at www.jneb.org

Podcast available online Research Brief

Do Food Blogs Serve as a Source of Nutritionally BalancedRecipes? An Analysis of 6 Popular Food BlogsElizabeth P. Schneider, MS, RD; Emily E. McGovern, MS, RD; Colleen L. Lynch, MS;Lisa S. Brown, PhD, RD

Nutrition DAddress foCollege, 30E-mail: Eli�2013 SOhttp://dx.d

696

ABSTRACT

Objective: To determine whether sampled food blogs provide nutritionally balanced recipes.Methods: Two entree recipes per season, per year (2010–2011) were selected from 6 highly rankedfood blogs (n ¼ 96). Food Processor Nutrition and Fitness software was used to analyze sodium, saturatedfat, and energy content. Analysis was separated by protein type (vegetarian, poultry, redmeat, and seafood).Results: Recipes met energy recommendations but were excessive in saturated fat and sodium.Vegetarian and seafood recipes were significantly lower in risk nutrients compared with red meat andpoultry recipes. Red meat recipes were not significantly different from poultry recipes for risk nutrientsstudied; poultry recipes were higher in sodium and energy compared with red meat recipes.Conclusions and Implications: The public should be aware of the nutritional limitations ofpopular food blogs; dietitians could assist in modifying blog recipes for individuals and partner withbloggers to improve the nutritional profile of recipes.Key Words: food blogs, social media, recipe modification, saturated fat, sodium (J Nutr Educ Behav.2013;45:696-700.)

INTRODUCTION

Dietitians and public health profes-sionals often recommend cooking athome to help individuals reducesodium, saturated fat, and total en-ergy intake, but doing so successfullydepends on access to a source ofrecipes that will assist in meeting tar-gets for these risk-related nutrients.1

Increasingly, individuals are seekingrecipes through Internet sourcesincluding food blogs.2 With the grow-ing popularity of social media,dietitians and other public healthprofessionals must become aware ofthe quality of recipe informationthat is available and commonlyaccessed on the Internet.

Despite recommendations thatAmericans limit saturated fat and so-dium intake and balance total energyintake, national data suggest thatAmericans do not meet recommendedtargets.3,4-6 Frequent consumption offood away from home contributes to

epartment, Simmons College, Bostr correspondence: Elizabeth P. Schne0 The Fenway, Boston, MA 02115;zabethSchneiderRD@gmail.comCIETY FOR NUTRITION EDUCoi.org/10.1016/j.jneb.2013.07.002

excess energy, fat, and sodium.7,8

However, data also suggest that intakeof fat, energy, and sodium exceedrecommended levels even for thosewho eat out less often.9 Completeunderstanding as to why those whoeat more frequently at home exceedrecommendations warrants further re-search. Hypothesized reasons includelarger portions and higher frequencyof snacking.9 One area that should belooked at more closely involves thevirtual space in which consumers areaccessing recipes, and the nutritionaladequacy of those recipes.

According to a recent survey fromBlogHer,10 the top female social me-dia network, 89% of the US onlinepopulation has used the Internet tofind recipes, and those who do soplace high importance on cookingdinner every night. With a growingnumber of individuals cooking athome compared with eating out,food blogs may markedly affectnutrient intake.11

on, MAider, MS, Nutrition Department, SimmonsPhone: (774) 238-7707; Fax: (617) 521-3002;

ATION AND BEHAVIOR

Journal of Nutrition Education and Beh

Several factors may influence thenutritional content of recipes postedon food blog sites. The use of colorfulphotographs to attract readershipmay emphasize appearance overnutritional value. Popular bloggersmay also have partnerships withfood companies, which can influencethe ingredients and therefore thenutritional value of the posted rec-ipes. In this way, food blogs may serveas a mode of food advertising inwhich the reader's degree of trust inthe blog may have an important rolein consumerism.2 The BlogHer studyalso found that 71% of young foodblog readers aged 18–44 years havemade purchases based on a blog'srecommendation.10 The purpose ofthis study was to determine whethersampled food blogs provide nutrition-ally balanced recipes for the public.

METHODS

The authors performed a contentanalysis to examine nutrient valuesof recipes found on popular foodblogs. The Web site BlogRank12 wasused to identify the top 6 foodblogs relevant to the study.BlogRank collects data for approxi-mately 20,000 blogs and uses an algo-rithm based on ranking factorsincluding Rich Site Summary (RSS)

avior � Volume 45, Number 6, 2013

Journal of Nutrition Education and Behavior � Volume 45, Number 6, 2013 Schneider et al 697

memberships, unique monthly visi-tors, search engine ranks, number ofincoming links, pages per visit, andlink-to-page ratio. The ultimate rankfilter, which incorporates all of thesefactors, was used to determine blogpopularity for the final selection.Blogs were required to have frequentpostings and contain entr�ee recipesthat could be generalized to thepublic, as opposed to blogs that caterto special diets such as gluten free orvegetarian. Table 1 provides a descrip-tion of the food blogs analyzed forthis study.

The recipe selection protocol speci-fied 2 entrees per season per year forthe years 2010 and 2011, for a totalof 96 recipes representing the 6 blogs.Season was specified in 3-monthincrements as fall (September, Octo-ber, and November), winter (Decem-ber, January, and February), spring(March, April, and May), and summer(June, July, and August). Recipes wereeligible for analysis if they were postedduring the specified season and year,and if they were labeled ‘‘maincourse’’ or ‘‘entr�ee.’’ Recipes wereselected chronologically as listed onthe blog Web site to avoid selectionbias. Recipes excluded from analysisincluded those labeled as beverages,desserts, and hors d’oeuvres. Blogsdesigned specifically for restricteddiets were also excluded because ofthe potential for the special dietaryneeds to alter nutrient content forthe nutrients of interest.

Table 1. Food Blog Author Locations and

Name Location Unique V

Pinch My Salt California, US 42Chocolate andZucchini

Paris, France 103

Simply Recipes California, US 2,005Smitten Kitchen New York, US 325Pioneer Woman Oklahoma, US 655Busy at Home Nebraska, US 7

NA indicates not available.aNumbers are as of July, 2012; bUnique vi(a count of unique individual people viscounted once, no matter how many timerepresent traffic from the US only; cGoogleregardless of paid advertising efforts; dBlorank blogs in any category. Some of the famembership, incoming links, Compete Alsites popularity.

The authors analyzed a subset of15 recipes to determine inter-raterreliability among the 3 coders. Allblogs were represented in this subset.An acceptable level of agreement wasset before analysis at a score of 0.90.An intra-class correlation score wascalculated based on 6 nutrients (totalcalories, protein, total fat, cholesterol,saturated fat, and sodium) with anaverage score of 0.972, which con-firmed high inter-rater agreement.Food Processor Nutrition and Fitnesssoftware (version 10.9.0, ESHAResearch, Inc, Salem, OR, 2011) wasused to calculate nutrient values inthe recipes. If a specific ingredientcould not be found in the FoodProcessor software, an online searchwas performed to determine theclosest relative, which was entered inthe code book and used by all codersto maintain consistency. Ambiguousmeasurements such as ‘‘pinch ofsalt’’ were quantified and compiledinto an analysis protocol, also toensure consistency.

The Food Processor program(database released: August 8, 2011)provided a complete list of nutrientinformation for the recipes. In addi-tion to the nutrient analysis, the USDepartment of Agriculture's MyPlatetool was selected within the FoodProcessor program to assess thenutrient breakdown compared withthe 2010 Dietary Guidelines.13 Theas-purchased weight of food and theedible portion weight were accounted

Popularity Rankingsa

isitorsb Google Rankc Blog Rankd

,213 4 7,505 21 6

,628 15 NA,687 5 3,219 37 7,554 > 100 18

sitors according to http://Compete.comiting the site per month). Visitors ares they visit a site in a month. CountsOrganic Search Rank is search resultsg Rank uses over 20 different factors toctors include Rich Site Summary (RSS)exa and Technorati ranking, and social

for within the Food Processor pro-gram.

One-way analysis of variancetested variations between the depen-dant variables of calories, saturatedfat, and sodium and the independentvariables of blog, season, year, andprotein type (meat, seafood, poultry,and vegetable). Analysis of variancetests were run independently foreach category and all data werenormally distributed as determinedby Kolmogorov-Smirnov testingwithin the SPSS program (version10.1, SPSS, Inc, Chicago, IL, 2012).Seafood recipes were initially left outof the analysis because of insufficientpower as a result of the small samplesize (n ¼ 8) but were later includedtogether with the vegetable recipesfor a post hoc analysis of a pescatarianrecipe category. Post hoc Tukey analy-sis was used when comparing $ 3 ormore variables, such as the casewith seasons and protein types, toinvestigate original relationships. Inaddition, overall nutrient meanswere calculated for each category andfor the entire sample. Statistical signif-icance was set at P < .05. All statisticaltests were run using SPSS software.

Institutional review board reviewwas not required for this studybecause human subjects were notinvolved, as per the US Departmentof Health and Human Services guide-lines.14

RESULTS

The final sample consisted of 96recipes, 38% of which were classifiedas vegetarian (pasta, salads, soups/stews, casseroles, etc, without meat,but including dairy and egg); 33% ofthe recipes featured red meat as theprotein source; 21% featured poultry;and 8% featured seafood in the formof either fin fish or shellfish. No signif-icant differences were found for totalcalories, saturated fat, or sodiumamong the 6 blogs or the individualyear categories.

Although no significant differencesin calories or saturated fat wereobserved by season, significantdifferences were found for sodiumacross seasons (P ¼ .04). Recipesposted from December to May con-tained 30% more sodium comparedwith recipes posted from June toNovember.

698 Schneider et al Journal of Nutrition Education and Behavior � Volume 45, Number 6, 2013

As shown in Table 2, significantdifferences were identified for risk-related nutrients across protein cate-gories. The protein categories differedinmean number of calories (P¼ .001),saturated fat (P ¼ .006), and sodium(P < .001), and all risk-related nutri-ents of interest were significantlylower in vegetarian recipes comparedwith red meat and poultry recipes.When seafood recipes were combinedwith the vegetarian category inpost hoc analysis to create a pescatar-ian category, the mean number ofcalories (P < .001), saturated fat(P ¼ .001), and sodium (P < .001)remained significantly lower com-pared with recipes that containedeither red meat or poultry as theprotein source.

As shown in Table 3, mean valuesfor the risk-related nutrients of inter-est found that overall, the meannumber of calories across all 96 blogrecipes was within dietary guidelinerecommendations (516 vs 667 kcalrecommended for one third of thedaily value of a 2,000-kcal diet),whereas sodium and saturated fatexceeded recommendations (855 mg

Table 2. Nutrient Means of Food Blog Rec

NutrientsAll Recipes(n ¼ 96)

Calories, kcal 516

Calories, % Rcmdc 26

Saturated fat, g 9.4

Saturated fat, % Rcmd 47

Sodium (mg) 855

Sodium, % Rcmd 36

Trans fatty acid, g 0.19

Protein, g 29

Protein, % Rcmd 57

Carbohydrates, g 42

Carbohydrates, % Rcmd 14

Dietary fiber, g 4.3

Dietary fiber, % Rcmd 17

Fat, g 26

Fat, % Rcmd 40

Omega 3 fatty acid, g 0.2

Omega 6 fatty acid, g 1.7

Rcmd indicates recommended.aRed meat recipes included beef, pork,percentage of the recommended daily va

of sodium vs no more than 767 mgrecommended, and 9.36 g saturatedfat vs no more than 7 g recom-mended).3 This analysis did not takeinto account beverages, side dishes,and desserts, which could have con-tributed additional calories, sodium,and fat to themeal. Also, daily sodiumand saturated fat targets could havebeen met with other meals through-out the day.

When protein categories wereseparated, vegetarian recipes weresignificantly lower in calories com-pared with red meat recipes (400 vs592 kcal; P ¼ .01), saturated fat (6.6vs 11.7 g; P ¼ .012), and sodium(605 vs 944 mg; P ¼ .014). Vegetarianrecipes were also significantly lowerin calories compared with poultryrecipes (400 vs 652 kcal; P ¼ .003),saturated fat (6.6 vs 11.7 g; P ¼ .035),and sodium (605 vs 1,252 mg;P < .001). Red meat recipes were notsignificantly different from poultryrecipes in terms of the risk nutrientsstudied, although poultry recipeswere higher in mean sodium andcalorie content compared with redmeat recipes.

ipes and Recommended Percentage Based

Red Meata

Recipes (n ¼ 32)Seafood

Recipes (n ¼ 8) R

592 398

30 22

11.7 6.6

59 33

944 631

39 26

0.28 0.20

40 24

80 47

34 39

11 13

4.1 2.9

17 11

32 16

49 24

0.2 0.5

1.6 0.9

and rabbit; bPoultry recipes included chicklue for a given nutrient.

DISCUSSION

This study compared popular foodblog entr�ee recipes with one third ofthe Dietary Reference Intakes similarto menu planning techniques usedin the National School Lunch Pro-gram.15 Using this proportion,a meal should have included appro-ximately 667 calories, 17 g protein,100 g carbohydrate, 22 g fat, 8.3 gfiber, 6.7 g saturated fat, and 767 mgsodium. Compared with these stan-dards, the overall mean nutrientcontent of these food blog recipeswas acceptable in calories, low incarbohydrate and fiber, and high insaturated fat and sodium. Althoughthe addition of side dishes, beverages,and other meal components mayalter the nutrient content of themeal, potentially raising carbohydrateand fiber into the target range,sodium and saturated fat would con-tinue to exceed recommendationsfor a meal featuring the analyzedentr�ee recipes.

To the knowledge of the re-searchers, this study was unique inits focus on online recipes from food

on Daily Value

Vegetableecipes (n ¼ 36)

Poultryb

Recipes (n ¼ 20)

400 652

20 33

6.6 11.7

33 58

605 1,252

25 52

0.11 0.20

14 39

27 79

45 48

15 16

5.0 4.0

20 16

19 32

29 49

0.2 0.2

1.1 3.3

en and turkey; c% Rcmd denotes the

Table 3. Post Hoc (Tukey) Results Comparing Nutrient Means Among Red Meat,Vegetable, and Poultry Recipes From Popular Food Blogs

Nutrient Protein Type Comparisons Pa

Calories, kcal Vegetable to red meatVegetable to poultry

.010

.003Saturated fat, g Vegetable to red meat

Vegetable to poultry.012.035

Sodium, mg Vegetable to red meatVegetable to poultry

.014< .001

aStatistical variance in nutrient means between different protein types was calcu-lated using 1-way analysis of variance.

Journal of Nutrition Education and Behavior � Volume 45, Number 6, 2013 Schneider et al 699

blogs. A key word search of thePubMed, Google Scholar, and Oviddatabases revealed that no studiesinvestigating recipes embedded with-in food blogs or otherwise posted onInternet sources. A limited numberof similar studies, however, assessedthe nutrient content of recipesdistributed by television chefs. A2012 study by Howard et al16 ana-lyzed the nutritional content of100 main meal recipes from 5 best-selling cookbooks authored by televi-sion chefs in the United Kingdom.The investigators found that the rec-ipes were excessive in total energy,protein, fat, and saturated fat com-pared with World Health Organiza-tion recommendations. Althoughsodium complied with recommenda-tions, the food industry in the UnitedKingdom adopted voluntary reduc-tion of sodium in processed foodsthat included canned or jarred ingre-dients commonly used in recipesstarting in 2006.17

A 2010 study by Silva et al18 ap-plied a healthy recipe index to 204recipes featured on 2 Food Networkshows and found that recipes wereranked as less than healthful by theindexmeasure. Further analysis foundthat the recipes analyzed were exces-sive in energy, saturated fat, andsodium based on a 2,300-kcal diet.

In this study, poultry recipes werenot significantly different in terms ofsaturated fat from red meat recipesand were high in sodium content.Post-analysis review to determine themain sources of saturated fat andsodium in the poultry recipes identi-fied that half of the recipes eitherfeatured skin-on pieces or werepoultry-based casserole dishes madewith butter, cheese, and/or cream

and other ingredients high in fat andsodium. Blog readers may be unawareof the unhealthy effects of addingthese ingredients, and may insteadassume that the dish is healthybecause it features chicken or turkey.

This study was a first step in deter-mining the relationship betweensocial media and nutrition. Futureanalysis should examine recipesearch tools and popular recipeWeb sites. Because of a limited sam-ple, this study did not compre-hensively investigate the nutrientcontent of seafood recipes. Furtherinvestigation may include an over-sampling of seafood recipes to deter-mine whether these recipes arenutritionally balanced or whetherthey contain excess amounts of risk-related nutrients.

IMPLICATIONS FORRESEARCH ANDPRACTICE

This study revealed an opportunity fordietitians and public health profes-sionals to improve recipes accessedon social media. Beyond helpingindividual clients modify recipesfound on these sites, dietitians andpublic health professionals couldaddress recipes posted online ina broader manner. Opportunitiesinclude encouraging more dietitiansto write their own blogs. Dietitiansmay also partner with existing foodblog authors to add more healthy op-tions or to create alternative, healthyversions of each recipe using modifiedingredients. Other possible interven-tions include designing a brandedicon for labeling recipes that meetspecific nutritional standards.

The sampled food blog recipes wereon average acceptable in terms oftotal calories but excessive in termsof saturated fat and sodium.When an-alyzed by protein type, there weresignificant differences among vege-tarian, poultry, and meat recipes;poultry and red meat recipes weresignificantly higher in sodium andsaturated fat. Vegetarian recipes werelower in calories, saturated fat, andsodium amounts compared withother recipes.

REFERENCES

1. Aub�e J, Marquis M. Attitudes andhabits of Canadians in relation to plan-ning and preparing meals at home.Can J Diet Pract Res. 2011;72:70-75.

2. Hui-Yi H, Chang PC. Influence ofmessage trust in online word-of-mouth on consumerism behavior byexample of food blog. 2010 InternationalConference on Electronics and InformationEngineering. 2010;1:395-399.

3. United States Department of Agricul-ture and US Department of Healthand Human Services. Dietary GuidelinesforAmericans, 2010. 7th ed.Washington,DC: US Government Printing Office;2010.

4. Hoy MK, Goldman JD, Murayi T,Rhodes DG, Moshfegh AJ. Sodium in-take of the U.S. population: what we eatin America, NHANES 2007-2008. FoodSurveys Research Group dietary data brief,no 8. October, 2011. http://ars.usda.gov/SP2UserFiles/Place/12355000/pdf/DBrief/8_sodium_intakes_0708.pdf. Ac-cessed July 29, 2013.

5. Wright JD,WangC-Y.Trends in intake ofenergy and macronutrients in adults from1999-2000 through 2007-2008. NCHSdata brief, no 49. Hyattsville, MD: Na-tional Center for Health Statistics; 2010.

6. Oh K, Hu FB,Manson JE, et al. Dietaryfat intake and risk of coronary heartdisease in women: 20 years of follow-up of the Nurses’ Health Study. Am JEpidemiol. 2005;161:672-679.

7. Lachat C, Nago E, Verstraeten R, et al.Eating out of home and its associationwith dietary intake: a systematic reviewof the evidence. Obes Rev. 2012;13:329-346.

8. Powell LM, Nguyen BT. Fast-food andfull-service restaurant consumptionamong children and adolescents:effect on energy, beverage, and nutrientintake. JAMA Pediatr. 2013;167:14-20.

700 Schneider et al Journal of Nutrition Education and Behavior � Volume 45, Number 6, 2013

9. Briefel RR, Johnson CL. Secular trendsin dietary intake in the United States.Annu Rev Nutr. 2004;24:401-431.

10. BlogHer. Food Facts 2012 BlogHerStudy. www.blogher.com/food-facts-2012-blogher-study. Accessed July 29,2013.

11. Tlapa C, Miller R, Washington K.Eating at Home: Restaurant, Food andBeverage Market Research Handbook.Loganville, GA: Richard K. Miller &Associates; 2010.

12. BlogRank. Top 50 food blogs. http://food.invesp.com.Accessed July 16, 2012.

13. US Department of Agriculture and USDepartment of Health and Human Ser-vices. Dietary Guidelines for Americans,2010. 7th edition. Washington, DC:US Government Printing Office; 2010.

14. United States Department of Health andHuman Services. Human subjectregulations decision charts. http://www.hhs.gov/ohrp/policy/checklists/decisioncharts.html#c1. Accessed July 29, 2013.

15. National School Lunch Program. http://frac.org/federal-foodnutrition-programs/national-schoollunch-program/.AccessedMarch 18, 2013.

16. Howard S, Adams J, White M. Nutri-tional content of supermarket readymeals and recipes by television chefs inthe United Kingdom: cross sectionalstudy. BMJ. 2012;345:e7607.

17. Vogel L. Voluntary sodium reductionsfar from ‘‘uncharted.’’ CMAJ. 2010;182:E575-E576.

18. Silva N, di Bonaventura E, Byrnes C,Herbold NH. Are 30-minute cookingshows on the food network an optionfor dietitians to recommend duringcounseling? Topics in Clinical Nutrition.2010;25:70-74.