SAT vs ACT: the differences that matter (format, pace, content, scoring), and how to decide with one official practice test of each

The SAT and the ACT are both accepted by the colleges that ask for a test score, so the choice is about which one your skills fit. The SAT is a digital, adaptive test with two sections and more time per question; the ACT has more sections, a faster pace and an optional science section that is mostly reading data. The way to decide is not a quiz: take one official practice test of each, under real conditions, a week apart, and compare the error logs. Here are the differences that matter and how to read the two logs.

The two tests side by side, in our words

Every row was read on the two test-makers' own pages on the day this page was last reviewed, and it is described, not reproduced. Both tests changed recently, the SAT when it went digital and the ACT in its 2025 redesign, so a comparison table from an older page is likely wrong in several rows. The SAT page and the ACT page carry each test's own facts and the practice tests to use.

	SAT	ACT
Format	Digital, in the test-maker's app; adaptive by module, so the second module of each section is harder or easier depending on the first	On paper or on screen; since 2025 shorter, with science optional
Sections	Reading and Writing, then Math	English, math, reading; science optional; an optional essay
Time and questions	Reading and Writing 54 questions in 64 minutes; Math 44 in 70; 2 h 14 min of testing, plus a 10-minute break	English 50 in 35; math 45 in 50; reading 36 in 40; 125 minutes for the three required sections; science adds 40 questions in 40 minutes, the essay another 40 minutes
Seconds per question	About 71 in Reading and Writing, about 95 in Math	42 in English, about 67 in math and reading, 60 in science
Scoring	Each section 200–800; total 400–1600	Each section 1–36; the composite the average of English, math and reading, rounded
Wrong answers	No penalty	No penalty
Calculator	Built into the app and allowed for the whole Math section	Allowed on the math section only; bring your own, or use the one built into the online test
Reading	Short passages, one question each	Longer passages with several questions each
Science	No section; a graph or table appears inside the other sections	An optional section of data reading, scored on its own, outside the composite

The seconds-per-question line is computed from the counts in the row above it: 64 × 60 ÷ 54 ≈ 71; 70 × 60 ÷ 44 ≈ 95; 35 × 60 ÷ 50 = 42; 50 × 60 ÷ 45 ≈ 67; 40 × 60 ÷ 36 ≈ 67; 40 × 60 ÷ 40 = 60. The ACT gives less time per question in every section. That one line is most of what the rest of this page is about.

Which is easier?

Neither, in general. The question is which is easier for you, and the two tests reward different things.

The SAT gives you time to think and a calculator for every math question. Its passages are short and each has one question, so the reading load is spread thin. A student who is careful, who likes to check, who reads a problem twice before starting, tends toward the SAT.

The ACT rewards speed and stamina. Its passages are long and carry several questions each; its math has fewer seconds a question and some of it must be done by hand; its science section is a reading test with data in it. A student who reads fast, who makes decisions quickly and does not tire of long passages, tends toward the ACT.

Those are tendencies, and a tendency is not a decision. The practice tests decide. A student who thinks of themselves as slow may find that the ACT's pace is fine once the two-read skip rule is in place; a student who reads fast may find that the SAT's harder second module is where the marks go. Take both and read the logs.

Do colleges prefer one?

Colleges that require or consider a test score accept either. That is as far as the general statement goes; which test a particular program's applicants mostly send, and whether it wants a score at all, is the program's own policy, and it changes. Many colleges do not require either test. Check the admissions page of each program you are applying to, this year, and take its word over any page like this one.

For Canadian readers: most Canadian universities do not ask for the SAT or the ACT from students in a Canadian curriculum. Some programs ask students applying from a US curriculum for one. Check the program.

How to decide: one official practice test of each

Two weeks, one test each.

Week 1, the SAT. One full official practice test in the test-maker's own app, in one sitting, the real timing, the official break, in the morning if the real test is in the morning, phone in another room. Grade it with the official scoring the same day and build the error log by skill: for every miss and every guess, the skill, the kind of error (did not know, slipped, ran out of time) and the fix. The practice-test method is the hub's version of this.

Week 2, the ACT. One of the test-maker's free full-length tests, under its own timing, at the same time of day, with a watch on the desk. The same log, with the ACT's fourth column: whether you reached the question with time to think.

Compare the scores once. The two test-makers publish a joint concordance table that says which scores on the two scales correspond. It was built in 2018, before both tests changed, so treat a match within a couple of ACT points as a tie. Look up your two totals, see which sits higher on the shared scale, and then put the scores away. The table says where you stand today on each test. It cannot say which test you will improve on, which is the question you are trying to answer.

Compare the logs. This is what decides. Count each log's misses by kind. Pacing is trainable in weeks: a student who learns to skip a two-read question and come back, who reads the stems before the passage, who knows the seconds a question, loses far fewer time marks a fortnight later. Content takes months: a skill you did not know needs worked examples, cold redos and spaced review, one skill at a time. So a log heavy in "ran out of time" on the ACT and "did not know" on the SAT says the ACT would rise more with a fortnight of pacing work. The reverse pattern says the SAT.

A tie goes to the test whose dates fit your calendar and whose official practice material you have more of left.

A Grade 10 who has sat the PSAT already has half of this read: the PSAT score report is an SAT error log a year early.

Worked example: two logs, one decision

Our own numbers, not a real student's.

	SAT log	ACT log
Misses	9	13
Did not know	6 (nonlinear functions, systems of equations, transitions)	4 (two in trigonometric ratios, two in science data reading)
Slipped	2	2
Ran out of time	1	7, all in the last third of math and the last third of reading

The totals say SAT: 9 misses against 13. The logs say the opposite. Seven of the ACT's 13 misses sit in one column, time, and all seven are at the ends of two sections, which is the signature of a pacing problem and not a content one. A fortnight of pacing drills, timed single sections with the two-read skip rule and the stems read first, is predicted to remove most of those seven. The SAT's six content misses are six skills to rebuild, which is six rounds of worked example, cold redo and sibling, each spaced over a week or two.

So the decision is the ACT, with pacing as the first drill. The arithmetic to check: 7 of 13 is a little over half the ACT's misses in the trainable column; 6 of 9 is two thirds of the SAT's misses in the slow column.

The counter-case, in one sentence: if the ACT's four "did not know" rows had been in a skill the SAT does not test, the call would still hold, because those rows are only four; if they had been seven and the time rows four, the call would flip. The rule is the proportion in the trainable column, not the total.

Say it as a prediction, because it is one. Two weeks later the second ACT test is the measurement: if the time column has not moved, the prediction was wrong and the SAT log gets a second look.

After the decision: the plan

Go to the chosen test's page for its calendar: a diagnostic you have now taken, then one official full test every two to three weeks with drilling between, the last about a week out, two to four tests in all. Carry the error log forward; the skills from the deciding test are the first fortnight's drills. Drill them cold and mixed: Rohrer and Taylor (2007) found that students who practised problem types shuffled did worse during practice and better on a later test than students who practised them in blocks, and both tests shuffle. Each full test under the real conditions is also retrieval, which Roediger and Karpicke (2006) found produced better recall a week later than rereading did. The last seven days are chapter 9 of the free playbook, laid out as a calendar; the test-anxiety guide has the evidence that practice under the real conditions is what steadies the nerves.

The decision, on one screen

• Format: the SAT digital and adaptive; the ACT on paper or on screen, shorter since 2025.
• Pace: the SAT about 71 and 95 seconds a question; the ACT 42, about 67 and 60.
• Scoring: 400–1600 against 1–36, related by the test-makers' joint table; no penalty on either.
• Calculator: the whole SAT Math section; the ACT math section only.
• Reading: short passages with one question on the SAT; long passages with several on the ACT.
• Science: none as a section on the SAT; optional data reading on the ACT, outside the composite.
• The method: one official practice test of each, a week apart; compare the scores once, then the logs.
• The rule: the test whose misses sit in the trainable column wins.