Individual researchers may interpret responsible conduct of research (RCR) in various ways, especially given the diversity of research personnel in modern science. Therefore, understanding individuals’ RCR-related misconceptions is important, as it can help RCR instructors customize their lessons to target learners’ incorrect and incomplete ideas. In this vein, this study aimed to explore whether Taiwanese and American graduate students differ in their perceptions and misconceptions regarding RCR-related concepts and, if so, to determine these differences. A diagnostic assessment, the Revised RCR Reasoning Test (rev-RCRRT), was developed to pursue the intended goal. The rev-RCRRT is a two-tier test capable of diagnosing whether a student’s justifications of ethical acceptance in relation to specific RCR-related concepts are based on incorrect or incomplete ideas of RCR. The current results indicated that, first, participating graduate students’ test performances dropped drastically between the first- and second-tier items, suggesting that they were able to judge the ethical acceptability of given RCR-specific scenarios but lacked the advanced knowledge required to explain their judgments. Second, in general, American students achieved significantly better scores on the rev-RCRRT than Taiwanese students. Third, the two groups held different RCR-related misconceptions centered around various RCR topics. Specifically, Taiwanese students’ misconceptions involved concepts related to authorship, piecemeal publication, and human-subject protection. However, American students sometimes misjudged the issues regarding duplicate submissions and publication and the reproducibility of research and replication of ideas. In summary, through a cross-national comparative method, this study not only suggests that graduate students from different national backgrounds interpret RCR differently but also provides substantial evidence for the employment of a two-tier test approach in RCR-specific contexts. The implications of the current findings for future research and the utility of using two-tier tests in RCR instruction are also discussed in this article.