Skip to main content

Table 5 Sensitivity analysis—spatial probit model (health service use)

From: A bootstrapping approach for generating an inverse distance weight matrix when multiple observations have an identical location in large health surveys

 CoefSECI (lower)CI (higher)  
Women      
 Original dataa      
  splag4.3100.4203.4865.134  
  Wealth− 0.0300.011− 0.052− 0.008  
  Education0.0760.0240.0280.124  
  _cons− 2.3190.248− 2.805− 1.832  
 CoefSEBoot CI (lower)Boot CI (higher)Coverage probability (%)MSE
5000 simulation
 Wealth− 0.0120.029− 0.0700.04594.60.001
 Education0.0680.065− 0.0600.19597.00.004
10,000 simulation
 Wealth− 0.0120.030− 0.0710.04794.40.001
 Education0.0660.065− 0.0610.19396.80.004
 CoefSECI (lower)CI (higher)  
Men
 Original dataa
  splag5.4190.2914.8485.990  
  Wealth− 0.0310.008− 0.046− 0.016  
  Education0.0530.0170.0200.086  
  _cons− 3.0500.187− 3.417− 2.683  
 CoefSEBoot CI (lower)Boot CI (higher)Coverage probability (%)MSE
5000 simulation
 Wealth− 0.0320.033− 0.0970.03296.20.001
 Education0.0570.073− 0.0860.19995.70.005
10,000 simulation
 Wealth− 0.0320.033− 0.0960.03296.30.001
 Education0.0570.074− 0.0880.20195.70.005
  1. aRow normalised