Skip to main content

Table 4 Bivariate regression

From: The modifiable areal unit problem (MAUP) in the relationship between exposure to NO2 and respiratory health

 

Natural Neighbourhood (n = 95)

Aggregated (n = 95)

Census Tract (n = 184)

 

Bivariate RESP

Bivariate RESP

Bivariate RESP

R-squared

0.0307

0.0494

0.0492

Adjusted R-squared

0.0202

0.0392

0.0440

Sum squared residual

55810400.0000

118990000.0000

424291000.0000

Sigma-square

600112.0000

1279470.0000

2330000.0000

S.E. of regression

774.6690

1131.1400

1526.8500

Sigma-square ML

587478.0000

1252530.0000

2305000.0000

S.E of regression ML

766.4710

1119.1700

1518.5300

F-statistic

2.9462

4.8363

9.4154

Prob(F-statistic)

0.0894

0.0303

0.0025

Log likelihood

-765.7700

-801.7310

-1608.9800

Akaike

1535.5400

1607.4600

3221.9500

Schwarz criterion

1540.6500

1612.5700

3228.3800

Intercept coefficient

869.8488 (0.0007)

1634.0250 (< 0.0001)

1379.7439 (< 0.0001)

Mean NO 2 coefficient

76.2276 (0.0894)

134.3078 (0.0303)

179.136 (0.0024)

Jarque-Bera

6.3023 (0.0428)

26.1702 (< 0.0001)

923.152 (< 0.0001)

Breusch-Pagan Test

5.1764 (0.0228)

1.5039 (0.2200)

0.1623 (0.6870)

Koenker-Bassett test

4.1263 (0.0422)

1.0065 (0.3157)

0.0270 (0.8693)

Moran's I (error)

-0.0039 (0.8723)

0.2454 (< 0.0001)

0.1080 (0.0096)

Lagrange Multiplier (lag)

0.0003 (0.9842)

12.2297 (0.0004)

5.4663 (0.0193)

Robust LM (lag)

0.0112 (0.9155)

0.1264 (0.7221)

0.0668 (0.7959)

Lagrange Multiplier (error)

0.0025 (0.9599)

1.0239 (0.3115)

5.4253 (0.0198)

Robust LM (error)

0.0133 (0.9079)

13.2537 (0.0013)

0.0257 (0.8724)

  1. Results of the bivariate regressions (OLS) between the Mean NO 2 concentration (mean NO2) and the Respiratory health outcome rate (RESP) for the three spatial structures under study. p-values are in parentheses.