Date of Award

Fall 2017

Document Type


Degree Name

Doctor of Philosophy (PhD)


Business Administration

Committee Director

Ling Li

Committee Member

John Doukas

Committee Member

Harris Wu


It is a life or death matter for a firm to observe its environment and identify new threats or opportunities quickly. Information technology has increased firm’s speed and agility in responding to environmental changes. Social media offers a vast and timely source of environmental information that firms can readily use gauge public sentiment. Twitter is a high-speed service that allows anyone to “tweet” a message to any interested parties. Firms can access near instantaneous changes in the public mood about any topic by using Sentiment Analysis. These topics range from predicting equities prices to predicting election outcomes. A gap exists in the literature because researchers discard tweets without any theoretically sound reason for doing so. We propose a framework that provides a theory-based justification for discarding data. We then explore the framework results using high frequency equity market prices. By examining the results of three case studies encompassing 57,600 OLS regressions and 1,887,408 tweets, our results indicate the framework yields higher quality results as measured by better R2 fits.