Abstract: Owing to the extensive use of online media and the fragmentation of its information and its other characteristics, crisis information has been sharing widely under the influence of herd effect after the brand crisis appears. In order to grasp the herd effect factors and action mechanism of brand crisis information sharing behavior, and formulate relatively effective monitoring and prevention measure, the impulse response and marginal impact of conformity situation representation and brand crisis information sharing behavior in cyberspace is studied in the paper. Sina Weibo in China is taken as the online platform research sample, 66 crisis events covering brands of seven industries such as food and medicine, which occurred between 2010 and 2019, are selected as the source of brand crisis information， and the official API and web crawler method are used for data collection to obtain 376492 combination of valid data on the total number of reposts, comments, and the number of reposts and comments. Vector autoregression (VAR) and state space model (Sspace) are used to analyze the impulse response and marginal impact between conformity situation representation and brand crisis information sharing behavior. It is found that: (1) There is significant positive causal relationship between conformity situation representation and information sharing behavior in cyberspace; (2) More obvious lagging effect is exerted by conformity situation representation on comment behavior than on reposting behavior, but there is a long lag period on reposting behavior,that is, there is a long tail effect; (3) Greater influence is exerted by the total number of information reposts than by the total number of comments on the user’s reposting behavior, greater influence is exerted by the total number of information comments than by the total number of reposts on the user’s commenting behavior,that is, there is greater herd effect between similar behavior;(4)There are structural differences in the herd effects of brand crisis information sharing behaviors among different fields and among different years.Through model construction and analysis, it is shown that the network conformity situation representation in the brand crisis has a long-tailed volatility effect on the crisis information sharing behavior, which can provide a theoretical basis for the discrimination, classification and handling of the network user’s conformity behavior, thus preventing the conformity situation representation from promoting repost and comment behavior which can lead to the rapid spread of crisis information and the formation of public opinion risks.

Key words: brand crisis； conformity context； information sharing； impulse response； marginal influence