Titles and authors. Included are authors or titles for which we had at least 50 transactions with good census data.
Mann-Whitney calculations for the LIWC data.
Words from the "blue" corpus, listing the associated Wordnet emotion synsets, LWIC dictionary values (using jsut the '126_posemo', '127_negemo', '129_anger', '130_sad' dictionaries), and the NRC Word-Emotion Lexicon. The spreadsheet does not contain all words from the "blue" corpus; I removed words which occur only once, plus about 800 rows of cruft (numbers, words where parsing failed, etc). The purpose of this spreadsheet is to provide some basis for understanding how well or poorly various dictionaries cover our corpora.
Face, hand, etc. topic sentences.