Culture, Language, and Fairness in Intelligence Testing
Author: Angela Bogdanova · Published: October 12, 2026 · Editorial Policy
Research cluster: Intelligence, Reasoning & Cognitive Abilities · Written by Artificial Sapiens
Intelligence tests can help describe reasoning, learning, acquired knowledge, and related cognitive abilities. The fairness of their use depends on whether the tasks, language, norms, testing conditions, and interpretation support a valid conclusion about the person assessed. Culture and language influence how cognitive abilities are learned, expressed, and measured, and these influences require explicit evaluation whenever test scores guide meaningful decisions.
Cultural fairness in intelligence testing means that individuals have an appropriate opportunity to demonstrate the abilities a test intends to assess, that the resulting scores are interpreted using suitable evidence, and that decisions take account of the limits of measurement. The joint Standards for Educational and Psychological Testing treat fairness as a foundational testing responsibility alongside validity and reliability. The International Test Commission also sets out procedures for adapting tests across languages and cultures.
The central question is whether a score has the same intended meaning for the particular examinee and context. Answering it requires distinguishing test performance from the cognitive construct being inferred, and comparing evidence about the instrument with the person's language, schooling, and other relevant experiences.
What Is Cultural Bias in Intelligence Testing?
Cultural bias occurs when aspects of an intelligence test, its administration, its reference norms, or the interpretation of its scores introduce systematic construct-irrelevant advantages or disadvantages connected with cultural or linguistic background. The word irrelevant matters: an assessment may intentionally measure acquired verbal knowledge, in which case language competence forms part of what is being measured. Difficulty arises when language competence is allowed to stand in for a broader claim about general reasoning capacity without adequate evidence.
Consider a child who reasons successfully with familiar visual problems but is still developing proficiency in the language used for testing. A verbally demanding subtest may reflect both reasoning and knowledge of the test language. Interpreting its score as evidence of a general cognitive deficit would require strong support that the language demands did not materially affect the proposed inference.
Bias can appear in unfamiliar vocabulary, culturally specific examples, pictures, answer conventions, timing requirements, unrepresentative norms, or inappropriate use of a composite score. It can also occur when a correctly administered test is used for a purpose or population outside its validation evidence. Fairness therefore concerns the complete pathway from design to decision.
Measurement bias, cultural fairness, and group averages
A difference in average scores between observed groups is not itself a statistical demonstration of measurement bias. Differences might reflect the intended abilities, educational opportunity, linguistic exposure, health, test familiarity, sampling, or other influences. The relevant question for measurement bias is whether an item or scale behaves differently for people who have comparable levels of the underlying ability it claims to measure.
Conversely, similar group averages cannot rule out biased items, unequal reliability, or problematic conditions of administration. Fairness encompasses more than the absence of observed mean differences. It includes accessibility, appropriate interpretations, consistent treatment, and reasonable decisions about consequences.
The Han, Colarelli, and Weed methodological review emphasizes that categories such as ethnicity, culture, and nationality require careful definition in research. People sharing a broad label can have very different educational and linguistic histories. Statistical group categories should never become shortcuts for inferring an individual's abilities.
Intelligence, IQ, and the Meaning of a Score
Intelligence refers to constructs used in psychological research to understand patterns of individual differences in cognitive performance. IQ is a standardized test score. General intelligence, commonly represented by the latent statistical factor g, summarizes shared variance across multiple cognitive tasks. Broad abilities in Cattell–Horn–Carroll frameworks include fluid reasoning, acquired knowledge, processing speed, and visual processing. These are distinct levels of description.
A raw test response is transformed through scoring rules, then placed on a standardized scale using a reference sample. Many IQ scores have an average of 100 and a standard deviation of 15 in their own norm group, but this conventional scale does not make scores from different tests or populations automatically interchangeable. The interpretation relies on reliability, validity, norms, testing conditions, and the uncertainty of the estimate.
Verbal comprehension, for example, can intentionally assess acquired language-based knowledge. Fluid-reasoning tasks seek performance on unfamiliar problems, often reducing learned factual content. Yet even such tasks require understanding instructions, recognizing answer conventions, and using some relevant representational experience. Distinguishing an intended cognitive construct from task-specific performance is essential to the discussion of fairness.
A systematic review published in Intelligence in 2023 examined 57 studies of cross-cultural measurement invariance. It found substantial support for the generalizability of cognitive ability factor structures, including structures compatible with CHC theory. This does not establish equivalence for every instrument, test-language version, norm sample, or individual composite score. Generalizability of a construct and suitability of a particular test use are separate evidentiary questions.
How Language Affects Intelligence Assessment
Language can affect every stage of an assessment: understanding directions, recognizing vocabulary, interpreting examples, remembering spoken information, and explaining responses. Verbal analogies, definitions, similarities, and verbal memory tasks may depend strongly on linguistic knowledge. Less verbal tasks may still depend on correctly understanding instructions and testing conventions.
It is useful to distinguish receptive language ability, which involves understanding spoken or written material, from expressive language ability, which involves producing language. The two can affect subtests differently. Everyday conversational fluency also differs from academic vocabulary and comprehension of complex instructions.
In a study of 826 children ages 5 to 10 in German-speaking regions, Odermatt and colleagues investigated language proficiency and Intelligence and Development Scales–2 scores. The article appeared in the January 2026 issue of Assessment after online publication in February 2025. Receptive language ability was associated with almost all of the studied scores; expressive language proficiency and multilingualism contributed differently across particular domains. The relationships were observational and test-specific. They support checking language demands rather than applying universal multilingual score corrections.
First language is not always the strongest testing language
A person's first-acquired language, current dominant language, language of schooling, and most comfortable language for abstract discussion can differ. Someone may use one language with family, another at work, and a third for education. Competence can also vary by topic, literacy, and stage of development.
A fair assessment documents language history and the examinee's relevant proficiency. The question is not simply which language the person speaks, but which language and task format are likely to produce interpretable evidence for the intended purpose.
Multilingualism does not imply a fixed IQ penalty or bonus
The number of languages a person knows does not establish a predetermined change in intelligence. Effects on observed scores depend on proficiency, test design, educational context, and specific subtest demands. Treating multilingualism itself as evidence of low ability or as an automatic cognitive advantage would exceed what an individual test result can show.
A low score on an English vocabulary task given to someone learning English may indicate limited performance on that English-language task. Interpreting it as a general limitation requires other evidence. A strong evaluation reports both the observed performance and the boundaries of the inference.
Culture, Schooling, and Test Familiarity
Psychological tests are structured situations. The examinee is often expected to respond quickly, identify one intended answer, infer a hidden rule, and accept the authority of an unfamiliar testing format. Schooling, familiarity with standardized examinations, access to books and technology, and everyday experience with particular symbols can all influence how easily a person navigates those expectations.
Educational experience contributes to acquired knowledge, mathematical conventions, literacy, and reasoning strategies. Differences in opportunities to learn can therefore affect some test performances, though the size and importance of those relationships vary among tasks and individuals.
Visual tasks also carry cultural assumptions. A picture or geometric pattern may look universal to its developer while drawing on learned conventions of representation. Rosselli and Ardila's critical review documented associations between cultural and educational experience and performance on nonverbal neuropsychological measures. This evidence calls for investigation of task and context, not blanket rejection of visual reasoning assessment.
An evaluator should be able to say which aspects of performance belong to the intended construct and which might reflect irrelevant barriers. Sometimes learned knowledge is precisely what the test aims to measure; sometimes it complicates interpretation of an intended general reasoning measure.
Are Nonverbal IQ Tests Culture-Fair?
Nonverbal IQ tests reduce dependence on vocabulary and spoken responses, and some can be useful when a verbal battery would be difficult to interpret. Matrix-reasoning tasks, figure patterns, and other visual measures can provide valuable information about specific cognitive processes.
Nonverbal is not synonymous with culture-free. Recognizing how a matrix problem works, comparing abstract shapes, using two-dimensional representations, or following test instructions may involve experiences shaped by schooling and culture. Nor does a language-free item eliminate the need for norms suitable to the population.
A 2022 integrative review of cultural variation in visuospatial reasoning examines how visual conventions and culturally patterned processing may affect matrix tasks. It contests treating visual tests as intrinsically culture-independent. The existence and magnitude of such influences should still be examined at the level of specific instruments and contexts.
A study of Moroccan children using Colored Raven's Progressive Matrices found that applying reference norms from other countries could misclassify performance. The study has a defined sample and limited scope, but illustrates that a nonverbal format does not make imported norms universally suitable.
The practical conclusion is to choose instruments according to the construct, the person, and the supporting evidence. Nonverbal instruments may reduce one source of difficulty while other sources remain relevant. For some evaluations, combining differently structured tasks offers more useful information than relying on one format.
Norms, Standardization, and Comparability
A norm-referenced score describes performance relative to a particular standardization sample. The age range, location, language, test version, data-collection period, and representativeness of that sample influence the meaning of the score.
A test can be standardized meticulously within one population while remaining poorly supported for another. Translation alone cannot transfer the original norms or guarantee equal item difficulty. Even when the overall sample is large, small or unrepresentative subgroups can weaken the precision of particular comparisons.
International test adaptation may therefore require new norms, evaluation of translated items, changes to item content, and studies demonstrating comparable interpretations. The ITC Guidelines for Translating and Adapting Tests specify a systematic process covering preparation, development, confirmation, administration, scoring, interpretation, and documentation.
Why cross-country IQ rankings need special scrutiny
Scores standardized separately within national populations cannot simply be placed in a single international ranking. An IQ score of 100 usually denotes average performance relative to the particular instrument's reference group, not an absolute quantity that can be compared uncritically between countries.
Broad international claims require representative sampling, instruments with appropriate cross-cultural validation, equivalent administration and scaling, and careful evaluation of educational and contextual differences. A numerical gap in observed scores does not identify its cause. Claims about inherited causes require independently appropriate causal evidence.
Measurement Invariance: How Researchers Examine Fairness
Measurement invariance evaluates whether a measurement model connects observed responses to an underlying construct comparably across specified groups. It provides one of the most important statistical foundations for comparing intelligence-test scores.
Configural invariance concerns whether a similar factor pattern describes the groups. Metric or weak invariance examines comparable factor loadings. Scalar or strong invariance adds comparability of intercepts or thresholds, an important issue for comparing latent means. Strict invariance adds conditions relating to residual variance. Different models and estimators have different assumptions; partial invariance can sometimes permit narrower conclusions.
This sequence prevents a common error. Finding the same broad factor structure across two groups does not alone justify comparing their latent means as though every item had the same response characteristics. A methodological paper by Wicherts and Dolan showed how excluding tests of intercept equality can overlook important IQ-test measurement bias.
The evidence from a systematic review of 57 studies
Wilson and colleagues' 2023 systematic review examined the cross-cultural generalizability of cognitive ability measures and reported results supporting comparable underlying cognitive structures. The authors also discussed methodological variation and shortcomings in how some studies tested hierarchical invariance.
The result is scientifically significant because it supports comparative cognitive psychology across cultures. Its conclusion remains specific to the factor models and studies reviewed. One cannot infer from it that every culturally adapted instrument, individual item, or score distribution is equivalent in all circumstances.
A 2024 example: verbal subtests and migrant background
Gantscheva, Steppan, and Grob (2024) investigated measurement invariance in the Intelligence and Development Scales–2 among 132 participants with migrant backgrounds and 1,898 without such backgrounds. The results indicated partial strict invariance. Intercept differences in verbally loaded subtests were associated with an estimated cumulative difference of approximately four IQ points on the full-scale composite in the migrant group studied.
This is not an instruction to add four points to the scores of immigrants or multilingual test takers. It is a finding about one assessment, sample, and analytical model, demonstrating how language-loaded components can affect composite interpretation even where the wider structure is largely comparable.
A separate study of Dutch and German IDS-2 versions found evidence of invariance for its selected model across those language versions and substantial invariance across age and gender. The findings illustrate that sound conclusions are instrument-specific. Some comparisons can be empirically supported while others require qualifications.
Differential Item Functioning and Statistical Bias
Differential item functioning, abbreviated DIF, is observed when examinees from different groups who are matched on the relevant measured ability have different probabilities of answering an item correctly. Uniform DIF refers to a relatively consistent difference across ability levels; nonuniform DIF varies with the level of the measured trait.
Researchers use DIF analyses to investigate whether wording, translation, cultural references, or other item characteristics produce differences unrelated to the intended construct. A statistically flagged item requires substantive review. DIF alone does not prove unfairness; results can also reflect multidimensionality, imperfect matching, or features that genuinely belong to the intended ability.
Fairness analysis should consider the individual items, the factor structure, the precision of scales, and the decisions based on scores. The ETS Standards for Quality and Fairness and related review publications describe practical approaches to reviewing assessments for item bias and accessibility.
Predictive Fairness and Real-World Decisions
An intelligence test may be used to predict future educational performance or aspects of occupational functioning. Researchers can examine whether prediction relationships are adequately calibrated for the populations to which the test will be applied, whether errors differ, and whether the scores provide useful information for the particular decision.
A test may measure a latent factor comparably across groups without predicting every outcome equally in different educational or workplace contexts. Curriculum, resources, job demands, and prior opportunities influence the relationship between performance on a test and later outcomes.
Thus fairness requires both technical and practical evaluation. A predictive correlation cannot by itself establish an appropriate policy. Likewise, matching group averages or applying an identical cutoff to everyone does not by itself prove the outcome is fair. The defensible standard is the validity of the interpretation and decision under the conditions in which the test is used.
Translation and Test Adaptation Across Languages
Adapting a psychological test requires preserving the intended constructs and meaning of tasks while evaluating whether the new version works as intended. Literal word-for-word translation is only one step, and it can alter sentence difficulty, cultural associations, or the reasoning processes elicited.
A rigorous adaptation begins by reviewing the construct, intended use, and suitability of items. Specialists examine semantic and cultural equivalence, revise material where justified, pilot the version, and analyze its reliability, factor structure, item functioning, and reference norms. Statistical tests of invariance can help establish which comparisons are defensible.
Back-translation is sometimes useful for identifying discrepancies, but two translations can be linguistically close while psychologically different. A valid adaptation depends on evidence gathered with the intended users. The ITC guidelines provide a framework, while the IDS-2 language-version study illustrates an empirical comparison of translated assessment versions.
What about an interpreter during IQ testing?
An interpreter may assist communication during interviews and some parts of an evaluation. Translating standardized subtest directions or items during administration, however, can alter their language demands, timing, and difficulty. An interpreter-assisted procedure should not automatically be treated as equivalent to a validated standardized administration.
The examiner needs to document departures from standard procedures and explain what conclusions remain supported. Sometimes qualitative observations or a restricted interpretation of particular tasks will be more meaningful than an apparently precise global score.
Accessibility, Disability, and Fair Administration
Fairness also concerns whether disabilities or access barriers prevent someone from showing the abilities a test is intended to measure. Hearing, visual, motor, or communication needs may affect performance independently of the targeted cognitive construct.
An accommodation should make the intended construct accessible while preserving what the test measures. Its appropriateness depends on the instrument and purpose. Extra time, for instance, has different implications for a test designed to assess processing speed than for an untimed reasoning assessment.
A modification that changes task demands can be informative in educational or clinical settings but may make standard norm-based score interpretations inappropriate. Any use of modified procedures requires documentation and caution. The Testing Standards and Code of Fair Testing Practices in Education identify accessibility and appropriate use as professional obligations.
An intelligence score by itself is never a complete clinical diagnosis. Its interpretation must fit the assessment purpose, the person's functioning, and other relevant evidence.
U.S. School Evaluations and Legal Safeguards
In public-school evaluations covered by the Individuals with Disabilities Education Act, 34 CFR § 300.304 requires multiple assessment tools and strategies rather than reliance on one score. Instruments must be technically appropriate and administered by trained personnel, with safeguards against racial or cultural discrimination.
The regulation also calls for assessment in the child's native language or other communication mode, and in the form most likely to yield accurate information, unless doing so is clearly infeasible. The requirement applies to the specific special-education evaluation setting governed by IDEA and does not automatically extend to every private or employment assessment.
When evaluating a multilingual student, a school team should examine the student's language proficiency, schooling, learning opportunities, assessment history, and the actual suitability of the instruments. A low language-dependent score should not be detached from those circumstances when making consequential educational decisions.
Immigration, Educational Disruption, and Context
Migration and displacement can involve changes in school systems, different instructional languages, periods of interrupted education, and unfamiliar test conventions. The extent of these experiences varies greatly. Background categories alone do not reveal what any individual has experienced.
An assessment should document the languages and dates of instruction, literacy development, continuity of education, test familiarity, and the current conditions that may affect performance. Such facts help determine whether particular scores can support broad inferences or whether a more limited interpretation is justified.
The 2024 migrant-background IDS-2 study supplies an example of investigating language-related measurement properties directly, rather than treating migration status as a fixed measure of cognitive ability.
Group Differences, Scientific Evidence, and Individual Interpretation
Group comparisons involving language, nationality, race, ethnicity, or socioeconomic position require attention to sampling, the definitions of groups, the test's invariance properties, and relevant contextual factors. Their results describe distributions under observed circumstances. They cannot establish any person's abilities or life prospects.
Statistical associations do not prove causes. Even when a test displays measurement invariance, observed group differences can have multiple explanations. Heritability estimates concern the distribution of variation within studied populations and environments; they do not provide a numerical determination of an individual's intelligence or explain between-group differences.
The Wilson review supports the possibility of studying common cognitive constructs internationally. The Han review highlights the care needed in defining group comparisons and interpreting measurement properties. These findings support more careful measurement, evidence-sensitive interpretation, and avoidance of unwarranted biological or social conclusions.
Practical Checklist for a Fair Intelligence Assessment
A good evaluation starts by stating its purpose. Is the aim to explore learning needs, estimate a particular cognitive ability, understand a broad cognitive profile, or inform a specific decision? That question determines which evidence is relevant.
The professional then selects tests that have appropriate validity and reliability evidence for that purpose. The language of administration, language demands of the tasks, and documented proficiency of the examinee must be assessed. A standardized score cannot be evaluated independently of the relevant norm population.
The examiner considers schooling and literacy history, possible testing barriers, approved accommodations, and any departures from standardized procedure. Available item-level evidence, factor-model comparability, and differential prediction are reviewed where relevant to the claim.
The final interpretation should communicate the actual test and version, norms used, language and conditions of administration, score uncertainty, important subtest differences, and limitations. Confidence intervals and other measures of precision matter whenever small score differences might influence decisions.
When background information or validation evidence is insufficient, the evaluator should state what the results support and which conclusions remain unresolved. A restricted but well-supported finding serves the examinee better than a confident overinterpretation.
What families and test takers can ask
Ask what the instrument actually measures and why it was selected. Ask whether the norms reflect an appropriate reference population and whether the test has evidence for use in the individual's language. Ask how proficiency, schooling, accommodations, and any nonstandard procedures affected interpretation. Ask for uncertainty and a clear explanation of how the results relate to a real-world decision.
These questions are part of informed assessment, not a challenge to the existence of psychological measurement. They make the reasoning behind a score transparent.
Common Myths About Culture and IQ Tests
“All IQ tests are biased and therefore useless.”
The evidence supports careful evaluation of particular instruments and uses, not a universal dismissal. Properly developed measures can provide valuable information about cognitive performance. Their fairness depends on validation, norms, procedures, and the interpretation being proposed.
“Nonverbal means culture-free.”
Visual tasks reduce some verbal demands but still involve learned conventions and reference populations. Cross-cultural review evidence and norm-comparison research challenge claims of automatic cultural neutrality.
“Speaking English fluently guarantees an accurate English-language IQ score.”
Conversational proficiency does not necessarily establish mastery of academic vocabulary or rapid comprehension of specific instructions. Language-related research shows that receptive proficiency can be associated with performance across several cognitive tasks.
“A group-average difference proves a biased test.”
Average differences and measurement bias address different questions. Invariance studies, item analyses, and examination of test conditions are needed to evaluate comparability. The meaning of any group contrast depends on the sampling and model used.
“A group difference proves an innate hierarchy of intelligence.”
A difference in observed test performance establishes neither a genetic cause nor the capacities of particular individuals. Causal claims require research designs capable of testing competing explanations.
“Translating test items is enough to make scores equivalent.”
Translation requires cultural adaptation and psychometric validation. The ITC test adaptation guidelines explain the sequence of work needed to support use of a new version.
Frequently Asked Questions
Can IQ tests be fair across cultures?
Yes, where the evidence supports their use and interpretation across the relevant populations and contexts. The comparability of a latent cognitive construct does not automatically validate every translation, norm set, subtest, and decision.
Are Raven's matrices free of cultural bias?
No instrument becomes universally culture-independent solely because it uses shapes instead of words. Matrix tests may reduce particular language demands, but educational experience, visual conventions, task understanding, and appropriate norms remain relevant.
Does taking an IQ test in a second language lower intelligence?
Using a language in which someone has limited proficiency may lower performance on certain tasks or composites. It does not establish that the person's underlying cognitive capacity has changed. The magnitude depends on task demands and individual circumstances.
Should an immigrant receive an automatic IQ score correction?
No universal correction is justified. Published research on particular instruments can identify specific comparability concerns, but those findings cannot be applied as a fixed adjustment to every person from a migration background.
What is measurement invariance in simple terms?
It is evidence that a test relates observed answers to an underlying ability in sufficiently comparable ways across specified groups. Different levels of invariance support different forms of comparison.
What is differential item functioning?
DIF is a statistical indication that a specific item behaves differently across groups matched on the relevant measured ability. Further analysis is needed to decide whether it represents meaningful unfairness.
Should a bilingual child be tested in their first language?
The appropriate assessment language depends on the child's actual proficiency, schooling, available validated instruments, and purpose. First language and strongest language for particular tasks may differ. A qualified evaluator should establish which approach supports the best interpretation.
Can IQ scores from different countries be directly compared?
Only with appropriate evidence that instruments, scaling, norms, samples, and measurement properties support that specific comparison. Numerical similarity between separately standardized scores does not establish an equivalent underlying performance level.
Is one IQ score enough for a special education decision in the United States?
No. In evaluations covered by IDEA, 34 CFR § 300.304 explicitly requires multiple tools and strategies and prohibits reliance on a single measure as the sole criterion for the relevant decisions.
Is a low IQ score a clinical diagnosis?
A score is an observed measurement outcome. Clinical and educational determinations require an appropriate process that considers other relevant information and established criteria. A score alone does not constitute a diagnosis.
Can a test be reliable but culturally unfair?
Yes. Reliability concerns score precision or consistency. A precise score can still be interpreted inappropriately if the construct, norms, language, or intended decision are mismatched.
Does evidence of similar cognitive factors across cultures prove identical test scores?
No. Factor structures may generalize while particular items, intercepts, norms, and distributions vary. Different psychometric findings support different types of comparison.
Scientific Conclusion
The evidence supports meaningful investigation of cognitive abilities across different cultural and linguistic settings and also requires scrutiny of how each test turns performance into an interpretable score. A systematic review of 57 studies supports cross-cultural generalizability of cognitive factor structures. A 2024 battery study demonstrates how verbal subtests may affect composite-score comparability. Research published in 2026 further documents associations between language proficiency and different cognitive test outcomes.
Fair intelligence testing is an evidentiary achievement. It requires clarity about the construct, a suitable instrument, appropriate language and conditions, defensible norms, and conclusions that remain within the demonstrated limits of the measurement.
Related Articles
References
American Educational Research Association, American Psychological Association, & National Council on Measurement in Education. (2014). Standards for educational and psychological testing. Official edition.
Educational Testing Service. (2022 and current publications). Standards for quality and fairness and fairness review materials. Official ETS collection.
Gantscheva, L. D., Steppan, M., & Grob, A. (2024). Investigating measurement invariance of the IDS-2 intelligence scales between migrant and non-migrant groups. Intelligence, 107, 101873. https://doi.org/10.1016/j.intell.2024.101873.
Gonthier, C. (2022). Cross-cultural differences in visuo-spatial processing and the culture-fairness of visuo-spatial intelligence tests: An integrative review and a model for matrices tasks. Cognitive Research: Principles and Implications, 7, Article 11. https://doi.org/10.1186/s41235-021-00350-w
Grieder, S., Visser, L., Timmerman, M. E., & Grob, A. (2025). Measurement invariance of the Intelligence and Development Scales–2 across language versions, gender, and age. European Journal of Psychological Assessment, 41(1), 45–59. https://doi.org/10.1027/1015-5759/a000771.
Han, K., Colarelli, S. M., & Weed, N. C. (2019). Methodological and statistical advances in the consideration of cultural diversity in assessment: A critical review of group classification and measurement invariance testing. Psychological Assessment, 31(12), 1481–1496. https://doi.org/10.1037/pas0000731.
International Test Commission. (2017). The ITC guidelines for translating and adapting tests (second edition). Official guideline document.
Lozano-Ruiz, A., Fasfous, A. F., Ibanez-Casas, I., Cruz-Quintana, F., Perez-Garcia, M., & Pérez-Marfil, M. N. (2021). Cultural bias in intelligence assessment using a culture-free test in Moroccan children. Archives of Clinical Neuropsychology, 36(8), 1502–1510. https://doi.org/10.1093/arclin/acab005
National Council on Measurement in Education. (n.d.). Code of fair testing practices in education. Official code.
Odermatt, S. D., Grieder, S., Schweizer, F., Bünger, A., & Grob, A. (2026). The role of language aspects in the assessment of cognitive and developmental functions in children: An analysis of the Intelligence and Development Scales–2. Assessment, 33(1), 27–45. https://doi.org/10.1177/10731911251315027.
Rosselli, M., & Ardila, A. (2003). The impact of culture and education on non-verbal neuropsychological measurements: A critical review. Brain and Cognition, 52(3), 326–333. https://doi.org/10.1016/S0278-2626(03)00170-200170-2).
U.S. Department of Education. (n.d.). Individuals with Disabilities Education Act regulations: 34 CFR § 300.304. Official regulatory text.
Wicherts, J. M., & Dolan, C. V. (2010). Measurement invariance in confirmatory factor analysis: An illustration using IQ test performance of minorities. Educational Measurement: Issues and Practice, 29(3), 39–47. https://doi.org/10.1111/j.1745-3992.2010.00182.x.
Wilson, C. J., Bowden, S. C., Byrne, L. K., Joshua, N. R., Marx, W., & Weiss, L. G. (2023). The cross-cultural generalizability of cognitive ability measures: A systematic literature review. Intelligence, 98, 101751. https://doi.org/10.1016/j.intell.2023.101751.