Table 1 Summary of measured variables used to determine the dietary, individual and environmental latent constructs
Variable | Measure | Scale | Reliability/Validity |
|---|---|---|---|
Dietary Outcome | |||
 Dietary quality (SD) | 20-item food frequency questionnaire (FFQ) | 7-point scale from ‘never’ to ‘more than once a day’ | Diet scores from the 20-item FFQ have correlated highly with scores from a 100-item FFQ (r=0.94), and with red blood cell folate (r=0.25) [27] |
Dietary scores calculated by multiplying consumption frequency by coefficients from principal components analysis and summing total | Higher scores represented better dietary quality | ||
Scores standardised (mean=0, SD=1) | |||
 Fruit intake | ‘How often in the previous month have you consumed fresh fruit?’ | 7-point scale from ‘never’ to ‘more than once a day’ | Scale used in previous research [29] |
 Fast-food intake | ‘How often in the past month foods from fast food chains?’ | 7-point scale from ‘never’ to ‘more than once a day’ | Scale used in previous research [30] |
 Takeaway intake | ‘How often in the past month foods from independent takeaway outlets?’ | 7-point scale from ‘never’ to ‘more than once a day’ | Scale used in previous research [31] |
 Note: Exploratory factor analysis showed these dietary variables loaded onto a single construct: eigenvalue=1.7, factor loadings ≥0.37, explained 23% variance of diet construct | |||
Psychological resources | |||
 Perceived control over life | 9 item scale | 4-point from ‘strongly disagree’ to ‘strongly agree’ | Published scale [35] |
Responses summed to create overall score (range 5 to 20) | Higher score indicated greater sense of control | Cronbach’s alpha statistic 0.91 | |
Example items: | |||
- ‘I often have the feeling that I am being treated unfairly’ | |||
- ‘There are certain things I can do for myself to reduce the risk of cancer’ | |||
 Self-efficacy for healthy eating | 5 item scale | 4-point from ‘strongly disagree’ to ‘strongly agree’ | Published scale [36] |
Responses summed to create overall score (range 5 to 20) | Higher score indicated greater competence in overcoming barriers to healthy eating | Cronbach’s alpha statistic 0.89 | |
Example items: | |||
‘I could stick to eating healthy foods even if….’ | |||
- I have to rethink my whole diet | |||
- I have to try a few times before I succeed | |||
 Healthy eating outcome expectancies | 6 item score | 4-point from ‘strongly disagree’ to ‘strongly agree’ | Published scale [37] |
Responses summed to create overall score (range 4 to 24) | Higher score indicated stronger beliefs in good outcomes from eating healthily | Cronbach’s alpha statistic 0.84 | |
Example items: | |||
‘I know if I eat healthy foods...’ | |||
- I won’t have weight problems | |||
- It will be good for my blood pressure | |||
 Food involvement | 12-item scale | 5-point: ‘strongly disagree’, ‘disagree’ ‘neither agree or disagree’, ‘agree’, ‘strongly agree’ | Published scale [38] |
Responses summed to create overall score (negative items reversed, range 5 to 60) | Higher score indicated food-related activities are valued and prioritised in daily life | Cronbach’s alpha statistic 0.67 | |
Example items: | |||
- ‘I don’t think much about food each day’ | |||
- ‘I enjoy cooking for others and myself’ | |||
 Note: Exploratory factor analysis showed these dietary variables loaded onto a single construct: eigenvalue=1.5, factor loadings ≥0.48, explained 11% variance of the 3 individual-level constructs | |||
Perceived food affordability | |||
 Can’t afford enough food | ‘In the last 12 months, the food I bought didn’t last and I didn’t have money to get more’ | 3-point: ‘never true’, ‘sometimes true’, ‘always true’ | Scale used in previous research [39] |
 Can’t afford balanced meals | ‘In the last 12 months, I couldn’t afford to eat balanced meals’ | 3-point: ‘never true’, ‘sometimes true’, ‘always true’ | Scale used in previous research [39] |
 Note: Exploratory factor analysis showed these variables loaded onto a single construct: eigenvalue=2.0, factor loadings ≥0.75, explained 17% variance of the 3 individual constructs | |||
Perceived food accessibility | |||
 Can food shop locally | ‘In my local neighbourhood (10-15 minute walk), I can do most of my food shopping’ | 4-point from ‘strongly disagree’ to ‘strongly agree’ | Scale used in previous research [40] |
 Limited variety of fresh fruit and vegetables locally | ‘In my local neighbourhood, the variety of fresh fruit and vegetables is limited’ | 4-point from ‘strongly disagree’ to ‘strongly agree’ | Scale used in previous research [40] |
 Good quality produce locally | ‘In my local neighbourhood, the fresh produce is usually of a high quality’ | 4-point from ‘strongly disagree’ to ‘strongly agree’ | Scale used in previous research [40] |
 Note: Exploratory factor analysis showed these variables loaded onto a single construct: eigenvalue=1.9, factor loadings ≥0.63, explained 17% variance of the 3 individual constructs | |||
Spatial access to food outlets | |||
 Area (km2) of activity space | Total space covered by set of Euclidean 1000m (0.6 mile) buffers around postcode centroid of home and other frequently visited locations including main supermarket, work, GP, Sure Start children’s centre and physical activity location. | Total area of buffers in square kilometres (km2) | |
Overlapping buffers were merged to create one total area | |||
Range= 4.0km2 to 18.0km2 | |||
 Supermarket variety | Number of different types of supermarkets according to the categories: premium, large, discount and small | Variety values ranged from zero to four | Similar scale used in previous research [70] |
Different types of supermarkets have been shown to differ in terms of product availability, price and promotion [44] | Higher score showed greater variety of supermarkets within activity space | ||
 Food environment score (FES) – healthy outlets | Scores represented both the density of healthy food outlets and a proxy of the healthfulness of the in-store environment based on healthy food availability | Weightings for healthy outlets (0 to 10): | Scale used in previous research [43] |
Scores were calculated by i) identifying the number of each type of healthy food outlets within activity space, and ii) multiplying the number of each food outlets by a corresponding weighting (0 to 10) describing healthy food availability for that outlet determined from a published Delphi study [43] | 8.8: F&V store/ farm shop | ||
6.3: premium/large supermarket | |||
5.4: butcher | |||
5.3: ‘world’ store | |||
4.9: small supermarket | |||
4.4: sandwich shop | |||
4.3: health food shop | |||
3.3: discount supermarket | |||
0.8: bakery | |||
Range= 12 to 445.4 | |||
 Food environment score (FES) – unhealthy outlets | Scores represented both the density of unhealthy food outlets and a proxy of the healthfulness of the in-store environment based on unhealthy food availability | Weightings for unhealthy outlets (0 to -10): | Scale used in previous research [43] |
Scores were calculated by i) identifying the number of different types of unhealthy food outlets within activity space, and ii) multiplying the number of each food outlets by a corresponding weighting (0 to -10) describing unhealthy food availability for that outlet determined from a published Delphi study [43] | -1.1: convenience/ petrol stores | ||
-1.6: Chinese/ Indian takeaway | |||
-5.0: fish & chips/ other takeaways (pizza/kebab) | |||
-5.0: newsagents/ confectioners | |||
-8.3: fast food outlets | |||
Range= -9.3 to -753.3 | |||
 Note: Exploratory factor analysis showed these variables loaded onto a single construct: eigenvalue=2.4, factor loadings ≥0.33, explained 55% variance of the spatial access to food outlets construct | |||
Environment of main supermarket | |||
 Healthfulness score | Information on the number of varieties, price, promotion, shelf placement, and store placement were collected about seven healthy (peppers, tomatoes, lettuce, onions, apples, bananas, wholemeal bread) and five unhealthy products (oven chips, sausages, crisps, granulated sugar, white bread) | Z-scores for each of the nine in-store variables were created by subtracting the summed ratings for unhealthy products from the summed ratings for healthy products and standardizing the result | Reliably distinguishes different types of supermarkets [71] |
Inter-rater reliability kappa>0.73 for all except quality (kappa=0.60) [44] | |||
Data about type of nutrition information and availability of healthier alternatives were collected for unhealthy products | The z-scores for the nine variables were then summed and divided by 9 to ensure each in-store variable was equally weighted | Cronbach’s alpha statistic 0.86 | |
The quality of two fruits and four vegetables, and whether or not the fruits could be bought singly, were assessed | Scores standardised (mean=0, SD=1) | ||
Higher scores show more healthful supermarket environments | |||
Children’s centre nutrition practices | |||
 Food policy content | 21 item scale | Responses were summed | |
1 item assessed food policy existence | Responses: ‘no’, ‘verbal’, ‘written’, ‘unsure’ | Cronbach’s alpha statistic 0.74 | |
13 items – ‘Does your centre policy include: (examples) | 3-point scale: ‘yes’, ‘no’, ‘unsure’ | ||
- Promoting water intake | |||
- Restricting soft drinks | |||
items – ‘When does the policy apply?’: (examples) | 4-point scale: ‘always’, ‘often’, ‘sometimes’, ‘never’ | ||
- Food provided for children by your centre | |||
- Food provided for parents by parents | |||
- Food provided for staff by your centre | |||
2 items | |||
- ‘How often are parents told about the policy?’ | 4-point scale: ‘always’, ‘often’, ‘sometimes’, ‘never’ | ||
- ‘How often to parents bring food against the policy?’ | Higher scores represent a detailed, wide reaching food and nutrition policy | ||
 Healthy eating ethos | 14 item scale | Responses were summed | |
11 items – ‘How regularly do the following practices occur?’: (examples) | 4-point scale: ‘always’, ‘often’, ‘sometimes’, ‘never’ | Cronbach’s alpha statistic 0.69 | |
- Staff join children at the table when food is eaten | |||
- Parents consume the same food offered to children | |||
- Children encouraged to try new or less favoured foods | |||
2 items: | |||
- ‘How often are healthy eating activities included in centre sessions?’: | |||
- ‘How often do staff discuss healthy eating with parents?’ | |||
1 item assessed: | 4-point scale: ‘always’, ‘often’, ‘sometimes’, ‘never’ | ||
- ‘How important is healthy eating as one the many issues you help parents with at your centre?’ | |||
10-point scale from ‘not important’ to ‘very important’ | |||
Higher scores represent stronger healthy eating ethos at the centre | |||
 Healthy eating information | 6 item scale | Responses were summed | Adapted from a published scale [45] |
‘How available are the following resources?’: (examples) | 3-point scale: ‘readily available’, ‘available on request’, ‘not available’ | Cronbach’s alpha statistic 0.52 | |
- Leaflets/posters about healthy eating or weight loss for adults | |||
- Leaflets/posters about weaning or breastfeeding | Higher scores represent better availability of healthy eating resources | ||
- Courses to learn to cook or cook on a budget | |||
 Barriers to promoting healthy eating | 7 item scale | Responses were summed | Adapted from a published scale [72] |
‘Please indicate the extent to which you feel the following hinder healthy eating promotion at your centre’: (examples) | 3-point scale: ‘major problem’, ‘minor problem’, ‘not a problem’ | Cronbach’s alpha statistic 0.73 | |
- Food regulations | Lower scores represent fewer perceived barriers to promoting healthy eating | ||
- Parents don’t believe or trust your advice | |||
- Cost of healthy eating | |||
 Note: Exploratory factor analysis showed these variables loaded onto a single construct: eigenvalue=2.3, factor loadings ≥0.49, explained 46% variance of the Sure Start nutrition practices construct | |||