Application Research Based On Mediation Mixed Model

It is of great practical significance to find out the potential subgroups in medi-ation analysis for solving social problems.This paper focuses on people's psycholo-gy and environmental protection,by summarizing and comparing the classification methods of mediation analysis,we choose the linear structural equation model with latent subgroup structure to study,and introduce it.This model is used to conduct empirical analysis on the relevant data of the above two social problems,reveal the interaction between variables,expand the application range of the linear structural equation model with latent subgroup structure.The main conclusions are as follows:First,for the experimental research data of the JOBS intervention project(JOBS II),we first use stepwise regression to determine that job-seek as an intermediate variable does have mediation effect.Secondly,using the linear structural equation model with latent subgroup structure,establish and analyze a mixed model contain-ing only one confounding factor(econ-hard)and mixed model with two confounding factors(econ-hard?sex),we got 3 subgroups with different mediating effects,they are positive mediation effect(2.9,the proportion is about 30%),no mediation effect(0,the proportion is about 30%)and negative mediation effect(-3.0,the proportion is about 40%),the mediating effect shows heterogeneity in the group.In short,job hunting as an mediation effect has different impacts on different groups of people.On the one hand,job hunting can inhibit people's psychological pressure after job training,on the other hand,job hunting may not have any impact on some people,job hunting can even increase people's psychological pressure after receiving training and affect their health.Second,analyzing data on the impact of financial development on carbon emis-sions(fd-CI),we test whether there is a mediating effect,and the regression results show that technological progress(tech)as an intermediate variable does have a mediating effect.Second,use the linear structural equation model with latent sub-group structure for analysis,simplifies the research process of previous researchers,reduces workload.The result is the same as before,the mediating effect of techno-logical progress is heterogeneous.There are three potential subgroups with different mediating effects(corresponding to 53.52,2.54,-43.66,respectively).Combining the distribution characteristics of the three major regions of my country's east,middle and west,we can classify the subgroups with difFerent mediation effects into the three major regions of our country.Technological progress in the central region,as an mediation effect,has obviously inhibited the effect of financial development on carbon emissions;Although the financial development of the eastern and west-ern regions promotes carbon emissions,they are different.Because in the eastern region,the improvement of technology has alleviated the promotion of financial development on carbon emissions;while the technological progress in the western region has increased the effect of financial development on the increase of carbon emissions.