Do reading tests measure the same construct in multiethnic and multilingual older persons?

Abstract

A critical focus of neuropsychological research is to identify unbiased ways to compare heterogeneous groups on background variables relevant to neuropsychological performance. Whereas recent work has pointed to single word reading as a less culturally biased measure of educational experience than years of education, the extent to which reading score captures a broad range of educational experience and does so consistently across ethnic and language groups is unknown. The current study evaluated reading in relation to years of education in English-speaking Whites, Blacks, and Hispanics, and Spanish-speaking Hispanic older persons (n = 342). Consistent with previous work, reading scores at each grade level were lower in English speaking ethnic minorities than in Whites, supporting the idea that variables related to lifetime educational experience are often confounded with ethnicity. Standardized reading scores were highest in the Spanish speakers; however, interpretation of this difference is limited because scores were necessarily derived using separate normative samples. Importantly, the slopes of reading score by years of education were comparable across all groups. That is, reading scores rose consistently with years of education independently of ethnicity or language, suggesting that such scores can be treated comparably for theoretical and statistical purposes in multiethnic and multilingual samples.

Figure 1. Model Derived Regressions of Reading Score on Education in Ethnicity/Language Groups

In order of highest to lowest y-intercept: Small dotted line: Hispanic speaking Spanish Large dotted line: English speaking Caucasian Thick solid line: English speaking African American Thin solid line: English speaking Hispanic