This figure shot up during the great recession (reaching a high of 14.9% of households in 2011) but has subsequently fallen a bit as indicated above, but still remains higher than was the case prior to 2008 when it was regularly in the 10 to 11% range.
I was curious how the sample of people I study every month in my Food Demand Survey (FooDS) matches up with these official government statistics. In the most recent April 2017 edition of FooDS, we added some questions (the short 6-item measure) based on work by the USDA to measure food insecurity. As an example, one of the questions is "'The food that I bought just didn’t last, and I didn’t have money to get more.' Was that often, sometimes, or never true for you/your household in the last 12 months?"
Data from FooDS reveals a strikingly high level of food insecurity - much higher than what the USDA reports. According to the criteria outlined at the above link, we found a whopping 46.7% of respondents were classified as having low or very low food security (22.9% of the sample had low food security and 23.8% had very low food security).
My first thought was that we must have made a mistake in how we asked the questions or in how we analyzed the data. We ruled out those possibilities. My second thought was that maybe my survey sample is really different from the U.S. population. After all, who is willing to sign up to take online surveys? Maybe people who really need the money and who are thus more likely to be food insecure. But, this couldn't be the complete answer because I use weights to force my sample to match the U.S. population in terms of age, education, gender, and region of residence, and the average income of my sample isn't much different from the average income of the country as a whole. Maybe the difference is that I used a 6-item measure of food insecurity rather than the full 18 items used by the USDA (but previous research has found strong agreement between the two).
When I mentioned this quandary to my friend Bailey Norwood, he knew immediately what was causing part of the the discrepancy, and I think it could have a big impact on how we fundamentally view the food security measures reported by the USDA.
In short, the USDA assumes that if you make enough money you can't be food insecure [*Addendum, this original sentence, as stated, was too strong. As the quote below suggests, you can't be classified as food insecure if you're high income AND if you answer two preliminary questions on food insufficiency in particular way. Some researchers in this area emailed me to note that about 25% of food insecure households have incomes at least 300% of the poverty line]. In their latest report, they indicate in footnote 5: