| Research on English Language Learners' (ELL) test performance in relation to their language use at home and attitudes towards mathematics have rarely been investigated. This study examines the effects of different test formats and constructs, learner characteristics, and attitudes towards mathematics on ELLs' performance in the Education Quality Accountability Office (EQAO) Grade 9 mathematics test. The participants include a total of 2867 ELLs in Grade 9 who participated in the EQAO Grade 9 mathematics test (1887 from the academic course and 980 from the applied course).;The overall results of the study supported the following conclusions. First, ELLs tend to score better in Multiple Choice Questions (MCQ) over Open Response Questions (ORQ), and language use at home has interaction effects on their performances on various test formats and constructs. Secondly, significant interactions occurred between language use at home and ELLs' performance of test formats and constructs. In addition, ELLs who speak another language (only another language) scored higher in performance in both test formats and constructs than ELLs who spoke only English at home. From the ELLs' questionnaire responses at both the academic and applied levels, two learner characteristics were derived from the section of the questionnaire measuring attitudes towards mathematics: perceived mathematics competency and mathematics interest. The questionnaire also revealed that ELLs in the academic level viewed mathematics in separate constructs (i.e., numerical and spatial mathematics), whereas ELLs in the applied level viewed mathematics as a whole. By extracting from learner characteristics, perceived mathematics competency was found to be the significant predictor for mathematics test performance of ELLs in the academic and applied levels. Moreover, perceived numerical and spatial mathematics ease are significant predictors for mathematics test performance in the academic level. On the other hand, perceived mathematics ease was found to be a significant predictor for the applied level. Results are discussed in relation to second language development, mathematics teaching and learning, and cultural differences. Limitations of the present study were presented where implications for research, test development, and practice are suggested for future research. |
