Detecting village‐level regional development differences: A GIS and HLM method

To respond to the problems that the previous research mainly targeted the poverty at larger scale and ignored individual effect or contextual effect during exploring poverty contributing factors, we attempt to use spatial cluster analysis and multilevel linear regression model to target the poverty at village level from the perspective of spatial poverty, so as to identify where the poor villages are, and why they are so poor, thereby targeting poverty interventions. Specially, we adopt four types of spatial cluster indices to detect the spatial aggregation distribution of villages, and design HLM model to examine the poverty contributing factors from both village level and county level. The case study from Wuling contiguous destitute area show that: (1) The overall distribution shows a spatial pattern of large scattered but small concentration, scatters‐polar core‐axis‐clump coexisted. (2) Poverty contributing factors at village level from high to low are: per cultivated area, safe drinking water access ratio, terrain type, suffered frequency of natural disasters, road access ratio, and distance from the nearest town’s bazaar. The contribution degree of county‐level factors to the villages’ poverty from high to low are: second gross enrollment ratio, per capita GDP. (3) 45.1% of the difference among the villages’ poverty degrees comes from the development differences among poverty‐stricken villages themselves, and 54.9% from that among counties they belong to. Contributing factors at village level account for 61.4% of the variation of village‐level independent variables, and factors at county level contributed to 65.3% of the variation of county‐level independent variables.