The argument usually runs without reference to the study that addresses most of it. ACT ran a linking study in June 2024 to put enhanced scores on the same 1 to 36 scale as the old ones, and the March 2025 report from that study compares the two formats directly.

On the question of whether the questions themselves are harder, ACT's finding is this: "Although there is a tendency for the enhanced forms to be slightly more difficult than the legacy form (especially for reading), the differences are small."

So: slightly, especially in Reading. And because the scale is equated, that difference is absorbed before it reaches your score. Section scores from the two versions are meant to be used interchangeably, including inside a superscore that mixes them.

Two other things did change, and they are more concrete than the difficulty question.

Ten of the fifty English questions do not count

The enhanced ACT embeds unscored field-test questions inside each section. ACT publishes both counts:

On the old ACT every question in the four scored sections counted, with field-test items sitting in a separate section of their own. Now nine of the thirty-six Reading questions and ten of the fifty English ones are being tried out on you and contribute nothing to your score.

You cannot tell which ones, so this changes nothing about how to take the test. It changes the arithmetic of what a single question is worth, and that turns out to matter.

A scored question is a bigger share of your raw score

Be careful with what follows, because it is easy to overstate and most people do.

Each correct scored answer adds one point to your raw score, and fewer scored questions means each one is a larger share of that raw total:

• English went from one of 75 to one of 40
• Reading went from one of 40 to one of 27
• Math went from one of 60 to one of 41
• Science went from one of 40 to one of 34

What that does NOT mean is that a question is worth twice as many scale points. ACT converts raw to scaled through a form-specific table that is not a straight line. On ACT's own practice test, one Reading raw point at the top of the scale moves you two scale points, from 34 to 32, while one raw point lower down moves you nothing at all. A question has no fixed scale-score value, so no ratio of question counts tells you what a miss costs.

Two careless errors used to be diluted across a long section, and in a 27-question scored Reading set there is much less to dilute them with.

The enhanced forms measure slightly less precisely

This is the least intuitive part, and the one most easily overstated, so here is exactly what ACT measured and what it did not.

ACT reports that reliability for the enhanced forms runs .83 to .88 for the test sections, "slightly lower than that for the legacy forms due to the reduction in test length". The conditional standard error of measurement rose "by about half a score point for each test section score", and Reading's standard error came out slightly above 2.5.

What that means: a section score on an enhanced form carries slightly more measurement error than one on a legacy form. In the two enhanced forms studied, Reading carried the highest standard error of the four sections and English showed the largest increase over the legacy form.

What it does not mean is that ACT watched students score differently across two sittings. The study assigned each student one form, legacy or enhanced, so it measured the precision of the instrument rather than the movement of individuals. And these are operational test forms, not the public practice tests you will actually sit at home, which the study did not examine.

ACT's own conclusion is that "the changes in measurement precision had a very small impact on score interpretation". Whether any of this carries over to the practice tests you sit at home is not something ACT has published, so we are not going to tell you it does.

You get more time per question, in every section

English and Math lost questions and lost proportionally less time. Reading lost four questions and gained five minutes. Science kept all forty and gained five minutes. Three different patterns, one result: the average time per question rose in every section.

Reading gained the most, about a quarter more time per question. Worth noting that Reading was not where students were most squeezed. In the linking study's sample, once examinees who rushed the final questions are excluded, ACT put legacy completion at 86 percent for Reading against 69 percent for Math and 63 percent for Science; on the unfiltered count the three were 93, 90 and 89. ACT notes the results depended on which definition was used and were mixed outside Science, so read this as one study's measure rather than a settled ranking. ACT reports some evidence of small positive effects on completion, which is a more careful claim than saying the test is now easy to finish.

Math also moved from five answer choices to four, so a blind guess goes from one in five to one in four.

What to do differently

Not much, which is the useful finding. ACT's position is that the enhanced test measures the same constructs, though it also rebalanced the reporting categories and changed the structure of the questions, so "nothing changed" overstates it. The full list of what did change is still shorter than the volume of discussion suggests. What moved is where your risk sits. More time per question is a fact; that it makes pacing less of a threat for you specifically is an inference, and ACT claims only some evidence of small positive effects on completion. And because the raw-to-scale conversion is nonlinear, a missed question does not carry a fixed cost either, so treat both halves of that trade as directional rather than measured.

Two adjustments follow. Practise on current-format material, because a 75-question English section trains a pace that no longer matches the test. And watch the trend across several tests instead of reacting to one, which was always the better habit and does not depend on any of the above.

The other change people misread as difficulty is Science leaving the composite, covered in what the three-section composite changes.

Where these numbers come from

Question counts, scored counts and section times are from ACT's published exam structure page and Table 1 of its Enhanced ACT Linking Study Report of March 2025. The difficulty comparison, the reliability figures and the measurement-error figures are from that same report, and the quotations are its wording. The five-to-four Math answer-choice change is documented separately, in ACT's Design Framework for the ACT Enhancements. The point about superscoring across versions comes from ACT's guidance on interpreting enhanced scores. Checked August 2026.