Well quasi experimental designs could be either prospective or retrospective. The problem is they don't use a probability sample so one can't really judge their results statistically. There are numerous references some are contained in the attachment. Best wishes, David Booth