Dimensions of test performance in English as a foreign language in different European settings: a two-level confirmatory factor analytical approach

• Published in: Educational research and evaluation 2015-01, Vol.21 (3), p.188-208
• Main Authors: Erickson, Gudrun, Åberg-Bengtsson, Lisbeth, Gustafsson, Jan-Eric
• Format: Article
• Language: English
• Subjects: Academic Achievement; Compulsory Education; Correlation; Correlation analysis; Cross Cultural Studies; Denmark; dimensionality; Discriminant analysis; Educational evaluation; Educational Sciences; English (Second Language); English language; European context; Factor Analysis; Finland; Foreign Countries; France; Goodness of Fit; internal structure of test; Language Patterns; Language Proficiency; Language Role; Language Tests; Listening Comprehension; Listening Skills; Lärarutbildning och pedagogisk yrkesverksamhet; Mathematical models; Norway; patterns of English proficiency; Reading Comprehension; Reading Skills; Second Language Instruction; Second Language Learning; similarities and differences; Spain; Statistical Analysis; Structural Equation Models; Student Surveys; Sweden; Teacher Education and Education Work; Teaching Methods; two-level SEM techniques; Utbildningsvetenskap
• ISSN: 1380-3611; 1744-4187; 1744-4187
• DOI: 10.1080/13803611.2015.1021259

The present study explored data from a survey of students' performance in English at the end of compulsory school in 6 European countries. The aim was to gain deeper knowledge of the internal structure of the test and to discuss similarities and differences between the different settings regarding patterns in language proficiency. The analyses, conducted by a factor analytic approach and 2-level structural equation modelling (SEM) techniques, indicated an overall English achievement factor at both student and school levels. Furthermore, an effect of format differences at the student level, constituted by a factor related to tasks demanding constructed response, was found. Three correlated modality factors related to listening, linguistic, and reading skills were identified but not further elaborated on as they improved the model only modestly. Considerable differences in between-school variation were found in the different settings. The results are discussed in relation to the instrument, student achievements, and the state of English in the participating countries.