In my first chapter, we investigate the use and utility of a family of statistical techniques called relative importance analysis. Relative importance analysis techniques are increasingly used in organizational research to assess the importance of various causes on their outcomes. These techniques are attractive because they promise to assess the importance of variables unambiguously and objectively. Regardless of their popularity, these techniques and their use have not been thoroughly scrutinized. We review the literature to see how relative importance techniques are used and critically assess if these tools are effective in addressing the problems they are used for. We find that researchers predominantly use relative importance techniques to assess the relative causal importance of variables, which these techniques are unsuitable for. We back our arguments with intuitive explanations, formal analysis of the relative importance techniques, and a Monte Carlo simulation comparing relative importance techniques against regression analysis. We show that regression provides a more accurate ranking of the relative causal importance of variables across a variety of scenarios. We conclude with practical guidelines on how to assess the relative causal importance of variables using the straightforward idea of comparable investments.
In the second chapter of my thesis, we seek to settle the decades old debate about the social construction of leadership. Societies are facing numerous grand challenges and leaders are increasingly counted on to provide solutions. But can top-level leaders affect outcomes that unfold via various pathways? Some research streams have suggested that organizational outcomes may not be caused by, but are simply ascribed to the leader; leadership may merely be a consequence of causal attributions. We provide a rigorous test to determine whether leaders matter by exploiting a very controlled, though unusual leadership context, where leader discretion is restricted; that of U.S. state governorships. This context allows us to estimate precisely what role top-level leaders and their teams may play in determining institutional outcomes, measured on a standard metric. We quantify the “leadership effect” in a sample of 500 governors across the 50 states of the U.S. and the district of Columbia. We implement state-of-the-science methodical advances in variance decomposition on a sample of 2,985 governorship-time observations, covering the periods 1963
to 2019, to explain variance in real yearly Gross Domestic Product (GDP) growth. After having partialed out time effects (0.47 of the variance in real yearly GDP growth), we show that governors and their administrations are responsible for 2.36% of the variation in real GDP growth, while state-effects only account for 0.77% of the variation in real yearly GDP growth. Our results contradict earlier research suggesting that top-level leadership may not matter (e.g., Salancik & Pfeffer, 1977)
Finally, in my third chapter, I sought to trace the history of management research, critique its development, and provide guidance for its advancement. Isaac Newton once said: “If I have seen further, it is by standing on the shoulders of Giants”. The idea that knowledge advances by building on the foundations laid by the ancients is not new. Building upon the shoulders of these giants requires an awareness of the newest theoretical and empirical developments in our field. In addition, the shoulders of these giants must be sturdy: our theories must be valid, and empirically testable (Rudner, 1966, p. 10) and our empirical tests must be capable of assessing causality (Antonakis, Bendahan, Jacquart, & Lalive, 2010). In this article, I conduct a co-citation analysis on all articles published in the top 50th percentile core- management journals from 1940 to 2022. This analysis allows me to summarize the development of the management literature, identify the major themes that constitute it, as well as the 20 publications that most influenced the field. I use these results to critically assess two of the most influential articles, which are representative of the broader literature, focusing on problems related to theory building and theory testing. I show that, despite being highly influential, these works contain circular theorizing and endogenous explanatory variables barring precise causal conclusions and hindering their ability to advance our knowledge of management phenomena. In all, the results of this analysis call for increased rigor in our efforts to both build and test theories.