A great deal of research has been carried out over the past two decades related to the effectiveness of computer-assisted teaching and learning. This large body of work, however, has not produced unequivocal findings, especially in the area of CALL. Our study takes a systematic look at what dedicated meta-research since 1991 might have contributed to this controversial field. Data was gathered from several hundred studies including ca 20,000 subjects. The paper highlights the benefits and limitations associated with this type of research, synthesizes findings and recommends models and strategies for future research in the context of an agreed research agenda. Our synthesis suggests consistent positive findings related to L1 spelling, writing and reading. While there is some support for this related to CALL per se, much work needs to be done before we might reach the levels of confidence enjoyed by some of the findings reported here.