Tuesday, January 20, 2015

Lying with Statistics: Education Division

When journalists grab information from biased parties, we get bad reporting. Case in point, a Washington Post article, "Majority of U.S. public school students are in poverty" which apparently unquestioningly took "allegations" as true without looking further. WAPO reported
The Southern Education Foundation reports that 51 percent of students in pre-kindergarten through 12th grade in the 2012-2013 school year were eligible for the federal program that provides free and reduced-price lunches. The lunch program is a rough proxy for poverty, but the explosion in the number of needy children in the nation’s public classrooms is a recent phenomenon that has been gaining attention among educators, public officials and researchers.
If you go visit The Southern Education Foundation's website ("Advancing Creative Solutions to Assure Fairness and Excellence in Education") you can find a link to the cited report, or a least a report on the report, "A New Majority Research Bulletin: Low Income Students Now a Majority in the Nation's Public Schools":
In 40 of the 50 states, low income students comprised no less than 40 percent of all public schoolchildren. In 21 states, children eligible for free or reduced-price lunches were a majority of the students in 2013.

Most of the states with a majority of low income students are found in the South and the West. Thirteen of the 21 states with a majority of low income students in 2013 were located in the South, and six of the other 21 states were in the West.
At this point, you are invited to download the "research bulletin" itself. You really need not bother, the "bulletin" merely states what the WAPO reported,
For the first time in recent history, a majority of the schoolchildren attending the nation’s public schools come from low income families. The latest data collected from the states by the National Center for
Education Statistics (NCES), evidence that 51 percent of the students across the nation’s public schools were low income in 2013.

The pattern was spread across the nation. Half or more of the public schoolchildren in 21 states were eligible to receive free or reduced-price lunches, a benefit available only to families living in poverty or near-poverty in 2013.1

In 19 other states, low income students constituted between 40 percent and 49 percent of the states’ public school enrollment. In other words, very high proportions of low income students were evident in four-fifths of the 50 states in 2013 ....

What the WAPO article and the Southern Education Foundation website fail to report is what the criteria are for qualifying for "free or reduced-price lunches" are. Instead in the Bulletin we are left with a footnote to "a benefit available only to families living in poverty or near poverty," which reads, in pertinent part:
***Students are eligible for free meals at public schools if they live in households where the income is no more than 135 percent of the poverty threshold. They are eligible for reduced-price lunches if their household income is no more than 185 percent. In 2013, for example, a student in a household with a single parent with an annual income of less than $19,669 was eligible for a free lunch or less than $27,991 for a reduced-price meal in a public school.
If you go to the Eligibility Manual for School Meals, you will find a USDA document of 108 pages in which the basic statement of "income eligibility" above is confirmed (p14). There is also this:
United States citizenship or immigration status is not a condition of eligibility for free and reduced price benefits.(p31)
Here are the U.S. Department of Agriculture income guidelines:

Let's take a family of 4, living in one of the 48 continental states. If the family has income of $23,850 or less, they are in "poverty." Income is defined as:
‘‘income,’’ as the term is used in this Notice, means income before any deductions such as income taxes, Social Security taxes, insurance premiums, charitable contributions and bonds. It includes the following: (1) Monetary compensation for services, including wages, salary, commissions or fees; (2) net income from nonfarm self-employment; (3) net income from farm self-employment; (4) Social Security; (5) dividends or interest on savings or bonds or income from estates or trusts; (6) net rental income; (7) public assistance or welfare payments; (8) unemployment compensation; (9) government civilian employee or military retirement, or pensions or veterans payments; (10) private pensions or annuities; (11) alimony or child support payments; (12) regular contributions from persons not living in the household; (13) net royalties; and (14) other cash income. Other cash income would include cash amounts received or withdrawn from any source including savings, investments, trust accounts and other resources that would be available to pay the price of a child’s meal.
‘‘Income,’’ as the term is used in this Notice, does not include any income or benefits received under any Federal programs that are excluded from consideration as income by any statutory prohibition. Furthermore, the value of meals or milk to children shall not be considered as income to their households for other benefit programs in accordance with the prohibitions in section 12(e) of the Richard B. Russell National School Lunch Act and section 11(b) of the Child Nutrition Act of 1966 (42 U.S.C. 1760(e) and 1780(b)).
You might note that not only are some federal benefits excluded from "income," at least one "anti-hunger" program, Supplemental Nutrition Assistance Program (SNAP) by which poor families may qualify for an "allotment" (read "cash payment") is a virtually "automatic qualifier for free or reduced cost school meals.

Let's take our family of 4 earning $23,850. If you follow the SNAP guidelines as set out here, our family could receive an amount of say $300 per month in SNAP benefits, which can be used to buy:
Households CAN use SNAP benefits to buy:

Foods for the household to eat, such as:
- breads and cereals;
- fruits and vegetables;
- meats, fish and poultry; and
- dairy products.
Seeds and plants which produce food for the household to eat.
You might note that this $300 per month, added to the family income, would, if not excluded by law, boost the family out of the "poverty" category. So would any "credit" given for reduced rent in public housing. So would any medical care covered by Medicaid. In any event, free school meals would seem to be a double benefit to the families receiving SNAP benefits.

Let's look at a family of 4 which has income of $44,123 per year. As you can see from the Eligibility Guide, the child of this family are eligible for "reduced" price meals.

What follows requires a knowledge of what a "median" is and what a "mean" is. For statistical purposes, here's a good lesson from Math is Fun. In a list of numbers, a median, then is the number above half of the remaining numbers are above it and half below. It is not an average. A "mean" on the other hand, is an average, as also set out in by Math is Fun:
The mean is the average of the numbers: a calculated "central" value of a set of numbers.

To calculate: Just add up all the numbers, then divide by how many numbers there are.
In his classic book, How to Lie with Statistics, Darell Huff covers how the differences between these terms can be used to confuse,

Let's have another government publication, the Census Bureau's Income and Poverty in the United States: 2013: Income and Poverty in the United States: 2013
What do we gather from this report? In 2013, the median income of American families was $51,939.

Which means what? That there were as many families with incomes above that number and there were families with incomes below it. Note that it wouldn't matter if every family above that number only had income of $1 more than the family before it. It only matters that of the 122 million households, there were 61 million above that number and 61 million below it. There are further break downs:
Households in the lowest quintile had incomes of $20,900 or less in 2013. Households in the second quintile had incomes between $20,901 and
$40,187, those in the third quintile had incomes between $40,188 and $65,501, and those in the fourth quintile had incomes between $65,502 and $105,910. Households in the highest quintile had incomes of $105,911 or more. The top 5 percent had incomes of $196,001 or more.
A quintile is one of five equal groups. So, each quintile referred to had roughly 24.4 million households in it.

It is also noted that families living in rural areas and outside of major metropolitan areas (i.e. much of the South and West) have lower incomes than people living in the major Northeastern metro areas. It is not noted that most people in the South and West pay substantially lower state taxes than do people in those metro areas. See page 8 of the Census report.

Well, okay, but what does the average American earn? According to the Social Security folks:
The average amounts of wages calculated directly from our data were $42,498.21 and $43,041.39 for 2012 and 2013, respectively.
Averages are deceiving, of course. If I work at a company where the average wage is $85,000, I must be doing great. Of course, if I earn $20,000 and my boss earns $150,000, the average ($20k+150k/2) is exactly $85,000.

Then there is this:
In 2013, there were 45.3 million people in poverty.
People, in this case, not households as we have been discussing. You should be aware this number can be skewed by young people working at jobs which pay little, because,
The medians for people are based on people 15 years old and over with income.

In Appendix B we get to "family" poverty. According to Table B-1, among all "families" 12.4% are in poverty. Note that this is different than being eligible for free or reduced fee school meals. In fact, it seems that the highest percentage of poverty exists in "Families with female householder, no husband present" where in 2013 some 33% of such households were in poverty. You can go examine the charts. They can be eye-opening.

It is also worth noting that "poverty" in the U.S. has little of a "static" component. From early in the report:

There is relatively little chronic poverty (oval) and people move up and down depending. For example, a high income earning family may drop down a quintile , say when retirement is reached and earned income decreases. A low income student family may move up when education is completed and a well-paying job is taken.

Enough. Let's go back to the free/reduced meal USDA chart and the 185% of the poverty eligibility for reduced cost meals.

Remember the "average" wage (see above) for 2013 was $43,041.39. No wonder the Southern Education Foundation is able to present such a skewed version of what has happened. It isn't just that we have that many more poor people, it is also driven by a very broad level of eligibility for this particular program. 

The WAPO should have done some work on this topic and examined those numbers and provided its readers with additional information necessary to determine how great the alleged problem is, instead of relying on numbers provided by a single source.

1 comment:

  1. ColoComment5:36 PM

    Another example of the truth of the statistics and damned lies quote. It's not only the measurement of income and "poverty" that's skewed, but also how they count students "eligible" for free lunch. It's too bad when your "study" is based on such a shaky foundation that it's blatantly misleading.

    From Breitbart:

    "The Southern Education Foundation study used the number of students eligible for free or reduced-price lunches as a proxy for actual poverty statistics. By this metric, the group concluded that 51 percent of children in 2013 now qualify for subsidized school food. The problem: Recent changes by the Department of Agriculture have instituted a policy known as the “Community Eligibility Provision” that allows schools that “predominantly serve low-income children” to give all students free meals, regardless of whether they are eligible for free meals."