Data selection and collection for constructing investor sentiment from social media

Abstract

Research based on investor sentiment in social media has been a hot topic of research in behavioral finance, and the reliability of investor sentiment mined from social media is a potential condition for the reliability of the results of these studies. In the past, scholars have often focused on using more reliable tools to track investor sentiment in order to get more reliable investor sentiment. However, less attention has been paid to another key factor affecting the reliability of investor sentiment on social media: the selection and collection of data. In this study, we systematically investigate the process of data selection and collection in relation to the construction of investor sentiment on social media. Our findings suggest that the process of creating a dataset from social media is a process that starts and ends with a research question. In this process, we need to overcome various obstacles to end up with an imperfect dataset. The researchers must take a series of steps to get close to the best dataset and acknowledge some of the shortcomings and limitations. We emphasize that the absence of accepted, reliable standards makes it particularly important to follow basic principles. This study is an important reference for social media-based behavioral finance research.