Skip to main content

Table 4 Individual variable and composite scores by store type

From: Measuring the healthfulness of food retail stores: variations by store type and neighbourhood deprivation

Variable

Premium supermarket

Large supermarket

Discount supermarket

Small supermarket

‘World’ store

Convenience store

Petrol store

Possible range f

 

Median (IQR) d

Min

Max

p-value

Composite score

1.7

1.7

0.3

0.9

-.04

-0.6

-0.9

-1.9

2.2

<0.001a

(1.2 to 1.8)

(1.5 to 1.9)

(0.1 to 1.0)

(0.6 to 1.1)

(-1.1 to 0.3)

(-1.0 to 0.2)

(-1.2 to -0.2)

Variety

8

8

-8

-6

-2

-9

-7

-25

35

<0.001a

(7 to 8)

(6 to 9)

(-10 to 1)

(-8 to -2)

(-5 to 0)

(-11 to -7)

(-8 to -5)

Price

0.03

-0.03

-0.01

0.02

0.04

0.03

0.02

Higher score is more healthful

0.002a

(0 to 0.05)

(-0.04 to -0.01)

(-0.04 to 0.01)

(0.01 to 0.05)

(-0.12 to 0.1)

(-0.06 to 0.08)

(-0.21 to 0.1)

Promotions

0.5

0

-1

-1

0

-1

-1

-5

7

<0.001a

(-1 to 1)

(-1 to 1)

(-3 to 0)

(-2 to -1)

(-0.4 to 0)

(-1 to 0)

(-1 to 0)

Shelf placement

5

7

7

5

0

-3

-5

-15

21

<0.001a

(4 to 7)

(5 to 9)

(4 to 9)

(4 to 7)

(-3 to 4)

(-6 to 3)

(-6 to 1)

Store placement

4

4

4

4

1

-4

-4

-15

21

<0.001a

(4 to 5)

(4 to 5)

(3 to 5)

(3 to 4)

(-4 to 5)

(-6 to 2)

(-6 to 1)

Quality

17

17

17

17

5

5

0

0

18

<0.001b

(17 to 17)

(16 to 18)

(15 to 18)

(15 to 17)

(0 to 10)

(0 to 13)

(0 to 8)

Healthier alternative

4

5

2

3

0

1

1

0

5

<0.001b

(4 to 5)

(5 to 5)

(2 to 3)

(2 to 3)

(0 to 1)

(1 to 2)

(0 to 2)

Nutrition information

13

14

13

15

4

11

9

0

15

<0.001b

(10 to 13)

(12 to 15)

(13 to 14)

(15 to 15)

(2 to 7)

(9 to 13)

(9 to 12)

Single sale of two fruitse

100%

97%

26%

87%

23%

19%

21%

0

2

<0.001c

  1. aOneway analysis of variance, bKruskal-Wallis test, cChi square test, dMedian and inter-quartile range (IQR) were provided for both parametric and non-parametric variables for ease of reading, ePercentage of two fruits available for single sale was provided because this variable was categorical, fPossible range of scores for each variable except composite score which shows actual range of composite score values.