<p>So at my school on the west coast those that do poorly on the SAT are encouraged to take the ACT. I’m wondering if this is a common practice, and then if it could have the potential to skew the tester pools in any significant way i.e. make ACT takers on average measurably dumber.</p>
<p>Also would be nice to know if the reverse happens: if people who do poorly on the ACT where its the norm, if they’re encouraged to take the SAT. </p>
<p>'Cause the thing is, I’ve been seeing friends hitting higher percentiles on their ACT then their SAT.</p>
<p>^Spammer.</p>
<p>I think the pools are about equal in ability. Most people tend to do better the second time they take a standardized test, so that might explain your observation.</p>
<p>You cannot compare percentile rankings across these two tests. Their testing pools are significantly different. Currently six states (CO, KY, IL, MI, TN, WY) require the test be taken by all high school grads (see [2010</a> ACT National and State Scores: Average Scores by State](<a href=“http://www.act.org/news/data/10/states.html?utm_campaign=cccr10&utm_source=data&utm_medium=web]2010”>http://www.act.org/news/data/10/states.html?utm_campaign=cccr10&utm_source=data&utm_medium=web)), as well as very high compliance (96+%) in LA and MI. In these states even non-college bound students take the exam. There is no state with similar requirements for the SAT, although ME comes close (<a href=“http://professionals.collegeboard.com/profdownload/cbs-2009-Table-3_Mean-SAT-CR-MATH-and-Writing-Scores-by-State.pdf[/url]”>http://professionals.collegeboard.com/profdownload/cbs-2009-Table-3_Mean-SAT-CR-MATH-and-Writing-Scores-by-State.pdf</a>). So, yes, you can assume more lower-ability students are taking the ACT.</p>
<p>You can, however, compare scores by concordance ([Estimated</a> Relationship between ACT Composite Score and SAT CR+M+W Score](<a href=“http://www.act.org/aap/concordance/estimate.html]Estimated”>http://www.act.org/aap/concordance/estimate.html)). This chart is not based on percentiles but rather on the comparative performance of 300,000 students who took both tests (in other words, the exact same pool). The ceilings on both tests turn out to be relatively similar.</p>
<p>descartesz: that makes sense. I was just guessing they correlated them by %-tiles. But of course that would be stupid, so this concordance comparison thing is a nice solution. Thanks for the reply+info.</p>
<p>hahalolk: good point. And I’m not sure what you meant by spammer…I thought my question was valid.</p>