Financial analysts use tweet analytics to prepare their forecasts, yet little information that describes how they do so exists. To address this gap, we scrutinize the associative relationships between tweets about a company’s service and the dispersion of analyst forecasts about the same company’s financial performance. We developed three sets of hypotheses. We extracted tweets related to airlines from the Twitter data from Archive Team and analyst forecast data from Institutional Brokers’ Estimate System Academic. We obtained airline-related tweets from nearly 200,000 individual Twitter users about 10 airlines during a 55-month study period and ran multiple regressions to test the associations between tweet characteristics and forecast dispersion. Our results suggest that, when more posters generate more tweets about a company’s service, analysts make less dispersed forecasts. In addition, negative (or non-verified) tweets reduce forecast dispersion to a greater extent than positive (or verified) tweets do. Theoretically, this paper confirms that Twitter can be a useful data source to provide analysts with additional information to prepare their forecasts. Practically, our findings provide empirical evidence about how Twitter data is associated with analyst forecast dispersion. We encourage stakeholders (such as analysts from small firms and individual investors) to extract data from Twitter as a supplement to market information when analyzing data.
Choi, K. W., Ho, S. Y., & Yang, F. (. (2019). Does Chatting Really Help? Tweet Analytics and Analyst Forecast Dispersion. Communications of the Association for Information Systems, 44, pp-pp. https://doi.org/10.17705/1CAIS.04431