Method · Reviewed 2026-09-23
How to calculate GAI
Official GAI is not FSIQ with a few points added. It is a composite from a named set of reasoning subtests. The publisher converts the sum of scaled scores with a table. This page explains that method and the only calculation we will run for you. We do not reprint the table.
If GAI is already on the report
Stop. You do not calculate it again. The examiner, or the scoring software, already converted raw scores to scaled scores and scaled scores to the composite. Your job is to read two standard scores, both on the 40–160 scale this calculator accepts, and subtract. Gap = GAI − FSIQ. That is the entire GAI vs IQ calculator.
Recalculating from subtests “to be sure” usually introduces a second, unofficial number. If the printed GAI and your hand total disagree, trust the report or ask the examiner. A website that offers a different total is not a correction of Pearson’s software.
How the publisher calculates GAI
The steps inside a Wechsler are ordinary, and the tables are not ours to copy. Each subtest starts as raw items correct. Raw scores become scaled scores, typically a mean of 10 and a standard deviation of 3, using age norms. GAI uses a specified subset of those scaled scores. You sum that subset. You look the sum up in the manual’s GAI conversion table. The result is a standard score with a mean of 100 and a standard deviation of 15, often with a percentile and a confidence interval beside it.
Which subtests enter the sum depends on the edition. On the WAIS-IV, GAI is built from the Verbal Comprehension and Perceptual Reasoning subtests (Similarities, Vocabulary, Information, Block Design, Matrix Reasoning, Visual Puzzles). On the WISC-V, GAI is Similarities, Vocabulary, Block Design, Matrix Reasoning, and Figure Weights. Visual Puzzles is in the Visual Spatial Index and not in that GAI. WAIS-5 keeps five primary indexes and still drops working memory and processing speed from GAI. Confirm the names on your protocol. The lists are summarized on GAI on WAIS and WISC.
We cite those manuals on the method page. We do not host the conversion tables, the item books, or the norms.
Why you should not average VCI, FRI, and VSI
It is tempting. GAI is “reasoning without memory and speed,” and VCI, FRI, and VSI are the reasoning indexes on a five-index Wechsler. Their arithmetic mean is a rough description. It is not the General Ability Index printed in the manual. Official GAI is normed from the sum of subtest scaled scores. Index scores have already been normed once. Averaging them again is a different statistic. It will often land near the real GAI and sometimes will not, especially when the indexes are far apart or when a subtest inside an index was substituted.
This laboratory uses that mean on purpose for one product only: the nine-task estimate. There, GAI is the mean of the verbal, fluid, and visuospatial domain scores, and FSIQ is the mean of all five. We call those analogs. The method states the formulas in the open. Do not paste an official VCI, an official FRI, and an official VSI into that story and call the average a WAIS GAI.
When the report never computed GAI
Examiners do not always derive ancillary composites. If the score summary has FSIQ, the indexes, and no General Ability Index, you do not have a GAI yet. Options that stay honest: ask the examiner to derive it, or describe the profile with the indexes you do have. The index calculator will show the peak, the load, and which index sits below the full-scale number. It labels those results as index means, not as GAI or CPI.
Do not reconstruct scaled scores from percentiles with a formula you found on a forum. Rounding and age bands make that reverse lookup wrong often enough to invent a gap. If you only have percentiles, leave the calculation alone.
How the nine-task estimate calculates its GAI
Each task produces a raw total. A provisional conversion places a moderate raw score near 100, clamped so the scale stays in a standard-score range. Domain scores group the tasks. The GAI analog is the mean of VCI, FRI, and VSI. The FSIQ analog is the mean of VCI, FRI, VSI, WMI, and PSI. CPI is the mean of WMI and PSI. The gap is GAI minus FSIQ. None of those steps uses a Pearson table. The exact conversion sentence is on the method page so it can be audited. It is not a norm.
That is why we will not add a “subtest scaled score” form to the calculator. Collecting six numbers and pretending to own the lookup would look like a Wechsler and would be a worse instrument than the report you already paid a psychologist to score.
The only worked calculation we publish
Printed FSIQ 109. Printed GAI 116. Difference 116 − 109 = +7. You can do that on paper. The calculator does the same thing and then points at the reading bands. It does not know which subtests produced 116. If you need that story, it is in the protocol, edition by edition, not in the subtraction.
Common questions
Can you calculate GAI from subtest scaled scores?
The publisher can, with the manual’s table. This site will not. If GAI is already a standard score on the report, paste it into the calculator.
Is GAI the average of VCI and PRI?
On the WAIS-IV the ingredients are the VCI and PRI subtests, converted as a sum, not a casual average of the two index scores. Later editions use a different list.
What is the formula on the nine-task test?
GAI analog = mean(VCI, FRI, VSI). FSIQ analog = mean of those plus WMI and PSI. See the method page. It is not an official Wechsler formula.
Why won’t you publish the conversion table?
It is the publisher’s. Reprinting it would pretend this site scores a WAIS or WISC. We subtract composites you already have.
Where to go next
The GAIvsIQ laboratory keeps one URL for each job. This note does not replace the instrument it points to.
Ready when you are
GAI already printed?
Paste the two standard scores. We subtract. We do not look up a table.