A replication of a previous study assessed the construct validity of a national test for academically oriented Swedish upper secondary students. The analysis consisted of a comparison of test results of 10 percent of the Swedish test population to results obtained from a sample of native English speakers of the same age. Analyses of the Swedish students' English written production skills were also performed. The test battery included subtests of vocabulary, phrases, grammar, reading comprehension, and listening comprehension. Results for the native English-speakers were significantly higher on all subtests but one, a reading test, suggesting that the test is a valid measure of English language proficiency. Another important observation was that the native and non-native score levels were unevenly distributed on the various subtests, interpreted as a sign of variable subtest validity and a need for modification of test content and format. Analysis of the students' English writing skills revealed little correlation between native and non-native average scores, interpreted as a difference in language ability in the two groups. Certain tasks appeared to be much more difficult for the non-native students. (MSE)