关键词: 分层空间自回归, 序数Probit, 贝叶斯, 空气质量, 生活满意度

Abstract: In the research of residents′ life satisfaction, more and more empirical studies take into account not only the demographic characteristics but also the macro environment of micro individuals, so a hierarchical model should be built for such data with a special hierarchical structure, but the traditional hierarchical statistical model does not consider the real spatial dependence. In this paper, four hierarchical spatial autoregressive probit models are constructed by combining the hierarchical model and the spatial autoregressive model. The model can be dealed with the ordered dependent variable with spatial dependence in hierarchical data, avoid the adverse effect of ignoring the real spatial dependence, and separate the variable with the spillover effects from the variable without spatial spillover effects, thus more advantageous in the interpretation of the model. The empirical study results show that air quality has a subtle effect on life satisfaction, and residents′ understanding of air quality and the request are not limited to the local place in isolation, but a comprehensive spatial result of the region.Compared with 2018 and 2016 model results, the welfare effect of air quality cannot be replaced by other livelihood welfare factors, and the highly open and wide spread statistical information on air quality helps the residents have a more comprehensive understanding of and attention to air quality.