Researchers and practitioners in the field of second language acquisition have come to realize the importance of non-cognitive skills such as self-efficacy and self-regulation in students’ learning of a second language. However, there has been limited systematic research on such measures in the second language context and the validity and reliability of their intended use and interpretation. The Questionnaire of English Self-Efficacy (QESE) was designed to measure English language learners’ self-efficacy beliefs. This study applied the Rasch model to evaluate the item-level psychometric properties of QESE and its adequacy for use in the Vietnamese context. The results demonstrated that the QESE items predominantly measured a unidimensional construct (i.e. self-efficacy beliefs in learning English as a second language). The scale was highly reliable, and its rating categories functioned effectively, reflecting varying levels of self-efficacy beliefs. Although the item difficulty hierarchical order was generally consistent with the anticipated hierarchy within the Vietnamese context, there were a few surprising results, contradicting what was previously found with Chinese students. For example, item 28 (writing diaries in English) was a very difficult task for Vietnamese students, but a previous study showed that it was a moderately difficult task for Chinese students. Item 5 (writing blogs on the Internet) was an easy task for Vietnamese students, but it was a difficult task for Chinese students. The findings provide directions for future research and have implications for helping professionals provide best practices that meet the unique educational needs of English language learners.