CogAT provides the same comprehensive assessment of reasoning abilities for ELL students as for their non-ELL classmates. The Level 5/6–8 tests use picture-based items like the text-based items that are used with older students. In order to ensure that the test items are fair, we developed items that accurately measured the ability assessed by the battery within different cultural groups of students and then selected only those items that worked well in all groups. This required much more time and effort than simply translating an English-language test.

We relied on the cooperation of many school administrators, teachers, professional item reviewers, and hundreds of students who diligently worked on the “puzzles” we asked them to solve. Our statistical analyses of the picture-based verbal subtests show they require the same verbal reasoning skills as the text-based verbal subtests used at higher grades, but the language-free picture-based subtests do not measure verbal abilities quite as well as the text-based verbal subtests used at grade 3 and higher. For this reason and because verbal abilities are so important for success in school, an optional Sentence Completion subtest is included in the Verbal Battery to help measure verbal abilities for students who speak English or Spanish. Pictorial item formats were also used for the primary-level quantitative subtests. Contrary to popular belief, the near-universal exposure of students to basic quantitative concepts at home and at school makes these subtests less sensitive to cultural differences than nonverbal subtests that use unfamiliar geometric shapes.

Following the standardization study and the development of national norms, we compared the scores of ELL and non-ELL students on the new tests. Since ELL students are more likely than non-ELL students to belong to a minority group and to live in poverty, we estimated the effects of ELL status after first statistically controlling for poverty (estimated by eligibility for the free or reduced-price school lunch) and ethnicity. For the kindergarten through grade 2 student sample, the average Verbal Battery standard age score for ELL students was only 2.2 SAS points lower than the average Verbal Battery SAS for non-ELL students when the Sentence Completion subtest was omitted. The average Quantitative Battery score for ELL students was 1.4 SAS points lower than the scores of non-ELL students on the Quantitative Battery. Both differences were smaller than the 2.7 SAS point difference between ELL and non-ELL students on the Nonverbal Battery. Taken together, the Form 7 and Form 8 picture-based subtests for young students proved remarkably effective in reducing the impact of language on test scores.

The number of students classified as ELL declines across grades. Therefore, our analyses have been restricted to those grades with sufficiently large numbers of students. At grades 3–6, differences between ELL and non-ELL students on the Quantitative Battery (3.3 SAS points) were similar to the differences on the Nonverbal Battery (2.7 SAS points). Combining the quantitative and nonverbal scores into a quantitative-nonverbal partial composite often provides a better estimate of ability for ELL students than either battery alone.

If ELL students in grades 3–6 take the Verbal Battery, it is usually best to compare their scores to those of other ELL students at their school. This can provide important information on the ELL students’ verbal reasoning abilities that is not apparent when national or even local norms are used.