Discussion

The results of the present study revealed that fluent Danish-speaking Arabic-Danish bilingual university students performed below both the Danish national mean, and expected scores based on their level of education, on most WAIS-IV indexes and subtests. This is extremely unlikely to reflect true intellectual functioning as all participants were university students, with the majority enrolled in highly demanding academic studies (e.g., medicine, engineering, law). In addition, the sample had high school GPAs well above the national mean and there is substantial evidence that general intelligence is the most powerful single predictor of academic achievement (Bergold & Steinmayr, Reference Bergold and Steinmayr2018). Despite all participants having grown up and completed their entire education in Denmark, and 88% had Danish as their primary language, performances on several WAIS-IV indexes and subtests were related to degree of bilingualism and acculturation, even after correcting for the effects of educational variables (high school GPA, father’s educational level). This is intriguing as it suggests that subtle cultural and linguistic factors, for example familiarity with specific task formats and language proficiency relevant to the items included, may affect WAIS-IV performances.

When compared to the Danish WAIS-IV normative data, participants on average performed approximately 2.5 SS points below the mean on FSIQ, approximately 3 SS points below the mean on VCI and PRI, and approximately 4.5 SS points below the mean on WMI. Only PSI scores, as well as Coding and Symbol Search scores were consistent with expected performances based on level of education (see Figure 1). These findings are consistent with previous studies in bilingual samples using similar Wechsler tests (Ardila et al., Reference Ardila, Lopez-Recio, Sakowitz, Sanchez and Sarmiento2019; Hasson et al., Reference Hasson, Wu and Fine2019; Razani, Murcia, et al., Reference Razani, Murcia, Tabares and Wong2007; Staios, Kosmidis, Tsiaras, et al., Reference Staios, Kosmidis, Tsiaras, Nielsen, Papadopoulos, Kokkinias, Velakoulis, March and Stolwyk2023) but the magnitude of variation from the expected scores was unexpected and, surprisingly, the lowest mean performance across all subtests was found on Block Design, often considered a culture-free test (Razani, Murcia, et al., Reference Razani, Murcia, Tabares and Wong2007). Also, index profiles were more uneven than expected. There were significantly more participants in our sample with abnormal differences between PSI scores and other index scores. Also, there were more participants in the sample with several abnormally large differences between pairs of index scores than expected based on the norms. These findings question the value of examining differences between WAIS-IV indexes in people who are culturally and linguistically different from the normative sample. Taken together, these findings support the first hypothesis that WAIS-IV normative data from the Danish majority population are not representative for Arabic-Danish bilingual university students.

Socioeconomic status, including parents’ educational level, is known to influence intelligence test performance (Sajewicz-Radtke et al., Reference Sajewicz-Radtke, Łada-Maśko, Olech, Jurek, Bieleninik and Radtke2025). Educational attainment has also been linked to test outcomes (Mayes et al., Reference Mayes, Calhoun, Bixler and Zimmerman2009; Spinks et al., Reference Spinks, Arndt, Caspers, Yucuis, McKirgan, Pfalzgraf and Waterman2007). Providing partial support for hypothesis 2, father’s education and the participants’ high school GPA showed weak to moderate correlations with several WAIS-IV indexes and subtests. The most robust associations were found between father’s educational level and participant’s high school GPA, and FSIQ, VCI, PSI, Similarities, Vocabulary, and Coding, explaining 13–22% of the variance in performances. In contrast to previous studies (Hasson et al., Reference Hasson, Wu and Fine2019; Razani, Murcia, et al., Reference Razani, Murcia, Tabares and Wong2007; Sajewicz-Radtke et al., Reference Sajewicz-Radtke, Łada-Maśko, Olech, Jurek, Bieleninik and Radtke2025), performance was unrelated to participant’s years of education and mother’s educational level, possibly due to limited variability in socioeconomic characteristics, i.e., all participants were university students and 46% had at least one parent with a bachelor’s, master’s, or equivalent degree. It can be argued that such higher socioeconomic factors may have resulted in access to more enriched environments, educational resources, and cognitively stimulating experiences, limiting the specific influence of mother’s educational level in WAIS-IV performance in this sample. In comparison, approximately 30% of the Danish WAIS-IV normative sample had at least a bachelor’s or equivalent degree (Wechsler, Reference Wechsler2011).

In support of our third hypothesis, several WAIS-IV indexes and subtests were related to degree of bilingualism and acculturation, even after correcting for the effects educational variables. Danish language use and proficiency was moderately correlated with FSIQ, VCI, WMI, PSI, and the Digit Span subtest, and weakly correlated with the Similarities subtest, while negative moderate correlations were found between Arabic language use and proficiency and the Vocabulary subtest. In fact, over and beyond educational predictors, Danish language use and proficiency continued to account for a significant proportion of the variance in FSIQ, VCI, WMI, and the Similarities, Vocabulary, and Digit Span subtests, explaining 12–17% of the variance in performances. These findings are intriguing as they suggest that, even among fluent Danish-speaking bilinguals who have grown up and completed all their education in Denmark, factors related to the bilingual experience affect performances on WAIS-IV indexes and subtests relying on verbal processes. For instance, while performances were not analyzed at the item-level, it was apparent that participants generally responded well to items requiring knowledge of more common everyday words but often struggled with items requiring knowledge of more abstract and complex terms. Previous studies have also reported bilingual effects on verbal intelligence tests (Ardila et al., Reference Ardila, Lopez-Recio, Sakowitz, Sanchez and Sarmiento2019; Hasson et al., Reference Hasson, Wu and Fine2019; Razani, Murcia, et al., Reference Razani, Murcia, Tabares and Wong2007) and various neuropsychological tests relying on verbal processing (Celik et al., Reference Celik, Kokje, Meyer, Frölich and Teichmann2022; Stålhammar et al., Reference Stålhammar, Hellström, Eckerström and Wallin2022) but in much more heterogeneous samples in terms of their language skills.

Contrary to our hypothesis, Danish acculturation was not significantly associated with WAIS‑IV index or subtest performance, except for a moderate correlation with Matrix Reasoning. These findings differ from prior studies linking WASI performance to acculturation toward US mainstream culture (Hasson et al., Reference Hasson, Wu and Fine2019; Razani, Murcia, et al., Reference Razani, Murcia, Tabares and Wong2007), potentially due to differences in sample profiles and acculturation measurement. Notably, in the present study 92% of the sample was born in Denmark and unlike most acculturation scales, VIA excludes language-related items. In contrast, Arabic acculturation demonstrated moderate positive associations with PRI and the Block Design subtest and accounted for a significant proportion of variance in FSIQ, PRI, and the Block Design and Digit Span subtests, explaining 9–12% of the variance in performance. Additionally, trend-level effects of Arabic acculturation were observed for the Matrix Reasoning and Similarities subtests. These findings were not anticipated and underscore the potential value of employing bidimensional acculturation measures in cognitive research, where such approaches remain relatively uncommon (Tan & Burgess, Reference Tan and Burgess2020). A possible interpretation is that a strong orientation toward the Arabic heritage culture, alongside successful engagement in mainstream educational contexts, may be associated with cognitive or contextual resources that support higher WAIS‑IV performance. However, further research is required to elucidate the mechanisms underlying this association and to assess its robustness and generalizability across populations.

The relationship between PRI, and the Block Design and Matrix Reasoning subtests, and Arabic acculturation supports previous reports that nonverbal intelligence tests are not independent of cultural factors (Gonthier, Reference Gonthier2022). In line with previous studies, performance on Matrix Reasoning approximated the expected scores (Hasson et al., Reference Hasson, Wu and Fine2019; Razani, Murcia, et al., Reference Razani, Murcia, Tabares and Wong2007). This contrasts with studies in less educated immigrant populations (Nielsen & Staios, Reference Nielsen and Staios2023; Staios, Kosmidis, Tsiaras, et al., Reference Staios, Kosmidis, Tsiaras, Nielsen, Papadopoulos, Kokkinias, Velakoulis, March and Stolwyk2023) but suggests that in Arabic-Danish bilinguals with educational experiences similar to the normative sample, Matrix Reasoning may be relatively robust against the influence of education, bilingualism, and acculturation factors.

This contrast between Block Design and Matrix Reasoning may reflect differences in culturally mediated cognitive demands required by the tests. Matrix Reasoning is a nonverbal abstract reasoning test that primarily depends on problem-solving abilities acquired through formal education (Nielsen & Staios, Reference Nielsen and Staios2023), whereas Block Design not only require problem-solving abilities, but also visuospatial reasoning abilities that may be influenced be culturally mediated experiences from outside the classroom. For instance, building with blocks or LEGOs are common childhood activities in Danish mainstream culture, while other (sub)cultures may prioritize other activities. Cultural differences in the perception, manipulation, and conceptualization of visuospatial materials have been documented in a series of studies comparing Western and Eastern cultures (Kitayama et al., Reference Kitayama, Duffy, Kawamura and Larsen2003; Miyamoto et al., Reference Miyamoto, Nisbett and Masuda2006; Nisbett & Miyamoto, Reference Nisbett and Miyamoto2005), and in studies on cross-Atlantic differences in WAIS cognitive profiles that partly attribute differences in PRI, PSI and WMI to differences in child rearing practices (Roivainen, Reference Roivainen2010, Reference Roivainen2019). This is supported by studies on the Wechsler Intelligence Scale for Children (WISC), reporting relatively higher scores on the Block Design and lower scores on the Digit Span and Coding subtests in European compared to US children (Roivainen, Reference Roivainen2019).

In contrast to VCI, PRI, and WMI, performances on the PSI, and the Symbol Search and Coding subtests, were associated with educational variables, but not degree of bilingualism or acculturation. Although a previous Danish study found processing speed to be affected by acculturation in middle-aged and older immigrants (Al-Jawahiri & Nielsen, Reference Al-Jawahiri and Nielsen2020), cultural differences in processing speed have mainly been attributed to different attitudes towards, and exposure to, timed tests in different cultures and educational systems (Agranovich & Puente, Reference Agranovich and Puente2007; Melikyan et al., Reference Melikyan, Puente and Agranovich2020; Rosselli & Ardila, Reference Rosselli and Ardila2003). This suggests that in Arabic-Danish bilinguals who have their education from the same educational system as the normative sample, PSI may be relatively robust against the influence of bilingualism and acculturation. Our results indicate that PSI may be a valid index in this population. Accordingly, when administering the full WAIS‑IV in similar populations, clinicians may place greater interpretive weight on PSI, as it is more likely to reflect true intellectual ability than the other indexes or the FSIQ.

A strength of this study is the cultural, linguistic, and educational homogeneity of the sample. Focusing on academically high-achieving fluent Danish speaking Arabic-Danish bilingual university students supports the interpretation that the results reflect test bias rather than true intellectual ability, highlighting the importance of considering acculturation and bilingualism when interpreting WAIS-IV results. The study also has some limitations that should be considered. The small sample size and limited range in acculturation, bilingualism, and sociodemographic variables may have obscured or diminished the effects of acculturation and bilingualism on WAIS-IV. Second, correcting for participants’ GPA and father’s education may not sufficiently have accounted for other explanatory factors such as assessor bias, stereotype threat, or influence of minority stress (Fujii, Reference Fujii2018; Ponton et al., Reference Ponton, Ardila and Uzzell2007). Third, the adopted normative data for WAIS-IV were collected in 2010 (Wechsler, Reference Wechsler2011) and may not be representative for young adults in 2025. However, as the Flynn effect predicts an increase in standardized intelligence test scores during the last one and a half decades (Trahan et al., Reference Trahan, Stuebing, Fletcher and Hiscock2014), this would be expected to overestimate WAIS-IV performances. Fourth, all participants were raised by immigrant parents and, on average, demonstrated a stronger orientation toward their Arabic heritage culture than toward the Danish mainstream culture. Consequently, it cannot be ruled out that cultural differences in the conceptualization of intelligence may have contributed to the present findings (Fujii, Reference Fujii2018). For example, within Islamic cultures, spiritual knowledge and moral understanding are often considered salient components of intelligence, dimensions that are not captured by the WAIS‑IV (Shaari & Matore, Reference Shaari and Matore2019). Finally, the restricted range of sociodemographic characteristics may limit generalizability. Nevertheless, it is reasonable to expect stronger effects of acculturation and bilingualism in individuals who differ more substantially from the WAIS-IV normative sample in terms of cultural, language, and educational backgrounds. This needs to be established in future research. Importantly, underestimating FSIQ may have more serious consequences for individuals in the lower ranges. For instance, a 10-point underestimation of FSIQ in someone with a true IQ of 82, considered within the low average range, would result in a measured FSIQ of 72, typically considered indicative of intellectual disability.

The findings from the present study provide further evidence of the influence of cultural and language factors on WAIS-IV performances and extend this to fluent Danish-speaking Arabic-Danish bilinguals who have completed all their education in Denmark. Despite being academically high-achieving university students, the sample performed below the national mean on most WAIS-IV indexes and subtests, and this was partly attributed to effects of acculturation and bilingualism. The results have important implications for clinical practice as they suggest that equitable and valid assessment in Arabic-Danish bilinguals may not be possible using the national standardized norm-referenced WAIS-IV. FSIQ results should be interpreted with great caution as it was the measure most significantly affected by educational, acculturation, and bilingual factors. Understanding such influences is of utmost importance as biased results from WAIS-IV or other standardized intelligence tests may result in erroneous clinical interpretations with important implications for the individual being assessed. When assessing Arabic-Danish bilinguals, psychologists should consider whether communicating a FSIQ based on the WAIS-IV adheres to the ethical code of conduct for their profession. Clinicians should also be cautious when administering and interpreting other indexes and subtests in this population, especially the Block Design subtest. However, clinicians assessing similar populations can have some confidence that PSI, and the Symbol Search and Coding subtests, might measure processing speed without significant acculturation or bilingual influences. The results of the current study are intriguing and raise additional questions for future research on the complex role of acculturation and bilingualism on intelligence tests and other test performances. In future studies, it will be important to determine how best to adjust for these factors and explore whether the findings generalize to other standardized intelligence tests and other groups, including groups with other cultural, language, and sociodemographic characteristics and children being assessed with WISC. As the sample predominantly included second-generation immigrants, it would also be worth repeating the study in a sample of third-generation Arabic-Danish bilingual university students to investigate generational effects.