SAT Preparation

How Much You Can Really Improve an SAT Score

Tens of points from a retake, not hundreds — and lower scorers gain about four times as much as higher ones. What the evidence says, and how it gets misquoted.

Short answer

How much can you improve your SAT score?

Expect tens of points from a retake rather than hundreds. The best causal evidence puts the average second-sitting gain at about 0.15 standard deviations, with lower scorers gaining roughly four times as much as higher scorers.

The short answer
Realistically, a retake with preparation moves a score by tens of points, not hundreds. The strongest causal study puts the average gain on a second sitting at about 0.15 standard deviations, with students in the lower half of the distribution gaining roughly four times as much as those near the top.

What changes the answer

A student whose first score came without preparation has far more room than one whose first score was the product of six months of work.
Gains at the bottom of the range come from unlearned content; gains near the top come from eliminating careless error, which is slower.
Where a university superscores, the effective improvement across two sittings is close to double the improvement in a single total.
The first retake is worth roughly twice as much as any subsequent one, so a third and fourth sitting rarely repay the time.

The two pieces of evidence, and what each can prove

Almost every claim about SAT improvement traces back to one of two studies, and they are not the same kind of thing. Keeping them apart is the difference between a realistic plan and a disappointed one.

The first is causal. Economists Joshua Goodman, Oded Gurantz and Jonathan Smith exploited the fact that students who score just below a round number are much more likely to retake, using that discontinuity to isolate the effect of retaking itself. Their finding: retake scores average 0.15 standard deviations higher than first-take scores. Because the comparison is quasi-experimental, this is a genuine estimate of what retaking does.

The second is associational. In 2017 the College Board and Khan Academy reported that 20 hours of practice on Official SAT Practice was associated with an average gain of 115 points between the PSAT/NMSQT and the SAT, and 6 to 8 hours with about 90 points. That is a real finding about a real dataset, and it is not a causal estimate: students who choose to do 20 hours of practice differ from those who do not, and the interval includes a year of ordinary schooling and growth. It also predates the digital SAT entirely.

The number everyone quotes is on a scale that no longer exists

You will see it written that retaking the SAT is worth 46 points on average. That figure is real, it comes from the study above, and it is almost always reproduced without the sentence that follows it: 46 points on a 2400 point scale.

The data covered a period when the SAT had three sections and a 2400-point maximum. The test has had two sections and a 1600-point maximum since 2016. Quoting the 46 directly against a modern score inflates it by roughly half, which is why students plan for a jump they then do not get.

The figure that survives the change of scale is the one in standard deviations: 0.15. Applied to the class of 2025, where the standard deviation on the total was 235, that implies an average second-sitting gain in the mid-30s rather than the mid-40s. We would rather give a student that number and be believed than the larger one and be wrong.

Low scores move more, and it is not close

The same study breaks the effect down by starting point, and the difference is the most useful thing in it. Students in the lower half of the distribution improved their second-take total by about four times as much as students in the upper half — 91 points against 22 on the old scale.

The mechanism is unsurprising once stated. A student at the bottom of the range has whole topics missing, and each one recovered turns a category of questions from wrong to right. A student near the top is already answering most questions correctly; what remains is the hardest material and their own residual slips, and there is simply less room above them.

Superscoring changes the picture in the other direction, and favours everyone. The same paper found that retakers' superscores rose about twice as much as their total scores, because a student rarely has a poor day in both sections simultaneously. If the universities on your list superscore, that is the number to plan against.

What we tell a student planning a retake

Sit it again. Only about half of SAT takers ever do, which means a great many students accept a first score that was never representative, and the evidence that retaking helps is about as solid as evidence in this field gets.

Then be specific about what changes between the sittings. A retake with no intervening work produces a small gain from familiarity alone. A retake preceded by eight to ten weeks aimed at one identified weak section — chosen by looking at the section scores, not the total — is where the meaningful movement comes from. Work the weaker section, because that is where the unlearned content is and therefore where the points are cheapest.

And plan to stop. Since the first retake is worth roughly twice a subsequent one, a third and fourth sitting are usually hours taken from essays and applications for a handful of points. Our SAT preparation runs in small groups on Zoom from PKR 6,000 / USD 25 a month, and we will say when a student has reached the point where more testing is not the constraint.

Average score gain from a second SAT sitting, in standard deviations: 0.15 SDAverage score gain from a second SAT sitting, in standard deviations[1]
How much more lower-half scorers gain on a retake than upper-half scorers: 4xHow much more lower-half scorers gain on a retake than upper-half scorers[1]

Common questions

Can you improve your SAT score by 200 points?

It happens, most often for students who started well below the mean and had unlearned content to recover, and it takes months rather than weeks. It is far above the average gain from a retake. Planning on 200 points as an expectation rather than a stretch is how students end up disappointed by a normal result.

How many times should you take the SAT?

Two, in most cases, with real preparation in between. The evidence shows the first retake is worth about twice as much as any that follow, so a third sitting has to justify itself with a specific reason — a scholarship threshold just out of reach, or a section score that was clearly unrepresentative.

Does SAT practice actually work?

Yes, though the headline figures overstate it. The widely quoted 115-point gain from 20 hours of practice is an association, not a causal estimate, and it measures change from the PSAT to the SAT — an interval that includes a year of school. Structured practice helps; it does not do what that number implies.

How long before the test should I start preparing?

Eight to ten weeks of consistent work for a targeted improvement in one section; considerably longer if the starting point is well below the mean and content is genuinely missing. Short bursts in the fortnight before a sitting mainly buy familiarity with the format, which is worth something but not much.

Is it worth retaking if my score is already high?

Usually only for a specific reason. Gains shrink sharply near the top of the scale — around a quarter of what lower-scoring students achieve — because there is little content left to gain and few questions left to get right. A scholarship threshold or a badly uneven split are the cases that justify it.

Sources

Take Two! SAT Retaking and College Enrollment Gaps (Working Paper 24945) — National Bureau of Economic Research
New Data Links 20 Hours of Personalized Official SAT Practice on Khan Academy to 115-Point Average Score Gains — College Board
2025 Total Group SAT Suite of Assessments Annual Report — College Board

Last updated August 21, 2026