Free and Reduced Lunches: an Overview

In 2018, 53.1 students were enrolled in a K-12 school. As of 2016 in the United States, 30.4 million students use the free and reduced lunch program, provided by the National School Lunch Program. This program is a subsidiary of the U.S. Department of Agriculture signed into law by President Harry Truman in 1946 to assist low income families with school lunches. The goal of this program is to provide healthy, nutritional breakfast and lunch to all students, regardless of income. All public schools K-12 and applicable charter schools are eligible to receive funding from this program. In order for households to qualify for assistance, they must meet the income standards: 185% below the poverty income line to qualify for reduced, and 130% to qualify for free lunches.

In 2020, when COVID-19 sent students home around the country, one of the biggest concerns was students who relied on this program to eat their (potentially) only healthy meal of the day. A bill was passed stating that all schools were able to provide free school breakfasts and lunches to students, regardless of income status, with a pickup and delivery method enacted to reach the maximum amount of households. This bill, called the Child Nutrition COVID-19 Waivers, expired on June 30th, leaving roughly 10 million students without free lunch who had previously utilized this program during the pandemic. A new bill has been introduced, but is currently making its way through Congress, with no current action in place.

Currently, roughly 1.54 million students pay full price for a meal they can’t afford because their households fall above the reduced meal line but still don’t make enough income to cover school meals. Often, these same families are unable to pack home lunches due to long work hours mandatory to make rent, other food payments, and to pay for other essential items. 74% of households have some form of student lunch debt, which, recently, have caused those same students to be taken into foster care or be denied meals at school.

Additionally, school meals can be the only source of nutrition for some students. The NSLP provides strict guidelines regarding what food should be served to children to meet nutritional needs and alleviate the concern of eating disorders and malnutrition. Studies have proven that when students are known to receive free lunches, they face stigma from their peers, and are less likely to eat and more likely to develop severe eating disorders.

The biggest opposition to providing all students with free meals are tax payers who believe their money will be wasted when students choose not to eat the meals provided. While this has basis on the surface level, looking closer into the issue reveals that schools already plan and adjust for what their anticipated numbers are–the number of students eating school meals are not likely to change, it is simply the number of students paying for those meals. Those same opponents claim that providing free meals will make students lazy and “expectant” of the world to cater to them. Others see it as an opportunity for all students to eat and be fed and focus during class.

Study Purpose and Hypothesis

In the spring of 2022, I took a class called “What is Education For?” where our final project was to examine a school issue we felt passionately about and create a research project. There’s a crystal clear memory in my mind of the stress surrounding school lunches, not knowing if my parents had put money into my account, if I would be denied certain food “bonuses” because of lunch debt. Even though my parents were fully capable of paying for my school lunch, there were times they forgot, or were too busy to pack me lunch, and I still had that awful, sinking feeling of my peers watching me put food back when I was in school lunch debt. I am a firm believer in the principle of if students are required to be somewhere for eight hours a day, we (as a broader, adult society), should be required to feed them.

My research led me to a data set by the National Center for Education that tracked the number of enrolled students per state and the number of eligible students for free/reduced lunches per state. After researching the number of students who currently use the program and how many used the program when it was available to them during the COVID-19 pandemic, I believe even more in the idea that all students should have access to free lunch, regardless of income status.

In this study, I will be evaluating the changes in eligibility for the free/reduced lunch program from 2000-2019 using geographical data and charts. This study does not evaluate how many students actively use the program; it looks at how many students, according to poverty and income levels, are able to use the program. My hypothesis, with all previous information in mind, is that most, if not all, states will have an increase of students eligible for the free/reduced lunch program, given the 2008 recession and other financial factors over the last 20 years. Additionally, I will evaluate the changes in population of students enrolled versus the number of students eligible for the free/reduced lunch program to determine if median income is increasing or decreasing in a given state. This study does not use median income data, but looks to see if there is a positive or negative correlation over the years.

Packages and Data Evaluation

For this project, I will be using the following packages:

library(tidyverse)
library(billboard)
library(tidytext)
library(readxl)
library(gridExtra)
library(cowplot)

The data set I am using for this project was found on the National Center for Education Statistics. Additionally, two Excel files were created (without adjusting any values, only structure of the tables) for ease of calculation and analysis.

nslp <- read_excel("nslp.xlsx")
adjusted <- read_excel("adjustedData.xlsx")
percentLunch <- read_excel("percentLunch.xlsx")

Mapping Proportional Change Over 19 Years

2000-2001

The data set for this project begins in 2000 and each data frame spans a school year. In order to get a scope for the size of the population, I’ve created a geographical visual displaying the enrolled K-12 student populations for each state in 2000-2001.

In this map, it’s clear states you’d expect to have higher populations – such as California, Texas, New York, and Florida all have higher student enrollments. For the sake of ease in visualization, Hawaii and Alaska were not mapped, but are included in the dataset for evaluation.

When looking at the proportion of students eligible for free/reduced lunches throughout the United States in the same time frame, population is less of a factor than income (although many would argue population density affects median income).

Below is the population proportion map for 2000-2001 enrolled students eligible for free/reduced lunches according to the NSLP household income standards.

Immediately, the map indicates that a large proportion of enrolled students in the southeast are eligible for free/reduced lunches, while most northeast and west states have lower percentages. Historically, this region does have a lower median income, so this data is not surprising.

2010-2011

In the 10 year gap, the significant change is not the states that have a higher proportion of eligible students – it is in the overall increase in eligible students. 2008 saw a recession and many households were likely still suffering financial effects in 2010-2011, causing this increase in eligibility as more and more household met NSLP standards of poverty.

The next grouping of years are several chronological school years, without large time gaps. Each of the maps are below, segmented by section headers, and a full discussion of the change between 2016-2019 will follow the 2018-2019 map.

2016-2017

2017-2018

2018-2019

Over these three school years, eligibility for the NSLP free/reduced lunch program increased sharply among nearly all states. States that already had a high proportion of eligible students increased, and states who’d started with a proportion lower than 50%, such as Texas, rose to above 60%.

Again, while these maps show eligibility, they do not show the proportion of eligible students actually using the program. However, with enrollment data and eligibility data, trend lines can be mapped to see if there is a positive or negative correlation throughout the years. As populations increase, does eligibility rise in the same proportion?

Holistic Evaluation

In looking at this data, it is important to understand the scope of change over time. Are the states that are increasing in eligibility the same as the states increasing in enrollment?

adjusted %>%
  filter(Status == "Enrollment") %>%
  ggplot(aes(Year, Students, color = State)) +geom_point(position = "jitter") +
  theme_classic() +
  labs(x = "Year",
       y = "Number of Students Enrolled",
       title = "Number of Students Enrolled in a K-12 School",
       subtitle = "By State")

This is a messy, jumbled map, but it does give us a good, baseline indication of enrollment statistics. There are a few outlier states, as seen in the 2000-2001 map, with significantly higher enrollment populations. These states are California, Texas, and New York, which are predictably denser.

adjusted %>%
  filter(Status == "Eligible") %>%
  ggplot(aes(Year, Students, color = State)) +geom_point(position = "jitter") +
  theme_classic() +
  labs(x = "Year",
       y = "Number of Students Eligible",
       title = "Number of Students Eligible for Free/Reduced Lunches in a K-12 School",
       subtitle = "By State")

Again, this looks very similar to the map above; illegible, cluttered, but indicative of densely populated states having higher eligibility numbers. However, the maps above looked at percentages of enrolled population eligible – not just numbers.

percentLunch %>%
  filter(!State %in% c("United States", "District of Columbia")) %>%
  ggplot(aes(Year, Percent, color = State)) + geom_point(position = "jitter") +
  theme_classic() +
  labs(x = "Year",
       y = "Percent of Students Enrolled",
       title = "Percent of Students Enrolled who are Eligible for Free/Reduced Lunches",
       subtitle = "By State, K-12 Schools")

Now, California is no longer at the top of the chart. Several states with overall lower student enrollments have higher student eligibility rates, indicating that more densely populated states do not necessarily have higher eligibility rates simply because they have more students. From 2000-2019, there is a sharp increase in the number of students eligible for free/reduced lunches, but there is not a sharp increase in enrollment population. Many of the states with higher proportions have lower overall population numbers–indicating possible higher levels of poverty.

Evaluation by Region

Based off map analysis, it is clear that the south and west are the two regions with the most increase in eligibility for free/reduced lunches. In order to look at change over time, I have split the states into four geographic regions: northeast, south, midwest, and west. These divisions were made based off standard U.S. regional divisions.

South

south <- adjusted %>%
  filter(State %in% c("Delaware", "Florida", "Georgia", "Maryland", "North Carolina", "South Carolina",
                      "Virginia", "West Virginia", "Alabama", "Kentucky", "Mississippi", "Tennessee",
                      "Arkansas", "Louisiana", "Oklahoma", "Texas"))

south %>%
  filter(Status == "Enrollment") %>%
  ggplot(aes(Year, Students, color = State)) +geom_point(position = "jitter") +
  theme_classic() +
  labs(title = "Enrollment by State",
       y = "Number of Students")-> s3

south %>%
  filter(Status == "Eligible") %>%
  ggplot(aes(Year, Students, color = State)) +geom_point(position = "jitter") +
  theme_classic() +
  theme(legend.position = "none",
        axis.title.y = element_blank()) +
  labs(title = "Eligibility by State")-> s4

percentsouth <- percentLunch %>%
  filter(State %in% c("Delaware", "Florida", "Georgia", "Maryland", "North Carolina", "South Carolina",
                      "Virginia", "West Virginia", "Alabama", "Kentucky", "Mississippi", "Tennessee",
                      "Arkansas", "Louisiana", "Oklahoma", "Texas"))

percentsouth %>%
  ggplot(aes(Year, Percent, color = State)) + geom_point(position = "jitter") +
  theme_classic() +
  labs(title = "Percent of Students Eligible") +
  theme(legend.position = "none",
        axis.title.y = element_blank()) -> s5

legend <- get_legend(s3)

s3 <- s3 + theme(legend.position = "none")

blankPlot <- ggplot() + geom_blank(aes(1,1)) +
  cowplot::theme_nothing()

grid.arrange(s3, s4, s5, legend, 
             ncol=4, widths = c(2.7, 2.7, 2.7, 1))

Delaware is the only state to decrease in population proportion of eligible students for free/reduced lunches from 2000-2019. Without a sharp decrease in either population for enrollment or eligibility, this decrease indicates that median income in the state increased during this time frame, leading to less students being eligible for free/reduced meals. Overall, however, the enrollment population did not have a drastic enough increase to match the proportion of eligible students, indicating that poverty in these states got worse during the selected time frame.

This region has the widest range of values in proportions of eligibility. Delaware is the second smallest state geographically, which lends itself to the lower percentage of eligible students. Additionally, the south has the highest overall proportion of students eligible per state. All but one state is above 40% of enrolled students eligible for the NSLP, which no other region has. The south also starts at a baseline in 2000-2001 higher than any other region, with all states having at least 30% of enrolled students eligible for the NSLP.

Midwest

midwest <- adjusted %>%
  filter(State %in% c("Illinois", "Indiana", "Michigan", "Ohio", "Wisconsin", "Iowa", "Kansas",
                      "Minnesota", "Missouri", "Nebraska", "North Dakota", "South Dakota"))

midwest %>%
  filter(Status == "Enrollment") %>%
  ggplot(aes(Year, Students, color = State)) +geom_point(position = "jitter") +
  theme_classic() +
  labs(title = "Enrollment by State",
       y = "Number of Students")-> m1

midwest %>%
  filter(Status == "Eligible") %>%
  ggplot(aes(Year, Students, color = State)) +geom_point(position = "jitter") +
  theme_classic() +
  theme(legend.position = "none",
        axis.title.y = element_blank()) +
  labs(title = "Eligibility by State")-> m2

percentmidwest <- percentLunch %>%
  filter(State %in% c("Illinois", "Indiana", "Michigan", "Ohio", "Wisconsin", "Iowa", "Kansas",
                      "Minnesota", "Missouri", "Nebraska", "North Dakota", "South Dakota"))

percentmidwest %>%
  ggplot(aes(Year, Percent, color = State)) + geom_point(position = "jitter") +
  theme_classic() +
  labs(title = "Percent of Students Eligible") +
  theme(legend.position = "none",
        axis.title.y = element_blank()) -> m3

legend <- get_legend(m1)

m1 <- m1 + theme(legend.position = "none")

blankPlot <- ggplot() + geom_blank(aes(1,1)) +
  cowplot::theme_nothing()

grid.arrange(m1, m2, m3, legend, 
             ncol=4, widths = c(2.7, 2.7, 2.7, 1))

There is a general trend of decrease of eligibility for the NSLP within the midwest states. These states generally have a much lower school enrollment, with all states having an enrollment of less than 200,000 students per state. As per the maps, these were the states with the lowest general enrollment. However, all of these states did have an increase in eligibility from 2000-2010, with a slight uptick in eligibility for several states.

In the eligibility proportion graph, the values are generally grouped in one range, from 35% to 55%, with the only outlier being North Dakota. North Dakota is the least populated state in this region, which additionally has the lowest eligibility proportion for the NSLP.

Northeast

northeast <- adjusted %>%
  filter(State %in% c("Connecticut", "Maine", "Massachusetts", "New Hampshire", "Rhode Island",
                      "Vermont", "New Jersey", "New York", "Pennsylvania"))

northeast %>%
  filter(Status == "Enrollment") %>%
  ggplot(aes(Year, Students, color = State)) +geom_point(position = "jitter") +
  theme_classic() +
  labs(title = "Enrollment by State",
       y = "Number of Students")-> n1

northeast %>%
  filter(Status == "Eligible") %>%
  ggplot(aes(Year, Students, color = State)) +geom_point(position = "jitter") +
  theme_classic() +
  theme(legend.position = "none",
        axis.title.y = element_blank()) +
  labs(title = "Eligibility by State")-> n2

percentnortheast <- percentLunch %>%
  filter(State %in% c("Connecticut", "Maine", "Massachusetts", "New Hampshire", "Rhode Island",
                      "Vermont", "New Jersey", "New York", "Pennsylvania"))

percentnortheast %>%
  ggplot(aes(Year, Percent, color = State)) + geom_point(position = "jitter") +
  theme_classic() +
  labs(title = "Percent of Students Eligible") +
  theme(legend.position = "none",
        axis.title.y = element_blank()) -> n3

legend <- get_legend(n1)

n1 <- n1 + theme(legend.position = "none")

blankPlot <- ggplot() + geom_blank(aes(1,1)) +
  cowplot::theme_nothing()


grid.arrange(n1, n2, n3, legend, 
             ncol=4, widths = c(2.7, 2.7, 2.7, 1))

The most notable trend in the north maps is the decline of enrolled students in New York with an increase of eligible students in New York. Additionally, Connecticut also had an increase in eligible students with a decrease in enrolled students. Overall, these trends in proportion of students eligible follow the trends of enrollment and eligibility populations.

There is a mild amount of distribution within the eligible population proportion, though from 2010-2019 the proportions all fall within a range of 30% of eligible students by state. The outlier of New York in terms of population is understandable, given the hub of New York City and the large population of New York.

West

west <- adjusted %>%
  filter(State %in% c("Arizona", "Colorado", "Idaho", "Montana", "Nevada", "New Mexico",
                      "Utah", "Wyoming", "Alaska", "California", "Hawaii", "Oregon", "Washington"))

west %>%
  filter(Status == "Enrollment") %>%
  ggplot(aes(Year, Students, color = State)) +geom_point(position = "jitter") +
  theme_classic() +
  labs(title = "Enrollment by State",
       y = "Number of Students")-> w1

west %>%
  filter(Status == "Eligible") %>%
  ggplot(aes(Year, Students, color = State)) +geom_point(position = "jitter") +
  theme_classic() +
  theme(legend.position = "none",
        axis.title.y = element_blank()) +
  labs(title = "Eligibility by State")-> w2

percentwest <- percentLunch %>%
  filter(State %in% c("Arizona", "Colorado", "Idaho", "Montana", "Nevada", "New Mexico",
                      "Utah", "Wyoming", "Alaska", "California", "Hawaii", "Oregon", "Washington"))

percentwest %>%
  ggplot(aes(Year, Percent, color = State)) + geom_point(position = "jitter") +
  theme_classic() +
  labs(title = "Percent of Students Eligible") +
  theme(legend.position = "none",
        axis.title.y = element_blank()) -> w3

legend <- get_legend(w1)

w1 <- w1 + theme(legend.position = "none")

blankPlot <- ggplot() + geom_blank(aes(1,1)) +
  cowplot::theme_nothing()

grid.arrange(w1, w2, w3, legend, 
             ncol=4, widths = c(2.7, 2.7, 2.7, 1))

Though California has the highest population of enrolled students and therefore, students eligible for the NSLP, New Mexico has the highest proportion of students eligible for the NSLP. New Mexico has 12% of all Native American tribe reservations in the U.S., a traditionally lower-income group due to racism and genocide. Native American students additionally struggle more with the standards of American education due to the differences in early childhood teaching and cultural “norms”.

Alaska and Arizona also have high Native American populations and have some of the highest eligibility proportions as well. California, with its high population density in both enrollment and thus eligibility, also has a high proportion of students eligible for the NSLP.

There is a large range of eligibility proportions in the West, though a majority of the lower eligibility proportions occur in 2000. By the 2018-2019 school year, all states had an eligibility proportion above 30% of all enrolled students per state.

Conclusion

Overall, the scales of the graphs give an indication of change over time, but are unable to show the minute changes that would demonstrate drastic change rather than minimal change. For states with small changes, it appears as though they have not increased/decreased much, if at all, but that is several thousands of students being impacted by every changed in increase or decrease.

More data and information is needed to formulate an accurate and concise conclusion. A preliminary conclusion states that enrollment populations in the United States did not increase at the same rate as eligibility populations, indicating financial strain was placed on many households as they became eligible for the NSLP by federal standards. The southern region of the United States has the highest proportion of students eligible for the NSLP; this does not follow the expected pattern of higher enrollments equating to higher proportions of eligible students due to income distribution among a larger population. Additionally, this data does not show how many eligible students actually use the NSLP. Despite this, up to 76% of all enrolled students in Mississippi qualify for free/reduced lunches. In an average Mississippi classroom of 20 students, 15 of those students are eligible. When a majority of the classroom is eligible for free lunches, it creates an unnecessary and potentially harmful gap between lower, middle, and upper class students. If students feel uncomfortable taking free lunch when they don’t believe their peers are doing the same, they are less likely to eat, and thus, less likely to learn.

This data demonstrates that a large proportion of United States students do qualify for free/reduced lunches. School lunches should be free for all students, regardless of income.

Further Studies

In order to fully understand the scope of this data and the impacts it has on American households, it is essential to include additional data. Further studies and evaluation would include: a. how many of the eligible students actually use the NSLP (as mentioned throughout), b. poverty and median income data to supplement the above maps and generate further understanding of what is causing such high eligibility rates in certain areas, and c. how many students per state used the program in the COVID-19 era that don’t normally qualify for it.

A. If 75% of students are eligible but only 30% are using the program, this indicates several potentials. It could be problems with the program and how people are accessing it; there’s a surrounding stigma that drives students away from using the program; only 30% of those students choose to use the program, potentially due to life situations where parents or grandparents may be able to provide home lunches.

B. Understanding poverty and median income throughout the states can reveal if the NSLP functions as it should – an alleviation for lower-income families. Additionally, if poverty and median income do not correlate with the prevalence of NSLP in America, it indicates that another factor does, potentially population density.

C. The argument that this program should be extended is founded heavily in the people it benefited during the COVID-19 lock down era. Knowing how many people who are normally in lunch debt who were relieved from that debt by access to free lunches is essential data to the argument for passing a bill to make school lunches free for all K-12 students.