How has proficiency in written English among middle school students changed over time?

The project, which is headed by Eva Olsson, aims to provide new and in-depth knowledge about Swedish middle school students’ writing proficiency in English. This will in turn provide vital information for use in future planning of English teaching.

English is increasingly used as the language of communication in higher education, in professional contexts and in social media. The ability to write English is thus a key skill in modern society. But what level of proficiency in written English do middle school students have, and how has this changed over time? There is currently a lack of large-scale and longitudinal in-depth studies of students’ proficiency in written English in Swedish middle school.

The researchers in this project will be using unique material: 4,000 student essays completed as part of the mandatory national test in English for ninth-grade students, collected between 2000 and 2022. Using modern language analysis tools based on Natural Language Processing (NLP), the researchers intend to carry out linguistic analyses of student essays to identify any patterns of change, for example in terms of vocabulary, syntactic complexity and textual coherence – aspects that are central to writing quality.

The findings will reveal whether, and if so in what way, student proficiency in written English has changed over the past few decades. Earlier research has shown that increased use of English outside of school may have a positive impact on students’ English vocabulary, and that boys tend to use English in their leisure time more than girls do. Among other things, the researchers will be examining any differences between boys’ and girls’ writing proficiency in English, and also whether this has changed over time.

These analyses will enable the researchers to identify and describe linguistic profiles for different proficiency and grade levels. The researchers will also examine the extent to which various textual and linguistic measures can predict teacher assessments, i.e. what language features that appear to be of importance for the overall assessment. The findings may also add to our knowledge of how automated text analysis could complement and support teachers’ holistic assessments of student essays, thereby strengthening the reliability and consistency of the assessment. The project is therefore of vital importance, not least given the forthcoming digitization of the national English tests.