The purpose of this study was to estimate pooled reliability values for the total and sub-scales of the Children's Test Anxiety Scale (CTAS) and to investigate the sources for the heterogeneity between the reliability estimates of the studies using the scale. In line with this purpose, a reliability generalization meta-analysis was applied, with 34 studies reporting 38 alpha coefficients. We also examined the induction rates of the CTAS. The pooled reliability values were estimated as .915, .867, .795, and .740 for the total scale, thoughts, autonomic reactions, and off-task behaviors sub-scales, respectively. Results of the meta-analysis showed that heterogeneity was detected in the total and subscales, and moderator analyses were conducted to determine the source of heterogeneity. The analyses revealed the statistical significance of some categorical (school level, continent, country, scale type, language, publication type) and continuous (mean age and mean scale score) variables. Reliability induction rate was determined to be 11.21%. This study represents the first reliability generalization analysis of the CTAS. The generalized reliability of both the CTAS and its subscales was established. Our findings endorse the CTAS as reliable instruments for assessing test anxiety in children. The limitations of the scale were mentioned, and suggestions were made to the researchers. Recommendations are offered for future research.