Bosma (1988) conducted a large scale study with a sample of some 1400 ABE literacy students over a two year period in Minneapolis. He was concerned with the characteristics of dropouts and persisters. Within the attrition category he was concerned with the differences between nonparticipants and dropouts. Within the persistence category he was concerned with differences between those who were still active in the program and completers. The study was shaped by the State of Minnesota's ABE reporting guidelines. Thus, “a nonparticipant” was someone who had less than 12 hours attendance at the program site. “A dropout” was someone who had dropped out after 12 hours of attendance in the program. An "active" learner was someone who had completed one module in the program and was still in the program. A "completer" was someone who had either advanced to another program, passed the GED, completed all his/her learning goals as specified on entry, or had completed one learning goal as specified on entry.

Bosma identified 36 independent variables in six categories for his study. He found that 17 variables were significant but only accounted for 6.5 percent of variance in attrition/persistence. He thus concluded that reasons for attrition are "unknown" and that 36 variables lacked the power to predict attrition and persistence. In trying to explain his results, he indicated that his sample consisted of a diverse, heterogeneous population which may have accounted for lack of relationships in the variables. He also indicated it was a large-scale study whereas results from other studies (e.g. Diekhoff and Diekhoff, 1984) used much smaller sample sizes. Another difficulty may have been the definitional categories imposed on the study as a result of the state's record-keeping requirements. The 12 hours of attendance to distinguish between nonparticipants and dropouts, for example, appears rather arbitrary. For other researchers and practitioners 12 hours would be considered indicative of participation!