I think that 80 children is too small. The different results could as been biased that there were different classes, so different (possibly) skill levels of the students.
For extrapolating the results to the entire Earth population? Let's see, that would require a cross-generational study across a number of different societies, across different schools and classrooms, from different socio-economic situations, and with an equal number of control groups.
So we're looking at a minimum of 4 (test moments across 25 years) * 4 (societies) * 5 (schools across socio-economic areas) * 2 (classes per school) * 2 (test + control group) * 30 (pupils per class) =~ 10000 test subjects.
How long have you been saving up this one-word response? Because the parent was addressing a complaint about small sample size, and basic reading comprehension would show you that your response was off-topic and did not logically follow.
