"CogAT" is searched by parents in the weeks before a school's gifted-programme screening and again when the score report arrives with four kinds of number on it. The test is straightforward once its structure and its scoring are laid out: three batteries, nine subtests, a level for each grade, and a chain of scores that ends in a percentile, a stanine and a profile code. This page gives all of that from Riverside's own materials, the SAS-to-percentile table from its norms guide, what districts usually treat as gifted, and how to prepare a child in the way that helps. The printable has the structure, the score table and the profile codes on two pages.
Download the CogAT sheet (PDF) (the levels by grade, the three batteries and nine subtests with what each asks, the scoring chain, the SAS to percentile and stanine table, the ability profile codes, the usual gifted cut-offs and how to prepare; two pages, free).
What it is
The Cognitive Abilities Test is published by Riverside Insights; the current versions are Form 7 (2011) and Form 8 (2017), built as equivalent forms so that a child can be retested on the other one. It measures, in Riverside's words, how well a child can "reason abstractly and identify patterns and relationships", as against an achievement test such as the Iowa Assessments, which measures how much of the curriculum has been learned. The two are often given together, and the gap between them is itself information: a child whose reasoning score is well above the achievement score is working below their ability.
Schools use the CogAT to identify children for gifted and talented programmes, to spot multilingual learners whose reasoning is strong though their English is not yet, and to adapt teaching to each child's pattern of strengths. Most districts test in grades 2 to 5; some test every child in one grade (universal screening), others only on referral.
The levels by grade
The level number is the typical age of the students who take it.
| Level | Grade | Format |
|---|---|---|
| 5/6 | kindergarten | pictures only; directions read aloud (English or Spanish); untimed in practice |
| 7 | grade 1 | pictures only; directions read aloud |
| 8 | grade 2 | pictures only (the optional sentence completion is read aloud) |
| 9 | grade 3 | words in the verbal battery from here on; timed sections |
| 10 | grade 4 | |
| 11 | grade 5 | |
| 12 | grade 6 | |
| 13/14 | grades 7 and 8 | |
| 15/16 | grades 9 and 10 | |
| 17/18 | grades 11 and 12 |
Levels 5/6 to 8 were rebuilt in Form 7 as a bilingual primary assessment: every subtest but one uses pictures only, with language in the directions, so that a child who cannot yet read, or reads in another language, can show reasoning. A school may give a child a level above or below their grade for a second opinion (out-of-level testing).
The three batteries and nine subtests
| Battery | Subtest | What the child does (primary levels in brackets) |
|---|---|---|
| Verbal (reasoning with words) | Verbal analogies (picture analogies) | big is to small as hot is to ?; at the primary levels the same with pictures |
| Sentence completion | choose the word (or picture) that best finishes a sentence read aloud or printed | |
| Verbal classification (picture classification) | three things that belong together; pick the fourth | |
| Quantitative (reasoning with numbers) | Number analogies | 2 goes to 4 as 5 goes to ?; at the primary levels, pictures of quantities |
| Number puzzles | make two sides balance: 4 = 2 + ? | |
| Number series | continue the pattern: 2, 4, 6, ? (beads on a string at the primary levels) | |
| Nonverbal (reasoning with shapes and space) | Figure matrices | a 2 by 2 grid of shapes with one missing; choose what completes the pattern |
| Paper folding | a paper is folded and punched; which unfolded sheet matches? | |
| Figure classification | three shapes that belong together; pick the fourth |
Each battery gives its own score, and the three combine into a composite (VQN). The Screening Form is one analogies subtest from each battery (picture or verbal analogies, number analogies, figure matrices) and gives a single score; districts use it for universal screening and then give the remaining six subtests, or the full test, to the children who screen in. Riverside's guidance is that a screening cut-off should be more generous than the full-test cut-off, because a short test is less precise.
The full test takes roughly two hours at the primary levels (given in sessions over several days) and about 90 minutes from Level 9, with each subtest timed; the school sets the schedule.
How it is scored
The chain, from Riverside's norms guide:
- Raw score: the number of items right in each subtest, added for each battery.
- Universal Scale Score (USS): the raw score converted to one scale that runs across all levels, so growth can be tracked from year to year.
- Standard Age Score (SAS): the USS compared with children of the same age (in one-month groups, from 4 years 11 months to 18 and over) in the national norming sample, on a scale with a mean of 100 and a standard deviation of 16, running from 50 to 160.
- Percentile rank and stanine: from the SAS (age norms) and from the USS by grade (grade norms). The percentile is the percentage of the comparison group scoring at or below; the stanine puts the same thing on a 1 to 9 scale.
- Local scores: the same comparisons against the district's own students, when the district orders them.
Riverside's groupings for the stanine: 1 to 3 below average, 4 to 6 average, 7 and 8 above average, 9 very high. About 4 percent of children are in stanine 9, 7 percent in stanine 8, 12 percent in stanine 7, and 20 percent in stanine 5.
SAS, percentile and stanine
From Table 37 of Riverside's 2017 norms guide (Forms 7 and 8), abridged:
| SAS | Percentile | Stanine | SAS | Percentile | Stanine | |
|---|---|---|---|---|---|---|
| 136 and above | 99 | 9 | 104 | 60 | 6 | |
| 132 | 98 | 9 | 100 | 50 | 5 | |
| 130 | 97 | 9 | 97 | 43 | 5 | |
| 128 | 96 | 9 | 94 | 35 | 4 | |
| 127 | 95 | 8 | 91 | 29 | 4 | |
| 125 | 94 | 8 | 88 | 23 | 3 | |
| 123 | 92 | 8 | 84 | 16 | 3 | |
| 120 | 89 | 8 | 80 | 11 | 2 | |
| 119 | 88 | 7 | 76 | 7 | 2 | |
| 116 | 84 | 7 | 72 | 4 | 1 | |
| 112 | 77 | 7 | 68 | 2 | 1 | |
| 108 | 69 | 6 | 64 and below | 1 | 1 |
So the 95th percentile is an SAS of about 127, the 97th about 130 and the 98th about 132, on any battery or the composite.
The ability profile
Every child who completes enough items gets an ability profile code, such as 5A, 7B (Q+) or 9C (V+ Nβ):
- The number is the median of the three age stanines: the child's overall reasoning level.
- The letter is the pattern across the three batteries: A, all three at about the same level (44 percent of children); B, one battery significantly above or below the other two (33 percent); C, a relative strength and a relative weakness, two batteries in contrast (12 percent); E, extreme differences, at least two batteries 24 or more SAS points apart (10 percent).
- The letters in brackets name the battery that stands out and the direction: V, Q or N with + or β.
Riverside's site (cogat.com) explains each profile with teaching suggestions; a 7B (Q+) child, for instance, reasons well overall and best with numbers, and a 6C (V+ Nβ) child reasons better with words than with shapes, which changes how a teacher presents new material.
What counts as gifted
There is no national cut-off; each district sets its own, usually in a board policy. The common patterns:
| Policy | Typical cut-off |
|---|---|
| Composite percentile | 95th, 96th or 97th and above (SAS about 127 to 130) |
| Any one battery | 97th or 98th percentile on verbal, quantitative or nonverbal (an SAS of 130 or 132), to catch children with one strong domain |
| Stanine | 8 or 9 |
| Combined with achievement | CogAT above a cut-off and an Iowa Assessments score above another; or a matrix of CogAT, achievement, teacher rating and work samples |
| Local norms | the top 5 or 10 percent of the district's or the school's own testers, which can be a different SAS from the national table |
Many districts use the nonverbal battery alone for English learners, and most retest a child who misses a cut-off narrowly, on the other form, rather than treating one morning's score as final.
How to prepare a child
- Use Riverside's free practice activities so the formats are familiar: analogies, series, matrices and paper folding are unusual tasks for a six-year-old, and a child who has seen them does better than one who has not, for reasons that have nothing to do with ability.
- Reasoning grows through play: pattern blocks, tangrams, sorting games, number talks ("what comes next, and why"), riddles, and reading aloud with questions about why.
- Sleep and a calm morning matter more than a last practice set at this age.
- Do not drill bought test banks for weeks: scores rise a little, the profile tells the school less, and a child placed in a programme on a drilled score can struggle in it. Riverside's own position is that teacher-led practice on the formats is valid and useful, and heavy coaching is not.
Practice: eight questions
- Which CogAT level does a grade 4 child take, and why that number?
- Name the three batteries and one subtest in each.
- An SAS of 100 is which percentile and stanine?
- An SAS of 128: percentile and stanine?
- A child's profile is 8B (N+). What does each part mean?
- What is the Screening Form, and why should its cut-off be more generous?
- Which score lets the school track growth across years?
- A district's cut-off is the 97th percentile on the composite. What SAS is that, about?
Answers: 1) Level 10; the number is the typical age. 2) Verbal (analogies, sentence completion, classification), quantitative (number analogies, puzzles, series), nonverbal (figure matrices, paper folding, figure classification). 3) 50th; stanine 5. 4) 96th; stanine 9. 5) Median stanine 8 (above average); one battery significantly different from the other two; the nonverbal battery is the strength. 6) Three analogies subtests, one per battery, for universal screening; it is shorter and less precise, so a stricter cut-off would miss children who would pass the full test. 7) The Universal Scale Score. 8) About 130.
Download the CogAT sheet (PDF)
Where to get help
A 1-on-1 online class for children from kindergarten to grade 8 in the US system, US after-school hours: reasoning built through maths and reading rather than test drills, the CogAT formats made familiar with Riverside's own practice activities, and, for a child already in a gifted programme, the acceleration the programme expects in mathematics and science. The recording and notes after every class, the same tutor every week, US$10 to $13 an hour. The subjects page lists the courses; the free 30-minute first class is a real lesson, usually on the work the child brings.
Questions parents ask
1What is the CogAT test?
2What is a good CogAT score?
3What are the CogAT levels?
4What is the CogAT ability profile?
5How do you prepare for the CogAT?
See how we teach this, 1-on-1 online β
Shobha
Founder and lead tutor, Science with Shobha
Teaching maths, science and English to children online since 2018. 500+ students, 20+ tutors, families in nine countries.
