Abstract: Mixed methods research (MMR) is a kind of research that combines the elements of bothquantitative and qualitative approaches, which not onlycombinesdifferent specific methods but also attempts to integrate theunderlying philosophies and theories. MMR was introduced in the late of 1950s, and became a distinctive methodology with the development of its underlying theoriesand applied processesduring the Paradigm War in the 1980s. Since the 1990s, MMR has developed into a relatively complete methodology as well as a popular research design. Advocates harshly criticize the viewpoint that different paradigms and methodsare notcompatible. They argue that: a) MMR reveals the continuity of paradigmsin that different paradigms do not oppose completely; b) even if different paradigms might beincompatible, specific methods still can be mixed; c) MMR has complementary advantages over quantitative or qualitative research. Thecriticisms provide space for the development of MMR. Moreover, pragmatism, as a widely acceptedparadigm, provides MMR with anunderlying philosophy. However, many researchers question the selection of pragmatism as the paradigm of MMR, as pragmatism seems to be a perfect excuse for researchers to escape reflectingits underlying philosophy. In practice, researchers should first respond to why they choose the design of MMR (DMMR). In particular, they should clarify the process and function of the integration.The integration tends tooccurwhere the qualitative research and the quantitative research joinor when researchers attempt to reachthe conclusions from different parts. It can promote the research or mutual attestation, complementation or innovative conclusions. Based on the sequence, and status of the qualitativeand quantitative research as well as the process and function of integration, DMMR can be divided intothree categories: parallel design, quantitative-qualitative sequence design, and qualitative-quantitative sequence design. Researchers may reorganize the three designs according toparticularresearch questions and research conditions. Currently, researchers still have to face the problems with data collection and analysis, data translation, conclusion integration, and the judgment of inference quality when they choose DMMR. In addition, they shouldfirst prove the rationality of the choice of MMR and provide a panoramic research process and conclusion in the articles. Finally, an investigation is conductedon35 MMRsfrom 330 educational doctoral dissertations, which reveals that few researchers prove the rationality of DMMR or try to integrate the conclusions from different approaches. Thedoctoral candidates seem to choose MMR because MMR has become a fashion, not that they need to integrate different solutions to their research questions more efficiently. In sum, Chinese researchers should pay more attention to addressing MMR in a moresystematic way.