How Do Standardized Tests Stay Consistent?

The Flynn Effect says IQs have been observed to increase from one generation to the next. But who's to say the IQ tests themselves haven't changed? How do we make IQ tests that are consistently equally difficult, especially down to the precision of tenths of an IQ point? The metric that should determine the difficulty of the test itself is, at the same time, the very metric the test is measuring. We have a causational loophole in standards. The same applies to reports on how today's American high school students are scoring on the SAT now compared with past students. The dual relativities of intelligence and intelligence tests make standards of comparison logically impossible.