Meta-Analysis procedures were employed to integrate existing science education research findings on the effectiveness of teaching strategy analysis training to: (1) determine overall magnitude of effect that can be expected from such training; (2) determine and compare relative magnitude of effects attributable to various approaches to such training (for example, self-analysis versus analysis of models, or microteaching versus no microteaching); (3) examine the relationship between the effect and the school level of the teacher (elementary, middle, or high school); (4) compare the effects when measured directly by recording teacher behaviors versus more indirectly when measuring student behaviors; and (5) calculate the effects assessed through behavioral versus cognitive or affective outcomes. Studies selected included those with sufficient statistical data to derive an effect size for treatment in the study (difference between baseline mean and treatment group mean in units equal to standard deviation of baseline data). Results indicated: (1) teaching strategy analysis is an effective training procedure; (2) all procedures presented had a positive impact on teachers' behavior, especially the use of models and providing systematic feedback (method of choice); and (3) positive effect sizes in cognitive and affective areas and across educational levels. (Author/JN)