Skip to main content

Table 2 Results of the balancing check for the generalized propensity score (GPS)

From: Assessing short-term impact of PM10 on mortality using a semiparametric generalized propensity score approach

Covariate

Class of exposure

Marginal t

GPS-adjusted t

Temperature (lag 0-3)

(0, 20)

5.60

-2.57

 

[20, 40)

14.15

1.68

 

[40, 70)

-0.39

-0.26

 

[70+)

-20.78

-3.00

Influenza epidemics

(0, 20)

-2.80

-0.44

 

[20, 40)

-7.47

-2.87

 

[40, 70)

1.32

3.43

 

[70+)

8.92

0.47

Summer indicator

(0, 20)

7.93

-2.19

 

[20, 40)

14.30

2.39

 

[40, 70)

-3.09

-0.77

 

[70+)

-18.46

-2.48

Holidays indicator

(0, 20)

1.91

-0.07

 

[20, 40)

0.02

-0.05

 

[40, 70)

-2.74

-2.81

 

[70+)

1.97

2.58

Humidity (lag 0)

(0, 20)

-5.78

-0.13

 

[20, 40)

-5.91

1.05

 

[40, 70)

-0.15

-2.78

 

[70+)

10.69

1.78

Weekend indicator

(0, 20)

4.92

2.27

 

[20, 40)

1.79

-1.33

 

[40, 70)

-1.39

-1.13

 

[70+)

-3.35

0.64

Heat episodes

(0, 20)

-2.25

-0.96

 

[20, 40)

1.77

-0.15

 

[40, 70)

3.48

2.43

 

[70+)

-4.73

-1.99

  1. Marginal and GPS-adjusted t statistics for the mean difference between one class of exposure and the others as a whole, for a set of selected covariates