Table 2.

Correlation matrix between all the variables used in the study

	Obesity	Log(Dist)	Income	Density	Rurality	Ethnicity	EL	UL
Obesity (%)†	1·000
Log(Dist)‡	−0·100*	1·000
Income§	−0·561*	0·113*	1·000
Density||	0·081*	−0·630*	0·052*	1·000
Rurality¶,**	−0·078*	0·701*	0·077*	0·449*	1·000
% Ethnicity‡‡	0·008	0·431*	0·074*	0·655*	0·288*	1·000
% EL§§	0·565*	−0·107*	0·816*	0·072*	0·111*	−0·044*	1·000
% UL||||	0·406*	−0·015	0·732*	0·153*	0·011	−0·184*	0·709*	1·000

UL = unemployment level, EL = education level, MSOA = middle super output area.

^*P < 0·01.

^†Proportion of overweight children (incl. obese) for MSOA 2013–16 (averaged) and collapsed to MSOA level.

^‡Road distance from postcode centroid to the nearest supermarket, the variable was log-transformed.

^§Total annual household income.

^||Number of persons per hectare.

^¶Urban–rural classification (0-urban, 1-rural).

^**Point-biserial correlation between the dummies and the other continuous variables.

^‡‡Proportion of households from the ethnic minority groups.

^§§Proportion of households with no qualification.

^||||Proportion of households with adults not in employment.