The Effect Nutrition Labeling Laws at Fast Food Chains has on Cholesterol Levels
A Cross-Sectional Analysis
Ryan Sailer
November 1, 2018
Abstract
High cholesterol levels have become an ever increasing problem in the United States. Many factors are to blame, but the most important is diet. In May of 2018, every fast food restaurant with at least 20 chains are required to implement a new labeling policy. This will force them to put the calorie count on their menu, as well as, have additional nutritional information readily available. 6 states have passed these laws prior to 2018 and this study is looking to see if that has had any effect on the cholesterol levels of the citizens of those states. The data used in this study is from the Behavioral Risk Factor Surveillance System (BRFSS) from 2017, with the policy data coming from the National Conference and State Legislatures (NCSL). While we did not find statistical significance showing a difference between states with and without the policy, we did find that people between the ages of 55-59, 60-64 and 65-69 showed statistically higher percentage of people having higher cholesterol than those between the ages of 18-24, within states that have passed NLL. This study does have some limitation however. The BRFSS does not include a question about daily diet. This did not allow me to control for one of the larger contributing factors of high cholesterol.Introduction
In the past few years, obesity has become a major problem in the United States. “The age-adjusted prevalence of obesity (ie, a body mass index [BMI] >=30) in the period 1999-2000 (the last official report by the National Health and Nutrition Examination Survey [NHANES III]) was 30.5% compared with 22.9% in 1980” (âObesity in America | Obesity | JAMA Surgery | JAMA Network,â n.d). Factors such as lack of exercise and diet have been major factors, but there is another major problem that often gets missed. High cholesterol is becoming a major problem for the American population. High cholesterol is a main factor in two of the main causes of human death, heart attack and stroke. A study done in 2012 states, “43.9% stated that they understood them very well, whereas only 27.2% achieved high scores”(Namba, Auchincloss, Leonberg, & Wootan, 2013). Many of these consumers just look at the calorie count and not the rest of the nutritional information. For the sake of cholesterol, the other nutrition facts are often more important to the health of a person. Saturated fat and trans fat are the main contributing factors that raise levels of cholesterol. Having some cholesterol is good for the body to function properly. Your “good cholesterol” is called High Density Lipoprotein or HDL, while your bad cholesterol is called Low Density Lipoprotein, or LDL. LDL is the cholesterol that builds up inside of your arteries, while the HDL takes that cholesterol and delivers it back to the liver.
Background and Rationale
The main point of this study is to see if the states who have passed nutrition labeling laws on fast food restaurants, has an effect on whether people in those states have higher levels of cholesterol. Fast food is not only high on calories, but also sodium, fat and cholesterol. In a prior study, it was found, “Mean calories per purchase decreased from 908.5 to 870.4 at 18 months post-implementation”(Krieger et al., 2013). The same article also found, “Awareness of labels increased from 18.8% to 61.7% in food chains”(Krieger et al., 2013). This is saying that consumers are in fact noticing the nutrition labels and are making a conscious effort to try healthier choices at these restaurants.The only problem with this is that many people do not know what the healthy options are. My research question is important because high cholesterol is becoming a major health problem in the United States. Diet is one of the main factors that raise the levels of LDL, which is the bad cholesterol.
Study Objectives
My study differs from others done because I am looking at the effects on cholesterol levels, not obesity. Studies have shown that fast food consumption is a main factor on why obesity has become such a problem in the United States. Meanwhile, high cholesterol is a major factor in causing heart attacks and strokes, which are some of the leading causes of death in our country. The main objective of this study is to see if providing nutrition information at fast food restaurants has an effect on the cholesterol levels of the population.
Methods
Study Design
This study uses cross-sectional variation in Fast Food Nutrition Labeling Laws to examine systematic differences in the cholesterol levels of individuals associated with passing nutrition labeling laws in fast food restaurants.
The main point of this policy is to have nutrition facts about each food item available to the customers of fast food restaurants. They must have the calorie count available on the menu, as well as, having additional nutrition facts readily available. There is 6 states that have enacted this law before the year 2018, with the other 44 states being required to implement this act by May of 2018. Figure 1 shows which states have passed Nutrition Labeling Laws in fast food chains.
Figure 1 States that passed Nutrition Labeling Laws at Fast Food Chains
Setting
California was the first state to pass these Menu Labeling Law back in 2008. Maine, Massachusetts, and Oregon passed the same law in 2009, with New Jersey and Vermont doing the same in 2010. The FDA recently passed legislation that is requiring every state to to abide by these laws. While this is not going to affect the current survey, May of 2018 is the date where all fast food restaurants are going to be required to abide by these new Menu Labeling Laws.
Participants
The target population of study is people who have gotten their cholesterol checked by a doctor. This will allow me to see if the states who have passed this law before 2018 have a lower percentage of the population with high cholesterol, than states who have not enacted this policy. I have decided to cut out those who have never had their cholesterol checked by a doctor. I am not able to include them because it is impossible to confirm whether they do or do not actually have high cholesterol levels. Many people go undiagnosed, so it would be improper to assign them as not having high cholesterol.
#Filter the BRFSS data to include only [fill in criteria]
#analysis <- brfss17 %>% filter(TOLDHI2 == 1 | TOLDHI2 == 2)
load("analysis.RData")Variables
My outcome variable is looking to see if the respondent has ever been told they have high cholesterol levels by a doctor. This variable is labeled as TOLDHI2 and is coded as either a 1 if they have been told they have high cholesterol, or a 0 for those who have not.
For my policy variable, I coded all of the states that have passed Fast Food Nutrition Labeling Laws before 2017 as a 1, and those who did not enact these laws before 2017 as a 0. This variable is labeled as NLL.
The first control variable is INCOME2, or income level. This variable is interval scaled, that places each respondent on one of 8 different categories. X.AGE5YR, or age, is another interval scaled variable that takes the respondents responses and places them into one of 13 different categories. EXERANY2 is an exercise variable, that asked the respondents whether they have exercised in the past 30 days. The respondents who have exercised are labeled a a 1, while those who have not exercised in the past 30 days are 0. The final control variable is SMOKE100, which asks the respondents if they have smoked at least 100 cigarettes in their life. Similar to the exercise variable, those who have smoked 100 cigarettes are coded as 1, while those who have not are 0’s. For all of the variables, I have omitted those who responded either don’t know or refused to respond.
#Select BRFSS Variables
#analysis <- analysis %>% select(X.STATE, TOLDHI2, INCOME2, X.AGEG5YR, EXERANY2, SMOKE100, X.LLCPWT)#Recode Outcome
analysis$TOLDHI2 <- as.character(analysis$TOLDHI2)
analysis <- analysis %>% mutate(TOLDHI2.cl = recode(as.numeric(TOLDHI2), "1" = "1", "2" = "0", "7" = "NA", "9" = "NA"))
#Recode additional variables
analysis <- analysis %>% mutate(INCOME2.cl = recode(as.factor(INCOME2), "77" = "NA", "99" = "NA"))
analysis <- analysis %>% mutate(X.AGEG5YR.cl =recode(as.factor(X.AGEG5YR), "14" = "NA"))
analysis <- analysis %>% mutate(EXERANY2.cl = recode(as.factor(EXERANY2), "7" = "NA", "9" = "NA"))
analysis <- analysis %>% mutate(SMOKE100.cl = recode(as.factor(SMOKE100), "7" = "NA", "9" = "NA"))
#Join your policy data with the BRFSS DataData Sources
PolicyData <- read.csv("PolicyData.csv")
analysis <- full_join(analysis, PolicyData, by = "X.STATE")The Behavioral Risk Factor Surveillance System (BRFSS), is a telephone survey that looks at different health effects and outcomes of the adult population of the United States. This survey is done yearly, which gives us continuously updated data (“Centers for Disease Control and Prevention”“, 2017). My policy data comes from The National Conference and State Legislatures (”Trans Fat and Menu Labeling Legislation,“n.d). ## Study Size I decided to filter out respondents who have never been checked for high cholesterol by a doctor. It is impossible to know what percentage of the diagnosed population has high cholesterol, so assigning them as with or without high cholesterol would be inappropriate. After filtering out some respondents, my total number of observations equals 418,738. This filtering took 31,904 respondents out of my analysis. ## Statistical Methods * Write out and explain your simple regression model. Explain potential sources of bias. In my simple regression model, I look to see if there is a percentage change in being told you have high cholesterol by a doctor for states who have passed a policy about Nutrition Labeling Laws at fast food chains. If the policy in fact is effective, than the percentage of people with high cholesterol should be lower in those states. There is some potential bias in this equation. One factor would be the access to health care. If a state has a higher average income, they would be more likely to see a doctor and have their cholesterol checked when compared to low income states. While those with a lower income may have a higher chance of having high cholesterol, but if they are never tested by a doctor, we have no way of confirming a diagnosis. Another problem is the lifestyle that these respondents are living. Are they eating health food, exercising regularly and abstaining from alcohol and tobacco. These are all key factors in causing people to have high cholesterol.
\[TOLDHI2{i} = \beta_{0} + \beta_{1}NLL_{i}+\epsilon_{i}\]
The first multiple regression model is doing the same test as the simple regression model, but we are controlling for some key variables. I control for 4 different variables, being income, age, exercise and smoking habits. Income is an important factor because those who have higher income, will have more disposable income to spend on healthier food options instead of fast food. Age is another factor, where I would expect those who are older to have a higher likelihood of being told they have high cholesterol. Exercise and smoking are important variables that are major causes of high cholesterol, so controlling for both is important.
\[TOLDHI2{i} = \beta_{0} + \beta_{1}NLL{i}+ \beta_{2}INCOME2{i} + \beta_{3}X.AGE5YR{i} + \beta_{2}EXERANY2{i} + \beta_{2}SMOKE100{i} + \epsilon_{i}\] Model 3 is the same as Model 2, where I control for Income, Age, Exercise and Smoking habits, except I allowed the Fast Food Nutrition Labeling Policy to fluctuate by age. I believe that those who are older will have a higher likelihood of having high cholesterol. Not only have they had longer time to allow build up in their arteries, but also they often have better access to health care. This would give them a higher chance of being tested for high cholesterol. The table below is showing that the proportion of people that have been told they have high cholesterol in my sample population.
\[TOLDHI2{i} = \beta_{0} + \beta_{1}NLL{i} + \beta_{2}INCOME2{i} + \beta_{3}X.AGE5YR{i} + \beta_{2}EXERANY2{i} + \beta_{2}SMOKE100{i} + \beta_{3}NLL{i}*X.AGE5YR{i} + \epsilon_{i}\]
Results
Simple <- lm(TOLDHI2 ~ NLL, data = analysis)
#summary(Simple)
Multiple1 <- lm(TOLDHI2 ~ NLL + INCOME2.cl + X.AGEG5YR.cl + EXERANY2.cl + SMOKE100.cl, data = analysis)
#summary(Multiple1)
Multiple2 <- lm(TOLDHI2 ~ INCOME2.cl + EXERANY2.cl + SMOKE100.cl + NLL*X.AGEG5YR.cl, data = analysis)
#summary(Multiple2)
Simplew <- lm(TOLDHI2 ~ NLL, data = analysis, weights = X.LLCPWT)
Multiple1w <- lm(TOLDHI2 ~ NLL + INCOME2.cl + X.AGEG5YR.cl + EXERANY2.cl + SMOKE100.cl, data = analysis, weights = X.LLCPWT)
Multiple2w <- lm(TOLDHI2 ~ INCOME2.cl + EXERANY2.cl + SMOKE100.cl + NLL*X.AGEG5YR.cl, data = analysis, weights = X.LLCPWT)#Include a table of descriptive statistics for your outcome variable and x variables. If your policy is discrete you can include separate columns for each value of your policy variable.
analysis$TOLDHI2.cl <- as.numeric(analysis$TOLDHI2.cl)
analysis %>% summarize(avg_TOLDHI2.cl = mean(TOLDHI2.cl, na.rm = TRUE))| avg_TOLDHI2.cl |
|---|
| 0.3794998 |
Results
#Include a table of regressions using stargazer to create a nice summary of your output
#simple <- stargazer(Simple, header = FALSE, title = "Table 1 Title", type = "html")
#multiple1 <- stargazer(Multiple1, header = FALSE, title = "Table 1 Title", type = "html")
#multiple2 <-stargazer(Multiple2, header = FALSE, title = "Table 1 Title", column.labels = c("OLS", "Weighted"), type = "html")
#stargazer table for use while writing
stargazer(Simple, Simplew, title = "Table 1 Results from Simple Regression", column.labels = c("OLS", "Weighted"), type = "text")##
## Table 1 Results from Simple Regression
## ==============================================================
## Dependent variable:
## ----------------------------
## TOLDHI2
## OLS Weighted
## (1) (2)
## --------------------------------------------------------------
## NLL 0.012*** 0.020***
## (0.002) (0.002)
##
## Constant 1.619*** 1.678***
## (0.001) (0.001)
##
## --------------------------------------------------------------
## Observations 408,958 408,958
## R2 0.0001 0.0003
## Adjusted R2 0.0001 0.0003
## Residual Std. Error (df = 408956) 0.485 11.004
## F Statistic (df = 1; 408956) 23.946*** 118.140***
## ==============================================================
## Note: *p<0.1; **p<0.05; ***p<0.01#stargazer table for knit version - comment out the one above and comment this one in
stargazer(Multiple1, Multiple2, header = FALSE, title = "Table 2 Results from Both Multiple Regresion Models", column.labels = c("Multiple1", "Multiple2"), type = "text")##
## Table 2 Results from Both Multiple Regresion Models
## =================================================================================
## Dependent variable:
## -------------------------------------------------------------
## TOLDHI2
## Multiple1 Multiple2
## (1) (2)
## ---------------------------------------------------------------------------------
## NLL 0.005** -0.016
## (0.002) (0.011)
##
## INCOME2.cl2 0.005 0.005
## (0.005) (0.005)
##
## INCOME2.cl3 0.026*** 0.026***
## (0.005) (0.005)
##
## INCOME2.cl4 0.041*** 0.041***
## (0.005) (0.005)
##
## INCOME2.cl5 0.045*** 0.045***
## (0.005) (0.005)
##
## INCOME2.cl6 0.045*** 0.045***
## (0.004) (0.004)
##
## INCOME2.cl7 0.051*** 0.051***
## (0.004) (0.004)
##
## INCOME2.cl8 0.061*** 0.061***
## (0.004) (0.004)
##
## INCOME2.clNA 0.060*** 0.060***
## (0.004) (0.004)
##
## X.AGEG5YR.cl2 -0.024*** -0.026***
## (0.005) (0.005)
##
## X.AGEG5YR.cl3 -0.064*** -0.065***
## (0.005) (0.005)
##
## X.AGEG5YR.cl4 -0.114*** -0.114***
## (0.005) (0.005)
##
## X.AGEG5YR.cl5 -0.165*** -0.167***
## (0.005) (0.005)
##
## X.AGEG5YR.cl6 -0.235*** -0.238***
## (0.005) (0.005)
##
## X.AGEG5YR.cl7 -0.293*** -0.295***
## (0.004) (0.005)
##
## X.AGEG5YR.cl8 -0.356*** -0.359***
## (0.004) (0.005)
##
## X.AGEG5YR.cl9 -0.405*** -0.408***
## (0.004) (0.004)
##
## X.AGEG5YR.cl10 -0.442*** -0.445***
## (0.004) (0.004)
##
## X.AGEG5YR.cl11 -0.467*** -0.469***
## (0.004) (0.005)
##
## X.AGEG5YR.cl12 -0.451*** -0.454***
## (0.005) (0.005)
##
## X.AGEG5YR.cl13 -0.396*** -0.398***
## (0.004) (0.005)
##
## X.AGEG5YR.clNA -0.273*** -0.274***
## (0.008) (0.008)
##
## NLL:X.AGEG5YR.cl2 0.014
## (0.016)
##
## NLL:X.AGEG5YR.cl3 0.004
## (0.015)
##
## NLL:X.AGEG5YR.cl4 0.004
## (0.015)
##
## NLL:X.AGEG5YR.cl5 0.013
## (0.015)
##
## NLL:X.AGEG5YR.cl6 0.028*
## (0.014)
##
## NLL:X.AGEG5YR.cl7 0.020
## (0.014)
##
## NLL:X.AGEG5YR.cl8 0.031**
## (0.013)
##
## NLL:X.AGEG5YR.cl9 0.027**
## (0.013)
##
## NLL:X.AGEG5YR.cl10 0.026**
## (0.013)
##
## NLL:X.AGEG5YR.cl11 0.021
## (0.013)
##
## NLL:X.AGEG5YR.cl12 0.025*
## (0.014)
##
## NLL:X.AGEG5YR.cl13 0.023*
## (0.014)
##
## NLL:X.AGEG5YR.clNA 0.014
## (0.024)
##
## EXERANY2.cl2 -0.039*** -0.039***
## (0.002) (0.002)
##
## EXERANY2.clNA 0.007 0.007
## (0.015) (0.015)
##
## SMOKE100.cl2 0.046*** 0.046***
## (0.002) (0.002)
##
## SMOKE100.clNA 0.047*** 0.047***
## (0.011) (0.011)
##
## Constant 1.862*** 1.864***
## (0.005) (0.005)
##
## ---------------------------------------------------------------------------------
## Observations 378,949 378,949
## R2 0.105 0.105
## Adjusted R2 0.105 0.105
## Residual Std. Error 0.460 (df = 378922) 0.460 (df = 378909)
## F Statistic 1,711.031*** (df = 26; 378922) 1,141.058*** (df = 39; 378909)
## =================================================================================
## Note: *p<0.1; **p<0.05; ***p<0.01Looking at Table 1, I have included the OLS and Weighted Least Squares regression model. There is a slight difference between the two, but I will be using OLS for all of my interpretations. This model shows the probability of having been told you have high cholesterol is 1.2 percentage points higher for states who have passed Nutrition Labeling Laws. This is saying that taking no other factors into account, having passed Nutrition Labeling Laws actually has a negative effect on being told they have high cholesterol. I expect that after I account for income, age, exercise, and smoking habits will have an effect on the probability of having high cholesterol.
In Table 2, we can see the results from both of my multiple regression models. For Multiple1, when the state has passed NLL, they have a .49 percentage point increase in the likelihood of being told that they have high cholesterol. This model is accounting for income, age, exercise and smoking habits. This model is statistically insignificant, so we are unable to confirm if there is a difference in the likelihood of being told you have high cholesterol for states with and without Nutrition Labeling Laws.
In Multiple2 of Table 2, when the state has passed NLL, they are 1.6 percentage points less likely to be told they have high cholesterol, than states who have not passed NLL. In this model, I have allowed the Fast Food Nutrition Labeling Policy to fluctuate by age. Of the 12 different age coefficients, every single one shows an increase in being told they have high cholesterol, when being compared to the 18-24 year group. Of those 12 different age coefficients, 3 have statistically significant evidence that the effect of this policy has an impact by age variation. The ages between 55-59, 60-64 and 65-69, show the most significant changes. For the 55-59 age range, we see a 1.5 percentage points increase, when compared than those in the youngest age group. For the 60-64 and 65-69 age range, we see a 1.1 and .7 percentage point increase for their respective group.
Robustness Checks
stargazer(Multiple1, Multiple1w, header = FALSE, title = "Table 3 Results from OLS vs. Weighted Least Squares", column.labels = c("Multiple1", "Multiple1w"), type = "text")##
## Table 3 Results from OLS vs. Weighted Least Squares
## ==============================================================
## Dependent variable:
## ----------------------------
## TOLDHI2
## Multiple1 Multiple1w
## (1) (2)
## --------------------------------------------------------------
## NLL 0.005** 0.008***
## (0.002) (0.002)
##
## INCOME2.cl2 0.005 0.010**
## (0.005) (0.005)
##
## INCOME2.cl3 0.026*** 0.032***
## (0.005) (0.004)
##
## INCOME2.cl4 0.041*** 0.043***
## (0.005) (0.004)
##
## INCOME2.cl5 0.045*** 0.048***
## (0.005) (0.004)
##
## INCOME2.cl6 0.045*** 0.050***
## (0.004) (0.004)
##
## INCOME2.cl7 0.051*** 0.042***
## (0.004) (0.004)
##
## INCOME2.cl8 0.061*** 0.057***
## (0.004) (0.004)
##
## INCOME2.clNA 0.060*** 0.050***
## (0.004) (0.004)
##
## X.AGEG5YR.cl2 -0.024*** -0.016***
## (0.005) (0.003)
##
## X.AGEG5YR.cl3 -0.064*** -0.067***
## (0.005) (0.003)
##
## X.AGEG5YR.cl4 -0.114*** -0.110***
## (0.005) (0.003)
##
## X.AGEG5YR.cl5 -0.165*** -0.168***
## (0.005) (0.003)
##
## X.AGEG5YR.cl6 -0.235*** -0.240***
## (0.005) (0.003)
##
## X.AGEG5YR.cl7 -0.293*** -0.297***
## (0.004) (0.003)
##
## X.AGEG5YR.cl8 -0.356*** -0.350***
## (0.004) (0.003)
##
## X.AGEG5YR.cl9 -0.405*** -0.400***
## (0.004) (0.003)
##
## X.AGEG5YR.cl10 -0.442*** -0.431***
## (0.004) (0.003)
##
## X.AGEG5YR.cl11 -0.467*** -0.458***
## (0.004) (0.004)
##
## X.AGEG5YR.cl12 -0.451*** -0.435***
## (0.005) (0.004)
##
## X.AGEG5YR.cl13 -0.396*** -0.374***
## (0.004) (0.004)
##
## X.AGEG5YR.clNA -0.273*** -0.260***
## (0.008) (0.007)
##
## EXERANY2.cl2 -0.039*** -0.037***
## (0.002) (0.002)
##
## EXERANY2.clNA 0.007 -0.017
## (0.015) (0.012)
##
## SMOKE100.cl2 0.046*** 0.042***
## (0.002) (0.001)
##
## SMOKE100.clNA 0.047*** 0.066***
## (0.011) (0.010)
##
## Constant 1.862*** 1.854***
## (0.005) (0.004)
##
## --------------------------------------------------------------
## Observations 378,949 378,949
## R2 0.105 0.129
## Adjusted R2 0.105 0.129
## Residual Std. Error (df = 378922) 0.460 10.218
## F Statistic (df = 26; 378922) 1,711.031*** 2,156.532***
## ==============================================================
## Note: *p<0.1; **p<0.05; ***p<0.01Table 3 looks at the difference in results using OLS versus Weighted Least Squares. Weighted Least Squares is trying to account for any variation in the data. There is a minimal difference between the two, which is why I chose to interpret my results with OLS.
Conclusion
Overview of Key Findings
This study looked at the effects that Nutrition Labeling Laws at Fast Food Chains had on high cholesterol levels. This is a topic that has not been tested, with most studies looking at how different factors affected levels of obesity. I found some statistically significant evidence that as people age, they are more likely to have high cholesterol. The ages between 55-59, 60-64 and 65-69 showed statistically higher percentage of people having higher cholesterol than those between the ages of 18-24 and both live in states who have passed NLL. While it did show some statistically significant evidence, there is no practical significance to any of my findings. No results differed by more than 5% from the average, which leads me to conclude that Nutritional Labeling Laws have no effect on the likelihood of having high cholesterol.
While I attempted to account for all bias, there is still some limitations to this study. One of the major factors of high cholesterol is diet. The BRFSS survey did not have any variables asking about the daily diet of the respondents, assuming they did not eat fast food for every meal. Another limitation is I have no idea how truthful the respondents were when answering each question. Since the respondents are being observed over the telephone, there is a possibility that they will give false information.
As for the policymakers, I would advise them to continue to enforce Nutrition Labels at not only Fast Food restaurants, but at every restaurant in the country. A prior study states, " Healthier food options increased from 13% to 20% at case locations while remaining static at 8% at control locations“(Sharf et al., 2012). Having these labels have encouraged restaurants to include more healthy options to their menu. Even if the actual labels have no effect on the customer, it has had an effect on providing healthier options. For future research, I would like to look at each respondents typical diet. Asking a question or two about how often people eat out and whether it is a sit down restaurant, or fast food chain. Since a lot of fast food is worse nutritionally than most home cooked meals, this could give us a better idea about the day-to-day eating habits of the respondents.
Works Cited
Centers for Disease Control and Prevention (CDC). Behavioral Risk Factor Surveillance System Survey Data. Atlanta, Georgia: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, 2017.
Krieger, J. W., Chan, N. L., Saelens, B. E., Ta, M. L., Solet, D., & Fleming, D. W. (2013). Menu Labeling Regulations and Calories Purchased at Chain Restaurants. American Journal of Preventive Medicine, 44(6), 595â“604. https://doi.org/10.1016/j.amepre.2013.01.031
Namba, A., Auchincloss, A., Leonberg, B. L., & Wootan, M. G. (2013). Exploratory Analysis of Fast-Food Chain Restaurant Menus Before and After Implementation of Local Calorie-Labeling Policies, 2005â“2011. Preventing Chronic Disease, 10. https://doi.org/10.5888/pcd10.120224
Obesity in America | Obesity | JAMA Surgery | JAMA Network. (n.d.). Retrieved October 30, 2018, from https://jamanetwork.com/journals/jamasurgery/fullarticle/394701
Sharf, M., Sela, R., Zentner, G., Shoob, H., Shai, I., & Stein-Zamir, C. (2012). Figuring out food labels. Young adultsâ understanding of nutritional information presented on food labels is inadequate. Appetite, 58(2), 531â“534. https://doi.org/10.1016/j.appet.2011.12.010
Trans Fat and Menu Labeling Legislation. (n.d.). Retrieved October30,2018, from http://www.ncsl.org/research/health/trans-fat-and-menu-labeling-legislation.aspx