Twitter-Inferred Character Strengths in Predicting CEOs' Impact on Firm Outcomes
Exploring CEOs' tweets, our project pioneers the artificial intelligence-based personality assessment. By inferring VIA character strengths from Twitter data, we seek to reveal the CEOs' impact on diverse firm outcomes.
Factsheet
- Lead school Business School
- Institute Institute for New Work
- Research unit Achtsamkeit und Positive Leadership
- Strategic thematic field Thematic field "Humane Digital Transformation"
- Funding organisation SNSF
- Duration (planned) 01.03.2024 - 30.06.2024
- Project management Prof. Dr. Dandan Pang
- Head of project Prof. Dr. Dandan Pang
- Project staff Prof. Dr. Dandan Pang
- Keywords AI, personality, character strengths, CEO, ESG
Situation
Considerable research has consistently highlighted the influential role of executive personal characteristics, especially those of CEOs, in shaping firm outcomes (Cannella et al., 2009; Carpenter et al., 2004). While much of the early work in this area focused on the influence of demographic or other observable executive characteristics such as age and tenure (Hambrick et al., 1996; Miller, 1991; Wiersema & Bantel, 1992), recent attention has shifted towards understanding how executives' psychological attributes, including cognitive and personality traits, affect firm outcomes. However, to assess CEO’s personality using traditional methods such as surveys are often too costly and time-consuming (O’Reilly et al., 2014). In response, an innovative approach to personality assessment has emerged, known as artificial intelligence (AI)-based personality assessment, achieving what is commonly known as content validity (Kosinski et al., 2013; Pang et al., 2020; Park et al., 2015). Therefore, the present project aims to investigate the feasibility of measuring CEOs' personalities, particularly focusing on VIA character strengths, through the analysis of Twitter data. By employing AI-based personality assessment techniques, the project will systematically explore whether Twitter-inferred personality scores can reliably indicate diverse firm outcomes. This forward-looking approach offers a cost-effective and streamlined means of capturing the nuanced personality traits of CEOs.
Course of action
The study's sample frame comprises CEOs of S&P 1500 firms sourced from ExecuComp. Twitter handles, if available, will be collected for the identified CEOs (we aim to collect around 500 twitter handles). Twitter Application Programming Interface (API) will be employed to query the most recent 3,200 tweets from each CEO. Ethical approval will be sought through the submission of an application to the Stanford University Institutional Review Board. The following firm outcomes will be collected through different databases: 1) shareholder returns: this pertains to the returns realized by shareholders from owning a firm's stock; 2) ESG Indicator: The Environmental, Social, and Corporate Governance (ESG) indicator, as assessed by the S&P rating, analyzing the ESG profiles of over 7,300 companies globally.
Result
Regression analysis will serve as a robust method to evaluate the predictive power of Twitter-inferred VIA character strengths of CEOs on various firm outcome metrics. Our hypothesis suggests that specific character strengths, such as creativity, curiosity, judgment, and love of learning, may emerge as superior predictors of financial returns. In contrast, other character strengths like zest, love, kindness, and social intelligence may prove to be more predictive of ESG scores. We will refine the already generated lexica, weighted lists of words derived from language-based assessments, to predict VIA character strengths. These lexica, alongside the software and language models developed for social media data analysis, will be openly shared, offering methodological insights for future research endeavors. The dissemination of our work through high-profile outlets aims to enhance awareness of the crucial link between CEOs' personalities and sustainable firm outcomes, advancing not only the understanding of leadership but also contributing to the methodology of personality assessment. The impact will be gauged through downloads, citations, and the adaptation of our resources by other researchers across diverse fields, underscoring the broad applicability and significance of our contributions to the academic community.
Looking ahead
We will refine the already generated lexica, weighted lists of words derived from language-based assessments, to predict VIA character strengths. These lexica, alongside the software and language models developed for social media data analysis, will be openly shared, offering methodological insights for future research endeavors. The dissemination of our work through high-profile outlets aims to enhance awareness of the crucial link between CEOs' personalities and sustainable firm outcomes, advancing not only the understanding of leadership but also contributing to the methodology of personality assessment. The impact will be gauged through downloads, citations, and the adaptation of our resources by other researchers across diverse fields, underscoring the broad applicability and significance of our contributions to the academic community.
