How many children are living in poverty?


Is a worm eating our brains: We hadn’t planned to conduct a daily discussion of the poverty rate.

But yesterday, it happened again! In the Washington Post, Michael Chandler reported a study about the disadvantages faced by kids from low-income homes.

Those disadvantages are very real, and they’re well worth studying. That’s why it’s maddening when journalists produce chaos like this:
CHANDLER (11/4/13): Nearly half of children 8 years old and younger are living in low-income households, according to the report. The poverty marker is defined as those with incomes below 200 percent of the 2012 federal poverty threshold, or $46,566 for a family of two adults and two children.

That includes 42 percent of children in the District, 34 percent of those in Maryland and 38 percent in Virginia.

Mississippi has the highest proportion of children 8 and younger considered low-income—63 percent.
Say what? According to Chandler, the study defined “the poverty marker” as “those with incomes below 200 percent of the 2012 federal poverty threshold.”

The poverty marker was defined as two times the poverty threshold? A lot of people simply stop reading when they meet chaos like that.

We decided to look at the study, which was conducted by the Annie E. Casey Foundation. Which groups of kids was the study comparing? We began to feel Chandler’s pain when we read this horrible explanation in the study’s Table 1:
NOTE Low-income households reflect those with incomes below 200 percent of the 2012 federal poverty threshold ($46,566 for a family of two adults and two children) and only include children who live in families for whom poverty status was determined, therefore it is not possible to replicate the low-income rate using the population estimate shown here as the denominator.
You’re right—that’s a “run-on sentence” or some such beast. Let’s tidy a bit, then proceed:
SAME NOTE, AFTER BEING TIDIED Low-income households reflect those with incomes below 200 percent of the 2012 federal poverty threshold ($46,566 for a family of two adults and two children) and only include children who live in families for whom poverty status was determined. Therefore, it is not possible to replicate the low-income rate using the population estimate shown here as the denominator.
We have no idea what that means. Tell the truth. Do you?

At first, that passage seems to say that the study concerns “low-income households,” defined as households with incomes below double the poverty rate.

That would be perfectly reasonable. We’d be comparing kids from those “low-income households” with kids from higher-income households—households earning more than double the poverty rate.

But alas! The passage then says that low-income households “only include children who live in families for whom poverty status was determined.” We have no idea what that means. And the last sentence in that NOTE seems wholly indecipherable.

This study examines important questions. What kinds of disadvantages face kids who come from low-income homes?

The study presents important information. All in all, it seems fairly clear that the study divided kids into two groups: 1) Kids from homes earning less than two times the poverty rate, and 2) kids from homes earning more than two times the poverty rate.

That makes perfect sense. But by the time the material reached the Post, the presentation was incoherent.

Here’s a dirty little secret: A lot of voters don’t believe official poverty figures. Their skepticism isn’t wholly unwarranted. The measure of poverty is a complex conceptual matter.

It doesn’t help when the liberal world and the mainstream press can’t even speak coherently about this topic. In just the last week, this is the third time we’ve encountered work in major publications in which the poverty rate 1) gets absurdly inflated or 2) is rendered incoherent.

For our first post, just click here. For our second such post, just click this.

Is a worm eating American brains? Is doesn’t seem hard to explain who that study actually studied, but the Washington Post couldn’t do it. In fairness, we began to feel Chandler’s pain after reading that chunk from the study.

Like Education Week and the New York Times, the Post wasn’t up to the simplest tasks involved in discussing low income and poverty. As with the Education Week mess, the problem began with bungled work inside a major study.

That study discusses important topics. Is a worm eating our brains?


  1. This study, like many others, wrongly assumes that correlation = causation. This bad reasoning has done enormous damage to disadvantaged American children. It justified policies that involve the government giving poor people money and other resources. That's a nice thing to do, but these policies may not have been good ways to improve the education of poor students.

    Having been raised in an immigrant family, I didn’t have some advantages possessed by wealthier, established families. However, my family came from a culture that valued education. I had a working father and a stay-at-home mother, who encouraged education and helped with coaching and tutoring, in the earlier years.

    We've seen wave after wave of poor immigrants come to this country and achieve high levels of success in 2 or 3 generations. Their culture was more important than their low income.

    Statistics show that poverty and bad education and marital status are highly correlated. Here's a hypothesis: Perhaps both bad education and poverty have the same causes, one of which is unmarried parenthood. If this is the case, the government ought to be encouraging people to marry -- especially poor people. Instead, the government has instituted many programs that have the unintended consequence of discouraging marriage.

    1. Perhaps bad education and poverty are caused by "unmarried parenthood"? I swear I don't know whether to laugh or cry, Mr. Correlation-Does-Not-Equal-Causation.

      So you think the government should be encouraging people to marry? You got your wish back in 2006, when the WPE, some of the brain-dead in Congress, and various religionists got together to take about $100M from welfare (TANF) to teach unmarried people to get married and to teach married people how to improve their marriages. The result? People who got the "marriage good, divorce and single parenthood bad" pitch were slightly less likely to get or remain married than the control groups, whose members weren't subjected to the propaganda.

      Ya know what did work? In Minnesota from 1994-1996, people who went from welfare to work got to keep some of their welfare money instead of being cut off. Divorce rates dropped significantly.

      Probably just correlation without causation, though. It was probably the cold winters in Minnesota that made the difference.

  2. Based on the wording Somerby presented, it sounds like it is saying that the figures shown in the table are for those families with children for which povery level could be determined (e.g., for whom income level could be discovered, verified, found out). Because they could not determine poverty level for all families in the sample, they could not figure out what the overall poverty level was (replicate the poverty level) for their group. I don't know anything about the study and thus cannot know why they couldn't verify income for all families, but sometimes it is because people don't want to say what they earn. This is opaque, but they study is presumably not written for the general public but for other researchers, who would be more likely understand such a note. It illustrates why journalists need some training in order to write coherently about topics drawing from the research literature.

    David, your comment about the effects of unmarried parenthood is one reason why liberals tend to be pro-choice. Marrying doesn't cure poverty, especially if young people drop out of school to do so. Programs that encourage unwed teens to finish high school have the opposite effect, whether they discourage marriage or not. Also, remember that correlation is not causation -- you cannot decide that unmarried parenthood causes poverty, just because the two tend to be correlated with each other demographically. I would be willing to believe that bad sex education causes unmarried parenthood, however.

  3. Lindy -- I'm pro-choice, too. However, since Roe v. Wade was first decided, the percentage of 1-parent families has exploded. So, other factors have more than offset the availability of legal abortions. This article gives some statistics

    deadrat -- I'm responding to your poverty-related point from an earlier post. One way my wife and I economized on food in 1966-68 was to use only powdered, non-fat milk. I don't recall the price, but a box of powdered milk was very cheap. My daughter didn't know what regular milk was until years later, when she got some at a friend's house. After that, she refused to drink powdered milk.

    1. Smart child! Powdered milk is vile. It was necessary during WWII because of lack of refrigeration. Two things that affected the divorce rate have been no-fault divorce and women entering the workforce. When women were dependent on men for financial support they were less likely to leave an unhappy marriage, especially if they had kids. There are lots of studies showing that unhappy marriages where parents stay together only because of the kids have a worse impact on children than non-poor single parent homes. The problem is that women who divorce tend to have very little money and that isn't good for anyone. In Scandinavian countries, single moms receive a stipend for their kids so marital status doesn't have such a big impact on their kids' chances in life. The welfare state works a lot better when we are not creating greater demands on it by disadvantaging people early in their lives, handicapping their coping later on. Funny how that doesn't seem to create a cycle of dependency in those countries.

  4. Mandatory abortions for pregnant teenagers would put a huge dent in poverty.

    1. Would it be as big a dent as the one left in your forehead when your mama dropped you on it?

  5. OMB (Dirty Secrets Revealed or Just a Muddled Mind?)

    "Here’s a dirty little secret: A lot of voters don’t believe official poverty figures."

    If BOB thinks he is revealing a secret, then he thinks everyone's head is as far up their behinds as his is. We also know, by reading this post, BOB needs to be concerned about worms and brains. His own.

    This is the third attack BOB's made on coverage of studies on low income children in just a few days.

    In the first BOB correctly blamed an Education Week writer for incorrectly using the word "poverty" to describe children in a study which used eligibility for the free and reduced lunch program to estimate the growing number of low income children in schools. But BOB went on to attack the study the report was based on and its author. That proved he didn't read the study, because they made no such definitional error. This was pointed out in the comments to that post, but he's repeated the same lie twice since.

    Next he accused an op-ed writer in the New York Times of inflating the poverty rate. To do this he had to drop sentences from the op-ed piece that indicated the author was doing nothing of the kind. This again was called to his attention in comments. He repeats the lie in this post.

    In this post the only problem with the Washington Post article is the reporter's selection of the word "poverty" as a substitute for the word "low- income." BOB turns it into "chaos" for the mythical reader.

    "A lot of people simply stop reading when they meet chaos like that."

    Of course in painting the picture of chaos he neglects to tell his readers what the WaPo readers turned off by the "chaos" who "simply stopped reading" were deprived of by this horrible piece of journalistic chaos. They missed the last two sentence in the article. One decribed the low income rate in the DC area. The other decribed the rate in Mississippi.

    He plows into the report from the Casey Foundation itself, and finds a footnote: two lines in tiny type at the bottom of a one page chart. It is literally the only thing from the report he states "presents important information" he quotes. He blows it up, attack its syntax, and make the proverbial "federal case" out of it, even though he also concedes the report's division of lower and higher income groups "seems fairly clear" and "makes perfect sense."

    BOB asks " We have no idea what that means. Tell the truth. Do you?"

    We do, BOB. We do. Because we saw the chart and the footnote above it, neither of which you mention. We know the chart is based in part on the US Census American Community Survey, where income is not reported by all respondents, so some families with children cannot be coded. And we know there is a column with the poverty rate which will not produce the same figures as the column with total population because of the explanation. The note has nothing to do with the accuracy of the report or the reporter's poor choice of "povery marker" instead of "low-income households." But that doesn't fit BOB's narrative of "liberalworld" collapse in a puddle of worm eaten brain death.

    BOB, most of us already knew lots of voters don't believe poverty figures. Many don't believe in other things most accept as fact. But the the choice of words in an article you otherwise find no fault with, or a footnote in a study, even if changed to your satsisfaction are not going to change their disbelief. Their disbelief is caused by something which may very well deter them from ever bothering to read an article like the one in the WaPo in the first place, much less the study, its charts, or a fine print footnote.

    BOB, you've spent over a month attacking a woman who wrote a book as an unqualified scam artist. But at best you have proven she manipulates data to further her tale. You lie to do so.


  6. Poverty line: 100% of the poverty line. The reason why commenters have been posting lengthy posts is that they can't keep it simple while at the same time pretending the poverty mark or line is above 100%. Journalists have to accurately report what they report. They cannot make up their own poverty mark. Is it 200% now? Why not 300%? Why not 250%?