<p>It’s called equating. They reuse questions from last year to figure out how competent this year’s class is to last year’s.</p>
<p>For example</p>
<p>This year’s class earns an average score of 60.7 points, normally distributed, with a standard deviation of 2.7</p>
<p>Last year’s class earns an average score of 65.4 with a Standard deviation of 3.4.</p>
<p>We have two possibilities:
A. Last year’s class was smarter.
B. This year’s exam was harder.</p>
<p>In statistics, this is called a confounding variable; We cannot tell in this situation whether hypothesis A, hypothesis B, or a combination of both is true. (Science-lovers, this is why you only manipulate one independent variable!)</p>
<p>To solve this problem they reuse some of the multiple choice questions from the past exam. They compare the scores of this year’s students with the scores of last year’s students.</p>
<p>Continuing the previous example:</p>
<p>This year’s class earns an average score of 12.4 on the 20 common multiple choice, with a standard deviation of .8</p>
<p>Last year’s class earns an average score of 11.8 on the 20 common multiple choice with a standard deviation of 1.3</p>
<p>Last year’s students did score higher. However, on the common questions, they did not. In this example, we can see that hypothesis B is correct (this year’s test was harder). Having a set of common multiple choice questions allows the college board to differentiate between the varying difficulty of the test and the varying ability level of the test-takers in different years.</p>
<p>To discuss the questions jeapordizes the whole system and discussing questions, especially putting them online, could ruin the validity of the whole AP system and make all our scores worthless!</p>
<p>So don’t do it :)</p>