Increasing Educational Achievement Via Self Concept Change

s do not always include full details on some aspects of the study, the overall results are usually reported. Dissertations are included here which contained measurements of both self concept and academic achievement variables, after an attempted educational innovation. Each dissertation also had to include comparison between one or more groups of students who received the intervention program, and other students who did not, although these groups were not always randomly selected. With these criteria, a total of eighteen dissertations were located which evaluated programs ranging from centered education for first-graders to Personal Growth Groups as a compensatory education component for college freshmen (Beckum, 1973; Bradford, 1973; Hale, 1973; Hopke, 1975; Jones, 1975; Kalunian, 1975; Kenemuth, 1975; Kosensky, 1975; Lawson, 1974; McGinnis, 1972; Mendenhall, 1974; Mynatt, 1973; Pine, 1975; Pokipala, 1975; Poudrier, 1976; J.A. Smith, 1975; M. Smith, 1971; Zirkel, 1972). The effects reported by these studies are strikingly simple to summarize: in no case were changes in achievement unambiguously associated with changes in self concept. None of these educational programs showed measurable effects on all target groups' self concept scores while at the same time increasing academic achievement. Only two studies found positive changes on both variables for even part of their target population. Lawson's (1974) study of versus graded schools for pupils in their first, third, or fifth year of schooling found higher reading achievement for the non-graded schools at all three levels and higher self concept scores in the same schools only for the fifth year students. Self concept change here appears to be an outcome of reading success rather than an intervening variable. Bradford (1973), comparing one school with an Individually Guided Education program with a school having self contained structure, found significantly greater gain scores in the experimental school on both self concept and arithmetic achievement, but not reading achievement. Few details are available for this study, but the use of only one school for each type of program confounds the potential program effects with other possible differences between the schools. The questionably positive results from these two studies thus do not overcome the main thrust of the pattern of negative results in all other studies. Further, a close examination of the results reveals that most interventions produced some effects on either self concept or achievement, but not both. Ten studies (Bradford, 1973; Hopke, 1975; Jones, 1975; Kalunian, 1975; Kosensky, 1975; Lawson, 1974; McGinnis, 1972; Pine, 1975; M. Smith, 1971; Zirkel, 1972) found significant effects on their self concept measures. Not surprisingly, these studies were mainly those emphasizing an or child-centered program, along with several vocational education programs for teenagers, which may also be more activity centered. Seven studies (Beckum, 1973; Hopke, 1975; Kenemuth, 1975; Lawson, 1974; Mynatt, 1973; Poudrier, 1976; J. A. Smith, 1975) produced effects on achievement variables. In several of these, the differences favored groups with traditional or academic oriented programs, (Hopke, 1975; Poudrier, 1976), or resulted from tutoring programs (Beckum, 1973; Kenemuth, 1975). Thus, the lack of evidence in these dissertations for a connection between self concept 139 SCHEIRER Smith & Bissell, 1970). As this paper has frequently noted, methodological problems are endemic in this type of research and frequently contribute to the difficulty of interpreting each individual study. Adequate control groups were not always available, particularly in the larger scale programs, although randomized assignment was often present in the dissertation research. Outcome measures were sometimes inappropriate for the treatment undertaken, as when IQ tests, usually viewed as assessing underlying academic abilities, were used in programs emphasizing increasing the child's motivations for academic achievement. Further, Wylie's (1974) recent review of self concept measures concluded that no well-constructed, well-validated measuring instruments are yet available, particularly for self concept in young children. The failure to find self concept changes in the programs reviewed here was often attributed by their analysts to inadequate measurement in this affec-