A short google brought up this:
http://www.ascd.org/ASCD/pdf/journals/ed_lead/el_199103_allan.pdfThe major problems with the research in the field appear to be that a) gifted students were excluded from the research, but results are applied to them indiscriminately, and b) studies are not carefully differentiated as to whether different ability groups received ability-differentiated instruction including differentiated learning objectives, and those objectives could be adequately demonstrated by testing.
Meaning that if you teach the same average ability curriculum and test for grade level standards, it doesn't matter whether you group or not : the high ability kids will do well, the average kids will do averagely and the low ability kids will struggle. In other words, does not really matter what kind of classroom your needs aren't met in.
But if you have an ideological axe to grind, you can use those poorly designed studies to advocate against grouping for the gifted and the high ability kids (defined in this article at some point as the top 33 %, who I think deserve ability leveled instruction as well.