I’ve been wondering about how schools report ACT and SAT scores on their common data set filings. If a school superscores one or both tests, wouldn’t their 25-75 percentile scores be inflated compared to a similar school that only takes the highest test score? My kids’ superscored results were 1-2 points higher than their highest individual result. I’m not saying that the end result is intentional, but are we really comparing apples to apples when looking at the vaunted 25-75 percentile range as a differentiating factor between schools?
Yes, of course superscoring schools inflate the range of reported scores That’s the reason superscoring exists.
My realization/concern is that there is no place on the common data set filings to indicate whether or not an institution superscores. All the school rating entities use the CDS as the basis for their reporting. Most consumers assume the range is the range, which is not really the case, and US News, Princeton, Fiske, etc. don’t differentiate.
It’s all a game. Test optional also inflates scores. “Consumers” (parents and students) are the ones driving the numbers game, by paying far too much attention to the numbers-driven national school rankings, such as US News.
I am personally less concerned that students don’t know the true ACT range of a school than I am that students would choose one school over another based solely upon incremental differences in rankings or average ACT scores. There is so much more to a college than that and it’s sad that schools feel they need to play these games to satisfy / attract the typical college “consumer”.
Colleges generally only care about component scores, not the aggregate scores. In fact, only component SAT scores are reported on CDS (ACT composites are reported, but ACT aren’t supposed to be superscored). The measures of the distributions of component SAT scores aren’t additive, so the measures of the distribution of aggregate SAT scores for any college aren’t even publicly available.
The genesis of this thread was an information session yesterday for a highly ranked LAC where we were told that ACT tests were superscored by the school. I don’t disagree with any of the prior comments, am just noting that two differing types of scoring are being reported as the same in the CDS, at least regarding the ACT. In doing so, the superscorers probably add a point or so to their ranges, which is not insignificant (for “top” schools) and might attract or deter students. I’m not making a judgement on whether this should or shouldn’t matter, just find it interesting.
Not sure I understand your concern. If they superscore and the top 25th percentile score, let’s say, 750+ M, 740+ CR, what’s it matter what the other scores were, on other test dates? As 1NJParent points out, they aren’t using the aggregate, but these two components. They’re allowing for the best section scores, not trying to say what best total top kids had in one sitting. For families, knowing the target score zones for both sections is a lot.
The ACT issue is most report only the Composite range, when they can look at sub scores. They can superscore by subscore, whether or not they officially state they do.
I live in a predominately ACT state, so I’m not familiar with the SAT. I shouldn’t have included the SAT in my premise.
However, if a school superscores the ACT test wouldn’t it stand to reason that they would report the superscored test composite as the input for the 25th and 75th percentile on the CDS? That would give an inflated score compared to schools that just take the top submitted composite score.
No. For colleges that look at subscores (mostly top schools,) the Composite is not the determinant. If a kid got a 36 in sci on one date and a 36 in Engl on another date, how would the Composite for either date reflect anything valid?
In superscoring, it doesn’t matter what date (or the Composite or whole SAT total on that date.) It’s just about the knowledge a kid did reach a high score in that component, at some point.
I suspect you’re coming at this from the wrong perspective. Try to just think of scuperscoring as about the best individual scores, across all dates.
Adding: yes, it’s confusing that top colleges tend to report the Composite ranges. But you have to work with what they do give us. And for top colleges, there’s so much more that goes into “match.”
Far more colleges allow super scoring of the SAT than the ACT. Far more. So you should absolutely include the SAT in your premise.
The schools that only take the highest test score tend to be reaches for all. The students applying to those schools largely have top test scores already. I personally don’t think it matters if a college superscores. If the college is hellbent on climbing the USNWR rankings, they will do so by means necessary, and eventually others follow. The best thing thing is to ignore rankings.
Agree. I’m not concerned with rankings. My only point is that I feel fairly sure that if my child’s best single sitting ACT score of 31 is raised to 32 after taking into account the best subscores of his other sittings, that the school might just consider the score to be 32 for purposes of determining averages for the CDS. This is a lively discussion! Forgive me for being foolish if I am so. As my late father told me once, “A fool doesn’t know he’s a fool.”
Interesting article today in Salon that describes some academic research into college rankings. Basically, regardless of what the ranking systems say about weightings of scores, SAT/ACT scores explained the lion’s share of actual school rankings.
Those are extraordinarily high correlations. Colleges have a tremendous incentive to inflate these scores.
https://www.salon.com/2019/08/26/college-rankings-might-as-well-be-student-rankings_partner/
Most of the criteria used by top colleges are high correlated because they are driven by a common set of factors: intelligence, work ethics, wealth, etc. Even if a top college only uses the test scores as one of the criteria or only for a cutoff, its student body is likely to end up with many high test scorers anyway.
Or like Chicago Bates bowdoin they don’t really matter at all for some.
As to anything provided by colleges for the ranking systems, you need to understand they are not necessarily gospel but merely guidelines that typically have margins of error (usually not intentional). Middle 50% SAT or ACT scores always need to be looked at with somewhat of a jaundiced eye and you need to check what is actually being reported by the college. If the college superscores, the range will be higher than if it does not superscore.
Moreover, those middle 50% ranges are provided for the entering freshman class, which means all those very high scorers who were admitted but who chose to go elsewhere are not even counted in that middle 50%, which could mean for many colleges that the middle 50% being reported is lower than what it is for those who were admitted. In addition, colleges can have varying middle 50% ranges among the majors, e.g., those who enroll in engineering may have much higher test scores than those admitted to most A&S majors; some colleges provide those varying middle 50% ranges but many do not. For some colleges, you may see ACT ranges that appear to be lower than the SAT ranges and guess, incorrectly, that the college favors the ACT, when, in fact, the difference usually exists simply because the college superscores SAT but not ACT.