Although College Board and ACT have worked together to create SAT-ACT concordance tables for previous versions of these tests, the newest concordance table reflects no such cooperation between these two competitors. Earlier this year, College Board issued concordance tables comparing not only Old SAT scores with the New SAT scores – but, purportedly, between the New SAT and the ACT. The two companies have been fighting for market share – and the ACT has been winning.
Marten Roorda, the ACT’s CEO, protested the CB’s action, stating that the SAT/ACT concordance was “produced by the College Board without our collaboration or the involvement of independent groups.” Criticizing the CB for issuing a concordance table before the New SAT had been broadly used for a year, he further noted that students who’d been willing to take the New SAT the first times it’s administered aren’t likely to be representative of the normal testing population. That means that early testing data – the data on which the concordance tables are based – should be viewed with healthy suspicion.
Roorda pointed out that the SAT-to-ACT concordance issued by the CB was based on something he referred to as “chained concordance” (more on that later). He saw some potentially weak links in that chain, one of which was the potentially unrepresentative initial testing population discussed above. Another is something the CB refers to as “equipercentiles,” which means that if 75% of the students taking the New SAT achieved that same score and 75% of the students who took the Old SAT achieved a different score, those two scores are “concorded” (equal for the purposes of comparison) — though how that can be so when the New SAT has changed so radically (no longer separate scores for reading and writing) truly escapes some CB critics.
Whether valid or not, equipercentiles are the CB’s answer to a question later posed by Roorda: “If you’re asking different questions using different rules and different scoring scales, how can you compare an old SAT score from last fall with a new SAT score from this spring?”
Which brings us back to “chained concordance”: After using a potentially unrepresentative sample to create “equipercentiles” that concord the New SAT and the Old SAT, the CB combined it with the existing concordance between Old SAT and ACT to arrive at a new concordance table that shows the relationship between New SAT and ACT scores.
The CB protested the accusation about lack of data by stating that it had “conducted two large-scale national concordance studies in December 2014 and December 2015,” but it takes more stretch than we’ve got to accept that those tests were similar enough to the current New SAT to make meaningful comparisons, “equipercentiles” and “chained concordances” notwithstanding.
The continuing inherently experimental nature of the New SAT lends support to its critics’ claims: After the CB for the first time purposefully introduced unnecessarily wordy questions into the New SAT math sections, its President, David Coleman, signaled an impending complete about face on that matter in response to a question posed by our owner, Judi Robinovitz. Here’s how Reuters reported the exchange on September 23rd of this year:
Judi Robinovitz, a Florida educational consultant, expressed concern about a Reuters report Wednesday that the College Board had ignored its own internal research showing that the math questions on the new SAT were too long…
Coleman said...that changes are in store for the new test…
“We are going to do everything we can to further simplify the mathematics section. Using superfluous words is superfluous,” he said, later adding, “Every extra word should go. Complex, distracting situations should go.”
So, what’s a fellow to do, especially if there are so many fellows – students, guidance counselors, educational consultants, test prep tutors, and college admissions officers?
Well, the current concordance tables might well have been constructed on shaky ground and seem destined for change – though probably not enough to end life as we know it – right now they’re all we’ve got. So our advice is that you use them, but remain a bit suspicious.
…and here’s a link to our concordance calculator reflecting CB’s score comparison between the tests.