ORCID

http://orcid.org/0000-0003-2787-2477

Date of Award

2020

Document Type

Dissertation

Degree Name

Psychology (Ph.D.)

Department

Psychology

First Advisor

Samuel O. Ortiz

Second Advisor

Dawn P. Flanagan

Third Advisor

Robin Wellington

Abstract

The National Center for Education Statistics (NCES, 2016) reported an increase in the number of non-native English-speaking students in U.S. public schools as well as a frequent coexistent correlation with low-SES and poverty, but not for all racial or ethnic minority groups. Because it is well known that SES and language difference play an important role in academic achievement, it is imperative that school psychologists attend to these variables when considering the validity of obtained test scores and their support for subsequent diagnostic conclusions, especially when current rates of ELLs in special education suggest that evaluations are not necessarily providing unbiased results (NCES 2013). This trend remains troublesome despite advances in psychometrics and test development based on theoretical models of intelligence (i.e., CHC, Luria). However, use of tests from varied theoretical camps provides an additional challenge, as not all batteries measure constructs in similar ways (i.e., construct equivalence).

As a result, this study evaluated the comparability of construct equivalence on neuropsychological measures across batteries and tests, the extent to which typical neuropsychological domains vary according to how much “language” is used in the measurement of each domain, and the equivalence of scores when domains are assessed in high SES monolingual and bilingual populations in a sample of 252 school-age individuals who underwent evaluations in a private clinic. Results indicated that there is variation in how domains are constructed on certain batteries, confirming that for some tests there is not construct equivalence; high SES bilinguals and monolinguals seem to perform just as well on language tests; and that linguistic demand impacts bilinguals’ performance. Post-hoc analyses indicated that the presence of a diagnosis sometimes indicated poorer performance on domain tasks. Implications include the need to consider the impacts of language, disability, and SES when evaluating bilingual students, as well as test selection during evaluation planning. Further research is needed to address the differences in performance for high and low SES bilinguals and address the possible presence of a “bilingual advantage.”

Share

COinS