||Interview and questionnaire data from previous studies have suggested that English-medium instruction (EMI) has enhanced students’ English proficiency but hindered content learning in non-English-speaking countries. This study seeks to verify these arguments using test-based evidence. Study participants were students enrolled voluntarily in either a Chinese/L1-medium instruction (CMI) or an EMI section of an engineering course. Data were collected from pre- and post-assessments of English language proficiency and the course midterm and final exams. Statistical analyses showed that both sections had significant gains in English listening but losses in reading. The EMI section significantly outperformed the CMI section in the listening and reading tests. However, no significant differences were found between the two sections in the English-written midterm and final exams. When controlling for reading, the EMI section had midterm scores significantly lower than the CMI section but slightly higher for the final exam. This study may justify the CMI and EMI parallel-instruction model, that is, offering CMI and EMI sections of the same course from which students can choose. It is hoped that this model will lead to smaller EMI sections with higher English proficiency, eventually leading to content learning outcomes equal to CMI sections.