Academic language competencies: Development and validation of a standardized test instrument

Academic language or the language used in schools has received increasing attention not only in academic discourse but also in social and political discussions during the last years. This is due to the fact that academic language proficiency has been proposed as an extremely important precondition for successful participation in school and for the acquisition of school-related competences (e.g., Bailey, 2007; Uccelli, Galloway, Barr, Meneses, & Dobbs, 2015). Empirical findings indicate that the comprehension of academic language is highly relevant to school achievements over and above the comprehension of everyday language (Heppt, Henschel, & Haag, 2016; Schuth, K?hne, & Weinert, 2017). Studies have also shown that, compared to everyday language, academic language demands constitute a hurdle for both students from monolingual majority language families and for students from multilingual or minority language families (Eckhardt, 2008; Heppt, Stanat, Dragon, Berendes, & Weinert, 2015; for summary, see Heppt, 2016). However, little is currently known about the prognostic impact of academic language skills on the development of scholastic competences. Furthermore, standardised reliable and validated test instruments for measuring academic language skills in schools and research were lacking.

Drawing on the developmental work undertaken by the projects BiSpra I and BiSpra II, the project BiSpra-Transfer aimed to:
(1) develop a standardized and validated test instrument for assessing academic language proficiency in primary school age based on items which had been designed to measure various aspects of academic language comprehension in earlier funding phases (listening comprehension of texts characterized by academic language features [BiSpra-Text], comprehension of connectives [BiSpra-Sentence], comprehension of general academic vocabulary [BiSpra-Word]);
(2) extend the cross-sectional results obtained in BiSpra II by longitudinal analyses of the developmental trajectories of academic language skills and their validity as predictors of school success;
(3) include selected test items in the framework of the project ‘ProSach’, a study investigating the efficacy of a subject-integrated approach to language support, to analyse the sensitivity of the test instrument to language support measures.
Within a standardization study, a sample of 3,625 children attending primary schools in six German Federal States were presented with and completed the test items. Based on these data, we constructed objective test norms that can be used to assess the academic language skills of students from Grades 2 to 4. To take different situations of language acquisition into account and to allow for meaningful assessments of academic language skills of children with different linguistic backgrounds, separate test norms were determined for three linguistic groups: monolingual majority language children (n = 1,667; 46 %); bilingual children who grew up with German and at least one other language before their third birthday (n = 1,222; 31 %); and children with German as a second language who started to learn German after their third birthday (n = 836; 23 %). In addition, (student) teachers were asked to evaluate the BiSpra 2-4 test handbook and test materials to ensure that both the test norms and the theoretical and practical principles of the test were presented in a comprehensible and easy-to-use way.

The internal consistencies of all three subtests revealed that the estimates of academic language skills are particularly reliable for monolingual majority language children and for children in higher school grades compared to children from multilingual or minority language homes or younger children. While almost all reliability coefficients for the subtests BiSpra-Text and BiSpra-Sentence fell within an acceptable to good range, this held true for BiSpra-Word from third grade onwards. Analyses of the intercorrelations between the three subtests and between students’ test results and their grades in reading, writing, maths, and science substantiate the construct and criterion validity of BiSpra 2-4.

Furthermore, in-depth analyses of the data collected within the validation study conducted in the project BiSpra II show that, on average, monolingual German-speaking children initially perform better on all three subtests compared to children from multilingual or minority language homes. Notably the differences between groups are not attenuated across primary school but instead partially increase during primary school age. While performance gaps between children from monolingual German and non-monolingual German-speaking families widened over time with respect to their comprehension of general academic vocabulary and – to a lesser extent – their listening comprehension of texts characterized by academic language features, the differences between groups in terms of comprehension of connectives remained rather stable over the course of Grades 2, 3 and 4. Besides, analyses showed that, for the most part, the families’ educational and socioeconomic background impacts students’ academic language proficiency and its development.

In order to study the effects of academic language comprehension on school performance, we analysed the interrelations between the three subtests BiSpra-Text, BiSpra-Sentence, and BiSpra-Word and achievements in reading and mathematics tests administered in the same grade and one year later. Without exception, positive correlations with medium to large effect sizes were identified. Even if the comprehension of basic grammar and general vocabulary were controlled for, substantial correlations between the subtests and the achievement tests remained, with particularly pronounced correlations between BiSpra-Sentence and the various school performance measures.

The results indicate that the evolving academic language skills assessed in primary school play a significant role in school success. Overall, BiSpra 2-4 allows for a reliable and valid assessment of school-related linguistic competences in children of primary school age. It may be used to evaluate academic language skills and approaches that aim to foster these skills. It may also help to identify children who may be in need of special language support.