Within labour economics, returns to education is an area of focused research. Moreover, amongst studies looking at emerging economies, China is the most widely studied economy. While there is a general consensus that returns to education are positive, studies use various datasets and methodologies and consequently present varying estimates of returns to education. We perform a meta-analysis of these estimates of the returns to education in China, addressing issues of heterogeneity in the existing literature and examining whether variations in reported estimates can be explained by study characteristics such as dataset and estimation methods, among others. The meta-regression results show that variations in reported estimates can be accounted for by study characteristics such as data source, estimation method and sample period, among others. The results support the college premium hypothesis and reveal that the returns to education for college graduates are higher than those for other (lower) levels of education.