Does SAT practice raise scores? College Board's evidence, 2017 to 2025

Key findings

• One full-length Bluebook practice test was associated with 25.7 more points than similar students with no practice test, in College Board's 2025 study. See figure
• Three or more practice tests were associated with 61.4 more points, but College Board notes this group may have had more adult guidance and academic support. See figure
• The widely repeated claim that 20 hours on Khan Academy equals 115 points comes from a 2017 College Board press release with no published method or statistical controls. See figure
• When College Board and Khan Academy adjusted for PSAT/NMSQT scores and demographics in 2020, 6+ hours of practice was associated with just 21 more points than non-users. See figure
• No College Board study gives an hours-per-points rate, so any "X hours for 100 points" rule has no primary source behind it. See figure
• Independent studies of paid coaching on the pre-2016 SAT found effects of roughly 15 to 56 points combined, depending on the study. They used a different test, so they don't compare point for point with the 2025 Bluebook results. See figure

What the 2025 Bluebook study found

In May 2025, College Board researchers looked at 148,154 students in the class of 2025 who took the SAT as juniors during the March 2024 School Day administrations. They compared students who completed full-length practice tests in Bluebook with similar students who completed none, matching on prior PSAT score, gender, race and ethnicity, parental education, earlier SAT experience and how students handled digital test setup (College Board).

• 1 practice test: associated with 25.7 more points.
• 2 practice tests: associated with 45.5 more points.
• 3 or more practice tests: associated with 61.4 more points.

Each extra test was tied to a smaller step up. The second test added about 19.8 points over the first, and the step to 3 or more was smaller again. About 50% of the sample took no practice test at all.

The gaps were larger for students who started lower. For students with a PSAT score of 900, 3 or more practice tests were associated with 79.0 more points. The sample also scored above the national picture, with an average SAT of 1160.

"Estimates are not causal." (College Board, footnote 4)

That line matters. Students who choose to take three practice tests are often the ones with more support and motivation, and the researchers say their design cannot separate that out.

How College Board's claims changed: 2017, 2020, 2025

Why 115 and 21 are so far apart. They measure different things. The 2017 figure is an average score gain among students who practiced, which includes the growth most students make anyway between tests (College Board Newsroom). The 2020 figure is the extra gain over comparable students who did not practice, after adjusting for prior scores (College Board and Khan Academy). Only about 8% of the 2020 sample both practiced 6+ hours and used a best practice, and the largest group, 246,325 students, did no practice on the site.

College Board's help center still repeats the 2017 numbers, 90 points for 6 to 8 hours and 115 for 20 hours, without naming a study or year (College Board).

Is there an hours-per-100-points rule?

No. We found no primary source from College Board, Khan Academy or peer-reviewed research that gives a rate of study hours per 100 points. The only hour-based figures are the 2017 press release (raw averages) and the 2020 report, which tied 6+ hours to 21 more points after adjustment. Neither supports stretching the numbers into a straight-line rule.

Treat any "study X hours, gain 100 points" promise as unsupported by published research. For a realistic plan, see how to study for the SAT, and use the SAT score calculator to read your practice test results.

What research says about paid coaching

Independent studies of coaching were done on the SAT used before 2016, so they do not map directly to today's digital test. Most are observational, and all used the pre-2016 SAT.

• Briggs for NACAC, 2009: a review of existing studies put the typical coaching effect at about 10 to 20 points in math and 5 to 10 in critical reading, as reported by Inside Higher Ed. The original NACAC report is no longer online, so treat this as a secondhand account.
• Domingue and Briggs, 2009: using national ELS:2002 data, coaching was associated with up to about 15 points on math, and only the math effect was statistically significant (Multiple Linear Regression Viewpoints).
• Briggs, 2001: after controlling for group differences in the NELS:88 sample, the combined coaching effect was about 20 points (Chance).
• Montgomery and Lilly, systematic review: coached students gained 32.7 points on math and 23.5 on verbal over control groups, about 56.2 combined (International Journal of Social Welfare).

Those effects are in the same range as the free practice findings above. If you are weighing a tutor, read is SAT tutoring worth it and what SAT tutoring costs.

Limits of this evidence

• Correlation, not cause. College Board's 2025 report says its estimates are not causal, and the 2020 report calls itself an observational design that cannot control for self-selection (2025, 2020).
• Who chooses to practice. Motivated students practice more and may also get more help at home or school.
• Narrow samples. The 2025 study covers weekday School Day juniors only. The 2020 study covers only students who linked their accounts.
• Old scales. The coaching studies used the pre-2016 SAT, so their point values are not a direct match for the digital SAT.
• No published method for 2017. The 115-point figure came from a press release, not a technical report.

What to do with this

The most recent and best-controlled data are consistent with taking full-length Bluebook practice tests, and more than one, though they cannot prove the tests cause the gain. The biggest step in the 2025 study came with the first test, with smaller steps after. Free options are listed on free SAT tutoring, and the wider plan lives on our SAT prep hub.

All figures on this page

Each row has its own link (for example does-sat-practice-raise-scores.html#F1), so you can point to one exact figure.

Download this data (CSV)

Methodology and limitations

Every figure comes from the original document: College Board's May 2025 report on digital SAT practice in Bluebook, the 2017 College Board press release, College Board's help center page, the 2020 College Board and Khan Academy technical report and its executive summary, and the peer-reviewed or published coaching studies linked in the text. Each number was read from the source at least twice. The 2009 NACAC figure is the exception: the original report is no longer online, so it is cited as reported by Inside Higher Ed and marked medium confidence.

Three figures are computed. The share with no practice test in the 2025 study is students with no test divided by all 148,154 students. The extra points for a second test is the 2 test estimate minus the 1 test estimate. The 2020 share with no practice is 246,325 divided by 545,640. Because the 2025 estimates come from three separate matched models, the step between them is descriptive, not a modeled effect.

The three College Board studies measure different outcomes (raw average gains in 2017, adjusted gains over non-users in 2020 and 2025), so their headline numbers should not be compared directly. None of them shows that practice causes higher scores.

College Board publishes practice research irregularly, so this page will be rechecked whenever a new report appears and at least every October.

How to cite this page

Copy this line, or link to a single figure with its row link from the table above.

Does SAT practice raise scores? College Board's evidence, 2017 to 2025. TutoringForSat.com. Last updated October 8, 2026. https://tutoringforsat.com/does-sat-practice-raise-scores.html