Corpus of Reading comprehension Exercises in German (CREG)
CREG is a corpus consisting of reading comprehension exercises. These exercises each comprise a reading text, questions to that text, possible target answers to these questions, and a large amount of answers by learners of German. The data are being collected in collaboration with Nina Vyatkina (University of Kansas) and Kathryn Corl (Ohio State University). Each learner answer is rated by two trained German teachers with respect to meaning: does it answer the question? In July 2012, the continuously growing corpus already contained 169 texts, 1470 questions 1916 target answers and 35211 learner ansers. The data are available for non-commercial scientific purposes on request.
Corpus of Reading comprehension Exercises in English (CREE)
CREE is a corpus consisting of 566 answers by ESL learners to 75 reading comprehension questions corresponding to 11 texts. The corpus furthermore contains one target answer and a list of keywords for each question. Each learner answer has been annotated by two persons with respect to correct meaning. (Does the student answer the question or does he not?) These date were originally collected by Stacey Baily at the Ohio State University. A revised version is available for non-commercial scientific purposes on request.