Skip to main content

Table 2 Moran’s I test for spatial autocorrelation for alternative models

From: Urban sprawl, obesity, and cancer mortality in the United States: cross-sectional analysis and methodological challenges

Covariates included in models

All cancers

Obesity related cancers

 

Male

Female

Male

Female

 

I

p

I

p

I

p

I

p

1. ln sprawl

0.099

≤0.0001

0.049

≤0.0001

0.067

≤0.0001

0.039

≤0.0001

2. ln sprawl + other covariates

−0.005

0.999

0.004

0.005

−0.006

0.999

−0.0001

0.154

3. ln sprawl + other covariates + census divisions

−0.003

0.992

0.001

0.051

−0.001

0.487

−0.0006

0.231

4. ln sprawl + other covariates + census divisions

        

+ selected interactions

−0.003

0.986

0.002

0.022

−0.002

0.669

0.002

0.03

5. ln sprawl + other covariates + census divisions

        

+ selected interactions + county centroids

−0.002

0.91

−0.002

0.588

−0.002

0.687

−0.001

0.436

6. ln sprawl + other covariates + census divisions

        

+ selected interactions + county centroids

        

+ ln sprawl * census division interaction

−0.002

0.827

−0.002

0.792

−0.002

0.687

−0.002

0.767

  1. Values in bold are associated with significant Moran’s I, p < 0.05.