R, facet, and MFRM analyses of the research data of the Academy of Finland -funded research project Broken Finnish. Contains e.g., interactions between assessment criteria rates (general proficiency; general, fluency, coherence, accuracy of the structures, idiomaticity of vocabulary, pronunciation criterion) on scale 1-6 given by NCLP 44 raters, L1s (Thai, Estonian, Finland Swedish, Arabic, Russian), recognition/assumption of the L1 of the examinee assumed by the rater (open q responses turned into numeric data). Added by background data/information from the longitudinal NCLP data: 1) the raters (e.g. education, age, rating experience in years); 2) examinees (e.g. L1, nationality, education).

2023