How Do Online Public Messages Affect Local Government Responsiveness in China? A Multilevel Analysis Based on the Message Board for Leaders

MBFL, an online platform, was founded in 2006 by the People’s Network for citizens to express their demands, problems, and suggestions for local political leaderships, allowing for the local leading cadres and local message processing units to handle and reply. People’s Network MBFL, website: http://​liuyan.​people.​com.​cn/​index.​html

The country’s multi-provincial institutionalization for netizens to leave messages, website: http://​leaders.​people.​com.​cn/​GB/​178291/​218130/​370359/​index.​html

There are 19 municipalities directly under the central government and sub-provincial cities in China, including 4 municipalities directly under the central government and 15 sub-provincial cities. However, due to the limitations of the capacity of the People’s Daily Online Leaders’ Message website, only up to 6,000 citizens’ messages to mayors or municipal party secretaries are displayed for each city, among which Beijing, Tianjin, Chengdu and Xi’an have a large number of messages. The limited capacity of the website also leads to the fact that no data from before 2017 are displayed for these four cities. In addition, the data collation showed that Jinan and Ningbo had 904 and 170 messages, respectively, in eight years from 2011 to 2017, but the number of government responses was zero, so these two cities were not included in the data analysis to ensure the validity of the statistical analysis of the data. After all exclusions, data from 13 cities were finally retained.

The specific treatments are as follows. First, “message reply,” “message processing” and other key words were searched directly through the People’s Network (people.cn) or the local government official website to establish whether a local-network message processing system is introduced to municipalities. Then, documents posted on the MBFL are checked to determine whether they are related to handling of messages. Finally, variable processing is carried out, in which the institutional variable for the city for that year and subsequent years is recorded as 1 if the city has the relevant official document; otherwise it is recorded as 0.

Shrout (1979) noted that the definition of ICC is equal to the ratio of the variance between groups and total variance. The formula of the total variance is $${VAR={\tau }_{00}+\sigma }^{2}$$, where $${\tau }_{00}$$ refers to the variance of the second-layer variable, and $${\sigma }^{2}$$ is the variance of the first-layer variable. Therefore, the proportion of the total variance that is explained by the variation of the second-layer variable is calculated as follows: