CEO Deception and Internal Motivation: An Application of Machine Learning in Upper Echelons

CEOs have an extremely important role in leading their firms. Unfortunately, they tend to be difficult to gain access to and consequently the impact of many important psychological constructs has been out of reach for scholars to study. The primary purpose of this dissertation is to introduce a methodological approach that can be used to capture these psychological constructs. The first essay introduces a theory driven approach to machine learning and provides a demonstration of its use. The second essay applies the machine learning approach to capture deception and examines its influence on financial analyst's recommendations. The results of this essay suggest that deception can increase the likelihood the firm will receive a superior recommendation. However, the positive influence of deception dissipates over time and is less influential on analysts who have lower reputations. The final essay applies the machine learning approach on a different construct, namely internal motivations. The results of this study show that a CEO's need for power and affiliation increases the likelihood they will engage in alliances but their need for achievement will decrease it.