Fast food breakfasts: Any better than burgers? A data science based approach

Introduction

Obesity in the United States of America

Obesity rates in the US have continued to increase (Stokes, Ni, & Preston, 2017). The last brief from Centers for Disease Control shows that in 2015–2016 the prevalence of obesity was 39.8% in adults and 18.5% in youth (Hales, Carroll, Fryar, & Ogden, 2017)

There are numerous factors contributing to this obesity epidemic. However, one of the main causes seems to be the easy availability of ultra-processed foods (Ziauddeen, Alonso-Alonso, Hill, Kelley, & Khan, 2015).

Even when there are several classifications and profiling of what constitutes as ultra-processed foods, the consensus points mainly to those types of food made from cheap ingredients and additives, usually low in nutritional quality, high in energy density, convenience, and highly palatable (Monteiro et al., 2017; Moubarac, Parra, Cannon, & Monteiro, 2014; O’Halloran et al., 2017).

It seems significant to note that the composition of fast-foods is typically energy-dense, low in micro-nutrients and fiber, and high in fat, salt, and sugar (Goyal & Singh, 2007).

Fast food restaurants and obesity

Food away from home accounts for nearly half of all U.S. consumer food expenditures (Okrent & Alston, 2012).

Fast-food restaurants offer high calorie and rewarding foods at relatively low prices (Chou, Grossman, & Saffer, 2004). The accessibility, ubiquity and proximity of fast-food restaurants has been linked to worse diets and higher rates of obesity among children (Alviola IV, Nayga, Thomsen, Danforth, & Smartt, 2014), adolescents (Davis & Carpenter, 2009) and adults (Jeffery, Baxter, McGuire, & Linde, 2006).

The fast-food industry has built a reputation for providing unhealthy food, but consumer trends in the United States continue to shift toward healthier options (Carden, Maldonado, & Boyd, 2018).

Information and decisions

Access to nutrition information (menu calorie labeling) in fast food restaurants might have a positive effect on some consumers; influencing them to make healthier food choices (Roberto, Schwartz, & Brownell, 2009). This certainly doesn’t work for all individuals (Swartz, Braxton, & Viera, 2011) considering more than 60% of people never read the information even when it’s available (Devine, Farrell, Blake, Wethington, & Bisogni, 2009); and evidence shows that taste is the most important factor to food choice (Kourouniotis et al., 2016).

However, nutrition information has been proven useful for consumers who base an significant part of their food decisions on health (Brissette, Lowenfels, Noble, & Spicer, 2013). Regardless of the vast amount of information available, people typically underestimate calorie (Block et al., 2013) and sodium (Moran, Ramirez, & Block, 2017) content of fast food restaurant menus.

Consumers and “healthy food” perception.

Various studies suggest that there’s a natural tendency to categorize food as either intrinsically healthy or unhealthy (Carels, Harper, & Konrad, 2006a; Carels, Konrad, & Harper, 2007; Chandon & Wansink, 2012)

This tendency might have its roots in the concept of heuristics, simple rules of thumb, proposed by Tversky & Kahneman (Tversky & Kahneman, 1974).

Because of this, rather than making informed and reasoned food choices, some people might rely on simplistic judgments to make their decisions. The mental categorization of certain foods as healthy may mean that some food can be eaten in greater amounts because it is assumed to contribute to health (Ross & Murphy, 1999)

Objectives and hypothesis

This study is based on the assumption that the concept of Fast food might be more related, in the mind of the consumer, to food and beverage items usually consumed as lunch or dinner meals (i.e, burgers, pizzas, soft drinks), than to items usually consumed as breakfast meals (i.e. eggs sandwiches, pancakes, coffee beverages, etc). Thus, I set to find out if items offered as Breakfast are nutritionally healthier than Lunch/Dinner items.

My main hypothesis is that Breakfast’s items might be as unhealthy as the items offered in lunch/dinner. However, “Unhealthy” is not an objective concept. Because of that, I will operationalize the concept using the Dietary Guidelines for Americans 2010–2015, which focus on limiting, on the one hand, overall Calorie content, and on the other, three specific Nutrients, which are: Saturated fat, sodium, and sugars. There is an additional consideration and it has to do with the fact that Breakfast menus comprise both Food items and Beverage items. These types of items should be analyzed separately because of different nutritional features.

Therefore, my hypothesis are:
1. Breakfast Food items have in average as much or more:
a. Calories
b. Saturated fat (g)
c. Sodium (mg)
d. Sugars (g) than Lunch/dinner food items and that

2. Breakfast Beverage items have in average as much or more:
a. Calories
b. Saturated fat (g)
c. Sodium (mg)
d. Sugars (g) than Lunch/dinner beverage items.

Design and methods

For this study, I chose 5 of the most popular fast food Restaurants worldwide: Mc Donald’s, Burger King, Subway, Starbucks and Dunkin Donuts.

Selection criteria were based on the following conditions:
-Nutrition information available on the website.
-Menus offering both Breakfast and Lunch/Dinner Items.

Data were obtained from the US official website of each Restaurant. Some of the items were excluded from the ‘Data cleaning procedure’. Explanations of that exclusions are included in the corresponding sections. Items of each menu were assigned either to be part of Breakfast or of Lunch/Dinner. Items of each menu were also assigned to a Food or Beverage category.

This study was done entirely using R Studio.
The ggplot2 package was used to create the visuals.
The data cleaning procedure is not included in this paper, but can be obtained by contacting me.

IMPORTANT CAVEAT
I’ve included the code used for the hypothesis tests. If you are not interested in code or statistical analysis, just stick with the text and graphs. That’s where the important stuff is.

Statistical analysis

Statistical analysis was performed in R Studio software.
Two sample T-test was used to comparisons between means. Equality of variances was performed previously to each T-test to decide the type of assumption to include in the T-test (assuming or not assuming equal variances).

Statistical significance was established when p values were < 0.01.

Descriptive statistics of the data.

The following is a basic exploration of the descriptive statistics of the dataset (just for data science freaks).

names(fast_food_restaurant)
##  [1] "Item"                        "Calories"                   
##  [3] "Calories % Daily Value"      "Total Fat (g)"              
##  [5] "Fat % Daily Value"           "Saturated Fat (g)"          
##  [7] "Saturated Fat % Daily Value" "Trans Fat (g)"              
##  [9] "Cholesterol (mg)"            "Cholesterol % Daily Value"  
## [11] "Sodium (mg)"                 "Sodium % Daily Value"       
## [13] "Carbohydrates (g)"           "Carbohydrates % Daily Value"
## [15] "Dietary Fiber (g)"           "Dietary Fiber % Daily Value"
## [17] "Sugars (g)"                  "Protein (g)"                
## [19] "Type of item"                "Breakfast_or_lunch_dinner"  
## [21] "Food_or_beverage"            "Restaurant"
str(fast_food_restaurant)
## Classes 'tbl_df', 'tbl' and 'data.frame':    2168 obs. of  22 variables:
##  $ Item                       : chr  "WHOPPER Sandwich" "WHOPPER Sandwich with Cheese" "Bacon & Cheese WHOPPER Sandwich" "BBQ Bacon WHOPPER Sandwich" ...
##  $ Calories                   : int  660 740 790 800 900 980 310 1150 550 220 ...
##  $ Calories % Daily Value     : int  33 37 40 40 45 49 16 58 28 11 ...
##  $ Total Fat (g)              : num  40 46 51 51 58 64 18 79 27 8 ...
##  $ Fat % Daily Value          : int  62 71 78 78 89 98 28 122 42 12 ...
##  $ Saturated Fat (g)          : num  12 16 17 17 20 24 5 31 12 3 ...
##  $ Saturated Fat % Daily Value: int  60 80 85 85 100 120 25 155 60 15 ...
##  $ Trans Fat (g)              : num  1.5 2 2 2 3 3 0.5 3.5 1.5 0.5 ...
##  $ Cholesterol (mg)           : int  90 115 125 125 175 195 40 240 95 30 ...
##  $ Cholesterol % Daily Value  : int  30 38 42 42 58 65 13 80 32 10 ...
##  $ Sodium (mg)                : int  980 1340 1560 1540 1050 1410 390 2150 1140 380 ...
##  $ Sodium % Daily Value       : int  41 56 65 64 44 59 16 90 48 16 ...
##  $ Carbohydrates (g)          : int  49 50 50 53 49 50 27 49 48 26 ...
##  $ Carbohydrates % Daily Value: int  16 17 17 18 16 17 9 16 16 9 ...
##  $ Dietary Fiber (g)          : int  2 2 2 2 2 2 1 2 2 1 ...
##  $ Dietary Fiber % Daily Value: int  8 8 8 8 8 8 4 8 8 4 ...
##  $ Sugars (g)                 : int  11 11 11 14 11 11 7 10 10 6 ...
##  $ Protein (g)                : num  28 32 35 35 48 52 13 61 30 11 ...
##  $ Type of item               : chr  "Whopper sandwiches" "Whopper sandwiches" "Whopper sandwiches" "Whopper sandwiches" ...
##  $ Breakfast_or_lunch_dinner  : chr  "Lunch_dinner" "Lunch_dinner" "Lunch_dinner" "Lunch_dinner" ...
##  $ Food_or_beverage           : chr  "Food" "Food" "Food" "Food" ...
##  $ Restaurant                 : chr  "Burger King" "Burger King" "Burger King" "Burger King" ...
summary(fast_food_restaurant)
##      Item              Calories      Calories % Daily Value
##  Length:2168        Min.   :   0.0   Min.   : 0.0          
##  Class :character   1st Qu.: 160.0   1st Qu.: 8.0          
##  Mode  :character   Median : 270.0   Median :14.0          
##                     Mean   : 289.7   Mean   :14.5          
##                     3rd Qu.: 390.0   3rd Qu.:20.0          
##                     Max.   :1350.0   Max.   :68.0          
##  Total Fat (g)    Fat % Daily Value Saturated Fat (g)
##  Min.   : 0.000   Min.   :  0.00    Min.   : 0.000   
##  1st Qu.: 0.000   1st Qu.:  0.00    1st Qu.: 0.000   
##  Median : 5.000   Median :  8.00    Median : 2.500   
##  Mean   : 9.219   Mean   : 14.26    Mean   : 3.991   
##  3rd Qu.:15.000   3rd Qu.: 23.00    3rd Qu.: 7.000   
##  Max.   :79.000   Max.   :122.00    Max.   :33.000   
##  Saturated Fat % Daily Value Trans Fat (g)     Cholesterol (mg)
##  Min.   :  0.00              Min.   :0.00000   Min.   :  0.00  
##  1st Qu.:  0.00              1st Qu.:0.00000   1st Qu.:  0.00  
##  Median : 13.00              Median :0.00000   Median : 10.00  
##  Mean   : 20.08              Mean   :0.09451   Mean   : 33.01  
##  3rd Qu.: 35.00              3rd Qu.:0.00000   3rd Qu.: 40.00  
##  Max.   :165.00              Max.   :3.50000   Max.   :575.00  
##  Cholesterol % Daily Value  Sodium (mg)   Sodium % Daily Value
##  Min.   :  0.00            Min.   :   0   Min.   :  0.00      
##  1st Qu.:  0.00            1st Qu.:  75   1st Qu.:  3.00      
##  Median :  3.00            Median : 160   Median :  7.00      
##  Mean   : 11.01            Mean   : 364   Mean   : 15.19      
##  3rd Qu.: 13.00            3rd Qu.: 400   3rd Qu.: 17.00      
##  Max.   :191.00            Max.   :3570   Max.   :149.00      
##  Carbohydrates (g) Carbohydrates % Daily Value Dietary Fiber (g)
##  Min.   :  0.00    Min.   : 0.00               Min.   : 0.000   
##  1st Qu.: 27.00    1st Qu.: 9.00               1st Qu.: 0.000   
##  Median : 42.00    Median :14.00               Median : 1.000   
##  Mean   : 43.14    Mean   :14.38               Mean   : 1.176   
##  3rd Qu.: 56.00    3rd Qu.:19.00               3rd Qu.: 2.000   
##  Max.   :184.00    Max.   :61.00               Max.   :21.000   
##  Dietary Fiber % Daily Value   Sugars (g)      Protein (g)     
##  Min.   : 0.000              Min.   :  0.00   Min.   :  0.000  
##  1st Qu.: 0.000              1st Qu.:  7.00   1st Qu.:  2.000  
##  Median : 4.000              Median : 24.00   Median :  7.000  
##  Mean   : 4.809              Mean   : 30.36   Mean   :  9.367  
##  3rd Qu.: 8.000              3rd Qu.: 46.00   3rd Qu.: 13.000  
##  Max.   :84.000              Max.   :174.00   Max.   :105.000  
##  Type of item       Breakfast_or_lunch_dinner Food_or_beverage  
##  Length:2168        Length:2168               Length:2168       
##  Class :character   Class :character          Class :character  
##  Mode  :character   Mode  :character          Mode  :character  
##                                                                 
##                                                                 
##                                                                 
##   Restaurant       
##  Length:2168       
##  Class :character  
##  Mode  :character  
               

How many items per Restaurant are there in the fast_food_restaurant dataset?

fast_food_restaurant$Restaurant %>% table()
## .
##   Burger King Dunkin Donuts     McDonalds     Starbucks        Subway 
##           126           559           977           370           136

Results

The first is going to be a really general approach, comparing all these nutrients by belonging either to the Breakfast category or to the Lunch/Dinner one.

Breakfast_vs_Lunch_dinner_nutrients_of_interest <- fast_food_restaurant %>% 
  group_by(Breakfast_or_lunch_dinner) %>% 
  summarise_at(c("Calories", "Saturated Fat (g)","Sodium (mg)",
                 "Sugars (g)"), funs(mean))
Breakfast_vs_Lunch_dinner_nutrients_of_interest
## # A tibble: 2 x 5
##   Breakfast_or_lunch~ Calories `Saturated Fat ~ `Sodium (mg)` `Sugars (g)`
##   <fct>                  <dbl>            <dbl>         <dbl>        <dbl>
## 1 Lunch_dinner            308.             3.52          493.         27.8
## 2 Breakfast               280.             4.24          295.         31.7Breakfast_vs_Lunch_dinner_nutrients_of_interest <- fast_food_restaurant %>% 
  group_by(Breakfast_or_lunch_dinner) %>% 
  summarise_at(c("Calories", "Saturated Fat (g)","Sodium (mg)",
                 "Sugars (g)"), funs(mean))
Breakfast_vs_Lunch_dinner_nutrients_of_interest
## # A tibble: 2 x 5
##   Breakfast_or_lunch~ Calories `Saturated Fat ~ `Sodium (mg)` `Sugars (g)`
##   <fct>                  <dbl>            <dbl>         <dbl>        <dbl>
## 1 Lunch_dinner            308.             3.52          493.         27.8
## 2 Breakfast               280.             4.24          295.         31.7

Mean Calories — Breakfast vs. Lunch/Dinner

However, this doesn’t provide enough information about the characteristics of the items that comprise each category.

Given that we deal with both food and beverage items, it might be more appropriate to account for that.

Furthermore, there are some Restaurants that don’t provide soft drinks and others that don’t provide an interesting offer of breakfast beverages (coffee, tea, etc). That’s why it makes more sense to have a more detailed view of Beverage and Food items separately.

means_nutrients_of_interest_by_Beverage_or_Food_and_Breakfast_or_Lunch_dinner_Items <-
  fast_food_restaurant %>% group_by(Food_or_beverage, Breakfast_or_lunch_dinner) %>% 
  summarise_at(c("Calories", "Saturated Fat (g)","Sodium (mg)",
                 "Sugars (g)"), funs(mean))
# I will arrange them by descending order based on the number of Calories.
means_nutrients_of_interest_by_Beverage_or_Food_and_Breakfast_or_Lunch_dinner_Items %>% 
  arrange(desc(Calories))
## Warning: package 'bindrcpp' was built under R version 3.4.4
## # A tibble: 5 x 6
## # Groups:   Food_or_beverage [3]
##   Food_or_beverage Breakfast_or_lunch_dinner Calories `Saturated Fat (g)`
##   <fct>            <fct>                        <dbl>               <dbl>
## 1 Food             Lunch_dinner                 447.                6.96 
## 2 Food             Breakfast                    413.                8.26 
## 3 Beverage         Breakfast                    239.                2.96 
## 4 Beverage         Lunch_dinner                 182.                0.389
## 5 Special          Breakfast                     59.5               3.53 
## # ... with 2 more variables: `Sodium (mg)` <dbl>, `Sugars (g)` <dbl>

However, this is just an average of the 5 Restaurants, and it is more interesting to observe the differences between these categories among different restaurants.

Tests and Results

Food items

How many food items per restaurant are in this dataset?

Food_items$Restaurant %>% table()
.
##   Burger King Dunkin Donuts     McDonalds     Starbucks        Subway 
##            63           191           215            95           136

What’s the difference between the Breakfast and Lunch/Dinner food items?

Food_Items_Breakfast_vs_Lunch_dinner <-
  Food_items %>% 
  group_by(Breakfast_or_lunch_dinner) %>% 
  summarise(avg_calories = mean(Calories),
            avg_sat_fat= mean(`Saturated Fat (g)`),
            avg_sodium = mean(`Sodium (mg)`),
            avg_sugar = mean(`Sugars (g)`)) 
Food_Items_Breakfast_vs_Lunch_dinner
## # A tibble: 2 x 5
##   Breakfast_or_lunch_dinner avg_calories avg_sat_fat avg_sodium avg_sugar
##   <fct>                            <dbl>       <dbl>      <dbl>     <dbl>
## 1 Lunch_dinner                      447.        6.96       947.      9.67
## 2 Breakfast                         413.        8.26       753.     13.4

Hypothesis 1.a. Food items. Calories.

First, I’m going to compare the variances between the two groups to see what kind of T test should be used.

var.test(Food_items$Calories[Food_items$Breakfast_or_lunch_dinner=="Breakfast"],
       Food_items$Calories[Food_items$Breakfast_or_lunch_dinner=="Lunch_dinner"])
##  F test to compare two variances
## 
## data:  Food_items$Calories[Food_items$Breakfast_or_lunch_dinner == "Breakfast"] and Food_items$Calories[Food_items$Breakfast_or_lunch_dinner == "Lunch_dinner"]
## F = 0.9796, num df = 338, denom df = 360, p-value = 0.8484
## alternative hypothesis: true ratio of variances is not equal to 1
## 95 percent confidence interval:
##  0.7941066 1.2092699
## sample estimates:
## ratio of variances 
##          0.9795976

Given that the variances are equal, I will perform a T test with pooled variances.

t.test(Food_items$Calories[Food_items$Breakfast_or_lunch_dinner=="Breakfast"],
       Food_items$Calories[Food_items$Breakfast_or_lunch_dinner=="Lunch_dinner"],
       var.equal = TRUE)
## 
##  Two Sample t-test
## 
## data:  Food_items$Calories[Food_items$Breakfast_or_lunch_dinner == "Breakfast"] and Food_items$Calories[Food_items$Breakfast_or_lunch_dinner == "Lunch_dinner"]
## t = -2.3232, df = 698, p-value = 0.02046
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -61.288000  -5.144231
## sample estimates:
## mean of x mean of y 
##  413.3628  446.5789

According to the p-value obtained (0.02046), there is evidence to say that the mean Calories of Breakfast are not equal. This result does not support my hypothesis, given that Breakfast food items have on average, a smaller amount of Calories compared to Lunch_dinner items.

Avg_calories_All_food_items_Breakfast_vs_Lunch_plot <- Food_items %>% 
  ggplot(aes(x= Breakfast_or_lunch_dinner, y=Calories, fill=Breakfast_or_lunch_dinner)) +
  geom_boxplot() +
  labs(title="Average Calories per Food Item - Breakfast vs. Lunch_Dinner", y="Calories", x="") +
   theme(plot.title=element_text(size=10, 
                                    face="bold", 
                                    hjust=0.5),
         legend.text = element_text(size=8, face="bold"))
Avg_calories_All_food_items_Breakfast_vs_Lunch_plot

Hypothesis 1.b. Food items. Saturated fat

First, I’m going to compare the variances between the two groups to see what kind of T-test should be used.

var.test(Food_items$`Saturated Fat (g)`[Food_items$Breakfast_or_lunch_dinner=="Breakfast"],
       Food_items$`Saturated Fat (g)`[Food_items$Breakfast_or_lunch_dinner=="Lunch_dinner"])
## 
##  F test to compare two variances
## 
## data:  Food_items$`Saturated Fat (g)`[Food_items$Breakfast_or_lunch_dinner ==  and Food_items$`Saturated Fat (g)`[Food_items$Breakfast_or_lunch_dinner ==     "Breakfast"] and     "Lunch_dinner"]
## F = 0.87358, num df = 338, denom df = 360, p-value = 0.2082
## alternative hypothesis: true ratio of variances is not equal to 1
## 95 percent confidence interval:
##  0.708161 1.078391
## sample estimates:
## ratio of variances 
##          0.8735764

Given that the variances are equal, I will perform a T test with pooled variances.

t.test(Food_items$`Saturated Fat (g)`[Food_items$Breakfast_or_lunch_dinner=="Breakfast"],
       Food_items$`Saturated Fat (g)`[Food_items$Breakfast_or_lunch_dinner=="Lunch_dinner"],
       var.equal = TRUE)
## 
##  Two Sample t-test
## 
## data:  Food_items$`Saturated Fat (g)`[Food_items$Breakfast_or_lunch_dinner ==  and Food_items$`Saturated Fat (g)`[Food_items$Breakfast_or_lunch_dinner ==     "Breakfast"] and     "Lunch_dinner"]
## t = 3.7091, df = 698, p-value = 0.0002245
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  0.6118664 1.9881793
## sample estimates:
## mean of x mean of y 
##  8.264012  6.963989

According to the p value obtained (0.0002245), there is evidence to say that the mean Saturated Fat of Breakfast and Lunch/Dinner are not equal.

This result supports my hypothesis, given that Breakfast food items have in average, a greater amount of Saturated fat compared to Lunch_dinner items.

Avg_sat_fat_All_food_items_Breakfast_vs_Lunch_plot <- Food_items %>% 
  ggplot(aes(x= Breakfast_or_lunch_dinner, y=`Saturated Fat (g)`, 
             fill=Breakfast_or_lunch_dinner)) +  geom_boxplot() +
  labs(title="Average Saturated fat per Food Item - Breakfast vs. Lunch_Dinner", 
       y="Saturated fat (g)", x="") +
   theme(plot.title=element_text(size=10, 
                                    face="bold", 
                                    hjust=0.5),
         legend.text = element_text(size=8, face="bold"))
Avg_sat_fat_All_food_items_Breakfast_vs_Lunch_plot

Hypothesis 1.c. Food items. Sodium

First, I’m going to compare the variances between the two groups to see what kind of T test should be used.

var.test(Food_items$`Sodium (mg)`[Food_items$Breakfast_or_lunch_dinner=="Breakfast"],
       Food_items$`Sodium (mg)`[Food_items$Breakfast_or_lunch_dinner=="Lunch_dinner"])
## 
##  F test to compare two variances
## 
## data:  Food_items$`Sodium (mg)`[Food_items$Breakfast_or_lunch_dinner ==  and Food_items$`Sodium (mg)`[Food_items$Breakfast_or_lunch_dinner ==     "Breakfast"] and     "Lunch_dinner"]
## F = 0.89387, num df = 338, denom df = 360, p-value = 0.2962
## alternative hypothesis: true ratio of variances is not equal to 1
## 95 percent confidence interval:
##  0.7246151 1.1034477
## sample estimates:
## ratio of variances 
##          0.8938739

Given that the variances are equal, I will perform a T test with pooled variances.

t.test(Food_items$`Sodium (mg)`[Food_items$Breakfast_or_lunch_dinner=="Breakfast"],
       Food_items$`Sodium (mg)`[Food_items$Breakfast_or_lunch_dinner=="Lunch_dinner"],
       var.equal = TRUE)
## 
##  Two Sample t-test
## 
## data:  Food_items$`Sodium (mg)`[Food_items$Breakfast_or_lunch_dinner ==  and Food_items$`Sodium (mg)`[Food_items$Breakfast_or_lunch_dinner ==     "Breakfast"] and     "Lunch_dinner"]
## t = -4.6274, df = 698, p-value = 4.414e-06
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -275.5684 -111.3866
## sample estimates:
## mean of x mean of y 
##  753.3923  946.8698

According to the p value obtained (4.414e-06), there is evidence to say that the mean Sodium of Breakfast and Lunch/Dinner are not equal.

This result does not support my hypothesis, given that Breakfast foods items have in average, a smaller amount of Sodium compared to Lunch_dinner items.

Avg_sodium_All_food_items_Breakfast_vs_Lunch_plot <- Food_items %>% 
  ggplot(aes(x= Breakfast_or_lunch_dinner, y=`Sodium (mg)`, 
             fill=Breakfast_or_lunch_dinner)) +  geom_boxplot() +
  labs(title="Average Sodium per Food Item - Breakfast vs. Lunch_Dinner", 
       y="Sodium (mg)", x="") +
   theme(plot.title=element_text(size=10, 
                                    face="bold", 
                                    hjust=0.5),
         legend.text = element_text(size=8, face="bold"))
Avg_sodium_All_food_items_Breakfast_vs_Lunch_plot

Hypothesis 1.d. Food items. Sugars

First, I’m going to compare the variances between the two groups to see what kind of T test should be used.

var.test(Food_items$`Sugars (g)`[Food_items$Breakfast_or_lunch_dinner=="Breakfast"],
       Food_items$`Sugars (g)`[Food_items$Breakfast_or_lunch_dinner=="Lunch_dinner"])
## 
##  F test to compare two variances
## 
## data:  Food_items$`Sugars (g)`[Food_items$Breakfast_or_lunch_dinner ==  and Food_items$`Sugars (g)`[Food_items$Breakfast_or_lunch_dinner ==     "Breakfast"] and     "Lunch_dinner"]
## F = 2.2167, num df = 338, denom df = 360, p-value = 1.803e-13
## alternative hypothesis: true ratio of variances is not equal to 1
## 95 percent confidence interval:
##  1.796950 2.736406
## sample estimates:
## ratio of variances 
##            2.21669

Given that the variances are not equal, I will perform a Welch T test.

t.test(Food_items$`Sugars (g)`[Food_items$Breakfast_or_lunch_dinner=="Breakfast"],
       Food_items$`Sugars (g)`[Food_items$Breakfast_or_lunch_dinner=="Lunch_dinner"],
       var.equal = FALSE)
## 
##  Welch Two Sample t-test
## 
## data:  Food_items$`Sugars (g)`[Food_items$Breakfast_or_lunch_dinner ==  and Food_items$`Sugars (g)`[Food_items$Breakfast_or_lunch_dinner ==     "Breakfast"] and     "Lunch_dinner"]
## t = 5.0024, df = 586.25, p-value = 7.493e-07
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  2.294602 5.261073
## sample estimates:
## mean of x mean of y 
##  13.44543   9.66759

According to the p value obtained (7.493e-07), there is evidence to say that the mean Sugar of Breakfast and Lunch/Dinner are not equal.

This result supports my hypothesis, given that Breakfast food items have in average, a greater amount of Sugars compared to Lunch_dinner items.

Avg_sugar_All_food_items_Breakfast_vs_Lunch_plot <- Food_items %>% 
  ggplot(aes(x= Breakfast_or_lunch_dinner, y=`Sugars (g)`, 
             fill=Breakfast_or_lunch_dinner)) +  geom_boxplot() +
  labs(title="Average Sugar per Food Item - Breakfast vs. Lunch_Dinner", 
       y="Sugar (g)", x="") +
   theme(plot.title=element_text(size=10, 
                                    face="bold", 
                                    hjust=0.5),
         legend.text = element_text(size=8, face="bold"))
Avg_sugar_All_food_items_Breakfast_vs_Lunch_plot

Beverage items

First, I’m subsetting the data to work entirely on Beverages

fast_food_beverages <-
  fast_food_restaurant %>% 
  filter(Food_or_beverage == "Beverage")
write_csv(fast_food_beverages, "Fast_food_beverages.csv")

How many items per Restaurant are there in the fast_food_beverages dataset?

fast_food_beverages$Restaurant %>% table()
## .
##   Burger King Dunkin Donuts     McDonalds     Starbucks        Subway 
##            63           368           762           265             0

It’s important to point out that Subway does not have Beverage information in their menus, so we can’t analyze them. Furthermore, while McDonald’s and Burger King offer soft drinks, Starbucks and Dunkin Donuts don’t. Therefore, just Burger King and McDonald’s will be considered for Lunch/Dinner beverage items.

Let’s look at the Nutrients of interest:

Beverages_means_nutrients_of_interest_and_Breakfast_or_Lunch_dinner_Items <- 
  fast_food_beverages %>% group_by(Breakfast_or_lunch_dinner) %>% 
  summarise_at(c("Calories", "Saturated Fat (g)", "Sodium (mg)", "Sugars (g)"), funs(mean))
Beverages_means_nutrients_of_interest_and_Breakfast_or_Lunch_dinner_Items
## # A tibble: 2 x 5
##   Breakfast_or_lunch~ Calories `Saturated Fat ~ `Sodium (mg)` `Sugars (g)`
##   <fct>                  <dbl>            <dbl>         <dbl>        <dbl>
## 1 Lunch_dinner            182.            0.389          79.3         44.4
## 2 Breakfast               239.            2.96          151.          37.8

Hypothesis 2.a. Beverage items. Calories.

First, I’m going to compare the variances between the two groups to see what kind of T test should be used.

var.test(fast_food_beverages$Calories[fast_food_beverages$Breakfast_or_lunch_dinner=="Breakfast"],
       fast_food_beverages$Calories[fast_food_beverages$Breakfast_or_lunch_dinner=="Lunch_dinner"])

##  F test to compare two variances
## 
## data:  fast_food_beverages$Calories[fast_food_beverages$Breakfast_or_lunch_dinner ==  and fast_food_beverages$Calories[fast_food_beverages$Breakfast_or_lunch_dinner ==     "Breakfast"] and     "Lunch_dinner"]
## F = 0.97653, num df = 1061, denom df = 395, p-value = 0.7657
## alternative hypothesis: true ratio of variances is not equal to 1
## 95 percent confidence interval:
##  0.8266179 1.1464975
## sample estimates:
## ratio of variances 
##          0.9765289

Given that the variances are equal, I will perform a T test with pooled variances.

t.test(fast_food_beverages$Calories[fast_food_beverages$Breakfast_or_lunch_dinner=="Breakfast"],
       fast_food_beverages$Calories[fast_food_beverages$Breakfast_or_lunch_dinner=="Lunch_dinner"],
       var.equal = TRUE)
## 
##  Two Sample t-test
## 
## data:  fast_food_beverages$Calories[fast_food_beverages$Breakfast_or_lunch_dinner ==  and fast_food_beverages$Calories[fast_food_beverages$Breakfast_or_lunch_dinner ==     "Breakfast"] and     "Lunch_dinner"]
## t = 6.5891, df = 1456, p-value = 6.168e-11
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  40.28953 74.44716
## sample estimates:
## mean of x mean of y 
##  239.3380  181.9697

According to the p value obtained (6.168e-11), there is evidence to say that the mean Calories of Breakfast are not equal. This result supports my hypothesis, given that Breakfast beverage items have in average, a greater amount of Calories compared to Lunch_dinner items.

Avg_calories_All_beverage_items_Breakfast_vs_Lunch_plot <- fast_food_beverages %>% 
  ggplot(aes(x= Breakfast_or_lunch_dinner, y=Calories, fill=Breakfast_or_lunch_dinner)) +
  geom_boxplot() +
  labs(title="Average Calories per Beverage Item - Breakfast vs. Lunch_Dinner",
       y="Calories", x="") +
   theme(plot.title=element_text(size=10, 
                                    face="bold", 
                                    hjust=0.5),
         legend.text = element_text(size=8, face="bold"))
Avg_calories_All_beverage_items_Breakfast_vs_Lunch_plot

Hypothesis 2.b. Beverage items. Saturated fat

First, I’m going to compare the variances between the two groups to see what kind of T test should be used.

var.test(fast_food_beverages$`Saturated Fat (g)`[fast_food_beverages$Breakfast_or_lunch_dinner ==
                                                   "Breakfast"],
                                                   fast_food_beverages$`Saturated Fat (g)`[fast_food_beverages$Breakfast_or_lunch_dinner ==
                                                   "Lunch_dinner"])
## 
##  F test to compare two variances
## 
## data:  fast_food_beverages$`Saturated Fat (g)`[fast_food_beverages$Breakfast_or_lunch_dinner ==  and fast_food_beverages$`Saturated Fat (g)`[fast_food_beverages$Breakfast_or_lunch_dinner ==     "Breakfast"] and     "Lunch_dinner"]
## F = 4.6976, num df = 1061, denom df = 395, p-value < 2.2e-16
## alternative hypothesis: true ratio of variances is not equal to 1
## 95 percent confidence interval:
##  3.976466 5.515254
## sample estimates:
## ratio of variances 
##           4.697616

Given that the variances are not equal, I will perform a Welch T test.

t.test(fast_food_beverages$`Saturated Fat (g)`[fast_food_beverages$Breakfast_or_lunch_dinner=="Breakfast"],
       fast_food_beverages$`Saturated Fat (g)`[fast_food_beverages$Breakfast_or_lunch_dinner=="Lunch_dinner"],
       var.equal = FALSE)
## 
##  Welch Two Sample t-test
## 
## data:  fast_food_beverages$`Saturated Fat (g)`[fast_food_beverages$Breakfast_or_lunch_dinner ==  and fast_food_beverages$`Saturated Fat (g)`[fast_food_beverages$Breakfast_or_lunch_dinner ==     "Breakfast"] and     "Lunch_dinner"]
## t = 16.29, df = 1396.1, p-value < 2.2e-16
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  2.264617 2.884724
## sample estimates:
## mean of x mean of y 
## 2.9635593 0.3888889

According to the p value obtained (2.2e-16), there is evidence to say that the means of Saturated Fat of Breakfast and Lunch/Dinner are not equal.

This result supports my hypothesis, given that Breakfast beverage items have in average, a greater amount of Saturated fat compared to Lunch_dinner items.

Avg_sat_fat_All_beverage_items_Breakfast_vs_Lunch_plot <- fast_food_beverages %>% 
  ggplot(aes(x= Breakfast_or_lunch_dinner, y=`Saturated Fat (g)`, fill=Breakfast_or_lunch_dinner)) +
  geom_boxplot() +
  labs(title="Average Saturated fat per Beverage Item - Breakfast vs. Lunch_Dinner",
       y="Saturated fat (g)", x="") +
   theme(plot.title=element_text(size=10, 
                                    face="bold", 
                                    hjust=0.5),
         legend.text = element_text(size=8, face="bold"))
Avg_sat_fat_All_beverage_items_Breakfast_vs_Lunch_plot

Hypothesis 2.c. Beverage items. Sodium

First, I’m going to compare the variances between the two groups to see what kind of T test should be used.

var.test(fast_food_beverages$`Sodium (mg)`[fast_food_beverages$Breakfast_or_lunch_dinner=="Breakfast"],
       fast_food_beverages$`Sodium (mg)`[fast_food_beverages$Breakfast_or_lunch_dinner=="Lunch_dinner"])
## 
##  F test to compare two variances
## 
## data:  fast_food_beverages$`Sodium (mg)`[fast_food_beverages$Breakfast_or_lunch_dinner ==  and fast_food_beverages$`Sodium (mg)`[fast_food_beverages$Breakfast_or_lunch_dinner ==     "Breakfast"] and     "Lunch_dinner"]
## F = 2.5457, num df = 1061, denom df = 395, p-value < 2.2e-16
## alternative hypothesis: true ratio of variances is not equal to 1
## 95 percent confidence interval:
##  2.154912 2.988807
## sample estimates:
## ratio of variances 
##           2.545716

Given that the variances are not equal, I will perform a Welch T test.

t.test(fast_food_beverages$`Sodium (mg)`[fast_food_beverages$Breakfast_or_lunch_dinner=="Breakfast"],
       fast_food_beverages$`Sodium (mg)`[fast_food_beverages$Breakfast_or_lunch_dinner=="Lunch_dinner"],
       var.equal = FALSE)
## 
##  Welch Two Sample t-test
## 
## data:  fast_food_beverages$`Sodium (mg)`[fast_food_beverages$Breakfast_or_lunch_dinner ==  and fast_food_beverages$`Sodium (mg)`[fast_food_beverages$Breakfast_or_lunch_dinner ==     "Breakfast"] and     "Lunch_dinner"]
## t = 13.836, df = 1123.8, p-value < 2.2e-16
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  61.71589 82.11222
## sample estimates:
## mean of x mean of y 
## 151.18173  79.26768

According to the p value obtained (2.2e-16), there is evidence to say that the means of Sodium of Breakfast and Lunch/Dinner are not equal.

This result does supports my hypothesis, given that Breakfast foods items have in average, a greater amount of Sodium compared to Lunch_dinner items.

Avg_sodium_All_beverage_items_Breakfast_vs_Lunch_plot <- fast_food_beverages %>% 
  ggplot(aes(x= Breakfast_or_lunch_dinner, y=`Sodium (mg)`, fill=Breakfast_or_lunch_dinner)) +
  geom_boxplot() +
  labs(title="Average Sodium per Beverage Item - Breakfast vs. Lunch_Dinner",
       y="Sodium (mg)", x="") +
   theme(plot.title=element_text(size=10, 
                                    face="bold", 
                                    hjust=0.5),
         legend.text = element_text(size=8, face="bold"))
Avg_sodium_All_beverage_items_Breakfast_vs_Lunch_plot

Hypothesis 2.d. Beverage items. Sugars

First, I’m going to compare the variances between the two groups to see what kind of T test should be used.

var.test(fast_food_beverages$`Sugars (g)`[fast_food_beverages$Breakfast_or_lunch_dinner=="Breakfast"],
       fast_food_beverages$`Sugars (g)`[fast_food_beverages$Breakfast_or_lunch_dinner=="Lunch_dinner"])
## 
##  F test to compare two variances
## 
## data:  fast_food_beverages$`Sugars (g)`[fast_food_beverages$Breakfast_or_lunch_dinner ==  and fast_food_beverages$`Sugars (g)`[fast_food_beverages$Breakfast_or_lunch_dinner ==     "Breakfast"] and     "Lunch_dinner"]
## F = 0.51656, num df = 1061, denom df = 395, p-value < 2.2e-16
## alternative hypothesis: true ratio of variances is not equal to 1
## 95 percent confidence interval:
##  0.4372608 0.6064694
## sample estimates:
## ratio of variances 
##          0.5165601

Given that the variances are not equal, I will perform a Welch T test.

t.test(fast_food_beverages$`Sugars (g)`[fast_food_beverages$Breakfast_or_lunch_dinner ==
                                          "Breakfast"],
          fast_food_beverages$`Sugars (g)`[fast_food_beverages$Breakfast_or_lunch_dinner ==
                                          "Lunch_dinner"], var.equal = FALSE)
## 
##  Welch Two Sample t-test
## 
## data:  fast_food_beverages$`Sugars (g)`[fast_food_beverages$Breakfast_or_lunch_dinner ==  and fast_food_beverages$`Sugars (g)`[fast_food_beverages$Breakfast_or_lunch_dinner ==     "Breakfast"] and     "Lunch_dinner"]
## t = -3.4828, df = 554.17, p-value = 0.0005354
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -10.31687  -2.87616
## sample estimates:
## mean of x mean of y 
##  37.82015  44.41667

According to the p value obtained (0.0005354), there is evidence to say that the mean Sugar of Breakfast and Lunch/Dinner are not equal.

This result doesn’t my hypothesis, given that Breakfast beverage items have in average, a smaller amount of Sugars compared to Lunch_dinner items.

Avg_sugars_All_beverage_items_Breakfast_vs_Lunch_plot <- fast_food_beverages %>% 
  ggplot(aes(x= Breakfast_or_lunch_dinner, y=`Sugars (g)`, fill=Breakfast_or_lunch_dinner)) +
  geom_boxplot() +
  labs(title="Average Sugars per Beverage Item - Breakfast vs. Lunch_Dinner",
       y="Sugar (g)", x="") +
   theme(plot.title=element_text(size=10, 
                                    face="bold", 
                                    hjust=0.5),
         legend.text = element_text(size=8, face="bold"))
Avg_sodium_All_beverage_items_Breakfast_vs_Lunch_plot

Summarizing Hypothesis for Food Items.

1.a. Calories. Evidence doesn’t support the research hypothesis.

1.b. Saturated fat. Evidence supports the research hypothesis.

1.c. Sodium. Evidence doesn’t support the research hypothesis.

1.d. Sugars. Evidence supports the research hypothesis.

That’s 2 supported against 2 rejected.

Conclusion: Breakfast food items are as “unhealthy” as Lunch/Dinner food items.

Summarizing Hypothesis for Beverage Items.

2.a. Calories. Evidence supports the research hypothesis.

2.b. Saturated fat. Evidence supports the research hypothesis.

2.c. Sodium. Evidence supports the research hypothesis.

2.d. Sugars. Evidence doesn’t support the research hypothesis.

That’s 3 supported against one rejected.

Conclusion: Breakfast beverage items are as or more “unhealthy” than Lunch/Dinner beverage Items.

Discussion

This study has severe limitations.
First of all, the methodological approach is neither rigorous nor representative of Fast food restaurants in general.
The selection of Restaurants might have been biased.

Moreover, the sample doesn’t represent the whole population of Fast food Restaurants.
Even though the information was obtained from official websites, the selection criteria left out a lot of items that might change these results.

Furthermore, the research question was based on a highly subjective concept (Healthiness), and thus the operationalization might not have been rigorous enough to measure what it was supposed to measure. Despite all the limitations, the study might shed some light over the nutrition facts of some of the most important fast food restaurants around the globe, an might raise awareness among consumers.

Author:
Ramiro Ferrando.
BS Human Nutrition
MS Nutrigenomics
MS Journalism Candidate
Former PhD Human Nutrition Candidate
https://www.linkedin.com/in/ramiroferrando/
www.piensoluegocomo.com

ramiroferrando@hotmail.com

References

Alviola IV, P. A., Nayga, R. M., Thomsen, M. R., Danforth, D., & Smartt, J. (2014). The effect of fast-food restaurants on childhood obesity: A school level analysis. Economics and Human Biology, 12(1), 110–119. https://doi.org/10.1016/j.ehb.2013.05.001

Block, J. P., Condon, S. K., Kleinman, K., Mullen, J., Linakis, S., Rifas-Shiman, S., & Gillman, M. W. (2013). Consumers’ estimation of calorie content at fast food restaurants: Cross sectional observational study. BMJ (Online), 346(7912), 1–10. https://doi.org/10.1136/bmj.f2907

Brissette, I., Lowenfels, A., Noble, C., & Spicer, D. (2013). Predictors of Total Calories Purchased at Fast-food Restaurants : Restaurant Characteristics , Calorie Awareness , and Use of Calorie Information. Journal of Nutrition Education and Behavior, 45(5), 404–411. https://doi.org/10.1016/j.jneb.2013.01.019

Cadario, R. (2016). The impact of health claims and food deprivation levels on health risk perceptions of fast-food restaurants. Social Science and Medicine, 149, 130–134. https://doi.org/10.1016/j.socscimed.2015.12.016

Carden, L. L., Maldonado, T., & Boyd, R. O. (2018). Organizational resilience : A look at McDonald ’ s in the fast food industry. Organizational Dynamics, 47(1), 25–31. https://doi.org/10.1016/j.orgdyn.2017.07.002

Carels, R. A., Harper, J., & Konrad, K. (2006a). Qualitative perceptions and caloric estimations of healthy and unhealthy foods by behavioral weight loss participants, 46, 199–206. https://doi.org/10.1016/j.appet.2005.12.002

Carels, R. A., Harper, J., & Konrad, K. (2006b). Qualitative perceptions and caloric estimations of healthy and unhealthy foods by behavioral weight loss participants. Appetite, 46(2), 199–206. https://doi.org/10.1016/j.appet.2005.12.002

Carels, R. A., Konrad, K., & Harper, J. (2007). Individual differences in food perceptions and calorie estimation: An examination of dieting status, weight, and gender. Appetite, 49(2), 450–458. https://doi.org/10.1016/j.appet.2007.02.009

Chandon, P., & Wansink, B. (2007). The Biasing Health Halos of Fast-Food Restaurant Health Claims: Lower Calorie Estimates and Higher Side-Dish Consumption Intentions. Journal of Consumer Research, 34(3), 301–314. https://doi.org/10.1086/519499

Chandon, P., & Wansink, B. (2012). Does food marketing need to make us fat? A review and solutions. Nutrition Reviews, 70(10), 571–593. https://doi.org/10.1111/j.1753-4887.2012.00518.x

Chou, S.-Y., Grossman, M., & Saffer, H. (2004). An economic analysis of adult obesity: results from the Behavioral Risk Factor Surveillance System. Journal of Health Economics, 23(3), 565–87. https://doi.org/10.1016/j.jhealeco.2003.10.003

Davis, B., & Carpenter, C. (2009). Proximity of fast-food restaurants to schools and adolescent obesity. American Journal of Public Health, 99(3), 505–510. https://doi.org/10.2105/AJPH.2008.137638

Devine, C. M., Farrell, T. J., Blake, C. E., Wethington, E., & Bisogni, C. A. (2009). Work Conditions and the Food Choice Coping Strategies of Employed Parents. Journal of Nutrition Education and Behavior, 41(5), 365–370. https://doi.org/10.1016/j.jneb.2009.01.007.Work

Dunkin Donuts — Nutrition Facts. (2018). Retrieved from https://www.dunkindonuts.com/content/dam/dd/pdf/nutrition.pdf

Goyal, A., & Singh, N. P. (2007). Consumer perception about fast food in India: an exploratory study. British Food Journal, 109(2), 182–195. https://doi.org/10.1108/00070700710725536

Hales, C. M., Carroll, M. D., Fryar, C. D., & Ogden, C. L. (2017). Prevalence of Obesity Among Adults and Youth: United States, 2015–2016. NCHS Data Brief, (288), 1–8. https://doi.org/10.1017/S1368980017000088

Jeffery, R. W., Baxter, J., McGuire, M., & Linde, J. (2006). Are fast food restaurants an environmental risk factor for obesity? The International Journal of Behavioral Nutrition and Physical Activity, 3, 2. https://doi.org/10.1186/1479-5868-3-2

Kourouniotis, S., Keast, R. S. J., Riddell, L. J., Lacy, K., Thorpe, M. G., & Cicerale, S. (2016). The importance of taste on dietary choice, behaviour and intake in a group of young adults. Appetite, 103, 1–7. https://doi.org/10.1016/j.appet.2016.03.015

Mc Donald’s Nutritional Facts. (2018). Retrieved from http://nutrition.mcdonalds.com/usnutritionexchange/nutritionfacts.pdf

Monteiro, C. A., Cannon, G., Moubarac, J. C., Levy, R. B., Louzada, M. L. C., & Jaime, P. C. (2017). Ultra-processing. An odd “appraisal.” Public Health Nutrition, 60(60), 1–5. https://doi.org/10.1017/S1368980017003287

Moran, A. J., Ramirez, M., & Block, J. P. (2017). Consumer underestimation of sodium in fast food restaurant meals: Results from a cross-sectional observational study. Appetite, 113, 155–161. https://doi.org/10.1016/j.appet.2017.02.028

Moubarac, J.-C., Parra, D. C., Cannon, G., & Monteiro, C. a. (2014). Food Classification Systems Based on Food Processing: Significance and Implications for Policies and Actions: A Systematic Literature Review and Assessment. Current Obesity Reports, 3(2), 256–272. https://doi.org/10.1007/s13679-014-0092-0

O’Halloran, S. A., Lacy, K. E., Grimes, C. A., Woods, J., Campbell, K. J., & Nowson, C. A. (2017). A novel processed food classification system applied to Australian food composition databases. Journal of Human Nutrition and Dietetics, 30(4), 534–541. https://doi.org/10.1111/jhn.12445

Okrent, A. M., & Alston, J. M. (2012). The demand for disaggregated food-away-from-home and food-at-home products in the United States. Economic Research Report Number 139, (139), 69. Roberto, C. A., Schwartz, M. B., & Brownell, K. D. (2009). Rationale and Evidence for Menu-Labeling Legislation. American Journal of Preventive Medicine. https://doi.org/10.1016/j.amepre.2009.07.015

Ross, B. H., & Murphy, G. L. (1999). Food for Thought: Cross-Classification and Category Organization in a Complex Real-World Domain. Cognitive Psychology, 38(4), 495–553. https://doi.org/10.1006/cogp.1998.0712

Starbucks — Nutrition Facts. (2018). Retrieved from https://www.starbucks.com/menu/catalog/nutrition?food=hot-breakfast

Stokes, A., Ni, Y., & Preston, S. H. (2017). Prevalence and Trends in Lifetime Obesity in the U.S., 1988–2014. American Journal of Preventive Medicine, 53(5), 567–575. https://doi.org/10.1016/j.amepre.2017.06.008

Subway’s — Nutrition Facts. (2018). Retrieved from http://www.subway.com/~/media/usa/documents/nutrition/us_nutrition_values.pdf

Swartz, J. J., Braxton, D., & Viera, A. J. (2011). Calorie menu labeling on quick-service restaurant menus: An updated systematic review of the literature. International Journal of Behavioral Nutrition and Physical Activity, 8, 1–8. https://doi.org/10.1186/1479-5868-8-135

Tangari, A. H., Burton, S., Cho, Y., & Thyroff, A. (2010). Article Summary : “ Weighing in on Fast Food Consumption : The Effects of Meal and Calorie Disclosures on Consumer Fast Food Evaluations .” Article Summary : “ Weighing in on Fast Food Consumption : The Effects of Meal and Calorie Disclosures on Consumer . Health (San Francisco), 4(3), 1–2.

Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty: heuristics and biases. Rationality in Action: Contemporary Approaches, 185, 171–188-NaN, 493.

Ziauddeen, H., Alonso-Alonso, M., Hill, J. O., Kelley, M., & Khan, N. A. (2015). Obesity and the neurocognitive basis of food reward and the control of intake. Advances in Nutrition (Bethesda, Md.), 6(4), 474–86. https://doi.org/10.3945/an.115.008268

Compartir este artículo

Leave a Comment

Tu dirección de correo electrónico no será publicada. Los campos obligatorios están marcados con *