Below is an interesting article that appeared in the Oregonian

Below is an interesting article that appeared in the Oregonian. However, somewhere in the article is an error of statistical interpretation. See if you can find it. Hint: the error has to do with statistical inference.

Just how hungry is Oregon?

05/12/03

STEVE WOODWARD

If you believe what you hear, Oregon is the country's hungriest state.

In late April, Gov. Ted Kulongoski kicked off Hunger Awareness Week by declaring, "For several years now, Oregon has had the highest hunger rate in the nation."

The Oregon Food Bank says, "Oregon remains first in the nation for hunger."

And the Oregon Hunger Relief Task Force asserts, "Oregon has the highest rate of hunger in the United States."

But it's not that simple.

Oregon hunger statistics have large margins of error that make comparisons with other states risky. And the problem is compounded by the way government studies define "hunger," which is substantially removed from the severe malnutrition and distended bellies associated with Third World hunger.

Researchers and critics have expressed doubts about the No. 1 ranking. But various interest groups have continued using it to move hunger higher on Oregon's agenda and to promote policies designed to cope with it.

Oregon earns its ranking Oregon first earned the "hungriest state" label in 1999. That's when a U.S. Department of Agriculture study first suggested Oregon had the nation's highest percentage of households with one or more members who went hungry at least once in the previous year.

Three later studies -- based on the same survey methods -- also suggested that Oregon led the pack in hunger.

Oregon may indeed be No. 1, but -- because of large sampling errors in the surveys -- it theoretically could have ranked as low as 13th in 1999. Far more likely, Oregon ranks among the top four or five states, says Mark Nord, lead author of the USDA study.

"What I'm pretty sure of, based on the statistics," Nord says, "is that Oregon is higher than the national average."

Because of the large error margins, Nord cautioned against comparing states. The researcher who ran the original study, in other words, says it shouldn't be used for hunger rankings. The USDA itself no longer issues rankings based on state-by-state comparisons, and the only other ranking came from a Brandeis University study in 2002.

Jeffrey Tryens, executive director of the Oregon Progress Board, regrets that the No. 1 ranking has played such a significant part in the discussion. "I have always said that Oregon has a significant hunger problem," Tryens says, "not that Oregon is the hungriest state in the nation. It's unfortunate that this 'Oregon is the hungriest state' issue has stuck."

Hunger definitions vary Statistics aren't the only issue when it comes to Oregon's No. 1 rating. The idea of "hunger" as some people use the word doesn't necessarily match the meaning of "hunger" as defined by the USDA study.

The USDA says people are hungry when they report that they have cut back, skipped meals or gone a day without eating because they didn't have money to buy food. In a quarter of U.S. households classified as hungry, that kind of hunger reportedly happens almost every month. And in the other three-quarters of hungry households, it happens at least once a year.

"The survey is not telling us we have thousands of people who are starving," says Michael Leachman, an Oregon sociologist who has studied the USDA data, "but we do have problems with people who at times have to go hungry . . . because they don't have enough money for food."

Nor do many children go to bed hungry, contrary to popular belief, according to the USDA numbers. In most cases, the USDA says, adults protect children -- especially younger children -- from hunger by going hungry themselves.

State collects own data Even with the USDA's limited definition of hunger, Oregon's apparent status as the nation's hungriest state mortified anti-hunger interest groups and state agencies.

"A lot of people were shocked," says Kim Thomas, public-policy manager for the Oregon Food Bank. "People were skeptical of the data."

Thomas says the food bank staff, which had seen a significant increase in the number of emergency food boxes it distributed, wasn't surprised by the ranking. Nevertheless, Oregon's high hunger numbers in the federal study -- now most likely 4.9 percent to 6.7 percent of the population -- puzzled researchers and policy-makers.

After all, Oregon is not a poor state. Its poverty levels are average, unlike Southern and Southwestern states classified as hungry.

Oregon's median income for four-person families was the nation's third-fastest-growing from 1996 to 2000.

Oregon's participation in the federally subsidized school-lunch program ranks near the bottom among states.

Yet the hunger numbers seemed unmistakably large. Statistically, it was a 90 percent certainty that hunger, as defined by the USDA study, was present in 64,000 to 87,000 Oregon households a year from 1999 to 2001. Those households probably represented a higher percentage of the population than in most other states.

The results might have been a fluke. The 90 percent confidence level meant there was a 1 in 10 chance the sample didn't accurately represent the larger population. And most studies are designed with a significantly higher 95 percent confidence level, which means they have only a 1 in 20 risk of a bad sample.

Nonetheless, later studies produced results similar to the original USDA findings, which suggested that the samples were good.

A Northwest puzzle National hunger studies first arrived on the U.S. scene in 1990, when Congress asked for more information on the subject. A coalition of federal agencies then spent three years devising a set of 18 questions that would determine where a household fell on a continuum from "food-secure" to "food-insecure with hunger." "Food insecure" essentially meant that survey respondents were worried about whether they would have enough to eat.

In 1995, Census Bureau surveyors began incorporating the questions into a nationwide survey of more than 40,000 households. The first year's results showed that about 88 percent of U.S. households were food-secure. Of the remaining food-insecure households, two-thirds avoided hunger by changing their diets or using federal food programs, community food pantries or emergency kitchens; one-third had at least one member who had experienced hunger in the previous year. That would account for roughly 4 percent of U.S. households.

The 1995 data also presented researchers with their first Northwest puzzle: Oregon and Washington, both moderate-poverty states, were among the highest hunger states. Their high hunger ranking persisted through all subsequent surveys, out of proportion to their poverty.

The burning question for Nord and the USDA is not whether Oregon has a high hunger rate, but why the rate is higher than expected.

The federal agency has contracted with researchers at Oregon State University to find an answer. Bruce A. Weber, one of the researchers, says the OSU team is looking at things such as the size of the income gap between the rich and the poor, migration into the state and the share of the population that spends more than 35 percent of its income on rent. The team will even consider Oregon's reputation as a state with low church attendance.

Weber doesn't expect preliminary results until late June.

The researcher who has done the most work on hunger's root causes in Oregon is Leachman, a policy analyst with the Oregon Center for Public Policy.

Leachman says several factors could account for Oregon's high ranking. He points to the restructuring of Oregon's economy in the 1990s, a decade that saw extraordinary economic growth, even as high-paying manufacturing jobs were replaced by lower-wage service-sector jobs.

But the newly created wealth, he says, was not equally distributed between the rich and the poor. Instead, the lowest-income households saw incomes stagnate. At the same time, the highest-income households saw incomes grow disproportionately fast, driving up housing prices and other living costs.

"That," Leachman says, "created these economic conditions where even though our poverty rate was fairly middling, costs were rising rapidly for poorer people, and their incomes weren't keeping up."

High rents a factor Median Oregon rents jumped from $408 a month in 1990 to $620 a month in 2000, wrote Leachman and fellow policy analyst Jeff Thompson in a report, "Boom, Bust, & Beyond: The State of Working Oregon 2002." They note that from 1990 to 1999, there was a 23 percent increase in Oregon households paying more than half their income in rent, resulting in a relatively high rate compared with other states.

Other potential factors, Leachman says, are Oregon's higher-than-average seasonal employment and rural unemployment rates.

Portland-based economist John Tapogna says his team has isolated three conditions that strongly correlate with high hunger rates: high unemployment rates, high population growth and a high percentage of residents who spend more than 50 percent of their income on rent. Oregon ranks high in all three conditions, says Tapogna, managing director of the Eugene consulting firm ECONorthwest.

Unemployment, more than chronic poverty, is a hunger indicator, Tapogna says, because it signals an abrupt and sometimes unmanageable change in a household's budget. As for high housing costs, the poor sometimes are forced to cut back on food to raise money for rent payments and avoid eviction. And population growth means that many residents are newcomers who lack a social safety net of family and friends or who don't know how to get government services.

John A. Charles, of the free-market Cascade Policy Institute in Portland, has other ideas about the factors that leave too little money for food. "Those factors," he writes in a position paper, "Does Oregon Have a Hunger Problem?" "could include lack of education, divorce, disability, unplanned pregnancy or any number of character flaws such as drug use or lack of a work ethic."

Charles, a critic of Oregon's No. 1 hunger ranking, contends that many of those people constitute a permanent underclass.

"There will be a certain number of people who will always be hungry because they continue to make bad decisions," he says.

Politics of being No. 1 Regardless of the reasons for Oregon's hunger, the state's high ranking has had its political and social effects: Everyday Oregonians have rallied to the cause, legislators are pushing anti-hunger legislation, and the governor has made hunger one of his top social priorities.

The news media picked up on the issue. Among others, The Oregonian has published news stories ranking Oregon "No. 1 in terms of the portion of the population at risk of hunger" and has editorialized about Oregon's "highest-in-the-country hunger rate."

The ranking also triggered a successful drive to sign up large numbers of Oregonians eligible for the federal food-stamp program, giving Oregon the nation's highest percentage increase in food-stamp enrollment from 1999 to 2001.

Charles maintains that anti-hunger groups want to preserve what he calls the "mythology" of Oregon as the hungriest state, because that label draws headlines and support.

"In the advocacy business," he says, "good news is bad news, and bad news is good news."

Patti Whitney-Wise, executive director of the Oregon Hunger Relief Task Force, disagrees about the use of the ranking.

"Well, it is ranked No. 1 in the studies that have come out," she says. "We don't try to flaunt it, but it certainly gets people's attention. . . . What we're trying to do is educate people that hunger is a problem in this state."

Others say the ranking doesn't really matter.

Pam Curtis, who served as human services and children's adviser in the Kitzhaber administration, recalls that agencies, community groups and anti-hunger advocacy groups looked long and hard at whether Oregon was really No. 1.

"You could spend years trying to pick that question apart," Curtis says. "What we did was say, 'You know, it doesn't matter. What matters is we have too many hungry people in this state.' "

Steve Woodward: 503-294-5134; stevewoodward@news.oregonian.com