The Johns Hopkins report counted what three SAT forms ask students to do. It did not measure whether the test became easier, and the Institute that published it has since said so directly, after coverage reported otherwise.

That leaves a more useful question than the headline one: what changed, what points the other way, and what happened when passage length was tested on its own.

Three forms, 46 years apart

The Institute for Education Policy at Johns Hopkins University published *The SAT Reading Section: A Comparative Analysis of the 1980, 1987, and 2026 Practice Tests* in late October 2026. The research was sponsored by Thomas Aquinas College, which commissioned it.

It compares the verbal side of three released forms: one from 1980, one from 1987, and a current College Board paper practice test. That is the entire sample, and one current paper form cannot stand in for every form in use today.

The current form in the comparison is a paper practice test. The operational digital SAT changes your second module based on how you did on the first, and a paper form cannot reproduce that. The paper practice form also carries 66 verbal questions where the operational digital Reading and Writing section carries 54, so the question counts below describe the forms studied, not the test you will sit.

What the counting found

Vocabulary moved the most. Each older form carried 60 vocabulary questions out of 85 verbal questions. Of those 60, some 45 were analogies and antonyms, which require you to arrive already knowing the word, and 15 were sentence completions that gave you context to work from. The current form carries 10 vocabulary questions out of 66, and all 10 supply context.

So the category that disappeared is not vocabulary. It is vocabulary without context, which went from 45 questions per form to none.

Those question types did not vanish recently. Antonyms were dropped in 1994 and analogies in 2005, both long before the test went digital.

Reading passages got much shorter. The report measures an average of 304 words in 1980 and 340 in 1987, against about 97 words on the current form. The older forms attached several questions to each long passage; the current one usually attaches a single question to each short passage.

The finding that points the other way

The report did not only find reductions, and this is the part the coverage mostly skipped.

On its measures of text complexity, the current passages came out above the 1980 form and close to the 1987 one. The report is careful about this, noting that readability measures get less reliable the shorter the text, which matters when the passages being measured average under 100 words.

Thomas Aquinas College's vice president and dean emeritus, John Goyette, summarized the balance this way in the sponsor's own announcement: "This report is a content analysis. It tells us only what the test asks. It does not, and was not designed to, tell us whether the test has become more or less difficult for students." He added that four of the report's five findings suggest the newer SAT demands less of the test taker, while the fifth, text complexity, suggests the opposite.

He speaks for the college that commissioned the study rather than for the research team. The Institute itself addressed the "easier" reading in a separate statement published in November 2026, after coverage drew that conclusion, and said the report does not make that claim. The report's own recommendation is the study that would answer the difficulty question: the same students answering questions from both the old and the current forms, with results placed on a shared scale.

Until somebody runs that, "the SAT got easier" and "the difficulty has not changed" both reach past the evidence.

What happened when passage length was tested directly

College Board tested the "shorter means easier" assumption in a July 2026 study. That is worth saying plainly before the results, because the test maker studying its own test is not a neutral source.

The research took two approaches.

The first followed students who sat the paper test in autumn 2023 and the digital test in spring 2024. Their paper-to-digital score correlation was about 0.87, against about 0.89 for a comparison group who retook the paper test with no format change. That tells you the two formats rank students very similarly. It does not tell you whether either was easier, because a uniformly easier test produces the same correlation. Treat it as evidence of similar score ordering across the two formats, and not as evidence about difficulty.

The second was a controlled experiment, and it is the one that speaks to length. It recruited 367 students to answer matched questions built from both a short and a long version of the same passages, with content, question skill and intended difficulty designed to be held constant so that length was what varied.

Students averaged 5.43 out of 8 on the short-passage questions and 5.89 out of 8 on the long-passage ones. The difference is statistically significant and favors the longer versions, with a small effect size. Broken out by subject, history and social studies accounted for most of the gap, and the humanities questions also favored the longer passages by a smaller margin. Science showed no significant difference at all. Literature ran the other way, with the short version scoring higher, though across all students that effect was negligible. Among the small group with the weakest prior reading results, the short-passage advantage in literature reached a medium effect.

The study's limits matter as much as its result. It rests on four passage pairs and 16 questions. Three of the four long versions were produced by prompting a generative writing tool to expand the short original, then edited by people for comparability. Students had about 45 minutes for the activity, which is far more time per question than the real test allows. Students with weak prior reading results were underrepresented.

What it supports is narrow and still useful: across these questions overall, making a passage shorter did not make them easier to answer.

Where difficulty sits on the current test

If length is not the thing separating hard questions from easy ones, the candidates are what a word count cannot show you. The rest of this section is our reading of the current section rather than a finding from either study.

On a short passage, two or three sentences carry all the claims, so the relationship between them tends to decide the question. The stem can ask for a precise inference where a looser one is available and wrong. Several answers can sit close to the evidence while only one is supported by it.

None of that is visible before you start reading, which is why a short passage is not a free question.

What to do before your next test

Skip analogy and antonym drills. Those question types have been gone for two decades. Vocabulary still matters, and the current test gives you the sentence and asks what the word is doing in it.

Practice reading charts and tables inside reading questions, and treat revision and transition questions as questions about meaning rather than about rules. Both are on the current section and neither was on the 1980 or 1987 forms.

When you review a timed section, sort your misses by what the question asked rather than by how the passage looked. Words in context, central ideas and details, evidence, inference, text structure and purpose, cross-text connections, transitions, rhetorical synthesis, and the conventions questions about punctuation and grammar are the categories the current section is built from. One section is a small sample, so look for the category that keeps costing you across several of them.

Primary sources: Thomas Aquinas College's announcement of the study, which carries the Goyette quote, the Johns Hopkins report, the Institute's November 2026 statement, and College Board's passage-length research brief.

For what else changed when the test went digital, see our guide to the adaptive second module.