In addition to the experimental study, data collection was car-ried out with the think-aloud method. The aim of this combination was to gain further insights into the subjective views of the target groups regarding the different language styles (cf. research ques-tion 4). For example, information should be obtained whether the computer is perceived as a social partner through personaliza-tion. Is personalized language perceived as natural communication, thereby creating the feeling of a dialogue between learner and computer? Using the think-aloud method, new factors and rela-tions between these should be extracted to explain personalization effects that go beyond previous explanatory approaches.