From Policy Guidance to Corporate Online Disclosure: Constructing and Validating a Policy-Guided ESG Textual Measurement Vocabulary
This study constructs and validates a policy-guided ESG textual measurement vocabulary for China's high-impact industries, demonstrating its effectiveness in capturing informal ESG disclosures on corporate WeChat accounts and revealing that social media communication relies heavily on contextual terms and prioritizes Social and Environmental dimensions over Governance.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Imagine the world of business as a giant, noisy marketplace where companies are constantly shouting about how good they are. For a long time, scientists studying these companies only listened to the official, polished speeches—the annual reports and formal sustainability documents. These are like the carefully rehearsed scripts actors read on stage: they are long, structured, and follow strict rules. But in recent years, researchers realized that companies also talk a lot in the "backstage" areas, like social media posts, where they share quick updates, photos of construction sites, and stories about helping neighbors. This is the messy, real-time chatter of daily life. The big question for scientists is: How do we measure what a company is actually doing in these casual chats? If we just use a generic dictionary to count words like "green" or "safe," we might get tricked. It's like trying to understand a complex movie by only counting how many times the word "run" appears; you might miss the plot entirely because "run" could mean running a race, running a business, or running away from a problem. To solve this, researchers need a special, custom-made dictionary that understands the specific language of heavy industries like construction and mining, distinguishing between a company actually fixing a safety issue and one just using fancy words to look good.
This study is about building and testing that very special dictionary for Chinese construction and mining companies. The researchers started with the official rulebooks—the government policy guides and industry standards—to find the "seed" words that truly matter for the environment, society, and governance (ESG). They then used a smart computer model, trained on financial news, to find related words and phrases that these companies actually use in their daily WeChat posts. But they didn't just trust the computer. They acted like strict editors, manually checking thousands of sentences to make sure the words they picked actually meant something real. They split their dictionary into two groups: "Core Terms," which are the obvious, no-doubt-about-it words (like "carbon emission" or "safety training"), and "Contextual Terms," which are words that only make sense as ESG when you look at the whole sentence (like "quality" or "team," which could just be about work, but in this specific industry, often mean safety and responsibility).
The team put their new dictionary to the test in two ways. First, they checked if it could spot real ESG actions in formal reports, finding that it was right about 75% of the time and that two independent human checkers agreed on the results 71.9% of the time. Then, they used it to scan the WeChat posts of 144 companies. They discovered something surprising: if you only used the "Core Terms," you would miss almost 30% of the ESG talk happening on social media. The "Contextual Terms" were essential for catching the real stories. When they looked at what was being said, they found that companies talked most about Social issues (like worker safety and community help), followed by Environmental issues, with Governance being the least mentioned. The study suggests that while companies do follow the official rules in their formal reports, they translate those rules into a more casual, story-telling style on social media, focusing heavily on the people and projects right in front of them. The researchers are careful to say this isn't a magic bullet that solves all measurement problems, but it is a much better, more accurate tool for listening to what these high-impact industries are really saying in their everyday conversations.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.