NAART vs. a Full IQ Assessment: What Each Can Tell You

What does the NAART observe, and what does it infer?

The North American Adult Reading Test (NAART) observes how a person performs on a task involving irregularly spelled words. That performance is the direct observation. A verbal-intelligence or IQ-related figure derived from it is a prediction: a value inferred through a relationship calibrated against another measure. The predicted figure is not the same observation as performance across a full cognitive assessment.

In the study titled “North American Adult Reading Test: Age Norms, Reliability, and Validity,” researchers examined 351 healthy adults aged 18 to 91. They described the NAART as an index used to estimate verbal intellectual ability and reported psychometric properties comparable to the WAIS-R Vocabulary test. That comparison supports a bounded interpretation as a verbal-ability signal in the studied sample. The study also found that age and education related to performance, a reminder that the task result is not independent of reader context. It does not show that the NAART samples every ability assessed by a full battery, or that its result is interchangeable with a full-scale score.

A focused signal may still be useful when the question is focused. It can inform an interpretation of verbal ability, provided the reader knows what comparison and prediction method produced the derived figure. It is not a complete account of current cognition. The first question is therefore not whether one result is a whole IQ assessment. Ask instead: what was directly observed, what target was predicted from it, and does that target match the decision at hand?

Sources: North American Adult Reading Test: Age Norms, Reliability, and Validity

How does a reading result become a predicted score?

A prediction equation maps a measured reading result to a criterion score observed in a calibration sample. The estimate depends on that sample, criterion, test version, and equation; it is not a second direct measurement of a person's full cognitive performance. Three quantities should remain distinct. Raw performance is the observed result on the reading task. A norm-referenced score compares performance with a defined reference group, but only when the applicable norms, version, and scoring method are used. A regression estimate predicts a value from the relationship between reading performance and a criterion measured in a study. Calling that predicted value an IQ score does not make it a directly administered full-scale result. The 2002 study “North American Adult Reading Test: age norms, reliability, and validity” examined 351 healthy adults aged 18 to 91. It describes the NAART as an index of verbal intellectual ability, with psychometric properties comparable to WAIS-R Vocabulary, and provides equations predicting NAART and WAIS-R Vocabulary scores based on age and education. The evidence concerns that sample and verbal criterion; it does not establish that every derived figure reproduces a full assessment for every reader. Regression can summarize a group relationship without making an individual prediction exact. A fitted line gives an expected criterion value for a reading result under the model. People with the same reading result can differ in the criterion score. Keep the equation, criterion, and evidence base attached to the number. Do not infer precision from displayed digits or assume its uncertainty is the same as that of a directly administered assessment; the values come from different procedures. “Prediction Limits of the National Adult Reading Test and Its Abbreviated and International Variants” examines 13 NART variants and explains that floor or ceiling effects and linear-regression limits can leave some criterion ranges unreachable for many variants. This is a reason to inspect a model's reach, not evidence of a specific NAART bound. When shown a NAART-derived number, ask which edition and equation produced it, which assessment was the criterion, and whether the reader fits the population and conditions behind that evidence.

Sources: North American Adult Reading Test: Age Norms, Reliability, and Validity; The prediction limits of the National Adult Reading Test and its abbreviated and international variants

What does a full adult assessment add?

A full adult cognitive assessment can add evidence about kinds of performance that an irregular-word reading measure does not sample. For example, Pearson describes the Wechsler Adult Intelligence Scale, Fourth Edition (WAIS-IV), as having ten core subtests across verbal comprehension, perceptual reasoning, working memory, and processing speed. Its Full Scale IQ (FSIQ) combines those subtests, while the assessment also yields index and subtest scores. The North American Adult Reading Test (NAART), by contrast, observes a narrower reading task; an estimate derived from it cannot directly show how someone performed on working-memory or processing-speed tasks. That difference concerns the evidence available, not which number sounds more authoritative. [Pearson, “WAIS–IV - Wechsler Adult Intelligence Scale | Fourth Edition”]

The extra breadth can matter when the question concerns present functioning across abilities or a pattern of strengths and difficulties. Separate task results let an examiner consider whether the overall composite summarizes performance adequately for that purpose. A total can conceal variation among components; a difference among scores is not, by itself, a diagnosis or explanation. Scores must be interpreted under the instrument's administration and evidence, alongside the reason for assessment and relevant context. The Standards for Educational and Psychological Testing state that intended interpretations and uses must be articulated and supported by relevant evidence, and caution that no interpretation is valid for every purpose or situation. “Full” therefore describes broader sampling, not an automatic answer to every clinical or practical question.

Edition matters. Pearson identifies WAIS-IV as the fourth edition and notes that WAIS-5 is now available. WAIS-IV is a documented example of a multi-domain battery, not a generic name for today's adult assessment or a recommendation that every reader needs that instrument. A professional should select an assessment to fit the referral question, intended population, and relevant evidence. Broader administration takes more time and calls for interpretation; breadth alone does not establish precision for every person or suitability for every decision.

The practical choice follows the question. If the need is a focused verbal signal or a bounded estimate of earlier functioning, a reading measure may contribute useful evidence when its version and comparison method fit. If the need concerns current performance across domains, an unexplained discrepancy, or a consequential decision, discuss a purpose-matched broader assessment with a qualified professional. Treat the NAART result as one piece of evidence, not a substitute profile. The longer battery earns its place when additional domains can change the interpretation or next step; otherwise, extra testing may add burden without answering the actual question.

Sources: WAIS–IV - Wechsler Adult Intelligence Scale | Fourth Edition; Standards for Educational and Psychological Testing

What does the NAART evidence support directly?

The direct evidence supports a specific claim: the North American Adult Reading Test (NAART) can provide an index of verbal intellectual ability in the population and comparison studied. It does not show that the NAART reproduces every domain in a full adult assessment, or that its result is a universally current full-scale estimate. In “North American Adult Reading Test: Age Norms, Reliability, and Validity,” a 2002 study administered the NAART to 351 healthy adults aged 18 to 91. The researchers examined performance in relation to age, education, and gender, and compared the NAART with the Vocabulary test from the Wechsler Adult Intelligence Scale–Revised (WAIS-R). They described the NAART as a measure of verbal intelligence with psychometric properties comparable to that Vocabulary test. The target is one verbal subtest, not the entire WAIS-R or the range of abilities represented in a full assessment. The study’s conclusion also had boundaries. Its participants were healthy adults in the studied sample; the criterion was a particular WAIS-R subtest, and the findings included relationships between NAART performance, age, and education. Those details are part of the evidence, not incidental footnotes. They tell a reader what was examined and caution against treating the reported relationship as a personal result independent of background or as a direct observation of every cognitive ability. The abstract also says the verbal-intelligence indexes helped predict or interpret performance on some, but not necessarily all, neuropsychological tests. That qualification argues against reading “valid” as “answers any question.” The title’s reference to “age norms” and the paper’s reported equations may sound broader than the evidence a reader needs. Norms and prediction equations are tied to the test, comparison, sample, and intended interpretation for which they were developed. A title does not establish that any given equation is appropriate for every current examinee, nor does an age-related norm turn a verbal indicator into a full-scale assessment. To apply a result responsibly, a reader needs to know which NAART version and equation were used, what criterion they predict, and whether the supporting population and conditions fit the case at hand. The Standards for Educational and Psychological Testing make the general principle explicit: validity evidence supports particular interpretations and uses, rather than conferring an unrestricted status on a test name. Applied here, the careful conclusion is neither that the NAART is useless nor that it is interchangeable with a full assessment. The 2002 evidence supports a bounded verbal-ability interpretation in its studied healthy sample, with a WAIS-R Vocabulary comparison. It does not establish a present-day full IQ score for all adults. Keep the claim at the scope the evidence tested, and use a broader, purpose-matched assessment when the question concerns a broader profile.

Sources: North American Adult Reading Test: Age Norms, Reliability, and Validity; Standards for Educational and Psychological Testing

Why use word reading to estimate prior functioning?

Reading may offer a clue about ability before illness or injury because word familiarity can remain available when other abilities change. This is a rationale for one kind of evidence, not proof that reading stays unchanged or captures all earlier ability. Premorbid means estimated functioning before an illness or injury began; it is not a score directly observed at that earlier time.

First, select a skill that may be relatively preserved after neurological conditions. Second, compare performance on that skill with a reference relationship between reading and a criterion measure. Third, treat the estimate as a hypothesis of prior functioning. Finally, compare it with current assessment results and history. Each link must fit the question and the person being assessed.

The 2024 article “The prediction limits of the National Adult Reading Test and its abbreviated and international variants” describes irregular-word pronunciation tests as common because word familiarity may persist across neurological conditions and correlate with intelligence. Their clinical use is to compare estimated prior ability with current evidence. But the article examined 13 NART variants and found many could not reach some higher or lower IQ categories because of floor or ceiling effects and limits of linear regression. This is a reason to inspect a test's calibration and range, not evidence of a specific unreachable range for NAART unless its own variant-level result supports that claim.

A boundary comes from the 2005 study “The use of word-reading to estimate ‘premorbid’ ability in cognitive domains other than intelligence.” Among 322 reasonably healthy adults, the British NART-R related strongly to concurrent verbal and full-scale IQ, but less strongly to 26 other cognitive measures. The authors concluded it had limited value as a proxy for prior ability in those other domains. This is contextual evidence about NART-R, not direct NAART validation; a healthy-adult comparison also does not show how accurately a test reconstructs ability after injury.

These findings show the appeal and boundary of reading estimates. Familiarity may preserve a useful signal, while the prediction equation and narrow task limit what it can support. A reading result can contribute evidence about focused verbal ability or a hypothesis about prior functioning; alone, it cannot establish a complete account of current cognition. Its value depends on the target it was calibrated to and how well that target fits the reader.

Sources: The prediction limits of the National Adult Reading Test and its abbreviated and international variants; The use of word-reading to estimate ‘premorbid’ ability in cognitive domains other than intelligence

Why can a reading proxy and a full assessment disagree?

They can disagree because they sample different performances. A reading proxy begins with a narrow word-reading task and uses it to estimate broader ability. A full cognitive assessment samples several kinds of performance directly. Even when reading performance relates to an overall IQ score in a study, that does not show it estimates every component equally well. One predicted number is not a complete profile.

The study “The use of word-reading to estimate ‘premorbid’ ability in cognitive domains other than intelligence” tested the National Adult Reading Test-Revised (NART-R), an abbreviated Wechsler Adult Intelligence Scale, and 26 other measures in 322 reasonably healthy adults. NART-R performance correlated strongly with concurrent verbal and full-scale IQ, with reported correlations of at least .72. Its correlations with the other cognitive measures were significantly lower. The authors cautioned against using NART-R to benchmark premorbid abilities in other domains. This is evidence about NART-R in that sample, not a direct finding about the North American Adult Reading Test (NAART); the participants’ current performance was used to examine the relationship.

A second study, “Comparison of methods for estimating premorbid intelligence,” involved 92 neurologically healthy British adults who completed the National Adult Reading Test (NART), Wechsler Test of Adult Reading (WTAR), and Wechsler Adult Intelligence Scale-Fourth Edition (WAIS-IV). Reading tests provided the most reliable and precise estimates of WAIS-IV full-scale IQ among the methods compared, with modest improvement from demographic data. However, relationships between NART/WTAR scores and individual WAIS-IV indices varied considerably. Reading measures were more useful for crystallized abilities represented by verbal-comprehension and general-ability indices than fluid abilities represented by working-memory and perceptual-reasoning indices. These were not NAART tests, and the sample was British and healthy; the findings cannot validate a NAART estimate for another population. They do show how an estimate can relate to an overall criterion while offering uneven information about its parts.

These findings do not make a broad assessment infallible or a reading proxy useless. They answer different questions with different evidence. An overall association describes how measures vary together across a sample; it does not establish that they are interchangeable for each person or describe the same pattern of abilities. A reading estimate may contribute a focused clue about verbal ability or prior functioning. A multi-task assessment can show performance across sampled domains, subject to its purpose, administration, and interpretive limits.

If a NAART-derived estimate and a broader assessment appear inconsistent, ask what each was intended to represent, which comparison instrument and edition supported the estimate, and whether the reading evidence fits the person and question. A mismatch is a reason to examine the evidence and its scope, not to declare one score the person’s true ability or identity. NART-R, NART, and WTAR findings help explain why reading proxies may be uneven across domains, but remain contextual until direct NAART evidence supports the same interpretation.

Sources: The use of word-reading to estimate ‘premorbid’ ability in cognitive domains other than intelligence; Comparison of methods for estimating premorbid intelligence

What do comparison studies establish—and what do they not?

Comparison studies can show when a reading score tracks a specified criterion in a sample; they cannot by themselves establish equivalence to a full assessment for a different test version, population, or use. The cited studies examine related reading tests, not NAART itself, so their findings do not establish NAART validity.

In “Comparison of methods for estimating premorbid intelligence,” 92 neurologically healthy British adults completed the National Adult Reading Test (NART), the Wechsler Test of Adult Reading (WTAR), and the Wechsler Adult Intelligence Scale, Fourth Edition (WAIS-IV). Reading tests provided the most reliable and precise estimates of WAIS-IV full-scale IQ ; adding demographic information gave modest improvement. Yet relationships with individual WAIS-IV indices varied considerably. The estimates appeared more useful for crystallized abilities represented by verbal comprehension and general ability than for working memory and perceptual reasoning. This shows that reading scores can predict a composite in that sample while representing its components unevenly. It does not directly validate NAART or show that each person’s reading estimate matches a full assessment. The study used British tests and a specific WAIS edition, which bound the conclusion.

“The use of word-reading to estimate ‘premorbid’ ability in cognitive domains other than intelligence” examined 322 reasonably healthy adults who completed the revised British NART, an abbreviated WAIS-R or WAIS-III, and 26 other cognitive measures. NART-R scores correlated at least .72 with concurrent verbal and full-scale IQ; correlations with the other measures were significantly lower. They concluded NART-R had limited utility for estimating other premorbid abilities. This study concerns NART-R and current performance in healthy adults, not NAART in a person assessed after injury. It illustrates a boundary: a reading test may track IQ-related criteria while serving as a weaker proxy for distinct abilities.

“The prediction limits of the National Adult Reading Test and its abbreviated and international variants” analyzed 13 NART variants. The authors reported that many could not produce predictions in some higher or lower IQ categories because of floor or ceiling effects and limits of linear regression. That finding applies to the variants analyzed as a group. Unless the NAART-specific analysis establishes a particular limit, it would be wrong to claim that NAART cannot reach a named range. The useful question is whether the exact version and prediction method can represent values relevant to its intended use.

Together, these comparisons support a practical rule, not a conversion: a reading score can inform the target for which it was calibrated while remaining insufficient for another target. A sample correlation does not demonstrate individual agreement, and findings for NART, WTAR, or NART-R do not automatically transfer to NAART. The comparison most likely to change the conclusion is direct NAART evidence against the exact current criterion, in a relevant population and setting. Until that match is established, verify test version, criterion edition, population, and intended use before applying an estimate.

Sources: Comparison of methods for estimating premorbid intelligence; The use of word-reading to estimate ‘premorbid’ ability in cognitive domains other than intelligence; The prediction limits of the National Adult Reading Test and its abbreviated and international variants

Which reader factors change how much the result can say?

Interpretation depends partly on whether a reader's language, word familiarity, education, reading history, age, pronunciation, and health context resemble the people and conditions behind the evidence. The NAART measures word-reading performance, so the result can reflect both the ability an examiner hopes to estimate and the reader's experience with the language and words involved. A mismatch weakens the basis for applying a prediction; it does not show the direction or size of any error. “North American Adult Reading Test: age norms, reliability, and validity” examined 351 healthy adults aged 18 to 91. Its abstract reports that age and education were related to NAART performance and describes the test as an index for estimating verbal intellectual ability. These findings make age and educational background relevant questions. They do not provide a correction for every individual or establish that the same relationship holds for every language history, health condition, or use. The healthy sample also matters: evidence from that context should not automatically be applied to every clinical population. Reading aloud raises another applicability question. Someone who learned English later, has less familiarity with the tested vocabulary, or has a reading or speech difficulty affecting pronunciation may perform differently for reasons beyond the target ability the estimate is intended to represent. That possibility is a reason to examine fit, not to guess the size or direction of an error. Similarly, the idea that word familiarity may remain relatively available in some neurological contexts does not mean reading is unaffected in every person. Evidence from neurological samples in “Performance of the North American Adult Reading Test in neurological patients” concerns its studied groups and conditions; it cannot establish a universal rule for an individual. The “Standards for Educational and Psychological Testing” call for evidence supporting an interpretation in its intended population, administration, and use. The standards do not validate the NAART for a particular reader; they help frame the applicability question. Ask the examiner which NAART version and comparison method were used, what population and target score the method relies on, and whether language, education, reading, age, or health factors mattered here. If the fit is uncertain, or the result will inform an important decision, widen the evidence through a qualified assessment. Do not apply an unverified conversion or treat the estimate as a personal ceiling.

Sources: North American Adult Reading Test: Age Norms, Reliability, and Validity; Performance of the North American Adult Reading Test in neurological patients; Standards for Educational and Psychological Testing

When is the NAART result relevant to the question you have?

The NAART is most relevant when the question concerns a focused verbal signal or a qualified estimate of prior ability. It is not sufficient when the question asks for a current multi-domain profile, a diagnosis, or a decision that depends on evidence beyond word reading. For low-stakes curiosity, the NAART may offer a narrow indication of performance on irregular-word reading and, under an appropriate reference method, a verbal-ability estimate. It cannot establish a complete account of present cognitive performance. A clinical hypothesis about functioning before an illness or injury is a more specific use. The study “Comparing the North American Adult Reading Test (NAART) and the Test of Premorbid Functioning (TOPF) to estimate premorbid Wechsler Adult Intelligence Scale - 4th edition FSIQ in a clinical sample with epilepsy” examined 101 patients with temporal lobe epilepsy. The researchers compared NAART-derived estimates with TOPF demographic predictions and reported that the NAART estimates may have been more accurate in that sample, where the TOPF estimates likely underestimated prior full-scale IQ. The result does not establish the same advantage for other conditions, populations, or decisions. The estimate remains one hypothesis to interpret with current assessment findings and other relevant history, not a directly observed score from before illness. For a consequential current evaluation, a reading estimate alone leaves too much unanswered. A decision may require evidence about several abilities, everyday functioning, or a specific referral concern. The Standards for Educational and Psychological Testing explain that evidence must support the intended interpretation and use for the intended population; the standards do not prescribe the NAART or certify a particular use. A qualified professional can select measures for the referral question and integrate their results. A broader assessment also depends on appropriate selection, administration, interpretation, and fit to the decision. This distinction keeps a narrower estimate from being treated as either useless or conclusive simply because it does not sample every domain. A practical scope rule follows. If the question is about a focused verbal signal or a carefully bounded estimate of prior functioning, ask what NAART version and prediction method were used, what criterion they target, and whether the supporting population and context fit. If the question concerns present abilities across domains, diagnosis, or a consequential choice, ask for a purpose-matched professional assessment and treat the NAART as one possible clue among other evidence.

Sources: Standards for Educational and Psychological Testing; Comparing the North American Adult Reading Test (NAART) and the Test of Premorbid Functioning (TOPF) to estimate premorbid Wechsler Adult Intelligence Scale - 4th edition FSIQ in a clinical sample with epilepsy

What should you do next with the result?

Start by asking what the result was meant to estimate: verbal ability, functioning before an illness or injury, or current performance across several abilities. Then ask which NAART version and comparison measure were used, and whether the reading and language background behind the result fits yours. The Standards for Educational and Psychological Testing say that evidence must support the particular interpretation and use; a test name alone cannot answer whether a result fits your decision.

If you need a current profile across abilities, or the answer may affect an important decision, discuss a purpose-matched assessment with a qualified professional. Ask how the reading result fits the other evidence and the referral question. A NAART estimate can contribute to that discussion, but it cannot supply a full account by itself.

For general curiosity, you can try Test IQ Free’s optional 50-question reasoning test and review its worked answers. Its displayed estimated IQ range is provisional, produced by an unvalidated conversion with no validated adult reference population. It cannot confirm or validate a NAART result. The optional paid report adds explanation, not validity. If you are unsure what your result was intended to support, ask the examiner: “What question does this estimate help answer, and what evidence would I need for the decision I’m facing?”

Sources: Standards for Educational and Psychological Testing; Test IQ Free 50-question reasoning test

Questions readers ask