Abstract
BACKGROUND There is a lot of health science information in social media. But most of the information lacks persuasive effect. Text features have a great influence on the information. Therefore, the text features of health science information in social media are studied. OBJECTIVE To investigate the text feature of health science information on social media and suggest ways to develop health science information on the basis of the research conclusions. METHODS Using Web crawler to collect the health science information data from WeChat official account, 4,645 myopia prevention texts of children in WeChat were collected, and 813 valid texts were included in the study. Through the analysis of text features, the text sentiment and the use of statistical evidence of health science information were extracted. RESULTS Results show that the use of commendatory words is obviously more than derogatory words (P < 0.01) and statistical evidence was used in over 96% of the text. Suggestions are put forward for the development of health science information in terms of text sentiment and statistical evidence. CONCLUSIONS This study investigated the emotional features of text and statistical evidence use features of health science information in social media. It is found that health information makers prefer positive emotions and use statistical evidence in most texts. Combined with previous research findings, we recommend greater usage of positive emotions and less of statistical evidence when formulating health information.