Last Updated on 18 November 2026 by Editorial Team
Last updated: 17 November 2026 · Applies to the 2026–2027 school year · Reviewed against Pearson and district score documentation
The short answer. The OLSAT 8 reports a School Ability Index (SAI), an age-based standard score with an average of 100 and a standard deviation of 16. An SAI of 116 is about the 84th percentile, 124 is the 93rd, and 132 is the 98th. Your report will also carry percentile ranks and stanines calculated two different ways — by age and by grade — and confusing the two is the most frequent error parents make with this test.
On this page
What the OLSAT 8 is ·
Verbal and nonverbal structure ·
Levels and grades ·
How scoring works ·
Age norms versus grade norms ·
SAI to percentile chart ·
Percentile to SAI chart ·
Stanine bands ·
Why an SAI is not an IQ ·
The Stanford 10 pairing ·
The New York City context ·
Preparation ·
FAQ ·
Sources
What the OLSAT 8 is
The Otis-Lennon School Ability Test is a group-administered, multiple-choice measure of abstract thinking and reasoning published by Pearson, now in its eighth edition. It covers kindergarten through grade 12 and is used mainly for two purposes: identifying children for gifted and talented programmes, and providing an ability measure to set alongside an achievement test so that schools can see whether a child’s attainment matches their reasoning.
The description Pearson uses — a test of a student’s ability to cope with school learning tasks — is worth taking literally. The OLSAT is not a nonverbal or culture-reduced instrument. It deliberately includes verbal reasoning, verbal comprehension and, at the upper levels, quantitative reasoning, because those are the abilities school work actually demands. That makes it a better predictor of classroom performance than a purely nonverbal test, and it also makes it more sensitive to language background and to preparation. Both of those properties follow from the same design decision.
The verbal and nonverbal structure
Every OLSAT 8 level produces a verbal score and a nonverbal score, which combine into a total from which the School Ability Index is derived. Across the levels there are 21 distinct question types. Among the things a child is asked to do: establish sequences, solve arithmetic problems, and complete analogies.
Broadly, the two halves break down as follows:
Verbal
- Verbal comprehension — following directions, understanding relationships between words, antonyms, sentence completion and arrangement. At the lower levels these are read aloud by the administrator.
- Verbal reasoning — logical selection, analogies, classification, inference and working out what must follow from given information.
Nonverbal
- Figural reasoning — patterns, series and matrices built from shapes.
- Pictorial reasoning — at the lower levels, classification and series using pictures rather than abstract figures.
- Quantitative reasoning — at the upper levels, number series, number matrices and numeric inference.
Ask for the split, not just the total. The verbal and nonverbal scores are reported separately, and the gap between them is often the most informative thing on the report. A large verbal-nonverbal discrepancy in either direction is worth discussing with the school: a much stronger nonverbal score in a child who is learning English usually says more about language exposure than about reasoning, while a much stronger verbal score can indicate a child who has had rich language input and may still need support with abstract and spatial material.
Levels and grades
The OLSAT 8 is published in several levels, labelled A through G, matched to grade bands rather than to individual grades in most cases. Level E is used in grade 4 and Level F in grade 5, as examples of the mapping. Your child’s school selects the level, and the complete level-to-grade table is in the examiner’s manual — the testing coordinator can tell you which level your child sat, and it is worth knowing, because the question types differ between levels.
At the lower levels the test is administered orally: the teacher reads each question aloud and the children mark their answers. From the middle levels children read for themselves. That transition matters when you compare a child’s score across years, because a child who scored well when questions were read aloud and less well once they had to read independently may have a reading issue rather than a reasoning one.
How OLSAT 8 scoring works
There are three stages between your child’s answer sheet and the headline number.
- Raw score. The number of correct answers in each section. Wrong answers are not penalised relative to blanks, so there is no reason to leave a question unanswered.
- School Ability Index (SAI). The raw score is compared with the scores of other children of the same age and converted to a normalised standard score. The SAI is age-based, with a mean of 100 and a standard deviation of 16. When the OLSAT is administered together with the Stanford 10, the SAI is derived from the verbal, nonverbal and total scores.
- Percentile ranks and stanines. These are reported both by age and by grade, along with normal curve equivalents.
Note the order carefully: the SAI is derived from an age comparison, and the percentile ranks are then derived from the scores. A percentile rank on this report is not a separate measurement — it is a restatement of the same performance on a different scale.
Age norms versus grade norms: the error that changes the answer
This is the most consequential thing on the page, and almost no parent is told about it.
The OLSAT 8 reports percentile ranks and stanines calculated two ways. Age percentile rank compares your child against other children of the same age. Grade percentile rank compares your child against other children in the same grade. For a child whose age is typical for their grade, the two are close. For a child who is young or old for their grade, they can differ substantially — and the difference runs in opposite directions.
| Situation | Age percentile | Grade percentile | Why |
|---|---|---|---|
| Child is young for the grade (autumn birthday, started early) | Higher | Lower | Against same-age children they look strong; against older classmates they look weaker. |
| Child is old for the grade (held back, or a summer birthday in a late-cutoff district) | Lower | Higher | Against same-age children they have no age advantage; against classmates they are more developmentally mature. |
| Child’s age is typical for the grade | Similar | Similar | The two reference groups largely overlap. |
Which one does your district use for its gifted threshold? That is a question you should ask explicitly, and ask for the answer in writing. For a child who is young for their grade, a threshold applied to the grade percentile instead of the age percentile can be the difference between qualifying and not — with no difference whatsoever in the child’s performance.
Two questions for your school. First: is the threshold stated as an SAI, an age percentile or a grade percentile? Second: does the district use national norms or local norms — comparing your child against the national sample, or against other children in your district? Those two answers determine the bar your child has to clear, and they vary from district to district. The score itself is only half the story.
SAI to percentile chart
The conversions below are calculated directly from the published SAI scale — a normalised distribution with a mean of 100 and a standard deviation of 16 — rather than copied from another source.
| School Ability Index (SAI) | Percentile rank | Standard deviations from average | Stanine | Descriptive band |
|---|---|---|---|---|
| 150 | 99.9 | +3.13 | 9 | Very superior |
| 145 | 99.8 | +2.81 | 9 | Very superior |
| 140 | 99.4 | +2.50 | 9 | Very superior |
| 136 | 98.8 | +2.25 | 9 | Very superior |
| 132 | 97.7 | +2.00 | 9 | Superior |
| 128 | 96.0 | +1.75 | 9 | Superior |
| 124 | 93.3 | +1.50 | 8 | Above average |
| 120 | 89.4 | +1.25 | 8 | Above average |
| 116 | 84.1 | +1.00 | 7 | Above average |
| 112 | 77.3 | +0.75 | 7 | High average |
| 108 | 69.1 | +0.50 | 6 | High average |
| 104 | 59.9 | +0.25 | 6 | Average |
| 100 | 50.0 | 0.00 | 5 | Average |
| 96 | 40.1 | −0.25 | 5 | Average |
| 92 | 30.9 | −0.50 | 4 | Low average |
| 88 | 22.7 | −0.75 | 4 | Low average |
| 84 | 15.9 | −1.00 | 3 | Below average |
| 80 | 10.6 | −1.25 | 2 | Below average |
| 75 | 5.9 | −1.56 | 2 | Well below average |
| 70 | 3.0 | −1.88 | 1 | Well below average |
| 65 | 1.4 | −2.19 | 1 | Well below average |
What those percentiles look like
Percentile to SAI chart
| Percentile rank | Equivalent SAI | Percentile rank | Equivalent SAI |
|---|---|---|---|
| 99 | 137 | 50 | 100 |
| 98 | 133 | 40 | 96 |
| 97 | 130 | 30 | 92 |
| 96 | 128 | 25 | 89 |
| 95 | 126 | 20 | 87 |
| 94 | 125 | 15 | 83 |
| 93 | 124 | 10 | 80 |
| 92 | 122 | 5 | 74 |
| 90 | 121 | 3 | 70 |
| 85 | 117 | 2 | 67 |
| 80 | 113 | 1 | 63 |
| 75 | 111 | — | — |
Stanine bands
A stanine is a standard score with a mean of 5 and a standard deviation of 2, reported as a single digit from 1 to 9. The conventional reading is that stanines 1, 2 and 3 are below average, 4, 5 and 6 are average, and 7, 8 and 9 are above average. Stanines are most useful for comparing a child’s profile across several subtests, where the wide bands stop you over-interpreting small differences.
| Stanine | Percentile range | Approximate SAI range | Share of children | Conventional label |
|---|---|---|---|---|
| 9 | 96–99 | 128 and above | 4% | Above average |
| 8 | 90–95 | 120–127 | 7% | Above average |
| 7 | 78–89 | 112–119 | 12% | Above average |
| 6 | 61–77 | 104–111 | 17% | Average |
| 5 | 41–60 | 96–103 | 20% | Average |
| 4 | 24–40 | 88–95 | 17% | Average |
| 3 | 12–23 | 80–87 | 12% | Below average |
| 2 | 5–11 | 72–79 | 7% | Below average |
| 1 | 1–4 | Below 72 | 4% | Below average |
Why an SAI is not an IQ, even though it looks like one
An SAI of 132 looks exactly like an IQ of 132. It is not the same thing, for two separate reasons.
The first is the scale. The SAI uses a standard deviation of 16. Most IQ scales, including the Wechsler family, use 15. The same number therefore sits at a different point in the distribution depending on which scale produced it. An SAI of 132 is two standard deviations above the mean, at the 97.7th percentile. A Wechsler IQ of 132 is 2.13 standard deviations above the mean, at the 98.4th percentile. The gap widens the further from average you go.
| SAI (SD 16) | Percentile | Wechsler-style IQ at the same percentile (SD 15) |
|---|---|---|
| 112 | 77.3 | 111 |
| 116 | 84.1 | 115 |
| 120 | 89.4 | 119 |
| 124 | 93.3 | 122 |
| 128 | 96.0 | 126 |
| 132 | 97.7 | 130 |
| 136 | 98.8 | 134 |
| 140 | 99.4 | 137 |
The second reason is what the test measures. The OLSAT is a group-administered school ability test taken in a classroom in under an hour. An individually administered IQ assessment is delivered one-to-one by a trained psychologist over a much longer session, with observation of how the child works, and it samples a wider range of abilities. The OLSAT is a screening instrument. It is appropriate for sorting a large group efficiently; it is not a substitute for individual assessment, and a district that makes a final placement decision on an OLSAT score alone is using it beyond its design.
The practical rule: when comparing scores from different tests, convert everything to percentiles first. Percentile rank is the common currency. Standard scores are not comparable across instruments with different spreads.
The Stanford 10 pairing
The OLSAT 8 is frequently administered alongside the Stanford Achievement Test Series, Tenth Edition — the Stanford 10 — and many score reports cover both. When the two are given together, the OLSAT’s School Ability Index is derived from the verbal, nonverbal and total scores earned in that combined administration.
The reason schools pair them is that the combination answers a question neither can answer alone. The Stanford 10 measures what a child has learned; the OLSAT measures reasoning ability. Put together, they show whether attainment is in line with ability.
- Achievement in line with ability. The ordinary case. The child is performing about as expected.
- Achievement below ability. The pattern worth investigating. Possible explanations include a specific learning difficulty, attendance, instruction that has not been matched to the child, motivation, or something going on outside school. This is the combination that most often leads to a referral.
- Achievement above ability. Usually means the ability estimate is an underestimate — a bad test day, a language barrier on the verbal section, or simply measurement error. Rarely a cause for concern in itself.
If your child’s report covers both tests, read them against each other rather than separately. The comparison is the point of giving both.
The New York City context
The OLSAT’s best-known use has been in New York City’s gifted and talented admissions, and that history is instructive about the test’s properties rather than just about one city.
Over the years the city’s education department combined the OLSAT with other instruments — including the Bracken School Readiness Assessment and the Naglieri Nonverbal Ability Test — and adjusted the weighting between them, reducing the weight given to the OLSAT. The stated reasoning was that the OLSAT is more susceptible to preparation than the NNAT is.
That assessment is worth taking seriously whatever district you are in, because it reflects something real about the instrument. A test with verbal comprehension, vocabulary and verbal reasoning content is a test where a child’s exposure to language, and to practice material, affects the score. That cuts two ways: it is a legitimate criticism of the OLSAT as a pure ability measure, and it is also why the OLSAT predicts classroom performance reasonably well, since classroom performance depends on the same things.
Preparation: an honest account
Unlike nonverbal matrix tests, the OLSAT does respond to preparation. That is a documented view of the authority with the most experience of it. Being honest about that does not mean recommending heavy coaching.
Genuinely useful
- Format familiarity. With 21 question types across the levels, a child who has never seen the formats will spend thinking time working out what is being asked. A short session with examples of each type removes that cost.
- Listening practice at the lower levels. If questions will be read aloud, practise following spoken multi-step directions. This is a real skill and a real source of lost marks for young children.
- Answer sheet mechanics. Practise with a separate answer sheet if one will be used.
- Vocabulary through reading. The verbal sections draw on vocabulary and verbal relationships. Reading widely over months and years is the preparation that actually works, and it is useful regardless of any test.
- A “move on” rule. No penalty for wrong answers means the right strategy is to eliminate, choose and move on rather than stall.
Not useful, or actively harmful
- Intensive short-term drilling. Gains tend to be narrow and short-lived, and an over-coached, anxious child often scores worse than a rested one.
- Treating practice percentiles as predictions. Commercial practice tests are not normed on a representative national sample, so a practice percentile is not comparable to an OLSAT percentile.
- Preparation that lifts the score past the child’s level. This is worth saying plainly. If coaching gets a child into a programme pitched well above where they can work comfortably, you have not done them a favour. The point of identification is placement in the right setting, not winning a place.
Frequently asked questions
What is a good OLSAT score?
Above an SAI of 100 is above average by definition. For gifted programme purposes, thresholds are set locally and commonly fall between the 90th and 98th percentiles, which is roughly SAI 120 to 133. There is no national qualifying score, and any site that publishes one is inventing it. Your district’s gifted identification policy is the only authoritative source.
What is the highest possible SAI?
Secondary sources commonly cite a practical maximum of around 150. The reported maximum depends on the level sat and the child’s age band, because the SAI is age-normed and each level has a ceiling. A child who reaches the top of their level has shown that the level could not measure them, not that their ability stops there. If that happens, ask whether a higher level or an individual assessment would be more informative.
My child’s age percentile and grade percentile are different. Which is right?
Both are correct; they answer different questions. The age percentile compares your child with children the same age, the grade percentile with children in the same grade. Neither is more valid. What matters is which one your district applies to its threshold, so ask.
Can my child retake the OLSAT?
Policy varies by district, and many allow testing only once per cycle. Where a retest is permitted there is usually a waiting period and a different form. Ask about the appeals process as well as the retest policy — some districts will consider a second measure, such as an individual assessment, instead of a retest.
Does a wrong answer count against my child?
No. The raw score is the number of correct answers, so a wrong answer and a blank score the same. Children should answer every question.
Is the OLSAT fair to children learning English?
Less so than a nonverbal test, by design — the verbal sections require English comprehension and vocabulary. This is exactly why districts with large numbers of multilingual children often pair it with, or substitute, a nonverbal instrument. If English is not your child’s first language, ask whether a nonverbal measure is available and how the district accounts for language background in its identification process.
Can I see the questions afterwards?
No. Test security means the booklets are not released. You are entitled to the score report and to an explanation of it, and you should ask for the full report — verbal and nonverbal scores, SAI, both percentile ranks, and stanines — rather than a single summary figure.
Where our numbers come from
All SAI-to-percentile conversions on this page are calculated from the published SAI scale, a normalised distribution with a mean of 100 and a standard deviation of 16. Scale definitions, score types and administration details are drawn from the publisher and district documentation below. These are technical and policy documents rather than journal articles and carry no DOIs; the publishing body is named so that each can be located.
- Pearson Assessments. Otis-Lennon School Ability Test, Eighth Edition (OLSAT 8) [product and technical documentation]. San Antonio, TX: Pearson Education, Inc. Source for the description of the test as a measure of abstract thinking and reasoning ability for kindergarten through grade 12, the verbal and nonverbal structure, and the derivation of the School Ability Index from verbal and nonverbal scores.
- A Beka Book. Student report: Introduction and score definitions [score report documentation]. Pensacola, FL. Source for the School Ability Index definition as an age-based, normalised standard score with a mean of 100 and a standard deviation of 16; for the statement that the SAI is derived from verbal, nonverbal and total scores when the OLSAT is administered with the Stanford 10; and for the stanine definition (mean 5, standard deviation 2; stanines 1–3 below average, 4–6 average, 7–9 above average).
- A Beka Book. Group report: Introduction and score definitions [score report documentation]. Pensacola, FL. Source for the score abbreviations used on combined OLSAT 8 and Stanford 10 reports, including normal curve equivalent and School Ability Index, and for the note that local norms may be used as a basis.
- Psychometric Success. OLSAT 8 Levels E and F [test information]. Source for the three-stage scoring sequence (raw score, School Ability Index, percentile) and for the level-to-grade examples of Level E at grade 4 and Level F at grade 5.
- TestPrep-Online. What is the OLSAT test? OLSAT 8 format and levels [test information]. Source for the count of 21 distinct question types across the OLSAT levels and for examples of the tasks involved, including establishing sequences, solving arithmetic problems and completing analogies.
- Pearson Education, Inc. Otis-Lennon School Ability Test — publisher background, including the test’s lineage through Harcourt Assessment. Source for the publisher history and the reporting of percentile ranks and stanines by age and by grade.
- New York City Department of Education, as reported in contemporaneous coverage of the city’s gifted and talented admissions process. Source for the combination of the OLSAT with the Bracken School Readiness Assessment and the Naglieri Nonverbal Ability Test, the subsequent change in weighting that lowered the OLSAT’s share, and the department’s position that the OLSAT is more susceptible to preparation than the NNAT.
A note on what we do not publish. You will find OLSAT “qualifying score charts by grade” online. The SAI is normed by age, and gifted thresholds are set by individual districts rather than nationally, so a grade-by-grade national qualifying table cannot be derived from the published norms — those charts are guesses presented as data. We publish the conversions that follow from the actual score scale and tell you which two questions to put to your district. If you spot an error here, write to us at the address below and we will correct it.
Found an error on this page? If a number here does not match what your school has given you, or you think something is wrong, write to the editor at Editorial@schooltest.org. We answer correction requests and we date every page so you can see when it was last checked.
