This study investigated the cognitive validity of two child English language tests. Some teachers maintain that these types of tests may be cognitively invalid because native‐English‐speaking children would not do well on them (Winke, 2011). So the researchers had native speakers and learners of English aged 7 to 9 take sample versions of two standardized English reading and writing tests: the Young Learners Tests of English, Bronze and Silver, administered by Cambridge Michigan Language Assessments. They videotaped the children taking the tests, had them draw pictures of how they felt during testing, and interviewed them. The tests reliably discriminated learners from native speakers. However, 3 of 25 items on the Bronze test and 5 of 40 on the Silver test proved more difficult for native speakers than for language learners. The researchers interviewed the children to uncover why; incorrect responses stemmed from lack of assessment literacy or age‐related cognitive limitations, not deficits in English. The researchers discuss whether this is a problem and conclude that standardized language tests for children, even those already psychometrically reliable and valid, can be improved upon by interviewing child test takers. They stress that parallel information from same‐age, native‐English‐speaking peers is informative in revealing construct‐irrelevant variance.