This study investigates how the macrostructure patterns (MSPs) of research articles (RAs) are distributed across different disciplines. The investigation is based on the Elsevier OA CC-BY corpus consisting of 76,835 RAs from 26 disciplines coming from Health Sciences (HS), Social Sciences and Humanities (SH), Life Sciences (LS), and Physical Sciences and Engineering (PE). In this study, we identify 72 salient patterns and show that a higher ratio of functional headings is found in HS, LS, and PE than in SH, which can be attributed, in part, to different emphases with respect to the knowledge mode and knower mode (Maton, 2009). In addition, we explore the connection between specific MSPs and particular fields. We also detail the wide distribution of the I_M_R_D pattern; it appears in the top 10 pattern list for 21 out of the 26 disciplines, and is particularly frequent in medical journals. In sum: this large corpus-based study provides a comprehensive picture of how MSPs vary across RAs from different disciplines and fields. [ABSTRACT FROM AUTHOR]