Skip to main content

Table 5 Cross classification by tertiles of food outlets between Google data and DENUE: by socioeconomic deprivation and municipality, 2021

From: Agreement between a web collaborative dataset and an administrative dataset to assess the retail food environment in Mexico

Density of food outlets (counts/km2)

n

Correctly classified (%)a

Misclassified (Google data underestimates) (%)b

Misclassified (Google data overestimates) (%)c

Weighted kappa (SE)

Total census tracts

 Socioeconomic deprivation

  Very low

474

60.8

4.4

34.8

0.52 (0.03)

  Low

413

76.5

9.7

13.8

0.63 (0.04)

  Medium

277

69.3

24.2

6.5

0.55 (0.05)

  High

154

59.1

36.4

4.5

0.46 (0.05)

  Very high

134

65.7

32.1

2.2

0.42 (0.06)

Hermosillo, Sonora

 Socioeconomic deprivation

  Very low

276

73.2

6.2

20.7

0.66 (0.05)

  Low

142

67.6

19.7

12.7

0.49 (0.07)

  Medium

41

56.1

34.1

9.8

0.29 (0.11)

  High

25

28.0

68.0

4.0

0.18 (0.10)

  Very high

46

50.0

37.0

13.0

0.24 (0.11)

Leon, Guanajuato

 Socioeconomic deprivation

  Very low

128

60.2

2.3

37.5

0.50 (0.06)

  Low

144

75.7

10.4

13.9

0.64 (0.07)

  Medium

165

79.4

13.9

6.7

0.67 (0.06)

  High

85

75.3

17.6

7.1

0.64 (0.08)

  Very high

70

70.0

28.6

1.4

0.45 (0.10)

Tlalpan, Mexico City

 Socioeconomic deprivation

  Very low

59

67.8

3.4

28.8

0.63 (0.10)

  Low

86

72.1

15.1

12.8

0.60 (0.09)

  Medium

27

88.9

11.1

0.0

0.87 (0.16)

  High

20

40.0

55.0

5.0

0.12 (0.13)

  Very highd

-

-

-

-

-

Oaxaca de Juarez, Oaxaca

 Socioeconomic deprivation

  Very low

11

63.6

9.1

27.3

0.55 (0.23)

  Low

41

70.7

4.9

24.4

0.55 (0.12)

  Medium

44

43.2

29.5

27.3

0.20 (0.11)

  High

24

60.5

37.5

0.0

0.47 (0.14)

  Very high

12

75.0

16.7

8.3

0.40 (0.28)

  1. Sample size: Total census tracts (n = 1452); Hermosillo, Sonora (n = 530); Leon, Guanajuato (n = 592); Tlalpan, Mexico City (n = 192); Oaxaca de Juarez, Oaxaca (n = 132)
  2. DENUE National Statistical Directory of Economic Units, SE Standard Error
  3. aCorrectly classified = % of census tracts with density of food outlets of DENUE and Google data in the same tertile
  4. bMisclassified (Google data underestimates) = % of census tracts adjacently or extreme opposite tertile classified when Google data underestimates
  5. cMisclassified (Google data overestimates) = % of census tracts adjacently or extreme opposite tertile classified when Google data overestimates
  6. d Too few sample size (n = 6) to rating categories