Because speaking appropriately and fluently is the ultimate goal in a foreign language learning process, the authenticity of dialogues in the ELT coursebooks becomes even more significant. However, it is rather doubtful whether the dialogues have the function of preparing students for the English language they are likely to encounter in daily life or not. For that reason, this study uses a corpus of ELT coursebooks to compare the language which the students encountered in school, so-called the term "school English" (Mindt, 1996, p. 232) with authentic English usage. The study investigates to what degree the items in the dialogues in the B1-B2 levels ELT coursebooks correspond to authentic language use by means of corpus results and native speaker's intuition. The "Spoken BNC2014" is used as a reference corpus and corpus data is analyzed through "Sketch Engine," an online corpus tool. Content analysis method is used to analyze qualitative data obtained from native speaker comments. Corpus based research results indicate that the dialogues in the coursebooks underrepresent the functional categories with pragmatic functions suggested by O'Keeffe et al. (2007). In qualitative research, the native speaker revealed a lot of inauthentic and erroneous language use in dialogues in the coursebooks. At the end of the study, it was found that there is a discrepancy between ELT coursebooks language and authentic language use. Lastly, the study emphasizes the significance of corpus methods and native speaker supports to material development for coursebooks' writers in the Ministry of National Education.