Reading the LSAT scoring chart: what a 5-question swing actually does to your 120-180 score

The LSAT scaled score is the single number every law school admissions committee reads first, yet the mechanics that produce it are hidden from candidates. The test contains roughly 75 to 76 scored multiple-choice items, and every examinee walks out of the centre with a different raw tally. That raw tally is never shown. Instead, the test maker applies a statistical procedure called equating, and the result is a scaled score on the familiar 120 to 180 scale, reported to the nearest integer. For candidates planning a preparation strategy, understanding the raw-to-scaled conversion matters because it controls how much each missed question costs, where the safest guessing thresholds sit, and how to interpret a score band rather than a single number. The remainder of this article walks through the conversion logic, the typical preptest curves, the role of unscored experimental items, and the score report fields that admissions readers actually use.

The architecture of a single LSAT section

Before the conversion itself makes sense, the structure of the test has to be visible. A standard LSAT administration is built from five scored sections plus a sixth, ungraded experimental section used by the test maker for future form calibration. The two Logical Reasoning sections together contribute around 50 to 53 questions, Reading Comprehension contributes roughly 26 to 28, and one Logic Games section contributes about 22 to 24. The experimental section is one of those three content types, and the candidate never knows which one is unscored. That hidden item is the first thing a serious preparation strategy has to absorb, because any decision to skip a question in a particular section assumes a risk that the section might be the throwaway. The raw score, in other words, is computed from a known count of questions out of a known count, and the candidate cannot tell in the moment which section counts.

Within each section the item order is fixed and the time pressure is tight. Logical Reasoning gives roughly 35 minutes for two scored sections combined, Reading Comprehension gives 35 minutes, and Logic Games gives 35 minutes. The pace most successful candidates settle into is just under 1 minute 30 seconds per Logical Reasoning question, a touch over 1 minute 15 seconds for Reading Comprehension, and close to 1 minute 35 seconds for Logic Games. That pace, multiplied by the question counts above, produces the raw score that feeds the conversion. For preparation strategy purposes the practical lesson is that pacing errors and not content errors are responsible for the largest swings in raw count, and therefore in scaled score.

What equating actually does to a raw score

Equating is a statistical adjustment that lets scores from different test forms be compared on a single 120 to 180 scale. Each LSAT form has a slightly different difficulty profile, and a raw count of, say, 60 correct on one form might represent the same ability as a raw count of 58 on another. The equating step solves that problem by mapping each form's raw distribution onto a common scale, using a reference group of test-takers who answered items shared between the forms. The mapping is monotonic, meaning the conversion never reverses the order of two candidates within the same administration, and it is non-linear, meaning the cost of a missed question varies by difficulty band.

For most candidates reading this, the practical consequence of non-linearity is that a question missed in the middle of the score range costs more than a question missed at the extreme top or bottom. If you are sitting near a 150, missing the next hardest question you would have got can pull you down by 2 or 3 scaled points once it is converted, because that question sits in a dense part of the score distribution. If you are already at 175, an additional missed question might cost you a single point, or none at all, because the curve is flatter at the ceiling. The reverse is also true: a guessed question that you got right at the low end can lift you by more than one point if it happens to sit in a sparsely populated band. This is why the question type by question type triage plan matters as much as raw accuracy.

Typical preptest curves and the question counts behind a target score

Test prep services publish so-called preptest curves that estimate the raw count required for a given scaled score on a particular released form. Those curves move form by form, but the average pattern across a large set of administrations is informative. For a target of 170, the typical raw count across all five scored sections combined is in the range of 88 to 92 correct out of the roughly 100 to 101 scored items, which is the rough LSAT question count once the experimental section is removed. For 165 the count rises to roughly 80 to 85, for 160 to about 73 to 78, for 155 to roughly 65 to 70, and for 150 to about 58 to 63. The exact number shifts on each form, and the test maker deliberately does not publish the mapping for a live administration, so the table below is an illustrative band rather than a guarantee.

The score band concept on the official report is the test maker's acknowledgement that the conversion is not a single point. Because equating introduces a small measurement error, the report shows a score band that is typically 3 to 4 scaled points wide and that reflects the range within which the candidate's true ability score lies with a stated confidence level. Law school admissions committees are trained to read the band rather than the headline number, and that single habit is one of the most important scoring ideas to teach. A 162 with a band of 159 to 165 is read as 159 to 165, not as 162, and the same is true for an applicant at 168 with a 165 to 171 band.

How the unscored experimental section distorts your guess at the curve

The sixth section is the silent variable in every candidate's mental model of the conversion. If a candidate scores 25 of 27 on what turns out to be the experimental Reading Comprehension section and only 18 of 26 on the scored Reading Comprehension section, the score report still reflects just the scored section. The conversion is therefore based on a raw count that is one section smaller than the count the candidate was keeping in the testing room. In a preparation strategy this has a direct consequence: stop tracking per-section raw scores during a timed section, and stop assuming any particular section is the live one. Two of the three content types appear twice across administrations, and a candidate who has prepared only one of the dual-appearance types is gambling on which slot is graded.

For the test maker, the experimental section also drives the next form's equating. Items that survive statistical screening on the experimental section enter the item bank and eventually appear as scored items on a future form, at which point the shared-item equating anchor is established. This is why curves shift between released preptests: every form has a slightly different anchor set, and the shared-item statistics on those anchors drive the conversion that turns raw into scaled. A preparation strategy that ignores experimental sections, or that uses the experimental section to dump guessing, is wasting a hidden opportunity. A serious candidate treats all five content sections identically and lets the curve do its work after the fact.

Score report fields, and what admissions readers actually use

The official LSAT score report contains more columns than most candidates ever look at. The scaled score is the headline. The score band is the next column, typically spanning 3 to 4 points. A separate column shows the percentile rank of the scaled score against a recent reference group of test-takers, usually drawn from the most recent three testing years. Some score reports also contain an average score per section, which is the scaled score weighted equally across the scored sections even though individual sections are not reported as separate numbers. Law school admissions committees, when comparing candidates, weight the scaled score more heavily than the percentile, because the percentile is just a translation of the scaled score into a population rank and carries no additional information beyond the underlying number.

Reading the report accurately is part of score report analysis, and it is a skill in its own right. A common mistake is to read a band like 162 to 166 as a guarantee that the candidate's true score sits in that interval, and to argue with an admissions officer about whether the true score is 162 or 166. The band is a statistical statement, not a binary on or off decision, and admissions officers are trained to take the whole interval into account. Another common mistake is to focus on the percentile when two candidates have the same scaled score; the percentile is the same in that case by definition, and the candidate with a wider band is in fact read more cautiously, not more favourably. For preparation strategy, the report teaches one final lesson: target the scaled score, monitor the band as a tolerance, and stop chasing the percentile as if it were a different number.

Common pitfalls and how to avoid them in score interpretation

Three pitfalls catch most candidates, and each one distorts the preparation plan that follows the test. The first is treating a single preptest curve as if it were the official curve. Preptest curves are useful for trend tracking, but a given form can be 2 to 4 scaled points stiffer or easier than the average, and a single administration does not generalise. The fix is to take at least 8 to 10 timed preptests, average the results, and treat the average as your stable ability estimate rather than the most recent score. The second pitfall is over-weighting the experimental section. Because the candidate never knows which section is the throwaway, treating a strong section as a free pass is a mistake. The fix is uniform effort across all five content sections. The third pitfall is reading the band as a single number. Admissions officers read the whole band, and a candidate who is mentally anchored to the midpoint is leaving a 1 to 2 point swing unclaimed on a competitive application.

A fourth pitfall deserves a separate note, because it appears in the room during the test rather than at the score report. A candidate who skips a question and returns to it later will not recover the time used in the skip unless the second pass is faster than the first. The conversion is indifferent to the order in which questions were answered, but the timer is not, and a slow second pass costs the candidate raw points across the rest of the section. The tactical fix is to mark and move with a hard cap, and to come back only if a fresh reading of the stem unlocks the question type. In my experience this single habit separates a 158 from a 163 more often than any content review.

Building a preparation strategy around the conversion mechanics

Once the conversion is understood, the preparation plan has a clean target. The goal is no longer to maximise the raw count in the abstract; it is to maximise the raw count on the questions that the conversion weights most heavily. In practice that means three tactical moves. First, build endurance for the fifth scored section, because the experimental section is sandwiched into the administration and the back half of the test is where pacing errors cluster. Second, practice skipping on hard items, and train a return cap of 30 seconds, so that any question that does not yield to a first read is flagged and revisited only if time allows. Third, drill the question types that have the highest per-item yield on the curve, which for most candidates is the 5-star Logical Reasoning arguments and the comparative reading passages in Reading Comprehension.

For a candidate aiming at 170, the strategic reading of the conversion is that the score band is narrow enough that one bad section is enough to knock the headline number down. A consistent 90 percent accuracy on five sections is far safer than a 95 percent peak on four sections with an 80 percent trough on the fifth. The conversion does not average away the trough, and the band is read as a whole. For a candidate aiming at 160, the strategic reading is the opposite: the band is wide enough that the score report can absorb a single off day, and the priority is moving the floor up rather than chasing the ceiling. These two strategies are different in target and identical in shape, which is one of the most useful generalisations the conversion teaches.

What to do on test day once the conversion is no longer in your control

The conversion runs after the candidate leaves the room, and no in-room decision can change it. The remaining tactical question is therefore one of pacing and guessing. A first pass should answer every question that the stem unlocks in under 90 seconds, mark and move on anything that takes longer, and reserve a second pass of 3 to 5 minutes per section for marked items. A first pass should also leave 1 to 2 minutes at the end of each section for a final blind review, which catches the small but high-frequency error of misreading the question stem on Must Be True, Must Be False, and Main Point items. These are the question types where a stem misread costs a raw point and where the conversion has the least sympathy.

Guessing has a different arithmetic. A blank answer is always wrong, and a guessed answer is correct on the probability floor set by the answer choice count, which is one in four for standard multiple choice and one in five for the comparative reading item type in Reading Comprehension. For a candidate below the median, a first pass guess on the hardest 10 percent of items has a positive expected value on the conversion, because the items are sparsely populated in the score distribution and a lucky correct answer can shift the scaled score by more than one point. For a candidate above 165, the arithmetic is similar but the upside is smaller, and the priority shifts to not letting a guess consume time that would otherwise go to a question the candidate can actually solve. The conversion rewards candidates who know when to stop.

Putting the pieces together

The LSAT scaled score is the product of three layers: a raw count of correct answers on a known set of scored items, an equating step that maps that count to the 120 to 180 scale, and a confidence band that describes the range within which the true score lies. The preparation strategy is to maximise the raw count on the items the conversion weights most heavily, to treat the experimental section as live, and to read the score report as a band rather than a point. Admissions readers do the same, and the candidates who understand the conversion before they take the test spend the weeks afterward arguing with the right number.

A short, focused next step is to take a recent official preptest under timed conditions, score it strictly, and plot the result on the table above. The exercise turns the conversion from an abstract statistic into a personal data point and reveals whether the current pacing and question triage plan is consistent with the target score. TestPrep İstanbul's diagnostic scoring walkthrough is a natural starting point for candidates who want to translate a preptest result into a concrete revision plan.

Frequently asked questions