ABSTRACT
Using a deep-learning-based textual analyzer provided by IBM Watson, this paper obtains scores measuring the overall sentiment and emotion of “joy” from the transcripts of conference calls and uses them as additional predictors of internal control material weaknesses (ICMWs), combined with other determinants of ICMWs suggested by prior literature (i.e., Doyle, Ge, and McVay 2007; Ashbaugh-Skaife, Collins, and Kinney 2007). The results indicate that, with the incorporation of the sentiment features (especially the score of “joy”), the explanatory ability of the model improves significantly, as compared to that of the baseline model that merely utilizes the major ICMW determinants.
2018
You do not currently have access to this content.