Reading comprehension is an essential skill for learning in general and in science classes. Problems with reading comprehension might hinder students' participation in learning science. Text in science includes specific language features that distinguishes it from narrative text, so should reading instruction be part of teaching science? The direct and inferential mediation (DIME) model of reading comprehension subsumes factors that influence reading comprehension. It was tested separately regarding narrative text as well as expository text in English; however, both have not been tested by directly comparing them to each other. In this study, we investigated to what degree general reading comprehension of narrative text is directly comparable to topic-specific reading comprehension of science text. Hence, first the applicability of the DIME model of reading comprehension in another language (i.e. German) was tested. Second, a general reading comprehension model was directly compared to a topic-specific model for reading comprehension of science text. Participants across the two studies were 704 German Grade 8 students who completed measures of comprehension and the DIME predictor variables. Results of two path analyses indicate the general applicability of the model for another language and additionally for both genres. However, some differences are highlighted that may be of importance in future science-specific studies as well as for teaching science.