IQ performance is increasing, but this is obscured by the fact that IQ tests are standardized so that the average score remains 100, regardless of raw score. Thus IQ test scores cannot be meaningfully compared over decades, because the meaning of 100 is always relative to what the average performance is at a given time.
One other point: IQ validation studies are done on broadly representative samples. The proportion of HS students taking the SAT has increased significantly over time. Thus historical comparisons in SAT performance are questionable. Also, there is no explicit attempt to maintain the same level of difficulty on the SAT test year after year; the primary goal is predictive validity (i.e., does this set of questions help us predict performance in college?). If putting more difficult (or easier) items on the SAT allows the test to better predict college performance, the SAT is revised.
In sum: we can track historical trends in SAT scores, but that tells us nothing important about the academic skills of current students compared to students in previous generations.
Miles