What the WAIS-IV Measures: Subtests Explained

The question in the title was first posed in print by Benson, Hulac, and Kranzler (2010), who ran the earliest independent factor analysis of the Wechsler Adult Intelligence Scale-Fourth Edition against Cattell-Horn-Carroll (CHC) theory. Sixteen years and a published successor later, the answer has matured into something the original publisher’s manual did not say: the WAIS-IV measures fewer cleanly separable broad abilities than CHC nomenclature suggests, and a single general factor accounts for most of the systematic variance among its subtests. The same pattern has now reappeared in the WAIS-5.

What Benson, Hulac, and Kranzler actually found

Working from the published WAIS-IV standardization correlation matrices, Benson and colleagues fit a series of confirmatory factor models that mapped the ten core and five supplemental subtests onto CHC broad abilities. They concluded that the WAIS-IV does deliver markers of crystallized ability (Gc), visual processing (Gv), short-term memory (Gsm), and processing speed (Gs), but that fluid reasoning (Gf) and quantitative reasoning (Gq) were not cleanly separable as distinct factors with the available subtests; they recommended treating Arithmetic and Figure Weights as joint Gf/Gq markers and offered formulas for CHC-aligned composite scores.

The more striking part of the 2010 result was negative: configural and metric invariance held across age, but stronger forms of measurement invariance did not. Factor loadings shifted systematically across the 16-69 age bands of the standardization sample, meaning that “the WAIS-IV does not measure the same constructs in the same way” across the adult life span (Benson et al., 2010). For practitioners, that finding undercuts a routine assumption: that an older client’s index scores can be interpreted on the same metric as a younger client’s.

The four-vs-five-factor debate

Pearson’s response came from Weiss, Keith, Zhu, and Chen (2013), who used the full 15-subtest standardization sample (N = 1,800) plus a clinical sample (N = 411) to compare the published four-factor model (Verbal Comprehension, Perceptual Reasoning, Working Memory, Processing Speed) against a five-factor CHC-aligned alternative that splits the Perceptual Reasoning Index into a visual processing factor (Gv: Block Design, Visual Puzzles, Picture Completion) and a fluid reasoning factor (Gf: Matrix Reasoning, Arithmetic, Figure Weights). Both models fit acceptably; the five-factor solution fit slightly better and was endorsed for CHC interpretation when supplemental subtests are administered.

This was the first official acknowledgment from the test’s publisher that the printed four-factor structure was a practical compromise rather than the best representation of the data. Independent reanalyses by Canivez and Kush (2013) flagged that the Weiss et al. comparison did not fully address bifactor alternatives in which a strong general factor absorbs most subtest variance, leaving the group factors with little unique interpretive weight—a critique that proved prophetic for the WAIS-5 a decade later.

Extending CHC interpretation to the elderly

The Benson et al. cross-age problem was substantially resolved by Niileksela, Reynolds, and Kaufman (2013), who restricted the analysis to the subtest battery actually administered to ages 70-90 (which omits Picture Completion, Letter-Number Sequencing, Cancellation, and Figure Weights). Within that reduced set, an alternative five-factor CHC model—Gc, Gv, Gf, Gsm, Gs—achieved partial strict measurement invariance across the full 16-90 age range, with only the Similarities subtest showing noninvariance. The takeaway for clinicians: when the WAIS-IV is interpreted using only the subtests common to all ages, CHC index scores are comparable across the adult lifespan; when the full 15-subtest battery is used in younger adults, that comparability is partially lost.

International replications (van Aken et al., 2017, on the Dutch WAIS-IV, N = 1,668) reproduced the five-factor CHC structure but again found that adding more group factors did not substantially reduce the dominance of the general factor in shared variance.

The WAIS-5 keeps the same question alive

When Pearson released the WAIS-5 in 2024, the index structure was revised explicitly to align with CHC theory: the old Perceptual Reasoning Index was split into a Visual Spatial Index (VSI) and a Fluid Reasoning Index (FRI), yielding five primary indices (VCI, VSI, FRI, WMI, PSI). The revision codified what Weiss et al. had argued for the WAIS-IV.

Independent factor analysis by Canivez, Watkins, McGill, and Dombrowski (2026), using the WAIS-5 standardization correlation matrix (N = 2,020) and 20 primary plus secondary subtests, did not vindicate the five-factor split. Their EFA with Schmid-Leiman extraction did not support separate Visual Spatial and Fluid Reasoning factors; a four-factor solution that re-merged VSI and FRI into a single perceptual factor—structurally similar to the WAIS-IV PRI—provided the better empirical fit. CFA results converged on a bifactor model with a dominant general factor and four group factors, with most reliable variance attributable to g rather than to the group factors. The authors note that the WAIS-5 Technical and Interpretive Manual did not report bifactor models, model-based reliability (omega hierarchical), or dimensionality estimates—an evidential gap their independent analysis was designed to fill.

Subtest-level CHC mapping

The CHC interpretive frameworks built by Benson et al. (2010), Weiss et al. (2013), and Niileksela et al. (2013) converge on the following primary and secondary CHC ability assignments for the WAIS-IV core and supplemental subtests. Single-factor purity is rare; most subtests load on a primary CHC factor with non-trivial secondary loadings reflecting the auxiliary cognitive demands of the task.

• Similarities, Vocabulary, Information, Comprehension → primary marker of Gc (crystallized knowledge); Vocabulary additionally loads on language development (VL) as a narrow stratum.
• Block Design, Visual Puzzles, Picture Completion → primary marker of Gv (visual processing); Block Design carries secondary Gf variance from the spatial reasoning demands of construction problems.
• Matrix Reasoning, Figure Weights → primary marker of Gf (fluid reasoning); Figure Weights additionally taps quantitative reasoning (RQ).
• Arithmetic → joint Gf/Gq/Gsm loadings; the Benson et al. analysis treats it as the WAIS-IV’s primary quantitative knowledge marker but its working memory demands inflate Gsm loadings.
• Digit Span, Letter-Number Sequencing → primary marker of Gsm (short-term memory); the backward and sequencing components extend into working memory capacity (WM).
• Coding, Symbol Search, Cancellation → primary marker of Gs (processing speed); Cancellation additionally carries visual scanning (P) and selective attention demands.

The mapping clarifies why the WAIS-IV’s named indices do not translate one-to-one into CHC abilities: the Working Memory Index combines a Gsm task (Digit Span) with a Gf/Gq/Gsm composite (Arithmetic), and the Perceptual Reasoning Index combines two distinct CHC abilities (Gv and Gf) into a single score. The WAIS-5 corrected the second of these by separating VSI from FRI, but the first—an index labeled Working Memory that is partly a Gf measure—persists.

What the WAIS-IV (and WAIS-5) actually measures

Synthesizing fifteen years of independent factor work, three claims are now supportable:

• The general factor is doing most of the work. Whether the model has four or five group factors, bifactor decompositions consistently show that g accounts for the bulk of common variance among Wechsler subtests; group factors retain limited unique reliable variance after g is partialled.
• The CHC labels on the indices are reasonable but not pure. VCI maps reasonably well to Gc; PSI to Gs; WMI to Gsm; the WAIS-IV PRI / WAIS-5 VSI+FRI carries Gv and Gf variance that is harder to disentangle than the index names imply.
• Cross-age comparability requires the elderly-eligible subtest set. Niileksela et al. (2013) gave practitioners a defensible CHC interpretive scheme that survives the lifespan; full-battery interpretations in younger adults trade that comparability for finer ability differentiation.

Practical interpretive guidance

For clinical and educational reports, the contemporary literature implies a hierarchy of confidence. The Full Scale IQ remains the most robustly supported number—it is the proxy for the general factor that bifactor analyses recover as dominant. Index scores are interpretable but their incremental information beyond FSIQ is modest, especially for VSI and FRI considered separately on the WAIS-5. Subtest profile interpretation—the longstanding clinical tradition of comparing Block Design to Matrix Reasoning, or Arithmetic to Digit Span—has the weakest empirical foundation and should be undertaken only when corroborated by external evidence.

The original Benson et al. (2010) paper is now seventeen years old and addresses a discontinued instrument, but the question its title posed has outlasted the test. The WAIS-5 inherits the same answer: a strong g, four broad ability domains that are reliably measured, and a fifth (the VSI/FRI split) that the publisher’s manual asserts and independent factor analysis still does not support.

Frequently asked questions

What does the WAIS-IV measure?

The WAIS-IV measures four published indices—Verbal Comprehension, Perceptual Reasoning, Working Memory, and Processing Speed—plus a Full Scale IQ that aggregates them. Independent factor analysis maps these onto five Cattell-Horn-Carroll broad abilities: crystallized knowledge (Gc), visual processing (Gv), fluid reasoning (Gf), short-term memory (Gsm), and processing speed (Gs). A general intelligence factor (g) accounts for the bulk of common variance among subtests in bifactor decompositions.

Is the WAIS-IV still used in 2026?

Yes, but it has been formally superseded by the WAIS-5 (Pearson, 2024). Many clinicians and researchers continue to use the WAIS-IV during the multi-year transition period, partly because of accumulated normative data and partly because comparison with prior WAIS-IV results is often clinically useful.

What is the difference between the four-factor and five-factor models of the WAIS-IV?

The four-factor model is the published index structure (VCI, PRI, WMI, PSI). The five-factor model splits the Perceptual Reasoning Index into a visual-processing factor (Block Design, Visual Puzzles, Picture Completion) and a fluid-reasoning factor (Matrix Reasoning, Arithmetic, Figure Weights). Weiss et al. (2013) found the five-factor model fits slightly better when all 15 subtests are administered.

How well does Full Scale IQ on the WAIS-IV measure general intelligence?

Bifactor analyses consistently show that the general factor accounts for the majority of reliable variance among WAIS-IV subtests. The Full Scale IQ is the most defensible single score from the test for representing general cognitive ability; index-level differentiation beyond FSIQ adds modest incremental information.

Did the WAIS-5 fix the structural issues identified in the WAIS-IV?

The WAIS-5 was redesigned with five primary indices (VCI, VSI, FRI, WMI, PSI) explicitly aligned with CHC theory. Independent factor analysis of the WAIS-5 standardization data by Canivez and colleagues (2026), however, did not support the separate VSI and FRI factors—a four-factor structure resembling the WAIS-IV PRI was the better empirical fit. The general factor remains dominant in either model.

What does it mean when a WAIS-IV score is interpreted across age groups?

Benson, Hulac, and Kranzler (2010) showed that strict measurement invariance does not hold across the WAIS-IV’s adult age range when all 15 subtests are used. Niileksela, Reynolds, and Kaufman (2013) showed that partial strict invariance does hold across the lifespan when only the subtests common to the elderly battery are used. Cross-age comparisons are most defensible using the elderly-eligible subtest set.