Via Kevin Drum, I heard about the FIMS, First International Mathematics Study of 1963-1967, the ancestor of the ongoing TIMSS (Third International ...). Here's a 1992 report (5.6 meg PDF) by Elliott A. Medrich on the results of FIMS and four other early international school achievement tests. Americans did pretty badly except in one science test.
I skimmed it but was mostly struck by how all-over-the-board early results were (e.g., Nigeria beat Sweden in certain math subtests). In the early tests, there didn't seem to be much correlation between national rankings on math subtests.
Math and science tests 13-year-olds seem to be particularly susceptible to the order in which subjects are taught to younger tests, while tests of high school seniors are much influenced by high school graduation rates (which were much higher in the U.S. in the 1960s) than elsewhere. In the 1960s, most country had only the smart fraction of their 17-year-olds still in school, so they beat our vast masses of high school seniors on a per capita basis.
One interesting result was that Israel came in tops in the FIMS of the mid-1960s (of course, they didn't give the test in Arabic). Today, Israel is thoroughly mediocre in, say, PISA. Some of that comes form testing Arabs, some from demographic changes involving both Arabs and the types of Jews, some from the old Zionist strategy of dumbing down the culture to produce fewer financiers and more farmers.
There's lots to be mined from this report for the enthusiasts.