Assuming there are collective agents that have beliefs, under what conditions are such beliefs justified? Specifically, how is the justifiedness of a group’s belief related to the beliefs of its members (and their justifiedness)? This under-explored question in social epistemology is starting to receive greater attention. One recent paper, Goldman’s “Social Process Reliabilism” (2014), advances a theory that parallels process reliabilism for individual justifiedness. A second paper, Lackey’s “What Is Justified Group Belief?” (forthcoming), includes an extended critique of Goldman’s theory, followed by a positive theory of her own. The present paper revisits the issues, replying to many problems Lackey raises. It is argued that the reliabilist approach, embedding aggregation as one of its prime elements, is not fundamentally unsound but needs clarification, elaboration, and tweaking of various sorts. One new refinement concerns the precise type of information transmitted when member beliefs are conveyed to the group. Lackey’s own positive theory will also be assessed, time permitting.