Research article introductions (RAIs) play a significant role in gaining publication, and therefore have been studied by many applied linguists. Research into RAIs published in Indonesia has begun to be developed (Adnan, 2009; Mirahayuni, 2001; Safnil, 2000), and generally conclude that Indonesian Humanities RAIs were structured differently from English RAIs. However, as these are early studies, their findings still awaits scrutiny, and little information on Indonesian RAIs especially in Education has been published. Several models describing discourse structure of research article introductions have been proposed, but they have been problematic when applied to analyse RAIs. This paper reports an examination of the applicability of two important models, the CARS (Swales, 1990) and the PJP model claimed to be an Indonesian model of Humanities RAIs (Safnil, 2000), using a selection of 21 Indonesian research article Introductions (RAIs) written by Indonesian academics in Education. It concentrates on the following questions: To what extent do these models fit the data and why? The examination found that none of the RAIs fit the CARS, and only less than half fit the PJP model for various reasons. Therefore a new model is proposed. This model fits most of the data. The paper claims that apart from national concerns, discourse patterns of RAIs are also affected by writing guides provided by the discipline. (Contains 2 figures and 3 tables.)