Minimum, first quartile, median, third quartile, and maximum word counts per child.

The first row shows statistics for total (i.e., non-unique) words, while the second shows those for unique words. We represented each child’s record as the unordered collection of his or her abstracted evaluations, which we treated as a single block of text (i.e., a document). We preprocessed the text by lowercasing all strings, removing stop words and special characters, and converting all words to their dictionary forms, or lemmas.