Skip to main content

Table 5 Fixed and random parameters of the social support multilevel models

From: Neighborhood differences in social capital in Ghent (Belgium): a multilevel approach

 

Model 0

Model 1a

Model 1b

Model 2a

Model 2b

 

β (SE)

β (SE)

β (SE)

β (SE)

β (SE)

Fixed effects

     

Constant

11.691 (0.169)***

12.465 (0.207)***

12.317 (0.160)***

12.382 (0.190)***

12.193 (0.117)***

Level 1

     

Female

 

−0.069 (0.191)

 

−0.064 (0.190)

 

65 years or older

 

−0.183 (0.246)

 

−0.197 (0.244)

 

Low education level

 

−0.102 (0.273)

 

−0.052 (0.266)

 

Single

 

−0.462 (0.221)*

−0.464 (0.217)*

−0.350 (0.219)

 

Not Belgian

 

0.153 (0.312)

 

0.336 (0.313)

 

Weak neighborhood attachment

 

−1.816 (0.224)***

−1.839 (0.220)***

−1.769 (0.219)***

−1.787 (0.216)***

<5 years of residence

 

−0.276 (0.224)

 

−0.149 (0.223)

 

Level 2

     

Deprived neighborhood

   

−0.115 (0.288)

 

Percentage of elderly

   

−0.175 (0.107)

 

Turnover

   

−0.567 (0.133)**

−0.618 (0.102)***

Physical disorder

   

−0.194 (0.134)

 

Random effects

     

Level 1

     

Constant

7.163 (0.378)***

6.707 (0.355)***

6.688 (0.353)***

6.677 (0.344)***

6.719 (0.354)***

Level 2

     

Constant

0.794 (0.261)**

0.381 (0.166)*

0.410 (0.172)*

0.000 (0.000)

0.055 (0.095)

Intraclass correlation

0.100

0.054

0.058

0.000

0.008

Log likelihood

3708,957

3608,963

3632,345

3576,124

3615,077

∆ Log likelihood (∆ df)

 

99,994

76,612

132,833

93,880

P

 

< 0.001

< 0.001

< 0.001

< 0.001

  1. Note: *p < 0.05, **p < 0.01, ***p < 0.001.
  2. Model 0: null model without level 1 and level 2 variables; Model 1a: model with only level 1 variables; Model 1b: parsimonious model with only level 1 variables; Model 2a: model with level 1 and level 2 variables; Model 2b: parsimonious model with level 1 and level 2 variables.