What is the Cognitive Proficiency Index? The Wechsler CPI explained

The Cognitive Proficiency Index (CPI) is an ancillary composite score on the Wechsler intelligence scales that combines the Working Memory Index with the Processing Speed Index into one number, reported on the same metric as IQ: a mean of 100 and a "standard deviation" of 15, which is a measure of how spread out scores are. It is not a separate IQ test, and it is not a diagnosis. Its job on a report is to sit next to the General Ability Index (GAI), so a psychologist can weigh what a person knows and can reason out against how efficiently that person handles information. This article covers which subtests feed the index on each Wechsler edition, how large a GAI-CPI gap has to be before it means anything, and what the score cannot tell you.

Pearson, the publisher of the Wechsler scales, states in its WISC-V interpretive report that high CPI scores "indicate a high degree of cognitive efficiency for manipulating and rapidly processing information." The score is built from four subtests drawn from two domains.

• Working memory: the capacity to hold information in mind and do something with it while it is there. Our explainer on what working memory is covers why cognitive batteries measure it.

• Processing speed: how quickly and accurately a person performs simple, well-defined visual and clerical tasks under a time limit. Both halves are about throughput, so neither asks what a person has learned.

Pearson's WISC-V report states that the CPI ranges from 40 to 160, and that "scores ranging from 90 to 109 are typically considered average." Because the scale follows a normal curve, the percentile anchors are fixed: 100 sits at the 50th percentile, 115 at roughly the 84th, and 130 at roughly the 98th. A percentile rank tells you the share of same-age peers a person equalled or outperformed. It is not the percentage of items answered correctly, and the distance between adjacent percentile points is not constant across the scale.

The name has survived three decades of Wechsler revisions while the ingredients changed, so the edition printed on a report matters.

• WISC-V (2014): the children's fifth edition builds the CPI from Digit Span, Picture Span, Coding and Symbol Search, the two working-memory and two processing-speed primary subtests. Our overview of the WISC test covers the battery as a whole.

• WAIS-IV (2008): the fourth adult edition paired Digit Span with Arithmetic for working memory, and Coding with Symbol Search for processing speed.

• WAIS-5 (2024): the current adult edition rebuilt the working-memory side. Pearson's guide to WAIS-5 ancillary index scores lists Cognitive Proficiency as Digit Sequencing, Running Digits, Coding and Symbol Search. Arithmetic is no longer a working-memory subtest at all, and the WAIS-5 framework places it under Fluid Reasoning. See our guide to the WAIS.

A related change matters when comparing an old report with a new one. The WAIS-IV and WISC-IV had a single Perceptual Reasoning Index, which the WISC-V and WAIS-5 split into a Visual Spatial Index and a Fluid Reasoning Index, changing what goes into the GAI on the other side of the comparison. On both current batteries the GAI is five subtests: Similarities, Vocabulary, Block Design, Matrix Reasoning and Figure Weights.

The index's status changed too. Pearson now lists Cognitive Proficiency among the WAIS-5 ancillary index scales, and the published sample report prints it with a percentile rank, a confidence interval and a standard error of measurement. On the WAIS-IV it had no such home: Pearson's own symposium slides listed WAIS-IV CPI norms, critical values and base rates as forthcoming publications rather than as manual tables.

The CPI was designed as one half of a pair. Pearson's interpretive report says so directly: the index "is most informative when interpreted as part of a comprehensive evaluation, together with its counterpart, the GAI." Dumont and Willis had proposed grouping the leftover working-memory and processing-speed subtests into a companion to the already-established GAI; Weiss and colleagues later calibrated norms for that score against the WISC-IV standardisation data and named it the Cognitive Proficiency Index.

Two cautions come from Pearson itself. The GAI is not simply Full Scale IQ with working memory and processing speed removed; the publisher's training materials call that a myth and note that the GAI still taps both abilities, only to a lesser degree. And with rare exceptions such as motor impairment or an uncomputable score, the GAI is not a substitute for Full Scale IQ, because working memory and processing speed are genuine parts of general intelligence rather than contaminants.

Two separate questions get confused here, and Pearson's reports answer them in separate columns.

• Is the difference reliable? Each comparison carries a critical value, the smallest gap that exceeds measurement error at the .05 level for that examinee's age. In Pearson's WISC-V interpretive report the GAI-CPI critical value is 10.18 points, so the sample child's 5-point gap counts as not significant. In the WAIS-5 sample report the critical value is 10.59 points, and the adult's 11-point gap clears it.

• Is the difference unusual? That is the base rate, the share of the reference group showing a gap at least that large. The same WAIS-5 report flags that significant 11-point difference with a base rate above 25%. Significant and rare are different things, and a gap can easily be the first without being the second.

• In which direction? Base rates are directional and reported by ability level, so a gap can be common at one GAI level and uncommon at another.

Pearson's own comparison of clinical and normative base rates shows how modest the separation is. In a WAIS-IV ADHD sample the average GAI-CPI difference was 6.2 points, present in 52.3% of that clinical group against 28.5% of the normative sample at matched ability. The publisher labels that table "for illustration only," noting that it applies group mean data to base rates designed for individuals. A GAI above CPI is more common in these groups and also common in people with no diagnosis at all.

All of these numbers carry measurement error. The CPI's standard error of measurement is about 4 points in both published sample reports, and its corrected retest stability in the WISC-V study ran from .82 to .91. That is why the score is printed with a confidence interval rather than as a single point, and why a difference between two imperfect composites is noisier still.

A large GAI-CPI split is often described as the signature of a learning disability or of ADHD. The evidence does not support using it that way on its own.

Sarah Devena and Marley Watkins tested the discrepancy as an ADHD marker across hospital and school samples. Against a simulated normative sample, the area under the ROC curve reached .64 and .63, which is weak. Against realistic clinical comparison groups, a referred but undiagnosed hospital sample and a matched school sample, the values fell to .46 and .50. An area of .50 is chance. They concluded that the method "has low accuracy in identifying children with attention deficit hyperactivity disorder."

David Giofrè and colleagues found more signal in specific learning disabilities, using WISC-IV profiles from 1,383 Italian children diagnosed in the ICD-10 F81 category. The GAI-CPI difference reached an area under the curve of .748, and the trade-off across cut-offs was stark: a 6-point threshold caught 70% of diagnosed children while wrongly flagging about a third of typical ones, while a 20-point threshold was 91.7% specific and detected only 36.3% of cases. Their conclusion was that discrepancy criteria alone are "not sufficient, as the predictive power is only moderate," though they can support a diagnosis alongside other evidence.

The structural evidence points the same way. Ryan McGill, Stefan Dombrowski and Gary Canivez ran the first confirmatory factor analyses of the WISC-V ancillary scores on the full normative sample of 2,200 children. While the GAI and CPI "appear to function reasonably as composite scores," they report, "they should not be interpreted as reflecting any psychological construct." They also note that the WISC-V authors themselves warn against treating a significant GAI-CPI difference as grounds for invalidating the Full Scale IQ, calling that a "very problematic practice."

None of this makes the CPI useless. It makes it a hypothesis generator. A low CPI beside a much higher GAI raises questions the rest of an evaluation has to answer: fatigue, anxiety, medication, motor difficulty, uncorrected vision, a slow but accurate working style, or a genuine processing weakness. What it does not do is sort people into learner types. The index describes a score, not a person.

The Cognitive Proficiency Index earns its place because it aggregates four subtests into something more stable than any one of them, and gives the General Ability Index a meaningful counterpart. It still describes one slice of cognition, measured on one day, with error attached.

For a broad picture of your own cognitive profile rather than a single efficiency number, the Reasoning and Intelligence Online Test is a professionally developed IQ test built for adults 18 and over by RIOT IQ with psychometrician Dr. Russell T. Warne. It runs 15 subtests across six cognitive indices covering verbal reasoning, fluid reasoning, spatial ability, working memory, processing speed and reaction time, and reports on the mean-100, standard-deviation-15 scale. It does not replace an individually administered diagnostic evaluation.

It uses the IQ metric, with a mean of 100 and a standard deviation of 15, but it is not an IQ. It summarises working memory and processing speed only, excluding the verbal and reasoning subtests a Full Scale IQ includes.

Pearson's WISC-V report treats standard scores from 90 to 109 as typically average, with the index ranging from 40 to 160. A score of 115 sits near the 84th percentile for the person's age group.

It means reasoning and knowledge outpaced processing efficiency on the day of testing. Pearson's reports pair every such difference with a critical value and a base rate, because a gap can be statistically significant while still being common.

No. Research on the discrepancy as an ADHD marker found accuracy at or near chance against realistic comparison groups, and work on specific learning disabilities found only moderate predictive power. The pattern is a hypothesis for an evaluation to test.

Yes. Pearson lists Cognitive Proficiency among the WAIS-5 ancillary index scales, built from Digit Sequencing, Running Digits, Coding and Symbol Search. Those differ from the WAIS-IV version, which used Digit Span and Arithmetic.

1. Devena, S. E., & Watkins, M. W. (2012). Diagnostic utility of WISC-IV General Abilities Index and Cognitive Proficiency Index difference scores among children with ADHD. Journal of Applied School Psychology, 28(2), 133-154. eric.ed.gov

2. Giofrè, D., Toffalini, E., Altoè, G., & Cornoldi, C. (2017). Intelligence measures as diagnostic tools for children with specific learning disabilities. Intelligence, 61, 140-145. researchonline.ljmu.ac.uk

3. McGill, R. J., Dombrowski, S. C., & Canivez, G. L. (2025). Beyond primary assessment on the WISC-V: An investigation of the structural validity of the ancillary scores. Psychology in the Schools, 62(7). ux1.eiu.edu

4. Pearson. (2024). A guide to WAIS-5 ancillary index scores. Pearson Assessments. pearsonassessments.com

5. Pearson. (2024). WAIS-5 score report (sample). Pearson Assessments. pearsonassessments.com

6. Pearson. (2025). Measure more with WAIS-5 [Overview brochure]. Pearson Assessments. pearsonassessments.com

7. Pearson. (2020). WISC-V interpretive report (sample). Pearson Assessments. pearsonassessments.com

8. Wechsler, D. (2014). WISC-V technical and interpretive manual supplement. NCS Pearson. pearsonassessments.com

9. Zhu, J., Raiford, S. E., Coalson, D., & Saklofske, D. H. The what, when, and how of the Wechsler General Ability Index [Presentation]. Pearson. pearsonassessments.com

Figure: original illustration created for RIOT IQ showing how the WISC-V primary index scores group into the General Ability Index and the Cognitive Proficiency Index. It is not a reproduction of any published test material.