Fig. 1From: Educational L2 constructs and diagnostic measurementRaw score distribution for the training and validation samplesBack to article page