Full-Scale IQ vs. Index Scores: What's the Real Difference

A psychoeducational or clinical report hands back a single headline number and calls it the IQ. Underneath that number sit several others from the same testing session, each measuring a narrower slice of reasoning, and in a meaningful share of reports those narrower scores do not agree with each other. When the disagreement is large enough, the single number at the top of the page is the least informative line in the entire document.

What the Full-Scale score actually is

A Full-Scale IQ is not a seventh, independently measured quantity. It is a composite built from the index scores the same battery produces, weighted and combined by a published formula, then placed on the familiar mean-100, standard-deviation-15 scale. Like any score on that scale, it carries a standard error of measurement and should be read as a range rather than a point — the arithmetic behind that is covered in our piece on why an IQ score is a band, not a number, and the same logic applies one level down, to every index discussed here.

Because it is an average, a Full-Scale score answers a specific question well — "broadly, where does this person sit against the norm sample" — and a different question badly: "is this person's profile even, or are there real peaks and valleys in it." Averaging is how the first question gets answered and exactly what makes the second one invisible unless someone looks past the total.

The index scores underneath it

Contemporary Wechsler-family and comparable batteries report a Full-Scale score alongside four or five index scores, each built from a small cluster of subtests aimed at one domain. The labels vary slightly by test edition, but the domains are stable across the field:

• Verbal Comprehension — reasoning with words: defining concepts, explaining similarities, drawing on acquired verbal knowledge.
• Visual-Spatial or Perceptual Reasoning — working with block patterns, puzzles and visual arrangements, sometimes split further into a separate fluid-reasoning index on newer editions.
• Working Memory — holding and manipulating information for a few seconds, such as repeating digits in reverse order or tracking a sequence mentally.
• Processing Speed — how quickly simple, low-difficulty visual tasks can be completed accurately under time pressure.

Each index is itself a composite of two or three subtests, put on the same mean-100, standard-deviation-15 scale as the total so they can be compared directly. That shared scale is what makes a gap between indexes measurable rather than just a feeling: a 15-point spread between two indexes is one full standard deviation, and that is a large, specific gap, not a rounding difference.

When the average stops being representative

Consider a report where Verbal Comprehension lands at 118, Working Memory at 112, Visual-Spatial Reasoning at 100, and Processing Speed at 84. Averaged and weighted through the standard formula, the Full-Scale IQ comes out close to 100 — "average," in the plainest reading of the term. But no single index in that profile actually reads as 100. The person is a strong verbal reasoner with markedly slow processing speed, and the composite erases both facts by landing exactly between them.

This is not a rare edge case. Meaningful scatter between indexes shows up often enough in clinical and educational testing that manuals publish tables for exactly this purpose — thresholds for how large a gap between two indexes has to be before it counts as unusual rather than ordinary variation, and separately, how large a gap has to be before it is considered clinically meaningful. Reading the total without checking those tables treats a real, measurable pattern as if it were noise.

The extreme version of this pattern

Index scatter has a well-documented extreme case: a profile with an ordinary or low Full-Scale score sitting alongside one narrow ability that is exceptional, sometimes far outside what the composite would predict. Our piece on savant syndrome and IQ tests works through that pattern in detail — it is the clearest illustration of why the total and the profile can tell almost opposite stories, and why an evaluator who only reads the headline number would miss the actual finding entirely.

Reading a report properly

None of this makes the Full-Scale score useless — for many people it is a reasonably faithful summary, because their index scores really do cluster close together. The point is that it needs checking, not assuming. A few habits make the check quick:

• Read the index scores before the total, and note the highest and lowest.
• Convert each index to a percentile rather than comparing raw numbers — our IQ percentile calculator does this instantly and makes a gap concrete rather than abstract.
• Check the confidence interval on each index individually, not only on the Full-Scale figure; an index typically rests on fewer items than the total, so its own band is wider.
• Ask which subtests were actually administered — a guide to reading a full report walks through the sections in order, including where the discrepancy tables usually sit.

A useful habit, once the indexes are in hand: treat the Full-Scale number as a hypothesis about the profile, not a substitute for looking at it. If the indexes are close together, the hypothesis holds and the total is doing its job. If they are far apart, the total is describing an average of two different people who happen to share one body, and the profile is where the real information lives.

The habit generalises past testing. Any summary statistic — a grade point average, a customer rating, a batting average — buys its simplicity by discarding the shape of what it summarises. An IQ report is unusual mainly in that it hands you the underlying components on the same page as the average, for free. Reading past the top line is the entire skill, and it costs nothing but looking one section further down the report.

Common questions