This article presents a three-step ISE selection procedure to identify and score companies based on their initiatives for social and environmental responsibility. The procedure involves a deep-learning framework to detect ISEs from reviews, a fixed generalized threshold of 0.31, and an ISE-specific threshold of 95% percentile of the ISE’s distribution. The article also discusses the relationship between a company’s ISE scores and its success in the forms of company ratings and stock growth, as well as the ISE scores variability across industry sectors.
