National Academies raise questions about measuring "hunger"

The Committee on National Statistics, part of the National Academies, has just posted online a pre-publication version of its new report on measuring food security. The report discusses and sometimes criticizes the survey methodology used to count how many Americans are "food insecure" and hungry. You have probably seen the resulting statistics reported by the media after they have been released each Fall in recent years (last year, conveniently, right after the Presidential election -- see this post and this one).

The expert panel that produced the report had commissioned a paper from me last summer about how the methodology is actually used by people in the media, government, and elsewhere, so I found the discussion of those issues especially interesting:

The USDA estimates, published in a series of annual reports, are widely used by government agencies, the media, and advocacy groups to report the extent of food insecurity and hunger in the United States, to monitor progress toward national objectives, to evaluate the impact of particular public policies and programs, as a standard by which the performance of USDA programs is measured, and as a basis for a diverse body of research relating to food assistance programs. Government agencies have also adopted the estimates as targets for performance assessment. The U.S. Department of Health and Human Services (DHHS) has included the food security measure to assess the performance of its Healthy People 2010 initiative. The Food and Nutrition Service of the USDA is using the measure as a target for its strategic plan to fulfill requirements of the Government Performance and Results Act of 1993 (Wilde, 2004a).

Despite the extensive use of the measure, some major questions related to the concepts themselves, the methodology, and their use, continue to be raised.

While the USDA annual reports define the concepts of food security and the three categories of food security that are estimated and reported (food secure, food insecure without hunger, and food insecure with hunger) and provide detail about how they are measured, the terms "food security" and "food insecurity" are relatively new to both policy makers and the public and are sometimes confusing. While the term "hunger" is not new, measurement of hunger and how hunger conceptually fits into food insecurity is not completely clear. As currently construed in USDA's food security measure, hunger is considered a severe level of food insecurity. This use of the term "hunger" has been questioned by some who believe that hunger is conceptually separate from food insecurity. Because the label "hunger" is a politically potent concept, the methods used to classify households as food insecure with hunger are particularly important.

Methodological and technical issues about the measure of food insecurity generally concern the clarity, appropriateness, and design of the CPS survey questions. Critics question:
-- using a relatively long (12-month) reference period,
-- mixing questions focused on the household with questions focused on individuals,
-- using the raw score on the module to categorize households into one of the three food security categories, and
-- using the same module to assess the food security of households with children and households without children.

Questions about the appropriate uses of the estimates of food security also have been raised. The primary use of the Food Security Supplement of the CPS is to estimate the prevalence of the categories of food security. The media and advocacy groups often interpret the prevalence estimates in language inconsistent with USDA usage. As currently measured, the estimates may not be appropriate for use in policy and program evaluations. Even if they are used, it would be helpful for their use to be consistent across federal government departments. The USDA strategic plan uses a food security target that differs from the DHHS Healthy People 2010 objectives, and the USDA annual performance reports omit the target altogether (Wilde, 2004a).

The report reaffirms the usefulness of survey methods for measuring food security, but it is likely to necessitate some revisions from the measure's sponsors at USDA. All for the best. It is worthwhile to make even a good measure better. My own number one suggestion is to keep the statistical methods as straightforward and transparent as possible.

The link to the report: