The aim of this paper is to examine a variety of features of research that might account for mixed findings of the relationship between teachers' subject matter knowledge and student achievement based on meta-analytic technique. Of several variables that might contribute to inconsistencies and variations in study findings, three features of research--measurement, data analysis techniques, and the samples of subjects--are especially explored. Relevant studies including 4 journal articles, 1 conference paper, 4 reports, and 18 dissertations provide 64 effect sizes to represent the association between subject matter knowledge. The overall mean effect size and its 95% confidence interval are obtained under random-effects model. Then, the effects of potential moderator variables on the relationship between teachers' subject-matter knowledge and student learning are examined using both weighted regression analysis and AN OVA-like categorical analysis. Result indicates that teachers' subject matter knowledge is positively related to student mathematics achievement but the magnitude of the estimated mean correlation coefficient across the 27 studies was very small under the random-effects model. The strength of the relationship between teachers' subject matter knowledge and student mathematics achievement was moderated by different types of indicators of subject matter knowledge and grade level taught, but none of moderators fully explains the variations among the correlations between teachers' subject matter knowledge and student achievement. In particular, under the random-effects model, the estimated mean correlation was .02 when subject matter knowledge was measured by educational background variables, but the estimated mean correlation was .11 when subject matter knowledge was measured by teacher tests. In addition, while the estimated mean correlation was .05 at the elementary level (grade 1-6), the estimated mean correlation was .07 at the l researchers or policy makers who make inferences about the relationship between teachers' subject matter knowledge and student learning and methodologists who want to carry out research synthesis in education or social science as well. However, it is required to explore more moderators to fully explain variations in the relationship between teachers' subject matter knowledge and student achievement. (Contain 3 figures, 10 tables, 2 appendices)