cognitive Academic Language Proficiency/CALP, using Cummins' (1979; 1991) terminology), e.g. during shared reading activities. We expected the observed children to have problems with these types of interactions as long as they lack the L2 proficiency necessary for actively taking part in these interactions. Preliminary results indeed show that the second language environment of the preschools did give the observed children the opportunity to acquire L2 skills for basic daily interactions, but still had difficulties in taking part in the teacher-initiated CALP-type interactions, when entering kindergarten.