There is no required or universally optimal number of LSAT attempts. LSAC's data show substantial repeat testing, but the answer depends on what is counted: test administrations, distinct test takers in a year, or a person's first, second or later score. Those are different measures. A population statistic cannot tell you that you personally need two attempts.
By Ali Unar, Founder, Lovare Institut. Updated November 5, 2026. This page examines LSAC Technical Report TR 26-01, especially Part 2. The arithmetic is disclosed below. It is not an admissions forecast or a claim about Lovare clients.
Table 1B, printed page 39, reports 172,819 tests administered and 123,973 test takers for 2024–25 within that analysis. Repeat takers are counted once in the test-taker total for the year, but each administration is counted in the test total. The report's methods exclude specified records, so these are not interchangeable with every other total in the report.
| Quantity | 2024–25 value | What it means |
|---|---|---|
| Tests administered, Table 1B | 172,819 | Administration events in the analysis. |
| Distinct test takers for the year, Table 1B | 123,973 | People counted once within that testing year. |
| Tests divided by test takers | 172,819 ÷ 123,973 ≈ 1.394 | About 1.39 administrations per person within the year, not their lifetime attempt average. |
| Tests beyond one per person | 172,819 − 123,973 = 48,846 | Additional administrations, not necessarily 48,846 different repeaters. |
For a simple hypothetical illustration, ten additional administrations could belong to ten people taking one extra test each, or five people taking two extra tests each. The same excess test count does not identify the number of people who repeated. That is why the last row cannot be converted directly into a percentage of people who retook.
The report also presents first-, second-, third- and later-time categories in Tables 2B and 3B. Its summary describes 49.1% as testing more than once. That should not be casually relabeled “49.1% of applicants took exactly two tests,” or “50.9% only ever test once.” Applicants and test takers are different populations, and a testing-year classification does not establish a lifetime history.
There is a further denominator caution: the 2024–25 category counts in Table 2B sum to 167,951, whereas Table 1B lists 123,973 distinct test takers. The category percentages therefore should not be presented as a mutually exclusive partition of that headcount without reconciling the underlying counting method. Our previous once-only/twice/three-times presentation did not make these distinctions clear enough. We have removed it rather than using it to prescribe an attempt plan.
Table 5B reports mean score changes of 2.39 points from first to second score and 2.03 from second to third for 2024–25. Figures 6B and 7B show smoothed gain distributions, including decreases. The report is descriptive: it does not establish the causal benefit of booking another test, a success probability for your plan, or the effect of repeating the same preparation.
Use these findings to reject a guaranteed-gain story, not to add an average to your score. The retake evidence worksheet focuses on your fresh practice, specific changes, capacity and costs instead.
If your concern is permission to test again, read the current limits and counting checklist. If it is the effect of waiting for another score, use the school-specific application worksheet. Neither policy question is answered by how often other people test.
Do not infer that a flat score looks unserious or that three attempts are harmless everywhere. LSAC says schools receive reportable scores and evaluate them differently. A sound next step follows from your circumstances and the applicable rules, not a supposed normal number of tries.
September 20, 2026
September 20, 2026