Model 1 | Model 2 | Model 3 | Model 4 | |||||||||||

Univariable logistic regression for presence of retailer | Univariable linear regression for log of retailer density | Multivariable logistic regression for presence of retailer | Multivariable linear regression for log of retailer density | |||||||||||

Variable | Estimate | 95% CI | P value | Estimate | 95% CI | P value | Cohens f | Estimate | 95% CI | P value | Estimate | 95% CI | P value | Cohens f |

Median income, €1000 | 0.866 | 0.834 to 0.9 | <0.001 | 0.907 | 0.896 to 0.918 | <0.001 | 0.496 | 0.542 | 0.481 to 0.608 | <0.001 | 0.956 | 0.929 to 0.984 | 0.002 | 0.506 |

% in lowest income category | 1.071 | 1.041 to 1.103 | <0.001 | 1.059 | 1.051 to 1.068 | <0.001 | 0.422 | 0.884 | 0.836 to 0.933 | <0.001 | 1.028 | 1.014 to 1.042 | <0.001 | 0.068 |

% unemployed | 1.186 | 1.142 to 1.234 | <0.001 | 1.037 | 1.027 to 1.047 | <0.001 | 0.214 | 1.063 | 1.01 to 1.12 | 0.020 | 0.976 | 0.964 to 0.987 | <0.001 | 0.117 |

% with higher education | 1.020 | 1.007 to 1.033 | 0.002 | 0.974 | 0.971 to 0.978 | <0.001 | 0.433 | 1.139 | 1.102 to 1.178 | <0.001 | 0.982 | 0.976 to 0.989 | <0.001 | 0.144 |

Population density, 1000 s per square kilometre | 2.903 | 1.843 to 4.909 | <0.001 | 1.033 | 1.004 to 1.063 | 0.023 | 0.068 |

Estimates are exponentiated and can be interpreted as multiplicative differences in OR and density, given 1 unit change in predictor.