This academic article investigates how user sentiment on Twitter influences the stock market performance of 24 major global companies. By utilizing artificial intelligence for data scraping and natural language processing, the researchers constructed time series to analyze the impact of news. The study focuses on the “asymmetry of information,” testing whether negative sentiment has a more significant effect on stock prices than positive sentiment, based on traditional economic theories of loss aversion.

The methodology involved three sophisticated stages: extracting millions of “top tweets” related to company tickers (such as $TSLA or $AMZN) over a ten-year period, calculating sentiment polarity through NLP, and applying advanced statistical models. Specifically, the authors used Effective Transfer Entropy (ETE) and EGARCH models to measure information flow and volatility, allowing them to isolate and compare the independent signals of positive and negative social media sentiment.

The findings confirm a strong asymmetric response, showing that negative news impacts daily stock prices more intensely and frequently than positive news. The results demonstrate that negative sentiment not only affects the specific company mentioned but also tends to spill over into the broader market. Ultimately, the research validates that investors react more strongly to pessimistic signals on social media, providing a deeper understanding of the relationship between digital public opinion and financial market behavior.

Download the full report