This study investigates the effect of teaching English parts of speech constructions to non-English major students at a Taiwanese university. Word-class constructions were taught with a pre-test and post-test comparison. A paired one-tail t-test was employed to examine students’ learning outcomes. Additionally, a questionnaire survey formed with a 5-point-Likert-type scale was distributed to collect participants’ ideas about the teaching. Mean scores and a one-way ANOVA were employed to analyse the results. The paired t-test result points out there is a significant difference (p<.05) between students’ pre-test and post-test scores, which suggests learning English parts of speech constructions can possibly improve students’ English learning achievement. However, the questionnaire results not only report that students do not think their English is improved but also that learning parts of speech construction is less helpful. Students’ lack of confidence may also influence their degree of agreement to the statements in the questionnaire survey.