Using nonverbal tests to identify gifted & talented ELLs

iStock_000016876858XSmallEnglish-language learners (ELLs) are often underrepresented in programs that serve gifted and talented students because of difficulties in identifying the abilities of these students. Schools increasingly are using nonverbal tests to evaluate the academic talent of ELLs, but how valid and reliable are these tests in identifying academically gifted ELLs? asks a recent study in Gifted Child Quarterly

Many educators assume that nonverbal reasoning tests level the playing field for ELL children, but in this study of about 1200 K-6 children, ELL students scored 8 to 10 points lower than non-ELL children on the 3 nonverbal tests examined in this study, researchers say.

“These differences are congruent with the conclusion that nonverbal tests do not see through the veneer of culture, education or language development,” the authors write.

The study compared the validity of 3 nonverbal tests used in identifying academically talented ELL students:

  • Naglieri Nonverbal Ability Test(NNAT),
  • Cognitive Abilities Test (CogAT, Form 6
  • Raven Standard Progressive Matrices (Raven)

“Separating test scores for all ELL and all non-ELL children is no more difficult than separating the scores for boys and girls or for third-graders and fourth-graders,” the researchers write.

To make inferences about talent or aptitude, educators need to compare children’s performance to that of other children who have had roughly similar opportunities to develop the abilities measured by the test, the authors write. This is true not only for verbal and quantitative abilities, but also for abilities measured by nonverbal tests.

Test norms

The researchers caution that they did not find the national norms on the tests to be of comparable quality.

“This controlled comparison of the Raven, the NNAT and the CogAT showed that the 3 tests differ importantly in the quality of their norms, in the reliability of the scores they produce, and in their ability to identify the most academically able ELL and non-ELL students,” the researchers write.

“On the NNAT, ELL students were much more likely to receive very low scores. On the Raven, non-ELL children were much more likely to receive very high scores. Only the CogAT Nonverbal showed normally distributed scores for both groups.”

Participants in the study were 1198 K-6 children in two elementary schools in a large Southwestern school district. The native language of almost all ELL students was Spanish. As well as the three nonverbal tests, students also took an achievement test, the Arizona Instrument to Measure Standards Dual Purpose Assessment (AIMS DPA).

The CogAT consists of three separate batteries that measure verbal, quantitative and nonverbal reasoning. In this study, the best predictors of achievement were students’ abilities to reason in the most relevant domains. “For reading comprehension, this was the CogAT Verbal score, whereas for mathematics it was a weighted combination of the three CogAT scores.” Another issue in the use of nonverbal tests to evaluate the academic talent of ELLs is whether in minimizing the role of language, these assessments fail to capture fluid reasoning ability.

“The question…is not whether nonverbal tests should be administered to ELL children,” the authors write. “All would agree that such tests can provide helpful information. Rather, the issue is whether nonverbal tests should provide the only estimate of ability or if other measures of ability should be used to provide additional information about a student’s academic aptitude.”

Nonverbal test scores should just be one part of a comprehensive identification approach incorporating a broad range of abilities and teacher ratings, they say.

“Identifying Academically Gifted English-Language Learners Using Nonverbal Tests, A Comparison of the Raven, NNAT, and CogAT,” by David Lohman et al., Gifted Child Quarterly, Volume 52, Number 4, Fall 2008, pp. 275-296.

 

Published in ERN February 2009, Volume 22, Number 2

Leave a Reply

  • (will not be published)