This paper presents a comprehensive picture of what has been investigated in terms of CALL effectiveness over the period 1981–2005 throwing light on why this question is still such a difficult one to answer unequivocally. The author looks at both strengths and weaknesses in this body of work, highlighting pitfalls and paradoxes in research procedures and providing valid design models. This includes the contribution of dedicated meta-analyses to this controversial field and a discussion of the benefits and limitations associated with this type of research. Substantial data, drawn from three extensive studies (Felix, 2005a, b; Felix, 2006a), allows the author to present for the first time synthesized findings relating to the impact of technologies on language learning. The paper concludes with strategies for future work in the context of a proposed research agenda.