This corpus-based research aims to identify frame markers (FMs) in master thesis abstracts written in English by English native and non-native postgraduate students and in Turkish by Turkish native students in terms of frequency and functions. To this end, 60 master thesis abstracts were compiled in a corpus. Data were analysed both quantitatively and qualitatively via AntConc.3.2.4 based on Hyland (2005)'s taxonomy of functions (sequencing, labelling stage, topic shift and announcing goal). Findings of the study yielded significant variation among three groups. Non-native English abstracts included the highest numbers of FMs and most diverse types whereas native English abstracts included the least numbers and very few types. Considering FM categories, items used to announce the goal of the writer abounded in number and types in all groups whereas items for indicating topic shifts almost did not occur. Only native speakers of Turkish employed FMs for the function of labelling stages to indicate the discourse act at the end paragraph of their abstracts. These findings are hoped to create awareness among native and non-native postgraduate students in thesis abstract organization via FMs and incorporating FM teaching into academic writing courses to help students structure their discourse better.