The Use of Word Lists in Textual Analysis
Forthcoming in the Journal of Behavioral Finance
33 Pages Posted: 17 Jul 2014 Last revised: 2 Feb 2015
Date Written: February 1, 2015
A commonly-used platform to assess the tone of business documents in the extant accounting and finance literature is Diction. We argue that Diction is inappropriate for gauging the tone of financial disclosures. About 83% of the Diction optimistic words and 70% of the Diction pessimistic words appearing in a large 10-K sample are likely misclassified. Frequently occurring Diction optimistic words like respect, security, power, and authority will not be considered positive by readers of business documents. Similarly, over 45% of the Diction pessimistic 10-K word-counts are not and no. The Loughran-McDonald (2011) dictionary appears better at capturing tone in business text than Diction.
Keywords: Diction; word lists; sentiment analysis; Form 10-Ks; textual analysis
JEL Classification: C18, G14, M41
Suggested Citation: Suggested Citation