<p>marite, </p>
<p>You understood my point correctly.</p>
<p>To summarize what the research has and has not shown:</p>
<ul>
<li> performance on AP exams is predictive (no surprise here) (of grades, graduation success, you name it…)</li>
<li> the mere taking of an AP course is not predictive.</li>
<li>there is much concern in the academy regarding the CONTENT of AP courses: improving the courses to better prepare for subsequent college material.</li>
</ul>
<p>Notice that none of the research separated the student from the course. To support Matthew’s implied hypothesis (maybe it is explicit. I don’t know), that AP courses help a broad range of students, one would need to compare a broad sample of students who took APs to a comparable sample of those that did not. Then you’d probably do subset analysis etc. Take a look at some of the methodology in the Early Admissions Game to get an idea of how to do this stuff properly.</p>