This dataset contains auditory classifications for prevocalic and postvocalic /r/ produced by German learners of English. The data originate from four different studies. The data by Pascoe (1987) include n = 26 speakers (instructional-setting learners from southern Bavaria; age: 14-16 years), who produced a total of n = 346 prevocalic and n = 303 postvocalic tokens. These data were elicited with the re-telling of a picture story. The data by Wunder (2012) include n = 30 speakers (instructional-setting learners from northern Bavaria; age: 11-20), who produced a total of n = 778 prevocalic and n = 658 postvocalic tokens. These data were elicited with the reading passage. The data by Soenning (2020) include n = 62 speakers (instructional-setting learners from northern Bavaria; age: 11-30), who produced a total of n = 1037 prevocalic and n = 2933 postvocalic tokens. These data were elicited with two reading tasks, a word list and short question-answer sequences. The data by Rank (2018) include n = 21 speakers (mostly university students from Bavaria), who produced a total of n = 1933 postvocalic tokens. These data were elicited with a word list, a reading passage, and a casual conversation with the author of that study. See ReadMe file for more details. Related publication: Soenning, Lukas. 2020. Phonological variation in German Learner English. University of Bamberg dissertation. DOI: 10.20378/irb-49135 Open access: https://fis.uni-bamberg.de/handle/uniba/49135