AbstractBackground:Previous research has suggested that fruits and vegetables aremore expensive and less readily available in more deprived communities. How-ever, this evidence is mainly based on small samples drawn from specific com-munities often located in urban settings and thus is not generalisable tonational contexts. The present study explores the influence of neighbourhooddeprivation and local retail structure on the price and availability of fruit andvegetables in a sample of areas representing the diversity of urban–rural envi-ronments across Scotland, UK.
Methods: A sample of 310 stores located in 10 diverse areas of Scotland wassurveyed and data on the price and availability of a basket of 15 fruit and vege-table items were collected. The data were analysed to identify the influence ofstore type and neighbourhood deprivation on the price and availability of fruitsand vegetables.
Results:Neighbourhood deprivation and store type did not significantly predictthe price of a basket of fruit and vegetables within the sample, although basketsdid decrease in price as store size increased. The highest prices were found inthe smallest stores located in the most deprived areas. Availability of fruit andvegetables is lower in small shops located within deprived neighbourhoodscompared to similar shops in affluent areas. Overall, availability increases withincreasing store size.
Conclusions:Availability of fruit and vegetables significantly varies by neigh-bourhood deprivation in small stores. Policies aimed at promoting sales of fruitand vegetable in these outlets may benefit residents in deprived areas.

Keywordsaccess, deprivation, diet, food store,neighbourhood, Scotland.

Introduction
    In recent years, there has been growing interest in therole of environmental factors in constraining an individ-ual’s ability to consume a healthy diet (Cummins &Macintyre, 2006). Early work on the phenomenon of‘food-deserts’, namely deprived areas where food is rela-tively more expensive and less readily available, leadresearchers to hypothesise that structural variations inneighbourhood access to food may be partly responsiblefor the production and maintenance of social and spatialinequalities in diet and diet-related chronic disease(Cummins & Macintyre, 2002). Subsequently, access tohealthy food at an affordable price has been recognisedby policymakers as a major barrier to healthy eating indisadvantaged communities (Department of Health,2004, Scottish Government 2004), with access beingdefined as including both global physical access to retailfacilities as well as access to healthy options within retailstores.
　　However, evidence for the existence of food deserts isvaried. Recent systematic and nonsystematic reviews haveconcluded that evidence for their existence is strong inthe USA, but equivocal and relatively weak in other highincome nations (Cummins & Macintyre, 2006; Beaulacet al. , 2009). In the UK, research in this area varies widelyin terms of quality and scope (White, 2007). Some studieshave found no association between food retail and indi-vidual fruit and vegetable intake (Pearson et al. , 2005;White, 2007) and minimal differences between deprivedand affluent areas in the availability of a balanced diet(Donkin et al. , 1999; Cummins & Macintyre, 2002) andthe presence of food retail outlets (Hackett et al. , 2008;Cummins & Macintyre, 1999). However, other studieshave demonstrated that there is poorer access to super-markets and grocery stores in deprived areas (Clarkeet al. , 2002), as well as increasing inequalities in grocerystore access (Guy & David, 2004). Qualitative work alsolacks consensus. Dibsdall et al. (2003) found that fewlow-income consumers reported any problems in access-ing supermarkets, despite transport difficulties, or per-ceived problems in the choice of fruit and vegetables.However, these findings are balanced by the discovery ofa complex relationship between concerns over age-depen-dent personal mobility, household economic constraintsand decision-making around the relative costs of trans-port options (walking, car use, public transport and pri-vate taxi hire) within deprived urban households (Whelanet al. , 2002). Qualitative work on food deserts alsoextends to rural areas, although this has been under-researched in the UK, with most work focused on theexperience and use of rural food retail provision. Distancetravelled to food retail shopping is greater in rural thanin urban areas, with 50% of consumers travelling between6 and 20 miles and 93% of trips being undertaken by car(McEachern & Warnaby, 2006). In rural areas, mostconsumers frequent their closest major supermarket toundertake major shopping trips on a weekly or monthbasis, with local convenience stores and small shopsserving as a source of secondary or top-up shopping(Scarpello et al. , 2009). Issues of perceived lack of stock,poor range of food items and a perceived higher price forfoods are cited as reasons why local provision is not com-monly used as a primary source of food shopping(Broadbridge & Calderwood, 2002).
　　In addition to this cross-sectional observational work,two studies have evaluated the effect of the introductionof large multiple-owned food supermarkets in deprivedurban areas with respect to fruit and vegetable intake,again with conflicting results. The Seacroft Study inLeeds, England, reported improvements in fruit and vege-table consumption, with the largest impact seen amongstthose with the lowest baseline intakes (Wrigley et al. ,2003). The second study, undertaken in Glasgow,Scotland, found little evidence for any effect on fruit andvegetable consumption patterns (Cummins et al. , 2005,2008).
　　Previous research undertaken in the UK has focused onsmall geographic areas such as one city or communityand has tended to focus solely on the experience ofdeprived urban neighbourhoods, failing to capture dataon other non-urban settings. Such an approach generatesevidence that is only applicable to the community settingsfrom which they originate and thus is not generalisable tothe national context. This may be one reason why con-sensus on whether ‘food deserts’ exist in the UK has notbeen reached. The present study aimed to investigatewhether the price and availability of a basket of fruit andvegetable items varies by store type as well as neighbour-hood deprivation in a national study in Scotland. Toensure that the varied settings in Scotland were ade-quately catered for, data were gathered from a sample ofareas that represent the full range of urban and ruralenvironments.
　　Materials and methods
　　Food price and availability data were collected using afood retail survey of 466 stores across Scotland. Thestores were located in 10 socio-environmentally diverseareas stratified by deprivation and urban/rural status andwere selected for inclusion to adequately represent the fullrange of social and physical environmental settings pres-ent across the country. The areas included islands, ruralareas, smaller towns and urban centres. Sites were initiallyselected by stratifying all available data zones in Scotlandby the Scottish Government’s Urban-Rural ClassificationScheme (SEUR). Data zones are the official small area sta-tistical geography used in Scotland and form the basis ofarea-based statistics generated from the Census. The siteswere selected by stratifying the 6505 data zones (meanpopulation of 778) by SEUR environment type and depri-vation. Deprivation was assigned to each data zone usingthe 2006 Scottish Index of Multiple Deprivation (SIMD),a publically available composite continuous score of rela-tive social and material deprivation (Scottish Govern-ment, 2006).
　　Sites were stratified into three categories of environ-ment: urban (SEUR 1–2), small town (SEUR 3–4) andrural (SEUR 5–6). Within each type of environment, adata zone from the top and bottom deciles of deprivationwas randomly chosen as the centre of a site. Additionaldata zones, contiguous to the randomly selected datazone, were then added to build up a study site that repre-sented a local community. This process resulted in sixresearch sites, although the sampling strategy failed toS. Cummins et al. Neighbourhood food price and availabilityª 2010 The Authors. Journal compilation ª 2010The British Dietetic Association Ltd. 2010 J Hum Nutr Diet , 23, pp. 494–501 495capture island communities, and the numbers of stores inrural and small towns were too small for valid statisticalanalysis. To enhance coverage of the range of settings andboost small numbers of observations, four additional sen-tinel sites were purposively selected to include two islandcommunities and a small town setting. Data from Ellonand Cupar were combined due to small numbers. Thefinal sample of sites included: urban affluent (BroughtyFerry, Dundee), urban deprived (Scotstoun/Drumchapel,Glasgow City), urban mixed (Inverness), small town afflu-ent (Ellon, Aberdeenshire, Cupar, Fife), small towndeprived (Kilbirnie, North Ayrshire), rural affluent(Haddington, East Lothian), rural deprived (Dornoch,The Highlands), and islands (Eilean Siar & Orkney). Intotal, 205 data zones were selected, which ensured fullcoverage of the environmental settings in Scotland(Table 1). Further details of the sampling process can befound in Smith et al. (2010).
　　Census of food retailing in the study sites
　　A comprehensive list of all retail food stores (excludingtakeaway, fast food and coffee shop outlets) in each sitewas generated. This list was compiled by combininginformation from a variety of sources, including industry(Institute of Grocery Distribution) and commerciallyavailable directories (Marketscan and Catalist), localauthority registers (Public Register of Food Premises),store websites (Tesco, Aldi, Lidl, Somerfield, Sainsbury,Asda, Morrisons), online retail directories (Yell.com) andwebsites of symbol groups (Spar, Londis, Budgens,Costcutter). Data from these various sources were com-bined, duplicate records removed and the data werecleaned to ensure that each street address matched theunit postcode. Postcode validity was ascertained by join-ing the retail data from Ordnance Survey code-pointinformation and identifying where postcodes could notbe grid-referenced. Postcode errors in the dataset werecorrected using the Royal Mail address/postcode checker.The final dataset of geo-coded stores included 457(98.1%) of the original sample. The stores were dividedinto three categories according to size of net sales area,measured in square feet: small (<3000), medium (3000–15 000) and large (>15 000).
　　Data on the price and availability of fruit and vegetables
　　Data on food item availability (available yes/no) andcheapest price (in pence) were collected by trained field-workers on fifteen fresh, tinned and frozen fruits and veg-etables (Table 2). These items comprised the ‘fruits’ and‘vegetables’ categories of the Healthy Eating IndicatorShopping Basket (HEISB) tool (Anderson et al. , 2007).The HEISB tool was developed using a literature searchto identify previous practice, information on current defi-nition of healthy foods by the UK Food StandardsAgency, and population-based dietary surveys to identifyculturally acceptable foods. Items were then appraisedwith respect to practical fieldwork considerations; forfurther details, see Anderson et al. (2007). Data on theselected food items were collected in person by fieldwork-ers through in-store visits to each of the food storesidentified in the food retail census. The data were col-lected during two waves, October/November 2005 andFebruary/March 2006.
　　Statistical analysis
　　Data analysis was restricted to food prices in stores thatsold the full range of surveyed items aiming to avoidproblems of missing data in the generation of summarybasket prices for each store. The majority of specialistfood stores and stores in which food is secondary wereremoved from this analysis because these shops did notTable 1 Descriptive characteristics of study sitesStudy site Population (2004)Percent of householdswith no car (2001) n data zonesDistribution of data zones by 2006income quintile*Q1 Q2 Q3 Q4 Q5Island mixed/deprived: Eilean Siar 18 683 30.2 25 0 3 12 10 0Island mixed: Orkney 12 365 20.5 17 5 10 2 0 0Rural affluent: Haddington 20 816 19.2 26 9 9 7 1 0Rural deprived: Dornoch 6679 20.7 9 0 4 3 2 0Small town deprived: Kilbirnie 13 223 30.4 17 3 3 2 5 4Small town affluent: Ellon and Cupar 7622 22.2 10 4 2 3 1 0Urban affluent: Broughty Ferry 13 535 26.8 17 8 3 3 1 2Urban deprived: Scotstoun/Drumchapel 41 992 53.9 51 2 2 5 12 30Urban mixed: Inverness 25 748 33.1 33 10 5 4 9 5*1 = least deprived; 5 = most deprived.Neighbourhood food price and availability S. Cummins et al.ª 2010 The Authors. Journal compilation ª 2010496 The British Dietetic Association Ltd. 2010 J Hum Nutr Diet , 23, pp. 494–501stock the entire contents of the fruit and vegetable bas-kets. Once these stores were removed, the reduced datasetconsisted of 113 stores in 87 data zones for the fruit bas-ket price analysis and 92 stores in 71 data zones in thevegetable basket price analysis. Large stores were morelikely to stock all of the fruit and vegetable items and aretherefore over-represented in this sample. Food priceswere standardised to a common weight to ensure compa-rability between food items.
　　Availability measures were calculated using the full dataset but excluding stores where fruit and vegetables wouldnot be routinely sold (such as butchers, fishmongers andbakers) to give a truer depiction of availability in ourstudy areas. This left 310 stores across 186 data zones eli-gible for the study. Each food item was coded as 0 (notavailable) or 1 (available). Specialist food shops and shopswhere food is secondary were removed from the analysisbecause there were too few observations for valid analysesto be undertaken. The resulting availability scores foreach basket were summed for every store to provide anoverall measure of store availability, with a higher scoredenoting increased availability.
　　Statistical analyses were performed using spss , version11.5 (SPSS Inc., Chicago, IL, USA). For each shop type,the median basket price and availability score werereported across quintiles of area deprivation based on2006 SIMD where the least deprived data zones were inquintile 1 and the most deprived areas in quintile 5.Because the distributions of both price and availabilityscores were non-normal, median values were used tosummarise the results. To test associations between quin-tiles of area deprivation and food price and availability innon-normal data, a nonparametric test (Kruskall–Wallischi-squared test) was used. Results were further stratifiedby store size defined on the basis of net sales area: small(<279 m2), medium (279 to 1394 m2) and large(>1394 m2).
　　Results
　　Tables 3 and 4 present the median price and availabilityof fruits and vegetable by store type (based on net salesarea) and quintile of area deprivation.
　　PriceTable 3 presents the data on median price and availabilityfor the fruit and vegetable baskets by store size and quin-tile of deprivation. Overall, for both fruit and vegetables,there is a general pattern of decreasing prices as the sizeof store increases in all deprivation quintiles, althoughthere is no clear pattern by deprivation. In all stores, themedian basket price for fruit is £1.37 more expensive inthe most deprived areas compared to the least deprivedareas, although this is not statistically significant( P = 0.4533). This pattern is mirrored for small stores,with medium and large stores showing inconsistent, non-significant, patterns by quintile of deprivation.Table 2 Food items included in the fruit and vegetable basketsFruit basket Vegetable basketApples (kg) Onions (kg)Bananas (kg) Carrots (kg)White grapes (kg) Broccoli (kg)Oranges (kg) Round lettuce (standard round)Orange juice (L) Red pepper (kg)Pineapple (227 g tin) Tomatoes (kg)Cucumber (whole singlestandard – 450 g)Baked beans (tinned 415 g)Frozen peas (standard bag 907 g)Table 3 Price of fruit and vegetables by store type and deprivation quintileStores (n )Quintiles of deprivation*Kruskall–Wallischi-squared P value 12345Median fruit basket prices (pence)All (113) 599.84 666.64 611.50 642.28 737.50 3.66 0.4533Small (80) 609.02 730.50 713.37 740.50 776.54 5.27 0.2611Medium (20) 532.11 607.00 539.00 509.00 532.98 4.45 0.3490Large (13) 454.23 453.39 492.62 499.52 485.54 1.65 0.7991Median vegetable basket prices (pence)All (92) 1130.55 1108.91 1258.32 1052.64 1089.25 1.96 0.7438Small (56) 1204.49 1181.45 1330.25 939.75 1538.09 6.57 0.1602Medium (23) 1017.08 932.60 1049.75 1158.22 1035.47 1.64 0.8010Large (13) 1008.91 1003.47 992.33 939.89 955.04 1.30 0.8610*1 = least deprived; 5 = most deprived.S. Cummins et al. Neighbourhood food price and availabilityª 2010 The Authors. Journal compilation ª 2010The British Dietetic Association Ltd. 2010 J Hum Nutr Diet , 23, pp. 494–501 497
　　For vegetables, in all stores, there is no evidence for adifference in median price by quintile of deprivation(P = 0.7438). Overall, the vegetable basket is cheapest inthe most deprived quintiles (4 and 5) compared to theleast deprived, although the magnitude of this differenceis relatively small. For small stores, the most expensivevegetable basket was found in the most deprived areas,although this is counterbalanced by the cheapest basketbeing found in quintile 4s. No clear patterns were presentfor medium and large stores and none of the tested asso-ciations were statistically significant.
　　Availability
　　Table 4 shows data on the availability of fruit and vegeta-bles by store type and deprivation quintile. Overall, thereis a clear pattern of increasing availability as store sizeincreases, with the largest stores stocking all of the sur-veyed items. Medium-sized stores stocked almost allitems, with only the fruit baskets in the most deprivedarea having a less than maximum score (6).
　　However, there were statistically significant differencesin availability for all stores by quintile of deprivation forboth fruit ( P = 0.002) and vegetable (P = 0.0002) baskets,with the poorest availability in the most deprived areas.This pattern was repeated for small stores (fruits:P = 0.0029; vegetables: P = 0.0001), although there is noclear trend with deprivation for all stores and smallstores. The availability analysis indicated that small storesmay be sensitive to the effects of deprivation because dif-ferences are confined to this category, with medium andlarge stores having almost perfect availability.
　　Discussion
　　The present study aimed to explore the role of store typeand area deprivation in determining the price and avail-ability of an indicative basket of fruit and vegetables innationally representative sample of settings in Scotland.The data presented here support previous research sug-gesting that residents of materially deprived areas maynot always pay more for healthy food compared to resi-dents of more affluent areas (Mooney, 1990; Cummins &Macintyre, 2002; White et al. , 2004; Winkler et al. , 2006).However, the size of store consumers’ chose to shop indetermined the price paid for those items, with thosewho shop in the smallest stores, in the most deprivedareas, paying the greatest price premium.The availability of fruits and vegetables did vary sig-nificantly by area deprivation for all stores but this wasdriven by patterns of availability found in smaller stores.The relatively poorer availability of fruit and vegetableproducts in small stores located within more deprivedareas may be a result of demand variability (and theshopkeeper’s perception of this), product characteristicsand shop capabilities. Skerratt (1999), as a result ofinterviews with rural Scottish shopkeepers, suggests thatconsumer buying patterns have a direct impact on thestocking and supply strategies of retailers. Bulk-purchas-ing and hoarding of certain food items by consumerscreates an uneven demand and thus creates an erraticsupply, with shops unable to consistently predict con-sumer need for some goods. The stocking of perishablefresh produce may involve a business risk because of thepotential loss of profit from fresh produce items, whichmay spoil before they are sold. There may also be prob-lems of storage capacity (such as chiller cabinets) forfresh produce items in smaller stores, which couldaccount for some of the differences in availability bystore type.
　　Residents of more deprived communities have a lowerintake of fruits and vegetables compared to residents ofwealthier areas (Wrieden et al. , 2006), although theresults presented here do not suggest that residents ofTable 4 Availability of fruit and vegetables by store type and deprivation quintileStores (n)Quintiles of deprivation*Kruskall–Wallischi-squared P value 1 2345Median fruit availability scores (minimum = 0, maximum = 6)All (310) 5 5 6 5 4 16.90 0.0020Small (268) 4 5 5 5 2 16.07 0.0029Medium (29) 6 6 6 6 5 5.27 0.2607Large (13) 6 6 6 6 6 0 1Median vegetable availability scores (minimum = 0, maximum = 9)All (310) 5.5 7 8 5 4 22.21 0.0002Small (268) 4 7 7 5 2 24.09 0.0001Medium (29) 9 9 9 9 9 0.84 0.9337Large (13) 9 9 9 9 9 0 1*1 = least deprived; 5 = most deprived.Neighbourhood food price and availability S. Cummins et al.ª 2010 The Authors. Journal compilation ª 2010498 The British Dietetic Association Ltd. 2010 J Hum Nutr Diet , 23, pp. 494–501more deprived areas face greater structural barriers, interms of physical access to food stores selling fruit andvegetables, than people living in more affluent areas.However, to benefit from the best price and the greatestchoice, residents of deprived neighbourhoods wouldneed to choose to routinely shop at medium and largerstores, although this may not always be the case.Williams & Hubbard (2001) have suggested that socio-economically disadvantaged consumers have a preferencefor ‘traditional’ facilities rather than newer stores orshopping. This may be because families may shop littleand often at local grocery stores because they do nothave the financial resources to commit income to bulk-buy from large stores or co-ops (Dobson et al. , 1994).Small stores, which make up the bulk of the retail net-work in more disadvantaged communities, thus remainan important source of food provisioning for low-income consumers. Strategies to support the stockingpolicies of these stores may be beneficial to local resi-dents because the evidence suggests that even the stock-ing of small ranges of fresh produce can providevaluable sources of anti-oxidant nutrients (Haleem et al. ,2008). For example, the Scottish Government is cur-rently evolving policy around improving diet throughthe promotion of fruit and vegetables, including initia-tives to encourage local small shops to stock more freshproduce such as assistance with chiller cabinets (ScottishGovernment, 2005). A recent evaluation of a similarsmall corner store intervention in disadvantaged neigh-bourhoods of Baltimore (MD, USA) (Song et al. , 2009)suggested that supporting independent store owners inincreasing their stock of healthy foods, combined withmarketing and promotion, can lead to increased sales ofthese items to local consumers. Such policies woulddovetail well with other policy initiatives such as the‘Healthy Start’ welfare foods scheme (Healthy Start,2009), which provides vouchers that can only be usedfor fresh, rather than canned, frozen and dried, fruitsand vegetables, emphasising the importance of goodaccess to these commodities for the poorest families.Qualitative research has suggested that fruit and vegeta-ble provision is poorer in remote rural areas (McKieet al. , 1998; Skerratt, 1999), although the residents ofthese areas report the highest fruit and vegetable con-sumption (Wrieden et al. , 2006). Thus, the potentialgains of small store interventions in deprived rural com-munities may be greater than those in urban settings.
　　Price and availability is not the only structural factorthat may influence food purchasing decisions. The per-ceived quality of available produce is also important toconsumers, and may be a mediating variable in the deci-sion to purchase fresh produce. Recent work has sug-gested that, even though the quality of fresh produce wasgenerally good, stores in more deprived neighbourhoodstended to stock lower quality fresh fruit and vegetables(Cummins et al. , 2009).
　　The present study has certain limitations. The studywas cross-sectional and ecological in design and does notlink the price and availability of fruits and vegetables withpurchasing or consumption behaviours. A cross-sectionalstudy such as this assumes that price and availability isstable over time when in fact it may vary as a result ofother external factors, such as wholesale supply, seasonali-ty and variations in transport and shipping costs (espe-cially important for stores in rural areas). Because weonly included stores that stocked all the items in the bas-kets (to avoid having missing data), larger stores wereover-represented in the price analysis, as a greater propor-tion of these stores stocked 100% of products. It may bethat the smaller- and medium-sized stores excluded fromthe analysis, on the basis of having one or more missingitems, stock a smaller range of fruit and vegetable itemsthat may be cheaper. The fruit and vegetable items drawnfrom the HEISB tool were only a selection of the possibleuniverse of available items, although they are the mostcommonly consumed fruit and vegetable items, based onnational data derived from The National Diet and Nutri-tion survey and other representative studies (Andersonet al. , 2007). Finally, price data was based on the cheapestprice available and did not reflect the extent of the pricevariation seen throughout the product category in eachstore.
　　Conclusions
　　The data presented here suggest that the price and avail-ability of a basket of fruit and vegetables does not vary byneighbourhood deprivation. However, small stores oftenhave lower availability of fruits and vegetables, and aremore expensive compared to medium and large stores.Structural policies to support the purchase and consump-tion of a healthy diet may involve supporting small storesin the stocking and promotion of a wider range of freshand other fruit and vegetable items.
　　Acknowledgments
　　SC wrote the paper and designed and directed analyses.DMS conducted analyses and contributed to the writing.JD, DM, ASA and LS contributed to project design andpaper drafting and editing. We would also like to thankZoe Aitken for preliminary data manipulation on thispaper and Matt Taylor for directing and conductingfieldwork. All authors critically reviewed the manuscriptand approved the final version submitted for publica-tion.S. Cummins et al. Neighbourhood food price and availabilityª 2010 The Authors. Journal compilation ª 2010The British Dietetic Association Ltd. 2010 J Hum Nutr Diet , 23, pp. 494–501 499Conflict of interest, source of funding andauthorshipThe authors declare that they have no conflict ofinterests.This work was supported by the Food Standards Agency(Scotland) as part of the project ‘Accessing Healthy Food:A National Assessment and Sentinel Mapping Study ofFood Retailing in Scotland’ (Grant Ref: S04005). DS issupported by the award of Philip Leverhulme Prize to SC.SC is also supported by a National Institute of HealthResearch Fellowship.
　　ReferencesAnderson, A.S., Dewar, J., Marshall, D., Cummins, S., Taylor,M., Dawson, J. & Sparks, L. (2007) The development of ahealthy eating indicator shopping basket tool (HEISB) foruse in food access studies – identification of key food items.Public Health Nutr. 10 , 1440–1447.Beaulac, J., Kristjansson, E. & Cummins, S. (2009) A system-atic review of food deserts, 1966–2007. Prev. Chronic Dis.6 ,3.Broadbridge, A. & Calderwood, E. (2002) Rural grocery shop-pers: do their attitudes reflect their actions? Int. J. RetailDistrib. Manag.30 , 394–406.Clarke, G., Eyre, H. & Guy, C. (2002) Deriving indicators ofaccess to food retail provision in British cities: Studies ofCardiff, Leeds and Bradford. Urban Stud. 39 , 2041–2060.Cummins, S. & Macintyre, S. (1999) The location of foodstores in urban areas: a case study in Glasgow Brit. Food J.101, 545–553.Cummins, S. & Macintyre, S. (2002) A systematic study of anurban foodscape: the price and availability of food in greaterGlasgow. Urban Stud. 39 , 2115–2130.Cummins, S. & Macintyre, S. (2006) Food environments andobesity – neighbourhood or nation? Int. J. Epidemiol. 35 ,100–104.Cummins, S., Petticrew, M., Higgins, C., Findlay, A. & Sparks,L. (2005) Large scale food retailing as an intervention fordiet and health: quasi-experimental evaluation of a naturalexperiment. J. Epidemiol. Commun. Health 59 , 1035–1040.Cummins, S., Findlay, A., Higgins, C., Petticrew, M., Sparks,L. & Thomson, H. (2008) Reducing inequalities in healthand diet: the impact of food retail development. Environ.Plan A.40 , 402–422.Cummins, S., Smith, D.M., Taylor, M., .Dawson, J., Marshall,D., Sparks, L. & Anderson, A.S. (2009) Variations in freshfruit and vegetable quality by store type, urban-rural settingand neighbourhood deprivation in Scotland. Public HealthNutr.12 , 2044–2050.Department of Health (2004)Choosing Health: Making Health-ier Choices Easier . London: Department of Health.Dibsdall, L.A., Lambert, N., Bobbin, R.F. & Frewer, L.J. (2003)Low-income consumers’ attitudes and behaviour towardsaccess, availability and motivation to eat fruit and vegeta-bles. Public Health Nutr. 6, 159–168.Dobson, A., Beardsworth, T., Keil, T. & Walker, R. (1994)Diet, Choice and Poverty: Social Cultural and NutritionalAspects of Food Consumption Among low-Income Families .Loughborough: Family Policy Studies Centre.Donkin, A.J.M., Dowler, E.A., Stevenson, S.J. & Turner, S.A.(1999) Mapping access to food at the local level.Br. Food J.101, 554–564.Guy, C.M. & David, G. (2004) Measuring physical access to‘healthy foods’ in areas of social deprivation: a case study inCardiff.Int. J. Consum. Stud. 28 , 222–234.Hackett, A.F., Boddy, L., Boothby, J., Dummer, T., Johnson, B.& Stratton, G. (2008) Mapping dietary habits may provideclues about the factors which determine food choice.J. Hum. Nutr. Diet. 21 , 428–437.Haleem, M.A., Barton, K.L., Borges, G., Crozier, A. & Ander-son, A.S. (2008) Increasing antioxidant intake from fruitsand vegetables: practical strategies for the Scottish popula-tion. J. Hum. Nutr. Diet. 21 , 539–546.Healthy Start (2009) Available at http://www.healthy-start.nhs.uk/ (accessed on 27 August 2009).McEachern, M.G. & Warnaby, G. (2006) Food shoppingbehaviour in Scotland: the influence of relative rurality.Int. J. Consum. Stud. 30 , 189–201.McKie, L., Clark, G.M., MacLellan, M. & Skerratt, S. (1998)The promotion of healthy eating: food availability andchoice in Scottish island communities. Health Educ. Res.13 ,371–382.Mooney, C. (1990) Cost and availability of healthy food choicesin a London health district. J. Hum. Nutr. Diet. 3, 111–120.Pearson, T., Russell, J., Campbell, M.J. & Barker, M.E. (2005)Do ‘food deserts’ influence fruit and vegetable consumption?– a cross-sectional study. Appetite45 , 195–197.Scarpello, T., Poland, F., Lambert, N. & Wakeman, T. (2009)A qualitative study of the food-related experiences of ruralvillage shop customers. J. Hum. Nutr. Diet. 22 , 108–115.Scottish Government (2004) Eating For Health: Meeting TheChallenge . Edinburgh: Scottish Government.Scottish Government (2005) Healthy Food at Corner Shops.Available at http://www.scotland.gov.uk/News/Releases/2005/02/07095154 (accessed 3 April 2010).Scottish Government (2006) Scottish Index of Multiple Depriva-tion 2006. Available at http://www.scotland.gov.uk/Topics/Statistics/SIMD/ (accessed 27 August 2009).Skerratt, S. (1999) Food availability and choice in ruralScotland: the impact of ‘‘place’’. Br. Food J. 101, 537–544.Smith, D.M., Cummins, S., Taylor, M., Dawson, J., Marshall,D., Sparks, L. & Anderson, A.S. (2010) Neighbourhood foodenvironment and area deprivation: spatial accessibility togrocery stores selling fresh fruit and vegetables in urban andrural settings. Int. J. Epidemiol. 39 , 277–284.Neighbourhood food price and availability S. Cummins et al.ª 2010 The Authors. Journal compilation ª 2010500 The British Dietetic Association Ltd. 2010 J Hum Nutr Diet , 23, pp. 494–501Song, H.J., Gittelsohn, J., Kim, M., Suratkar, S., Sharma, S. &Anliker, J. (2009) A corner store intervention in a low-income urban community is associated with increased avail-ability and sales of some healthy foods. Public Health Nutr.12 , 2060–2067.Whelan, A., Wrigley, N., Warm, D. & Cannings, E. (2002) Lifein a ‘food desert’. Urban Stud. 39 , 2083–2100.White, M. (2007) Food access and obesity.Obes. Rev.8,99–107.White, M., Bunting, J., Raybould, S., Adamson, A., Williams,L. & Mathers, J. (2004) Do Food Deserts Exist? A Multi-Level,Geographical Analysis of the Relationship Between Retail FoodAccess, Socio-Economic Position and Dietary Intake . London:Food Standards Agency, Final Report to The FoodStandards Agency.Williams, P. & Hubbard, P. (2001) Who is disadvantaged?Retail change and social exclusion.Int. Rev. Retail Distrib.Consum. Res. 11 , 267–286.Winkler, E., Turrell, G. & Patterson, C. (2006) Does living in adisadvantaged area entail limited opportunities to purchasefresh fruit and vegetables in terms of price, availability, andvariety? Findings from the Brisbane Food Study. HealthPlace12 , 741–748.Wrieden, W.L., Barton, K.L., Armstrong, J. & McNeill, G.(2006) A review of food consumption and nutrient intakesfrom national surveys in scotland : comparison to the Scottishdietary targets . Aberdeen: Food Standards Agency (Scotland).Wrigley, N., Warm, D. & Margetts, B. (2003) Deprivation,diet, and food-retail access: findings from the Leeds ‘fooddeserts’ study. Env. Plan. A35 , 151–188.S. Cummins et al. Neighbourhood food price and availabilityª 2010 The Authors. Journal compilation ª 2010The British Dietetic Association Ltd. 2010 J Hum Nutr Diet , 23, pp. 494–501 501