Gender differences are one of the most contentious issues in educational research. This study analyzes long-term changes in gender gaps in reading comprehension at the end of primary school in 63 education systems. It links test data from seven comparative studies that were conducted between 1970 and 2016 using a common achievement scale based on item response theory. We investigate whether mean gender gaps have widened or narrowed over time--controlling for changes in the sample of countries from measurement point to measurement point--using a system-level regression with fixed effects. We observe an advantage of girls over boys in reading in almost all countries, although the size of the gender gap varies considerably internationally. Further, we observe a significant increase in the international gender gap between 1971 and 2001 and a slight decrease since then.