Learning independence is very influential in improving student learning outcomes. But at this time, many students still need an independent attitude in learning, which impacts low student learning outcomes. This study aims to analyze the level of learning independence towards improving mathematics learning outcomes based on educational level. This type of research is meta-analysis. This study uses a meta-analysis method, where data is collected from relevant research results, including the Fisher's test, t-test, correlation test, number of participants, and classical completeness. Data collection and research instruments were taken from relevant journals in this study. The data collection technique that can be carried out in this study is by looking for articles that match the predetermined criteria. Data were analyzed using JASP software with each data's input effect size and standard error values. The results of data analysis showed that there was a significant influence between learning independence on mathematics learning outcomes at all levels of SD, SMP, SMA, and PT by 20% (sufficient category), besides that at the SD level by 17% (low category), at the SMP level by 22% (enough category), at the high school level by 25% (good category) and PT level by 11% (low class). So the effect of independent learning on student learning outcomes at the SMA level is higher than at the SD, SMP, and PT levels.