Why not try some different class sizes first?

As 60% of the group got one question correct and 70% the other what does this tell you about the percentage that got both right?