analysis of supermarket foods targeted at children
C. Elliott School of Journalism and Communication, Carleton University, Ottawa, ON, Canada Received 2 June 2007; revised 26 September 2007; accepted 1 October 2007 Address for correspondence: C Elliott, School of Journalism and Communication, Carleton University, K1S 5B6. E-mail: charlene_elliott@carleton.ca Summary This article provides a nutritional prole of foods targeted specically at children in the Canadian supermarket. Excluding confectionery, soft drinks and bakery items, 367 products were assessed for their nutritional composition. The article examines the relationship between fun food images/messages, product claims and actual product nutrition. Among other ndings, it concludes that approxi- mately 89% of the products analysed could be classied as of poor nutritional quality owing to high levels of sugar, fat and/or sodium. Policy considerations need to be made in light of the fact that fun food is a unique category that poses special challenges; as such, recommendations regarding food labelling and pack- aging are presented. Keywords: Childhood obesity, childrens foods, food marketing, nutrition. obesity reviews (2008) 9, 368377 Childhood obesity is a signicant public health problem and requires innovative public health solutions. Excess body weight affects up to 35% of children across Canada, the United States and Europe (1,2), and is linked to a range of comorbidities, including type 2 diabetes, hypertension, cardiovascular disease and some forms of cancer (3,4). Overweight children also suffer from psychological and social consequences (5,6) stemming from a pervasive and serious weight stigma in society (7). Various interventions in school and family settings seek to address the childhood overweight/obesity problem. Yet there is a weak evidence base as to what actually works (2,8,9) and, consequently, the need to explore a range of solutions (1013). The food industry is routinely identied as a major contributor to the problem of childhood obesity for promoting food to children that is high in sugar and/or fat and low in nutrients. However, most of the studies that assess food-related media messages targeted at children focus primarily on televised advertising (1420). As the US-based Institute of Medicine Committee on Food Mar- keting and the Diets of Children and Youth reports televi- sion advertising remains the dominant form of marketing reaching children . . . that is formally tracked (14). One key challenge the committee identies in its comprehensive review of studies related to food marketing to children and youth is that virtually all of the published scientic research has focused on advertising and television adver- tising in particular (14). In light of this, it proves necessary to look beyond television food advertising to other critical spaces, places and messages that comprise our food environment. Supermarket food and the packaging of fun Missing from the scholarly discourse on food messages and environmental cues is detailed knowledge of the foods tar- geted specically at children in the supermarket. While a substantial literature addresses television/advertising mes- sages and fast-food campaigns, little research focuses on the very basic mediascape of the supermarket. Many analyses pertaining to childhood obesity, moreover, focus squarely on confectioneries, sweets, fatty snack food and sugared sodas (routinely classied under the category of junk food) (1,21). Yet when one moves beyond this junk food classication, a clear knowledge gap exists. With the exception of one (dated) study (22), current research does obesity reviews doi: 10.1111/j.1467-789X.2007.00418.x 368 2007 The Author Journal compilation 2007 International Association for the Study of Obesity. obesity reviews 9, 368377 not examine the types of foods in the dairy, dry goods, produce and frozen-food categories, which constitute chil- drens fare or detail their nutritional value. Also lacking are data on the relationship between food product images, appeals and nutrition claims and actual product nutrition. This study provides such a nutritional prole and packag- ing assessment. Method The research uses content analysis to create a prole of the food products being marketed specically for and to chil- dren within the Canadian supermarket, and focuses on regular foods (within the dry goods, dairy, produce and frozen food categories) that have been repackaged to appeal to children. The study excludes junk food con- fectionery, soft drinks, cakes, potato chips, etc. as parents would likely expect this category of food to be of poor nutritional value. Canadas Weston-owned Loblaws Superstore was selected as the site/supermarket for food coding. George Weston Limited and Loblaw Companies Limited is the largest player in the Canadian food retail market, both in terms of revenue and in terms of number of stores. As such, it can be considered representative of the grocery food environment available to Canadians. Moreover, Loblaws market stores stock roughly 50 000 food and non-food items combined. The expansive product offering within Loblaws ensures that the study is comprehensive but also generalizable. Most stores carry the same national brands, and so the foodstuffs present in an Ontario-based Loblaws will be consistent with one based in Alberta. Fun foods those specically targeted at children were purchased in three separate trips made to the Loblaws Superstore in December 2005. Products were photo- graphed, stored and subsequently coded. Nutritional infor- mation was recorded for each product. Selections of fun food were made according to very specic criteria. Food products/packaging had to speci- cally speak to children; products signalled their status as childrens food and/or fun through the use of particular package shapes, colours, sizes, iconography, and graphics and language. Specic indicators of fun food, and criteria for inclusion in the sample, included a minimum of two of the following aspects: direct claims or allusions to fun/play on the package; cartoon iconography pointedly directed to children; tie-ins with childrens television programmes, mer- chandise or lms; the foregrounding of strange shapes, unusual colours or unconventional tastes; puzzles or games targeted at children. 1,2 A total of 367 products were purchased for coding and 36 variables were recorded for each product. Each case was identied in terms of brand, product name, food category and food type (i.e. breakfast, lunch, dinner, snack, beverage or mixed/variable). Twenty-six variables pertained to the packaging itself (i.e. the package semiotics), focusing on the graphics (i.e. use of cartoons to attract attention, cross- merchandizing appeals, etc.), nutrition claims and verbal claims. Coders also identied any unique product claims, detailed whether a game or activity adorned the package, and recorded all nutritional information. Two graduate students were involved in the product coding. Prior to the actual coding, they independently coded identical products until intercoder reliability reached higher than 0.80. Sixteen products (4.4% of the sample) were later randomly selected and coded for the intercoder reliability test, showing a 0.90 agreement level between coders. To assess the nutritional value of the products, we used the criteria for Poor Nutritional Quality (PNQ) outlined by the US-based Center for Science in the Public Interest (CSPI) for fat and sodium (23) and a unique set of criteria with regard to sugar. According to the CSPI, foods of PNQ meet one of the following criteria: >35% of total calories from fat, excluding nuts, seeds, and peanut or other nut butters; >35% added sugars by weight; >230-mg sodium per serving for chips, crackers, cheeses, baked goods, French fries, other snacks; 1 Cartoon images on the package do not suffice to signify childrens food. The Jolly Green Giant, Betty Crocker, Aunt Jemima and the Pillsbury Dough Boy are cartoon images, yet their mere presence does not make a product fun food. These brand characters, used to promote everything from corn to crescent rolls, require additional visuals to cue that the product is targeted at children. Similarly, products were not selected according to a predetermined or presumed classication of what com- prises childrens food. While adults may consider Kraft Macaroni and Cheese or Chef Boyardee Mini Ravioli as typical foods for children, such products were not selected as part of the sample unless the package adhered to the operationalization of fun food outlined above. 2 Other signicant exclusions pertain to both the type and package size of the food. As the research is interested in creating a prole of childrens food outside the category of junk food, candies, chips and sodas were excluded, along with the array of snack cakes (e.g. Twinkies, Joe Louis). Frozen novelties such as fruit and ice pops, however, were included in the study, as such products could not be dismissed as mere junk. As the sample was culled froma supermarket Superstore, all club pack products (bulk-sized versions of products) were excluded fromthe study. For similar reasons, as identical cereals (with identical boxes) might come in three sizes (425 g, 725 g and 1 kg), only one size was included in the sample. obesity reviews Assessing fun foods C. Elliott 369 2007 The Author Journal compilation 2007 International Association for the Study of Obesity. obesity reviews 9, 368377 >480-mg sodium per serving for cereals, soups, pastas and meats; >600-mg sodium for pizza, sandwiches and main dishes; >770-mg sodium for meals (23). Evaluating sugar in the selected food products posed a unique methodological challenge and requires some elaboration. CSPIs criteria state that, with regard to sugar, foods of poor nutritional quality contain more than 35% of added sugars by weight. As milk sugars and fruit sugars are naturally occurring, it is not possible to discern the added sugars by weight from the Nutrition Facts label (the label only displays total sugars). The quantity of added sugars must be provided by the manu- facturer. Although the research team contacted the manu- facturers of all products in which added sugars were in question (due to the presence of fruit/fruit juice concen- trate), virtually all manufacturers refused to reveal such information due to the condential nature of their for- mulas. As such, it was necessary to use another set of criteria for sugars. A similar study of food products con- ducted by Fitzhugh and Lobstein of the British Food Commission (22) classied foods as having high levels of sugar if they had more than 10 g sugar per portion or per hundred grams. However, a more nuanced calculation of sugar is presented in Harrison and Marskes nutritional assessment of the food products advertised to children on television (17). Harrison and Marske calculated the percentage of sugars per 200-calorie serving; any products that exceeded the 20% limit recommended by the American Heart Association were classied as PNQ. Since Harrison and Marske were interested in the percentage of sugars (rather than an absolute cut-off regardless of portion size), their criteria are used for this project. It is critical to underscore that, due to the proprietary nature of food manufacturing, it proves difcult to separate the natural content of the food (e.g. naturally occurring milk or fruit sugars) from the added sugars. Previous research (such as the Fitzhugh and Lobstein and Harrison and Marske studies) have had to read the Nutrition Facts label at face value; CSPIs criteria for dening food of PNQ equally acknowledge the problem of distinguishing fruit sugars from added sugars from the information pro- vided on the Nutrition Facts label. CSPIs compromise, as with the other studies, was to assess the nutritional quality of products such as fruit snacks using the amount of total sugars displayed on the label (23). Similar difculties extend to naturally occurring milk sugars. To reiterate, without actual data from the manufacturer, it is nearly impossible to determine precisely what percentage of sugars are added. So while it would be ideal to know how the products were made, we simply needed to go with the information available. This analysis both follows and builds upon other researchers methodology, updating and extending existing studies that classify products according to nutritional quality. Part two of our analysis examined whether foods meeting the poor nutritional quality criteria also con- tained nutrition claims on the front of the package e.g. a cereal tagged as of PNQ owing to its high sugar content might also contain a nutrition claimof high in iron or low in fat (thus appearing healthy owing to the nutrition label for iron/low fat while distracting attention away from the sugar content). The analysis also considered whether package images were consistent with or counter to the actual product nutrition. Methods of analysis Two types of quantitative analysis were used in this paper. First, univariate analyses in the form of frequencies were initially conducted to describe the nature of variables used in this study. Results of these are summarized in Tables 14. Second, bivariate analyses were conducted on our data; these include cross-tabulations with appropriate measures of association (chi-square, phi and Cramers V). The results of those tabulations are summarized in Tables 514. Where appropriate, statistically signicant relationships were agged by asterisk. Results Of the 367 products analysed, three failed to provide infor- mation related to the quantity of sugar in the product, three related to sodium and two related to fat. Approximately 89% of products (326 products) with nutritional data could be classied as of poor nutritional quality owing to high levels of sodium or an excessive proportion of calories coming from fat or sugar. Almost 23% of products were classied as poorly nutritious owing to an excessive quan- tity of calories coming from fat (>35% of total calories from fat), 17% had high levels of sodium and 69.5% are poorly nutritious owing to a high percentage of calories Table 1 Nutritional assessment of childrens foods* High level (%) Sugar 69.5 Total fat 22.7 Sodium 17.0 One or more of the above 89.3 Low levels of sugar, sodium and fat 10.7 *Excluding all the products that lacked information relevant to the amount of fat, sodium, and/or sugar in the product, we end up with a sample of 320 products. However, the statistics in this table are calculated from a sample of 367 products minus the missing cases relevant to the nutritious values under investigation. 370 Assessing fun foods C. Elliott obesity reviews 2007 The Author Journal compilation 2007 International Association for the Study of Obesity. obesity reviews 9, 368377 coming from sugar (20% or more of calories came from sugar). Conversely, just over 10% of products were low in sodium and did not contain an excessive amount of calories sourced from fat or sugar (Table 1). The most dominant representation of fun foods was found in the dry goods category (222 products; approxi- mately 61% of the sample). Refrigerated and frozen foods came second (63 products; 17.1%) and dairy, third (53 Table 2 Nutritional quality within dry goods Over 35% of calories coming from fat (poor nutrition) High sodium (poor nutrition) Over 20% of calories coming from sugar (poor nutrition) Type of food Frequency/% Frequency/% Frequency/% Cereal (1) 2.5 (37) 92.5 Crackers (12) 70.6 (11) 64.7 (2) 11.8 Cookies/biscuits (4) 13.3 (1) 3.8 (27) 90.0 Fruit snacks/applesauce (38) 88.4 Granola bars/cereal bars (1) 12.5 (8) 100.0 Pasta (boxed/canned)/soups (8) 88.9 (8) 88.9 Drinks/drink boxes (25) 86.2 Dressings/sauces/condiments (19) 100.0 Puddings/Jell-Os (13) 100.0 Dressing/sauces/condiments/toppings (2) 100.0 (2) 100.0 Peanut butters/jams/spreads (2) 18.2 (1) 10.0 Total (22) 10.0 (20) 9.1 (180) 81.8 Table 3 Nutritional quality within refrigerated and frozen foods Over 35% of calories coming from fat (poor nutrition) High sodium (poor nutrition) Over 20% of calories coming from sugar (poor nutrition) Type of food Frequency/% Frequency/% Frequency/% Frozen dinners/meals (2) 22.2 (7) 77.8 (1) 11.1 Pizza/pogos (17) 81.0 (14) 66.7 (1) 4.8 Fries/potatoes (3) 75.0 (1) 25.0 Packaged lunch (13) 72.2 (17) 94.4 (7) 38.9 Frozen breakfast foods/waffles/strudels/pancakes (4) 36.4 (5) 45.5 Total (38) 61.9 (38) 60.3 (15) 23.8 Table 4 Nutritional quality within dairy products Over 35% of calories coming from fat (poor nutrition) High sodium (poor nutrition) Over 20% of calories coming from sugar (poor nutrition) Type of food Frequency/% Frequency/% Frequency/% Milk* (1) 100.0 Yogurt (10) 100.0 Milk/yogurt drinks (14) 100.0 Cheese (12) 42.9 (16) 57.1 Total (12) 22.6 (41) 77.4 *It is rather misleading to see milk in the category of poorly nutritious foods. Note that this one product is 2% chocolate milk with extra sugars coming not only from the naturally occurring lactose but also from the chocolate sweetening. obesity reviews Assessing fun foods C. Elliott 371 2007 The Author Journal compilation 2007 International Association for the Study of Obesity. obesity reviews 9, 368377 products; 14.4%). Signicantly underrepresented was the produce category: fruits and vegetables comprised roughly 1% of the sample, being promoted under the two brands of Presidents Choice Mini Chefs and Tanimura & Antle. Worth noting is that childrens fare in the world of produce boils down to small apples and baby carrots no other fruits and vegetables were specically targeted at children. Dry goods Within the dry goods category, high levels of sugar were particularly evident. Just under 82% of the 222 products in this category were classied as foods of PNQ because of a high percentage of calories coming from sugar. Ten percent of dry goods products were of PNQ owing to a high proportion of calories derived from fat, and 9% of these products also fared poorly because of high sodium levels 3 (Table 2). Examples of culprits include granola bars/cereal bars (100% PNQ, S), puddings/Jell-Os (100% PNQ, S), pasta (boxed/canned) and soups (89% PNQ, Na), and drinks/drink boxes (86.2% PNQ, S). Not surprisingly, over 90% of the childrens cereal analysed was identied as PNQ owing to high levels of sugar. Refrigerated and frozen foods Sugar was less of an issue in this food category; however, many products contained high levels of fat or sodium. Over 60% of the refrigerated and frozen foods were PNQ foods owing to a high proportion of fat calories; another 60% were deemed of poor nutrition due to high sodium levels, and 23.8% were PNQ foods owing to a high proportion of sugar calories (Table 3). Examples include: packaged lunch (94% PNQ, Na and 72% F), frozen dinners/meals (78% Na), pizza/pogos (67% Na and 81% F) and fries/potatoes (75% F). Dairy Overall, dairy products fared well nutritionally when it came to sodium and fat. Of the 53 products analysed, none contained high levels of sodium, and only 22.6% of prod- ucts (all cheese) contained a higher proportion of calories coming from fat. Sugar, however, was still an issue as over three-quarters of products (77.4%) had a signicant pro- portion of calories derived from sugar (Table 4). Nutritional quality and nutrition claims Mandatory nutrition labelling is now required on most pre-packaged food in Canada; a standardized Nutrition Facts table provides information on calories and 13 nutri- ents. Along with the Nutrition Facts table, however, many food products highlight particular nutrition claims on the front of the box. This study revealed that fun food generally claims to be nutritious food, insomuch as 230 products (62.7%of prod- ucts analysed) make one or more nutrition claims on the front of the box (Table 5). Products contain a Smart Spot! or nutrition mark/seal (7.9%) claim to be low fat (6.5%), a source of calcium (6.8%), made with real fruit juice (6.0%) or trans-fat-free (4.4%). Other claims include: no articial avours/colours (7.4%), a source of X essential nutrients (6.5%), preservative-free (5.7%), peanut-free (3.2%), a source of whole wheat/bre (3%) or organic (1%). What proves interesting, however, is the relationship between nutrition claims and actual product nutrition. Using the established criteria for identifying foods of PNQ, it was found (as earlier noted) that 326 of the childrens fun foods were of PNQ. Of these 326 items, 202 (62.0%) have nutrition claims on the front of the package. 3 A total of 222 products comprised the dry goods category. One product did not display information regarding the amount of fat and/or calories in the product; another two did not reveal information regarding sodium; and two regarding sugar and/or calories. Thus, a valid sample of products from the dry goods category for analysis of nutritional quality with regard to fat is 221 products, to sodium is 220 products and to sugar is 220. Table 5 Overview of nutrition claims made on the front of the box* Poor nutritious quality due to either fat, sodium, and/or sugar Not poor nutrition Claim on front box Frequency/%* Frequency/%* Some sort of claim (202) 62.0 (27) 69.2 No claim (124) 38.0 (12) 30.7 Total (336) (39) This table provides a cross-tabulation of poor/not poor food and claims, not simple frequencies of claims. *c 2 = 0.79 (r < 0.38); Phi = 0.05 (r < 0.38). Table 6 Overview of nutritional claims made on front of the box of products of poor nutritional quality due to sugar Poor nutritious quality due to sugar Not poor nutrition Claim on front box Frequency/%* Frequency/%* No claim (81) 32.0 (55) 49.5 Some sort of claim (172) 68.0 (56) 50.5 Total (253) (111) *c 2 = 10.4 (r < 0.001); Phi = 0.17 (r < 0.001). 372 Assessing fun foods C. Elliott obesity reviews 2007 The Author Journal compilation 2007 International Association for the Study of Obesity. obesity reviews 9, 368377 Remarkably, over 75% of products (22 out of 29) that have a Smart Spot! or nutrition mark/seal on the front of the package also contain over 10 g sugar per serving the amount Fitzhugh and Lobstein classify as high levels of sugar (22). Almost all of the products (21 out of 22) that claim to be low fat also contain over 10 g of sugar per serving. Seventy-nine per cent of products (19 out of 24) that claim to be a source of calcium are equally a source of high sugar a claimthat is not emphasized. Made with real fruit juice, in around 83%of cases (15 out of 18), equally brings a high sugar content along with the juice claim. Similarly, products with nutrition claims may equally be high in fat or sodium. It is worth noting that Health Canada, under the Food and Drugs Act, states that a low fat food contains less than 3 g fat per serving (24,25), yet many childrens foods have higher percentages of fat. Moreover, the disparity between claims regarding fat and actual fat content may not be recognized by all consumers. For instance, 87.5% of products that claimto be trans-fat-free also have over 3 g of fat per 100 g serving size. Asimilar observation can be made for sodium. Health Canada identies as healthy products containing 480 mg of sodium or less per serving. However, a number of products in our analysis had high levels of sodium along with a nutrition claim (e.g. 30% of products claiming a source of calcium; 27.3%of products claiming no articial avours/colours). Sugar Of the 364 products that listed nutritional information on sugar, 253 (69.5%) were classied as of PNQ owing to a high proportion of calories coming from sugar. Yet 172 products (68.0% of the sample of products with high sugar) contained a specic nutrition claim on the front of the products package (Tables 6 and 7). Cereals, fruit snacks and granola/cereal bars and drinks/drink boxes were particularly likely to have nutrition claims along with high levels of sugar. Fat Of the 365 products that listed details on fat, 83 products (22.7%) were of PNQ owing to the >35% of total calories coming from fat. However, 37%(31 products) of the prod- ucts tagged as of PNQ due to fat made a specic nutrition claim on the front of the product package (Tables 8 and 9). Peanut butters mixed with chocolate claimed to be a source of six essential nutrients, for instance, or pizza pops (a pre-cooked, calzone-type snack available in Canada) highlighted a source of calcium. Several manu- facturers of crackers also made specic nutrition claims emphasizing no trans-fats. Table 7 Nutrition claims on products of poor nutritional quality due to sugar Less than 20% calories from sugar (not a poor nutrition) More than 20% calories from sugar (poor nutrition) Claim on front box Frequency/%* Frequency/%* No specic claim (53) 51.5 (82) 31.5 Smart spot/nutrition mark/seal (10) 9.7 (19) 7.3 No articial avours/colours (3) 2.9 (8) 3.1 Low fat (3) 2.9 (21) 8.1 Whole wheat/bre (1) 1.0 (4) 1.5 Source of calcium (6) 5.8 (18) 6.9 Non-hydrogenated oil (3) 2.9 No trans-fat (6) 5.8 (10) 3.8 Organic (1) 1.0 (2) 0.8 No preservative (2) 1.9 (6) 2.3 Source of x essentials (4) 3.9 (10) 3.8 Made with real fruit juice (22) 8.5 Peanut-free (2) 1.9 (6) 2.3 Other (9) 8.7 (52) 20.0 Total (103) (260) *c 2 = 34.3 (r < 0.001) ; Cramers V = 0.31 (r < 0.001). Table 8 Overview of nutritional claims made on front of the box of products of poor nutritional quality due to fat Poor nutritious quality due to fat Not poor nutrition Claim on front box Frequency/%* Frequency/%* No claim (52) 62.7 (83) 29.5 Some sort of claim (31) 37.3 (198) 70.5 Total (83) (281) *c 2 = 30.11 (r < 0.0001); Phi = -0.29 (r < 0.0001). obesity reviews Assessing fun foods C. Elliott 373 2007 The Author Journal compilation 2007 International Association for the Study of Obesity. obesity reviews 9, 368377 Sodium Of the 364 products with information regarding sodium content, 62 (17.0%) were PNQ due to high sodium. Of this, almost 34%(21 products) made a specic nutrition claim on the front of the package (Tables 10 and 11). Particular appeals were made regarding the foods calcium level (Source of calcium = 11.3%) or the fact that the product was free of trans-fats (No trans-fat = 9.7%). Crackers and pizza pops/pogos were particularly notable in this regard. 4 Package appeals and product nutrition Key indicators of fun food are often found through the food packagings font and graphics. As such, it is certainly not surprising that 84% of coded products used either a car- toonish script or a crayoned font (suggesting a childs hand- writing) to identify itself as childrens fare. Three out of four products also had a cartoon image on the front of the box predominantly an anthropomorphized animal or gure (47.8%) or the cartoon image of a human boy or girl (24.5%). Interestingly, almost one-fth of the sample (19.3%) pictured the cartoon image engaged in some type of physical activity, from skateboarding (3.9%) and snowboarding/skiing (3.6%) to soccer/basketball (1.8%) and biking (1.1%). Over 25% of the sample (28.6%) also showed cartoon images engaged in activity on the back or side of the box. Selling food through the use of cartoon gures either skateboarding, snowboarding or engaged in some type of sporting activity raises the question of the relationship between the active nature of the cartoon image speci- cally that found on the front of the package (as it is the most visible) and the actual product nutrition contained within the package. The association between PNQ and the presence of cartoons engaged in physical activity was found to be statistically insignicant. However, it is worth noting that one in ve products with active images on the front of the box were of PNQ. Almost one-third (31%) of products of PNQ owing to high sodium had a cartoon image engaged in some kind of physical activity displayed on the front of the package (Table 12). 4 Pizza pops are a calzone-type snack available in Canada, which are sold pre-cooked and frozen. Pogos are battered cooked frankfurters on a stick. The association between claims made on the front of the box and products actual nutritional quality (i.e. poor or not poor) is statistically signicant. The relationship is of moderate-strong strength, as indicated by the measure of Cramers V (equal to 0.39). Table 9 Nutrition claims on products of poor nutritional quality due to fat Under 35% of calories coming from fat (not a poor nutrition) Over 35% of calories coming from fat (poor nutrition) Claim on front box Frequency/%* Frequency/%* No specic claim (83) 29.5 (52) 62.7 Smart spot/nutrition mark/seal (27) 9.6 (2) 2.4 No articial avours/colours (11) 3.9 Low fat (22) 7.8 (2) 2.4 Wholewheat/bre (5) 1.8 Source of calcium (19) 6.8 (5) 6.0 Non-hydrogenated oil (3) 3.6 No trans-fat (11) 3.9 (5) 6.0 Organic (3) 1.1 No preservative (8) 2.8 Source of x essentials (14) 5.0 (1) 1.2 Made with real fruit juice (22) 7.8 Peanut-free (3) 1.1 (5) 6.0 Other (53) 18.9 (8) 9.6 Total (281) (83) *c 2 = 63.9 (r < 0.0001); Cramers V = 0.42 (r < 0.0001). Table 10 Overview of nutritional claims made on front of the box of products of poor nutritional quality due to sodium Poor nutritious quality due to sodium Not poor nutrition Claim on front box Frequency/%* Frequency/%* No claim (41) 66.1 (94) 31.1 Some sort of claim (21) 33.9 (208) 68.9 Total (62) (302) *c 2 = 27.01 (r < 0.0001); Phi = -0.27(r < 0.0001). 374 Assessing fun foods C. Elliott obesity reviews 2007 The Author Journal compilation 2007 International Association for the Study of Obesity. obesity reviews 9, 368377 In terms of fat, almost an equal proportion of foods of PNQ and not PNQ displayed cartoon images engaged in physical activity on the front of the package (23.5% and 18.1%, respectively) (Table 13). This also holds true for sugar, in which a similar proportion of foods of poor and not poor nutritious quality owing to sugar displayed cartoon images engaged in physical activity on the front of the package (19.5 and 18.2, respectively). As a whole, then, Table 11 Nutrition claims on products of poor nutritional quality due to sodium Low content of sodium High content of sodium Claim on front box Frequency/%* Frequency/%* No specic claim (94) 31.1 (41) 66.1 Smart spot/nutrition mark/seal (27) 8.9 (2) 3.2 No articial avours/colours (8) 2.6 (3) 4.8 Low fat (23) 7.6 (1) 1.6 Wholewheat/bre (5) 1.7 Source of calcium (17) 5.6 (7) 11.3 Non-hydrogenated oil (3) 1.0 No trans-fat (10) 3.3 (6) 9.7 Organic (3) 1.0 No preservative (8) 2.6 Source of x essentials (15) 5.0 (4) 4.4 Made with real fruit juice (22) 7.3 Peanut-free (7) 2.3 (1) 1.6 Other (60) 19.9 (1) 1.6 Total (302) (62) *c 2 = 51.8 (r < 0.0001); Cramers V = 0.38 (r < 0.0001). Table 12 Active activity of cartoon image on front of package (sodium) Low content of sodium High content of sodium Active activity of cartoon image Frequency/%* Frequency/%* No activity (202) 82.4 (22) 68.8 Some kind of activity (43) 17.6 (10) 31.3 Total (245) (32) *c 2 = 3.43 (r < 0.1); Phi = 0.11 (r < 0.1). Table 13 Active activity of cartoon image on front of package (fat) Under 35% of calories coming from fat (not a poor nutrition) Over 35% of calories coming from fat (poor nutrition) Active activity of cartoon image Frequency/% Frequency/% No activity (185) 81.9 (39) 76.5 Some kind of activity (41) 18.1 (12) 23.5 Total (226) (51) c 2 = 0.81 (r < 0.5); Phi = 0.05 (r < 0.4). Table 14 Active activity of cartoon image on front of package (sugar) Less than 20% calories from sugar (not a poor nutrition) More than 20% calories from sugar (not a poor nutrition) Active activity of cartoon image Frequency/% Frequency/% No activity (63) 81.8 (161) 80.5 Some kind of activity (14) 18.2 (39) 19.5 Total (77) (200) c 2 = 0.06 (r < 0.9); Phi = -0.02 (r < 0.8). obesity reviews Assessing fun foods C. Elliott 375 2007 The Author Journal compilation 2007 International Association for the Study of Obesity. obesity reviews 9, 368377 one cannot be certain that an image of activity on the package corresponds to a food product that is of good nutritional quality. High levels of sodium, fat or sugar are sometimes packaged along with the active image displayed on the box. Appeals to fun and value systems Fun foods, as noted, are explicitly coded as fun to chil- dren. Beyond the use of cartoon graphics and fonts, pack- ages may directly reference fun on the box (7.4%), with descriptors such as fun to eat or claims that whenever you eat them, youll agree they spell fun. Connections to fun are most frequently linked directly with the food, as in fun fried potatoes, fun, smile-shaped potatoes or nutrition and fun, all in one. Sometimes the very names of the foods are fun, as with Presidents Choice Mini Chefs Funshines Biscuits. Fun is equally connoted by the use of unusual product names or avours (38%) such as Bug-A-Licious pasta, Cha-Cha Cheezy macaroni and cheese or Choco- late Splat pudding. Finally, fun can be indicated by direct claims to the products unique characteristics: 18.3% of the sample verbally emphasized qualities including the foods interactivity (e.g. food is stackable, stretchable, peel- able, shred-able) or its transformative properties (e.g. food changes colour, size or shape or even glows in the dark). Reinforcing this theme is the use of both cross- merchandising claims and games/activities. One of every 10 products urges kids to collect points, enter a contest or use a code for a free download. Three of every 10 products display a game or activity on the package. Packages also encourage kids to interact with a web site (10.9%), while some products even provide a game to play using the food itself (3%). Remarkably, only 1.6% of all products coded had nutrition-related activities (such as games, word nds or matching exercises that have to do with teaching nutri- tion) on the back of the package. This low percentage of packages explicitly addressing nutrition afrms the weak correlation between fun and nutrition. Fun, in the world of food, is not about nutritional awareness. Comment Fun food is a rapidly category of foods that needs to be scrutinized in order to create a full picture of the food environment that can contribute to childhood health (and conversely, childhood obesity). Fun food products are no longer contained within the cereal aisle; they pervade the entire supermarket and are available for every eating expe- rience (breakfasts, snacks, lunches, dinners). Fun food mar- keting suggests that children should be served special foods made uniquely for them, that family meals should be recon- gured into foods for children and foods for adults, and that childrens fare should be more interactive, more colourful, increasingly removed fromthe natural. All of these sugges- tions can create a problematic relationship between food and children, working to create a space in which children become accustomed to the unnaturalness of food and learn to appreciate the value of food as fun, sport or distraction (instead of focusing on nutrition). As I discuss elsewhere (26), in consistently emphasizing a foods play factor, arti- ciality and general distance from regular foods, fun food marketing can work to create a particularly unnatural rela- tionship with food in children. Food becomes framed as entertainment, and this entertainment is both premised on and emphasizes the articiality of what is being consumed. (Remarkably, it is only in the world of childrens food that articiality is actually framed as a selling feature.) This study reveals that less than 1% of the foods speci- cally targeted at children in a Canadian supermarket are fruits and vegetables and that 89% of fun food products can be classied as of PNQowing to high levels of fat, sugar or sodium. Of these products, almost a quarter had a high proportion of calories from fat, seven out of 10 had a high proportion of calories fromsugar, and almost two out of 10 had high levels of sodium. As such, fun food is not neces- sarily nutritious food and this proves signicant as the obvious category of junk food was excluded from the sample. While caregivers are likely to purchase products that they hope their children will like, it clearly can result in a less nutritious diet than they might realize. More signicant, perhaps, is the fact that 63%of fun food products make one or more nutrition claims on the front of the package and that 62%of the foods classied as of PNQ(according to the specied criteria) have nutrition claims on the front of the package. Health Canada has strict regulations regarding the Nutrition Facts table; beyond this, it allows ve health claims to be used on food labels and is currently expanding its regulatory framework for other health claims (specically regarding whole grains and folic acid). However, policy attention needs to be directed towards the fact that the very existence of a nutrition claim (or claims) might be misleading. A nutrition claim (whether for low fat, source of iron, etc.) might also reasonably lead consumers to believe that the product as a whole is nutri- tious. However, this is not necessarily the case. Products claiming low fat might be of PNQ owing to high levels of sugar. Products claiming no trans-fats might be of PNQ owing to a high proportion of fat or high levels of sodium. Even the package images may create contradictory impressions regarding the product nutrition. While all foods provide a source of energy, not all foods are best choices for fuelling an active lifestyle. Products that display cartoon images engaged in activity on the front of the package create an association between activity and the particular food in question. When two in 10 products with active images on the front of the package prove to be of PNQ, there is a clear difculty. Consumers are left in a 376 Assessing fun foods C. Elliott obesity reviews 2007 The Author Journal compilation 2007 International Association for the Study of Obesity. obesity reviews 9, 368377 situation of communicative ambiguity in which certain fun food products displaying active images are healthy selections, while others are not. On 5 February 2007, Health Canada unveiled its newly revised food guide, Eating Well with Canadas Food Guide. This six-page booklet provides recommendations on food choices according to age and gender, and suggests a diet based largely on vegetables, fruits and grain products. It recommends that consumers read product labels, but it does not ag some of the perils of eating packaged goods. It does not, for instance, acknowledge the disjunct that might exist between nutrition claims and actual product nutrition. And while the guide recognizes the unique nutritional needs of children, it fails to acknowledge the expansive category of fun foods or to provide guidance for navigating through this selection of foods. All of these issues require redress, so as to help consumers make more informed choices in pur- chasing and assessing food products targeted at children. Conict of Interest Statement No conict of interest was declared. Acknowledgements Funding for this research was generously provided from the Canadian Institutes of Health Research. The author would like to thank Anton Maslov for his exceptional work in providing the statistical analysis. Dr. Andre Turcotte gra- ciously assisted in double-checking the results. Thanks, also, to graduate students Melissa Fennell, Lee-Anne Peluk and Cathy Allison for coding assistance. References 1. Canadian Institute for Health Information. Improving the Health of Canadians: Promoting Healthy Weights 2006. Canadian Population Health Initiative of the Canadian Institute for Health Information: Ottawa, 2006. 2. Crawford D, Jeffery R (eds). Obesity Prevention and Public Health. Oxford University Press: New York, 2005. 3. World Health Organization. Diet, Nutrition and the Preven- tion of Chronic Diseases. World Health Organization: Geneva, 2003. 4. Library of Parliament. The Obesity Epidemic in Canada. (Pre- pared by Starky S). PIRS: Ottawa, 2005. [WWW document]. URL http: / / www.parl.gc.ca/ information/ library/ PRBpubs/prb0511-e. htm#11 (accessed March 2007). 5. Berg FM. Underage and Overweight: Americas Childhood Obesity Crisis. Hatherleigh Press: New York, 2004. 6. Lobstein T, Baur L, Uauy R. Obesity in children and young people: a crisis in public health. Obes Rev 2004; 5: 4104. 7. Brownell KD, Puhl R, Schwartz MB, Rudd L (eds). Weight Bias: Nature, Consequences, and Remedies. Guilford: New York, 2005. 8. Lobstein T, Dibb S. Evidence of a possible link between obe- sogenic food advertising and child overweight. Obes Rev 2005; 6: 203208. 9. Mendelson R. Think tank on school-aged children: nutrition and physical activity to prevent the rise in obesity. Appl Physiol Nutr Metab 2007; 32: 495499. 10. Elliott C. Childhood obesity and our toxic environment: suggestions for future research. J Fam Consum Sci Educ 2005; 23: 4751. 11. Elliott C. The strange, the bizarre and the edible. Articiality and fun in the world of childrens food. In: Greenberg J, Elliott C (eds). Communication in Question: Competing Perspectives on Controversial Issues in Communication Studies. Thomson-Nelson: Toronto, 2007, pp. 8188. 12. Committee on Prevention of Obesity in Children and Youth/ Institute of Medicine. Preventing Childhood Obesity: Time for Action. National Academy Press: Atlanta, 2005. 13. Malik VS, Schulze M, Hu FB. Intake of sugar-sweetened beverages and weight gain: a systematic review. Am J Clin Nutr 2006; 84: 274288. 14. McGinnis JM, Gootman J, Kraak V (eds). Food Marketing to Children and Youth: Threat or Opportunity? National Academies Press: Washington. DC, 2006. 15. Botterill J, Kline S. Flow, branding and media saturation: towards a critical analysis of promotional marketing in childrens television. Paper Presented at Child and Teen Consumption, 2006. Copenhagen Business School, Copenhagen, April 2728, 2006. 16. Chamberlain L, Wang Y, Robinson T. Does childrens screen time predict requests for advertised products? Cross-sectional and prospective analyses. Arch Pediatr Adolesc Med 2006; 160: 363 368. 17. Harrison K, Marske A. Nutritional content of foods adver- tised during the television programs children watch most. Am J Public Health 2005; 95: 15681574. 18. Wilcox B, Kunkel D, Cantor J, Dorwick P, Linn S, Palmer E. American Psychological Association. Report of the APA Task Force on Advertising and Children. 2004. [WWW document]. URL http://www.apa.org/pi/cyf/advertisingandchildren.pdf (ac- cessed March 2007). 19. Consumers International. A Spoonful of Sugar: Television food advertising aimed at children, an international comparative survey. Consumers International Programme for Developed Economics: London, 1996. 20. Gantz W, Schwartz N, Angelini JR, Rideout V. Food for thought: television food advertising to children in the United States. Kaiser Family Foundation Report. 2007. [WWW document]. URL http://www.kff.org/entmedia/upload/7618.pdf (accessed May 2007). 21. Nestle M. What to Eat. North Point Press: New York, 2006. 22. Fitzhugh K, Lobstein T. Childrens food examined. an analysis of 358 products targeted at children. The Food Commission. 2000. [WWW document]. URL http://www.foodcomm.org.uk/ childrens_food_examined.htm (accessed March 2007). 23. Center for Science in the Public Interest. Dening foods of poor nutritional quality. [WWW document]. URL http: / / www.cspinet.org/ new/ pdf/ foods_of_poor_nutritional . pdf (accessed March 2007). 24. Health Canada. Guide to food labelling and advertising. [WWW document]. URL http://www.inspection.gc.ca/english/ bureau/labeti/guide/7-0-63.html (accessed March 2007). 25. Health Canada. Food and drugs act and regulations (SOR/ 2003-11, s. 20). [WWW document]. URL http://laws.justice.gc.ca/ en/F-27/C.R.C.-c.870/232590.html (accessed March 2007). 26. Elliott C. Marketing fun foods: a prole and analysis of supermarket food messages targeted at children. Can Public Policy (in press). obesity reviews Assessing fun foods C. Elliott 377 2007 The Author Journal compilation 2007 International Association for the Study of Obesity. obesity reviews 9, 368377