America's Second Harvest, a national network of food banks and food rescue organizations, released its quadrennial Hunger in America survey on February 23, 2006,  only a few months after the U.S. Department of Agriculture (USDA) released its annual food insecurity and hunger study.  The Second Harvest report makes an important contribution to understanding the extent of hunger and food insecurity in America.
However, the popular press has misrepresented the results of the Second Harvest survey, partly as a consequence of the report's flawed executive summary. Specifically, the Associated Press and other media outlets have suggested that food bank utilization increased by 9 percent between 2001 and 2005 when in fact the numbers indicate no such change.
Mathematica Policy Research (MPR), the statistical contractor for Second Harvest, estimated that food banks and other emergency food providers served 22.8 million to 28.1 million people in 2001, compared to 23.7 million to 27.0 million individuals in 2005-statistically equivalent ranges. Indeed, MPR notes, "Our 'bottom line' estimate is that overall the range of participation estimates for 2005 is similar to that in 2001." 
Notwithstanding this major concern, the Second Harvest survey provides a great deal of welcome information on the role played by its network of food banks and food rescue organizations as part of the nation's social safety net.
Hunger in America 2005 Survey
About every four years, America's Second Harvest surveys tens of thousands of individuals who use the services of food pantries, soup kitchens, and shelters in the Second Harvest network. Because some 80 percent of all emergency food providers participate in the Second Harvest network,  it is uniquely positioned to conduct this survey. While the food banks and other organizations in the Second Harvest network conducted many of the individual surveys, Mathematica Policy Research, a survey research organization located in Princeton, New Jersey, administered the overall effort.
In early 2005, more than 52,000 surveys were collected from clients of emergency food providers to glean not only basic information on utilization, but also information on the characteristics of individuals who receive bulk food and/or meals.
Utilization. The Second Harvest network served some 4.5 million individuals on a weekly basis in 2005, which is down considerably from 2001, when about 7.0 million people were served weekly.
Extrapolated to an annual figure, MPR estimated that between 23.7 million and 27.0 million individuals were served by the Second Harvest network of providers in 2005. This range is similar to the 2001 Second Harvest survey, which estimated that between 22.8 million and 28.1 million individuals were served, even though the weekly figure is much lower.
Which are better: the weekly or annual estimates? By construction, weekly figures are easier to tabulate. As MPR remarked:
Computing annual estimates unavoidably required asking survey respondents to report on their use of the emergency food system over a significant amount of time-a year in some instances. This long reporting span undoubtedly increases reporting error. In contrast, the weekly estimate requires only that respondents be able to report on their use of the system during the week of the survey- a considerably less exacting requirement.
Different turnover rates explain why the estimate of weekly utilization in 2005 is so much lower than the 2001 estimate while the annual numbers are so similar. In 2001, MPR estimated that only 5 percent of clients were new; that is, they had begun using pantries, soup kitchens, and/or shelters in the previous month and had not used them in the previous 12 months. In 2005, however, the turnover rate increased to 14 percent.
Lower turnover means that fewer new people are entering the system on a monthly basis; therefore, fewer individuals are served over the course of a year. Higher turnover means that more individuals are utilizing the services for the year. In other words, in 2001, fewer clients were new in a given month than in 2005, when far more clients were new.
Food Insecurity. Not surprisingly, most clients in the Second Harvest survey were "food insecure" as defined by the USDA. Some 70 percent of clients were food insecure, but only about one-third were considered food insecure with hunger, which is what the USDA means by hunger.
This distinction is important because it underscores the extent of hunger in America. While Second Harvest has done a remarkable job of putting together a fine survey, some of the results have been reported inaccurately. For example, the report's executive summary-written by America's Second Harvest, not MPR-appears to confuse hunger with emergency food utilization. It implies that 25.35 million hungry people are using their network's services when in fact only one-third of that number are hungry. Most of those remaining are more accurately categorized as food insecure. (For definitions, see the text box.)
What Is Food Insecurity and Hunger?
Advocates and analysts frequently use the two concepts "food insecurity" and "hunger" interchangeably when they mean different things. The USDA uses the following definitions:
U.S. Department of Agriculture, Economic Research Service, "Food Security in the United States: Measuring Household Food Security," updated November 19, 2004, at http://www.ers.usda.gov/briefing/FoodSecurity/measurement (March 15, 2006).
S. A. Andersen, ed., "Core Indicators of Nutritional State for Difficult to Sample Populations," Journal of Nutrition, Vol. 120 (1990), pp. 1557S-1600S.
U.S. Department of Agriculture, "Food Security in the United States."
Comparison of the Surveys
Because the Second Harvest survey and the annual USDA survey on food insecurity ask many of the same questions about food insecurity and hunger, their results are comparable, at least in part. However, because the two surveys deal with different populations (clients of food banks and other providers versus a general survey of all households), the comparisons are imperfect.
According to the USDA's survey of over 60,000 households:
- Only 3.7 percent of individuals in the United States were hungry during 2004, according to the USDA definition of hunger. This percentage, while it has fluctuated somewhat over the decade, is the same as the rate in 1998.
- Only 0.7 percent of children were hungry in the United States. This rate has varied only slightly during this decade and is the same as the rate in 1999.
America's Second Harvest study of the over 52,000 clients found that:
- About 4.5 million individuals received meals or packaged food in a given week in 2005. This is down substantially from 2001, when an estimated 7.0 million individual clients were served in a given week.
- Although fewer people were served on a weekly basis, food pantries and other providers found that they were serving about the same number of people annually. In 2005, between 23.7 million and 27.0 million people were served by the network, compared to 22.8 million to 28.1 million in 2001. Because the two ranges include roughly the same numbers of people, there is no significant difference in emergency food utilization between 2001 and 2005.
- Accounting for this apparent discrepancy between the weekly and annual estimates is that more clients reported being new to the Second Harvest system in 2005 than in 2001. A higher turnover means fewer clients on a given week but more (unique or unduplicated) clients served over the course of a year.
The results of the Second Harvest survey compare very favorably to the USDA survey of food insecurity and hunger in America. According to the USDA, just under 10.7 million individuals in the United States were hungry, or about 3.7 percent of the total U.S. population. While this proportion has fluctuated somewhat over the past several years, the percent of the U.S. population deemed hungry is the same as it was in 1998.
Thankfully, most of the U.S. population considered hungry by the USDA definition goes to Second Harvest's network of providers for emergency food. About 8.45 million people (midpoint estimate) are hungry and are Second Harvest clients, compared to the 10.7 million hungry individuals nationwide, or just under 80 percent.
What the Surveys Mean
Perhaps the most problematic finding reported in the press is that hunger has increased by 9 percent between 2001 and 2005. This "finding" appears in the report's executive summary, which Second Harvest wrote. However, it is not supported by the full report as written by MPR: "Our 'bottom line' estimate is that overall the range of participation estimates for 2005 is similar to that in 2001." The study results indicate that utilization did not increase or decrease to a statistically significant degree, although external data seem to rule out a decrease.
Between the two ranges of plausible estimates developed by MPR, there is no statistically significant difference. In layman's terms, the utilization rates in the Second Harvest network between 2001 and 2005 are sufficiently close that they can be considered statistically equivalent after the possibility of errors is taken into account.
Nevertheless, the Associated Press and other media outlets have reported erroneously that food utilization has increased and have ignored the fact that any differences between the 2001 and 2005 estimates could have been due entirely to chance.
Notwithstanding this concern, the Second Harvest survey illustrates two important features of the charitable food system in the United States. The first is that these private organizations serve a vital function in the social safety net. As Jean Daniel from the USDA noted to the Associated Press, "We have said all along that the government cannot do this alone, nor should it. Their efforts dovetail very nicely with ours." Furthermore, these private food and service organizations are likely to know (and serve) the needs of their clients better and more effectively than government bureaucracies can.
In addition, the Second Harvest survey squarely illustrates the importance of faith-based organizations in providing emergency food services. For example, faith-based agencies run nearly 75 percent of all food pantries. With this faith component, it is not surprising that volunteers are critically important to the food distribution effort, with 90 percent of pantries using volunteers. In fact, two-thirds of all pantries are staffed wholly by volunteers. This shows that these organizations not only provide needed services, but also are a labor of love.
America's Second Harvest has again issued an extremely useful report on the role that its network of food banks and food rescue organizations plays in the nation's social safety net. While Second Harvest's national report has rightfully received much of the attention, many local and regional reports are also available. For example, the Capital Area Food Bank report documents the efforts of dozens of local agencies in the Washington, D.C., metropolitan area that serve over 69,000 individuals every week.
These reports showcase the generosity of Americans in providing aid and assistance to the needy in their local communities. Their generous and extensive work shows why there are few more worthy charitable organizations to which to volunteer one's time and talents.
Kirk A. Johnson, Ph.D., is a Senior Policy Analyst in the Center for Data Analysis at The Heritage Foundation and served as a member of the Technical Advisory Group for Second Harvest's 2005 survey.
Policy Research, Hunger in America 2006: National Report
Prepared for America's Second Harvest, draft report, MPR
Reference No. 6100-500, February 2006, at
about_the_study/A2HNationalReport.pdf (March 15, 2006).
Mark Nord, Margaret Andrews, and Steven Carlson, "Household Food Security in the United States, 2004," U.S. Department of Agriculture, Economic Research Service Economic Research Report No. 11, October 2005, at http://www.ers.usda.gov/publications/err11/err11.pdf (March 15, 2006).
See Associated Press, "Study: More Americans Are Relying on Food Banks," February 23, 2006, and Steve Inskeep, "Food Assistance Used by 25 Million Americans," National Public Radio News, February 23, 2006.
Mathematica Policy Research, Hunger in America 2006, p. 67.
James Ohls and Fazana Saleem-Ismail, "The Emergency Food Assistance System-Findings from the Provider Survey," U.S. Department of Agriculture, Economic Research Service Food Assistance and Nutrition Research Report No. 16, June 2002, at http://www.ers.usda.gov/publications/fanrr16-1/fanrr16-1.pdf (March 15, 2006).
Mathematica Policy Research, Hunger in America 2006, p. 46, Table 4.2.1. MPR also adjusted for those individuals who visit different agencies (e.g., receive bulk food at a pantry but also a meal at a soup kitchen).
Mathematica Policy Research, Hunger in America 2006, p. 51.
L. O'Brien and Halley Torres Aldeen, "Hunger in America 2006:
Executive Summary," 2006, at
about_the_study/FinalHungerStudyExecutiveSummary.pdf. The authors titled one of the charts "Hunger on the Rise" when in fact it discusses utilization figures, not hunger as defined by the USDA.
The 25.35 million figure cited is the midpoint between the lower and upper ranges of the MPR estimate of annual utilization. If 33 percent of those individuals are deemed hungry, then only about 8.5 million clients are hungry, not 25.35 million.
Nord et al., "Household Food Security in the United States, 2004," p. 6, Table 1.
Mathematica Policy Research, Hunger in America 2006, p. 67.
See Associated Press, "Study: More Americans Are Relying on Food Banks," and Inskeep, "Food Assistance Used by 25 Million Americans."
Associated Press, "Study: More Americans Are Relying on Food Banks."
Smaller proportions of shelters (43 percent) and soup kitchens (65 percent) are run by faith-based organizations.