ACT has a reputation for stodginess. Its eponymous test hasn’t had any substantial changes since 1989. Today, ACT just blew up that reputation. It announced superscore reporting, online testing on national test dates, and most radically, section retesting. The changes would go into effect starting in September 2020. EdWeek has been first to share reactions, and ACT has provided a detailed FAQ. But a wide range of questions remain unanswered, and we will have to wait and see how colleges respond. Here is our take:
Transforming the ACT one section at a time
Allowing students to retake individual sections rather than sitting for the entire exam may be the biggest change the ACT or SAT has seen. Each exam has evolved over the last 70 years, but the fundamental concept has always been three to four hours in a room on a Saturday morning. And with the exception of the foundering essay exams, the rule has been all or nothing. Starting next year, a student wanting to increase an English score need only retake English. A student wanting to raise Math and Science scores need only retake Math and Science.
Essential takeaways on section retesting:
- Before retaking a section, students must have earlier completed a full ACT.
- Section retesting will only be available online on national test dates. Section retesting is not yet available for state or school district testing.
- Section retesting is not available for paper-and-pencil testers, although a student’s original ACT can be either paper-and-pencil or online.
- No costs have been been determined for section retesting, and voucher policies have not yet been changed.
The rationale for section retesting is at odds with ACT’s research justifying the availability of section retesting. ACT claims that its research shows that section order does not impact student scores. Its research also shows that superscores are just as valid as traditional Composite scores. It concludes, then, that students will perform similarly whether they take one section or all four and that scores can be used interchangeably. However, if order and specialization don’t matter, why should a student intentionally forego the opportunity to raise scores in other sections? ACT is effectively encouraging bad testing decision-making. Or it is not being completely forthcoming about its research? A student with a 36 in English is an example where section retesting—skipping English—would clearly be beneficial. But what if the student has a 34? Or a 32? Or a 26?
ACT has heavily marketed full-exam retesting to students by emphasizing that, for the majority of students, scores improve. Students can’t always predict where that improvement will come, yet that’s what section retesting expects of them.
ACT has researched test order and superscoring, but it has not researched how section retesting will play out when combined with, for example, section prepping. Will scores and predictive validity change when students can focus on a subset of the test? As tempting as the prospect of shorter testing is to families, the benefit of section retesting—as opposed to complete retesting—is far from established.
Acceptance of section retesting by colleges
ACT is avoiding any claims about how colleges will accept section retesting. A score improvement is useless unless a college will incorporate it. Will section retesting finally convince non-superscoring colleges to reconsider policies, or will colleges simply turn up their noses at section retesting scores? Is it ethical to provide section retesting in the absence of clear policies from colleges? How are families and counselors expected to navigate the complex decision tree involving available test dates, section scores, college policies, and retesting options? At its core, ACT’s argument is that choice is good. A counterpoint is that chaos is bad.
ACT will begin next fall to provide students the ability to report their superscores—the best individual section performances across test dates. Specifically,
“ACT will supply [colleges] at least one full composite score with each superscore, plus all of the scores from the test events that are part of the superscore composite.”
It’s not clear what can be accomplished through superscore reporting that couldn’t be accomplished by reporting all scores. ACT seems to be trying hard to one-up College Board in an area where it has long been at a disadvantage.
ACT’s historical emphasis on the Composite score hampered it as more and more colleges shifted to superscoring. Of 360 highly competitive colleges and universities tracked by Compass, 61 still don’t superscore the SAT, while 134 don’t superscore the ACT. More than a dozen additional schools only superscore-lite for the ACT, looking at the highest section scores but not calculating a new Composite score. For decades, ACT encouraged this line of thinking. Only in the last couple of years has the organization (a) realized the competitive disadvantage being perpetuated and (b) done the research to show that superscoring can be at least as predictive as highest Composite scoring.
The competitive disadvantage has clearly been weighing on the organization. On its superscore FAQ, it argues for fairness and consistency between SAT and ACT policies no fewer than 4 times. Compass agrees with this sentiment—at least when considering full-exam testing and retesting. There is no published research indicating that superscoring is more valid for the SAT than for the ACT. The distinction lives on as a historical artifact. However, ACT seems to be using students as unpaid lobbyists. It is rolling out a policy before getting sign-off from colleges. It neglects to consider that even colleges that superscore often reserve the right to “consider all scores.” Superscore reporting sounds student-friendly, but it is decidedly unfriendly if students don’t know how to properly use it or if college applications and electronic data layouts can’t handle it.
Who will benefit from superscoring and section retesting?
ACT’s FAQ includes the question, “Do you expect to see an increase in superscores overall with the advent of subject retesting?” Tellingly, it has neglected to provide an answer.
With the zeal of the converted, ACT may be overstating the case for superscoring. It is making strong claims about how superscoring does not disadvantage subgroups of students, while neglecting to consider the interplay of section retesting and superscoring. The proportion of low income students taking the ACT only a single time is almost twice that of high income students. High income students, as a group, start testing earlier, end testing later, and take the test more frequently than low income students. The lack of uniform superscoring has acted as something of a check. Amherst might superscore, for example, but a student applying to Georgetown would still have to contend with an “all scores” disclosure policy.
Online testing rollout
ACT previously rolled out computer-based testing (“Online Testing” in ACT parlance) for some state and school district testing (2015) and for international students (2018). As of September 2020, it will begin making online testing an option on national test dates.
Online testing is computer-based, but is otherwise identical to paper-and-pencil testing.
- The content will be the same.
- The number of questions will be the same.
- The timing will be the same.
- Online testing is done only on secure, school-administered computers at testing centers.
The key selling points to students are:
- You can take the test in the mode in which you are most comfortable.
- You get your multiple-choice scores back in two days versus two weeks.
- Section retesting will only be available online.
ACT would like to establish itself as the leader in computer-based testing for college admission. ACT and College Board both see a future where paper booklets can be abandoned. Booklets are expensive to print, expensive to ship, expensive to keep secure, and expensive to return and grade. They are all too easy to steal, scan, and put on the web. ACT and College Board both re-use some test forms in order to stretch their R&D dollars. Breaches can cost ACT and College Board millions of dollars.
It is unlikely that ACT plans to turn over any of these savings to students. It has not yet announced pricing for online testing. Compass expects online testing centers to be stretched thin. ACT has often found it difficult to provide enough seats for paper-and-pencil testing that can be administered in virtually any classroom. Online testing can only be done in sites with the necessary infrastructure. Students looking to do section retesting will be competing for space with those taking full tests. ACT has provided no specifics on how it will convince schools to undertake the complexity of section retesting.
Choice versus confusion
At first blush, these announcements seem radically pro-student. Choice, choice, and choice. There are a number of things that have to come into alignment, though, if ACT is to avoid confusion, confusion, and confusion. It is not clear that students will benefit from section retesting, and they could, in fact, be missing out on important gains. To the extent that students do benefit from section retesting, it is not clear that the benefit will be uniform across subgroups. It will be interesting to see how College Board responds.
Superscore reporting seems like a clever marketing campaign geared toward students on one side and universities on the other. If the new reporting encourages colleges to coalesce around a set of standards, it will benefit students. At present, it is one more item upon which parents and students can perseverate.
ACT has produced research that shows comparability of results for paper and computer-based tests and has tried to show that issues as varied as screen-size or processor speed or digital divide will not skew results. Limited seating is likely to slow the rollout of online testing, and in this case, a slow rollout is for the best.
The new score reporting and testing options are clearly designed to encourage more testing and to produce more revenue for ACT. Whatever the immediate reactions are to these changes—and we’ve provided some of our immediate reactions—the long-term consequences are uncertain. Section retesting, in particular, could dramatically alter how standardized tests are evaluated and trusted.