Skip to main content

Table 3 Results using footpath-accessible network and street-side buffers

From: Characterising urban green space density and footpath-accessibility in models of BMI

  Network BufferStreet-side Buffer
  (1)(2)3)(4)(5)(6)(7)(8)
  Marginal Effect
(SE)
Marginal Effect
(SE)
Marginal Effect
(SE)
Marginal Effect
(SE)
Marginal Effect
(SE)
Marginal Effect
(SE)
Marginal Effect
(SE)
Marginal Effect
(SE)
Share of Footpath-accessible Green Space
1600 m Network BufferNon-Urban Settlement0.478*
(0.272)
0.722***
(0.238)
      
Quintile 1A (lowest)0.800**
(0.325)
0.722**
(0.323)
      
Quintile 2A0.533*
(0.298)
0.486
(0.299)
      
Quintile 3A[ref.][ref.]      
Quintile 4A0.366
(0.320)
0.435
(0.320)
      
Quintile 5A (highest)0.452
(0.319)
0.532*
(0.318)
      
800 m Network BufferNon-Urban Settlement  0.0764
(0.259)
0.269
(0.226)
    
Quintile 1B (lowest)  0.166
(0.308)
0.128
(0.306)
    
Quintile 2B  −0.199
(0.300)
−0.226
(0.298)
    
Quintile 3B  [ref.][ref.]    
Quintile 4B  −0.0724
(0.311)
−0.0200
(0.312)
    
Quintile 5B (highest)  −0.170
(0.315)
−0.0739
(0.313)
    
1600 m Street-side BufferNon-Urban Settlement    0.285
(0.278)
0.466*
(0.240)
  
Quintile 1C (lowest)    0.141
(0.329)
0.0909
(0.322)
  
Quintile 2C    0.195
(0.302)
0.201
(0.301)
  
Quintile 3C    [ref.][ref.]  
Quintile 4C    0.0848
(0.315)
0.148
(0.312)
  
Quintile 5C (highest)    0.334
(0.338)
0.449
(0.331)
  
800 m Street-side BufferNon-Urban Settlement      0.256
(0.253)
0.427**
(0.214)
Quintile 1D (lowest)      0.215
(0.300)
0.193
(0.297)
Quintile 2D      −0.0314
(0.294)
−0.0408
(0.294)
Quintile 3D      [ref.][ref.]
Quintile 4D      0.0550
(0.300)
0.128
(0.300)
Quintile 5D (highest)      0.190
(0.321)
0.309
(0.311)
N 58045807580458075804580758045807
  1. Standard errors in parentheses. * p < 0.1 ** p < 0.05 *** p < 0.01.
  2. The results in column (1) refer to the results of the full model with green space footpath-accessibility expressed in terms of a 1600 m network. The results in column (2) are those of the parsimonious specification. The parsimonious models have three more observations due to the medical insurance variable being dropped. The results in Column (3) are from full model with green space footpath-accessibility expressed in terms of an 800 m network buffer, while the results in Column (4) refer to the more parsimonious specification of the models. Columns (5) to (8) are the equivalent models using green space footpath-accessibility using street-side buffers.