Table 2 Psychometric properties of measures included in the review

Boles et al., 2014, US [30]

Availability

Kappa 0.74-0.84 between independent raters

Parental reports compared to direct observation by trained independent raters - lower value for kappa range varied between 0.029 and 0.336 and upper value was 1.00, 62/131 items found to have kappa less than 0.61 and were removed % agreement varied between 92% and 100% for food categories for which it was computed (n = 4)

Dewar et al., 2012, US [31]

Availability

Chronbach’s alpha: 0.79, factor loadings (0.44–0.86)

ICC (measurement 2 weeks apart): 0.81 (0.75–0.86)

Confirmatory factor analysis performed with fit indices that were a good or exact fit of the hypothesized model - RMSEA-0.00, CFI-1.00, GFI-1.00, AGFI-0.99

Benarroch et al., 2011, Spain [32]

Accessibility

Factor loadings of 0.40–0.68

Correlations with dietary behaviors: 0.085 (cheese)–0.248 (vegetables)

Vyduna et al., 2016, US [33]

Availability

Chronbach’s alpha 0.71

Pearson r (2 weeks apart): 0.80

Petty et al., 2013, Brazil [34]

Availability

Chronbach’s alpha:0.69 Factor loading 0.70–0.78

Pearson r (2 weeks apart): 0.80

Convergent validity using pearson r: 0.60 (other parent living with the family was asked to answer questions according to how they believed their partner answered the questionnaire previously) significant associations between availability measure and the consumption of fruits, vegetables, soft drinks and sweets (B = -0.92-6.23 for association with weekly frequency of consumption of foods)

Hearst et al., 2012, US [35]

Availability and accessibility

Criterion validity assessed comparing family reports with those of staff scores kappa for availability: 0.16 to 0.85 (2 items with kappa <0.3); spearman r: 0.20–0.88 (3 items with r <0.3) kappa for accessibility ranged from 0.26 to 0.51 (2 items with kappa <0.3); spearman r ranged from 0.25 to 0.52 (3 items with kappa <0.3), Kappa for obesogenic food availability score was 0.57, spearman r was 0.78

Rimkus et al., 2013, US [36]

Availability

Average kappa was 0.84, proportion of overall agreement was 0.95, ICC was 0.82 for product availability, 52 of the measures (93%) had a kappa ≥0.61, 2 measures had kappa <0.4

Nepper et al., 2014, US [37]

Availability and accessibility

HFA instrument: for healthy and unhealthy foods and beverages: chronbach’s alpha 0.94 for healthy, 0.91 for unhealthy and 0.90 for total items; for fruit and vegetable items alpha was 0.82 and 0.80. HFES instrument: chronbach’s alpha for healthy and unhealthy food items: 0.83

HFES only (availability), 1 week apart: ICC for unhealthy foods and beverages: 0.79–0.96, ICC for healthy foods and beverages: 0.07–0.93 (all except one item had ICC >0.4)

HFA instrument only: assessed comparing parental report to those of research staff. healthy and unhealthy foods and beverages: for availability, kappa 0.08–1.00, 2 items had kappa <0.3; for accessibility, kappa -0.02 to 1.00, 3 items had kapp <0.3; for fresh fruits, kappa ranged from 0.41–1.00 for availability and 0.38–1.00 for accessibility; for fresh vegetables, kappa ranged between 0.42 and 1.00 for availability and between 0.24 and 1.00 for accessibility, 2 items had kappa <0.3

Boles et al., 2013, US [38]

Availability and accessibility

For healthy and unhealthy foods and beverages: kappa ranged between -0.07 and 1.00, 4 items had kappa values less than 0.60 for availability; kappa ranged between -0.07 and 1 for accessibility, with 1 item with kappa less than 0.60; items with inadequate kappa were subsequently removed; fresh fruits and fresh vegetables: all items had kappa values >0.60

HHE instruments were examined between obese and non-obese children: healthy and unhealthy food items: there was no difference between groups; fruits and vegetables: families of obese preschoolers were significantly less likely to have fresh vegetables accessible in the home compared with healthy weight families; families of obese preschoolers were significantly less likely to have fresh vegetables available compared with healthy weight families;

Nathan et al., 2013, Australia [39]

Availability

Results compared to observations made by pre-service teachers

Percent agreement ranged from 52 to 95 (3 items with % agreement <60), kappa ranged from 0.02 to 0.9, with 13/26 food items having kappa values >0.6, 6 items had kappa <0.3; food sold through fundraisers had higher kappa values than food sold in canteens; no vending machines were reported in the schools, so percentage agreement and kappa scores were 100% for these items

Lee et al., 2014, US [40]

Availability

Compared to direct observation by trained observers (based on 175 meals served, snack consumption of 528 children), for weekly OSNAP-OPAT, r = 0.43 (whole grains) - 0.84 (fruit and vegetable), for daily OSNAP-OPAT, r = 0.32–0.66

Dodds et al., 2014, Australia [41]

Availability

Compared to direct observation by trained research assistants; kappa 0.45 to 1.00; percentage agreement 73% to 100%

Izumi et al, 2014, US [42]

Availability

Kappa values ranged from 0.48 to 1.00, two items had kappa of < 0.60

Krukowski et al., 2011, US [43]

Availability

% agreement values ranged from 0.38 to 0.99 for school lunch menu assessment with 1 value <0.60, and from 0.41 to 0.95 for observation of foods offered in cafeteria during lunch, with 1 value <0.60

Ding Ding et al., 2012, US [44]

Availability

Chronbach’s alpha ranged from 0.60 to 0.75 for adolescent report, 0.65 to 0.73 for parent reports for adolescents and from 0.40 to 0.67 for parent report for children (1 item had alpha less than 0.6)

ICC (measurement 2–4 weeks apart) ranged from 0.47 to 0.58 for adolescent report, 0.72 to 0.78 for parent reports for adolescents and from 0.70 to 0.88 for parent report for children

Association with fruit and vegetable intake using partial correlations: -0.18–0.31 for adolescents, -0.17 to 0.27 for parent report for adolescents and 0.15 to 0.34 for parent report for children

Singh et al., 2011, six European countries (Belgium, Greece, Hungary, the Netherlands, Norway, Spain) [45]

Availability

1 week apart, for availability of fizzy drinks or fruit squash ICC = 0.74, fruit juice ICC = 0.67, and breakfast products at home ICC = 0.42

Face-to-face interviews used: for availability of fizzy drinks or fruit squash ICC = 0.52, fruit juice ICC = 0.57 and breakfast products at home ICC = 0.25, % agreement 80

Ward et al., 2015, US [46]

Availability

1 day ICC varied between 0.06–0.55, with all values for beverages <0.40 and 1 value for beverages being <0.40. 4 day ICC varied between 0.20 to 0.83, with only 1 value being lower than 0.4

Criterion validity with comparison using obervation by trained data collectors over 4 days, r: 0.25–0.85, with 1 value being less than 0.3

Anzman-Frasca et al., 2015, US [47]

Availability

Criterion validity assessed comparing the program leaders’ responses to observations (2 visits) by trained research assistants and trainers in the first sample and by using digital photography in the second sample. % agreement ranged from 61.5 (water) to 93.9 (sweetened beverages), kappa ranged from 0.08 (sweet snacks) to 0.62 (milk), with 4 items having kappa values less than 0.3. Spearman r = 0.13 (water)-0.56 (salty snacks), only 1 item with r < 0.30

Fulkerson et al., 2012, US [48]

Availability

Criterion validity assessed by comparing results to those of research staff trained to use the screener, kappa ranged between 0.52 and 1 for major food categories served/not served (1 value less than 0.6), kappa ranged between 0.74 and 0.87 for food subcategories served/not served (averaged across foods within same subcategory)

Hua et al., 2014, China [49]

Availability

% agreement between 67.8 and 99.3 for foods in restaurants; between 80.4 and 93.7 for drinks in restaurants; between 86.5 and 98.9 for foods in stores