About these results

Here’s how we calculated the score-improvement statistics on our homepage.

3.2× the average score improvement

Programs with at least 20 hours of tutoring showed an average SAT-equivalent score gain 3.2 times the growth predicted for comparable starting scores using College Board research.

6.3× more likely to reach 1500

Among programs with at least 20 hours of tutoring and starting scores below 1400, students reached a 1500 SAT-equivalent score at 6.3 times the rate predicted by our model using College Board growth estimates.

These outcomes include SAT scores and ACT scores converted to SAT equivalents.

Why we measure results this way

Two of the most common ways tutoring companies report results are average score increases and the percentage of students who reach a particular score. We think both have a serious flaw: they can tell you as much about the students a company enrolls as the progress those students make.

A student starting at 1000 has more room to gain points than one starting at 1450. And a student starting at 1450 is much more likely to reach 1500 than one starting at 1200. Those differences can make comparisons between tutoring companies misleading.

We prefer to compare our students’ results with the growth predicted for similar starting scores. It accounts for an important difference that a simple average gain or high-score percentage leaves out.

The comparison behind the numbers

The benchmark comes from College Board’s Student-Level Growth Estimates for the SAT Suite of Assessments, which follows students from the fall of 11th grade to the fall of 12th grade. We estimated section scores by dividing starting totals equally between reading/writing and math. The 1500 calculation uses a simulation with an assumed correlation of 0.7 between section outcomes.

What to keep in mind

These are historical, reported program outcomes, not a guarantee of future scores. The analysis includes repeat students and both SAT and converted ACT results. Ending scores are the highest reported official scores during tutoring and may include superscores.

The benchmark describes observed growth in a national research sample; it is not a control group of students who received no tutoring. Program lengths vary, and this comparison does not isolate the effect of tutoring.