Listening Comprehension Tests of Immigrant Students’ First Languages (L1) Russian and Turkish in Grade 9: Extended Report of Test Construction and Validation
NEPS Working Paper, (2015)Abstract
In large-scale studies, immigrant students’ first language (L1) proficiency is typically measured with subjective instruments, such as self-reports, rather than with objective tests. The National Educational Panel Study (NEPS) addresses this methodological limitation by testing the L1 proficiency of the two largest immigrant groups in Germany, namely students whose families have immigrated to Germany from the area of the Former Soviet Union or Turkey. Listening comprehension tests in Russian and Turkish were developed for this purpose. The current paper describes the general framework and requirements for testing first language proficiency within the NEPS and describes the construction of the L1 tests for 9th grade students. Subsequently, the paper reports the item difficulty and reliability of the tests as well as analyses of measurement equivalence indicating that the Russian and Turkish test assess the same construct (configural equivalence). The ability scores and their correlations with other variables are, however, not directly comparable. Analyses of construct validity confirm the unidimensional structure expected for the test. In addition, the L1 test scores correlate with other indicators of L1 proficiency as well as with factors regarded as crucial for L1 acquisition, such as exposure to L1, in the expected way (convergent validity) and they are not substantially related to measures of general cognitive abilities (discriminant validity). We conclude that the listening comprehension tests developed in the NEPS are valid measures of L1 proficiency.