Machine Learning for Pattern Discovery in Management Research

Prithwiraj  Choudhury; Ryan  Allen; Michael G. Endres

Working Paper | HBS Working Paper Series | 2020

Machine Learning for Pattern Discovery in Management Research

by Prithwiraj Choudhury, Ryan Allen and Michael G. Endres

Abstract

Supervised machine learning (ML) methods are a powerful toolkit for discovering robust patterns in quantitative data. The patterns identified by ML could be used as an observation for further inductive or abductive research, but should not be treated as the result of a deductive causal test. To demonstrate the application of ML for pattern discovery, we implement ML algorithms to study employee turnover at a large technology company. We interpret the relationships between variables using partial dependence plots, which uncover surprising patterns that may have gone unnoticed using traditional methods. To guide readers evaluating such pattern discovery, we provide standards for evaluating model performance, illustrate human decisions in the process, and warn of common misinterpretation pitfalls. An online appendix provides code and data to implement the algorithms demonstrated in the paper.

Keywords: Machine learning; theory building; induction; Decision Trees; random forests; k-nearest neighbors; neural network; p-hacking; Data and Data Sets; Analysis;

Language: English

Format: Print

48 pages

SSRN Read Now

Citation:

Choudhury, Prithwiraj, Ryan Allen, and Michael G. Endres. "Machine Learning for Pattern Discovery in Management Research." Harvard Business School Working Paper, No. 19-032, September 2018. (Revised May 2020.)

Working Paper | HBS Working Paper Series | 2020

Food Security and Human Mobility During the COVID-19 Lockdown

Prithwiraj Choudhury, Wesley W. Koo, Xina Li, Nishant Kishore, Satchit Balsari and Tarun Khanna

During the COVID-19 crisis, millions of migrants around the world face food insecurity. This could force migrants to travel during the pandemic, exposing them to health risks and accelerating the spread of the virus. Anecdotal evidence demonstrates the importance of enforcing food security policies to tide the crisis. However, the effects of these policies on containing mobility during the crisis remain unknown. Using mobility data from Facebook, we demonstrate that a policy to guarantee food security that has attracted attention from the Supreme Court of India—portable ration cards—is related to lower mobility during India’s COVID-19 lockdown. Intra-state portable ration cards, which give migrants access to food when they move within their state, are associated with 12% lower intra-state mobility. This effect is particularly strong for states that have fully implemented the policy compared to states that have partially implemented it. However, inter-state portability of ration cards is not related to reduced inter-state mobility, suggesting frictions in implementing the policy across state borders. We also find that food distribution activities by ration shops and civil society actors at the local level are associated with reduced mobility both within states and across state borders. Our study provides generalizable lessons for policymakers around the world: food security policies are essential for helping migrants restrict their travel during the pandemic. Policy implementation requires lead time; therefore, central governments need to coordinate with local actors to increase food distribution to migrants in the immediate term.

Keywords: COVID-19; migrants; food security; mobility; Health Pandemics; Food; Distribution; Policy; Global Range;

Citation:

Choudhury, Prithwiraj, Wesley W. Koo, Xina Li, Nishant Kishore, Satchit Balsari, and Tarun Khanna. "Food Security and Human Mobility During the COVID-19 Lockdown." Harvard Business School Working Paper, No. 20-113, May 2020. View Details

Cite View Details SSRN Read Now

Although previous work has examined how audiences evaluate category-spanning organizations, little is known about the conditions under which such spanners affect audience evaluations of other organizations. We posit that the emergence of atypical categorical combinations, pioneered by startups, signals an altered future state—and seeds doubt about certain incumbents’ future prospects within a re-ordered industry. We test this hypothesis using startup funding announcements as an information shock to investors evaluating incumbents in the financial services sector from 2010 to 2017—a period marked by significant category-spanning entry from non-traditional Fintech startups. We find that announcements by startups that spanned unusual combinations of industry categories led to lower cumulative average returns for incumbents, compared to entry by typical startups. Our theory and results contribute to research on categorization in markets and to theories of disruptive innovation, industry evolution, and technology change.

GitLab is arguably one of the world’s largest “all-remote” companies. Started in 2011 and with more than 1,000 employees at present, it has no physical offices and all employees, including the entire C-Suite, work remotely from all parts of the world. The case highlights the value created for companies and employees from the adoption of the all-remote model. It also sets up a debate for how GitLab should customize organizational processes related to communication, coordination, and socialization as the firm scales up.

Faculty & Research

Machine Learning for Pattern Discovery in Management Research

Abstract

Citation:

About the Authors

More from these Authors