Analyzing the distribution of researchers in China: An approach using multiscale geographically weighted regression

Abstract

We employ a Multiscale Geographically Weighted Regression (MGWR) model to examine the spatial variation of researchers in China in 2015 and its determinants. It is found that the distribution of researchers is driven by the economy in urban centers, public services, natural areas of recreation, urban consumption, and work‐related facilities. Results from the MGWR model conclusively identify significant spatial non‐stationarity in the determinants measuring scientific researchers' distribution. Other factors such as per capita GDP, the number of hospital beds per 10,000 inhabitants, and public financial expenditures influence the distribution of researchers more in the Western cities, and their effect decreases steadily from West to East. Air pollutants affect researchers negatively in the eastern part of the country, while the percentage of available green land is a strong driver in the central and western parts. The factor measuring the effect of the number of universities is seen to affect researchers more strongly in the northwestern and northeastern cities. Other findings further suggest that the economy and public services in cities have a consistent effect on the distribution of researchers across the country whereas, natural areas of recreation, work‐related facilities, and urban consumption have a significantly varying effect across cities.

Administrative level, city tier, and air quality: contextual determinants of hukou conversion for migrants in China

“…Significant coefficients are those which pass the significance test at 0.05 level settlement intention usually used in China (Chen & Fan, 2016;Hao & Tang, 2015;Li & Liu, 2019;. Following previous studies using CMDS data (Gu, Jie, et al, 2020;Gu, Yu, et al, 2020;, settlement intention in this paper refers to migrants who would like to stay for a long duration (more than 5 years) in the destination. A different definition of settlement intention may produce different results.…”

“…This research attempts to capture the geographically varying relationships between contextual factors and migrants' willingness to hukou conversion at the city level. Compared with the global regression focusing on spatial homogeneity, MGWR (Multiple Geographically Weighted Regression) could capture the spatial heterogeneity and improve data fitting (Fotheringham, Yang, & Kang, 2017;Oshan, Li, Kang, Wolf, & Fotheringham, 2019), which has been applied to the study of migration pattern and migrants' hukou transfer intention in previous studies (Gu, Yu, Sachdeva, & Liu, 2020;Lao & Gu, 2020). The Traditional Geographically Weighted Regression (GWR) is a spatial regression technique to evaluate a local model of the process and model spatially nonstationary relationships (Fotheringham, Brunsdon, & Charlton, 2002).…”

Multiscale Impact of Environmental and Socio-Economic Factors on Low Physical Fitness among Chinese Adolescents and Regionalized Coping Strategies

“…Compared to GWR, MGWR has three main important improvements. First, allowing different levels of spatial smoothing for each covariate addresses the shortcomings of geographically weighted regression models [ 33 , 75 ]. Second, these covariate-specific bandwidths can be used as indicators of the spatial scale at which each spatial process acts.…”

Spatiotemporal Dynamic of COVID-19 Diffusion in China: A Dynamic Spatial Autoregressive Model Analysis

“…Compared with other spatiotemporal analysis methods, the DSAR model is more explanatory in mechanism analysis. Methods for detecting spatial heterogeneity have also been used in the analysis of COVID-19, such as multi-scale geographically weighted regression [10,[43][44][45][46][47][48]. However, multi-scale geographically weighted regression currently has no expansion of time-space analysis.…”