(Guest Post by Matthew Ladner)
Ryan McMaken notes that international childhood poverty measures use national median incomes as their base. Unicef defines you as poor if you fall below 60% of the median income. What goes unmentioned however is that median incomes vary- a lot.
Having your Rawlsian druthers, you rather not be born a poor child in a country with a relatively low median income-say Mexico or Turkey. If you are going to be born a poor child- much better to be born in a country with a high median income like the United States- or tiny little international banking havens like Luxembourg or Switzerland (ancestral homeland of the Ladner clan btw-we were exiled to the new world when caught smuggling salt into France) or a tiny Nordic country sitting on a whole bunch of oil (Norway).
So the above chart plots Unicef childhood poverty rates across the y-axis but family incomes for those living in relative poverty cross the x-axis. As you can see, the American childhood poverty rate is relatively high, but so to is the American childhood poverty income. In fact a poor child in America has an income four to five times greater than a poor child in Mexico.
It is not clear to me whether the American number includes transfer income or not. President Obama very helpfully noted that if you include transfer income (many poverty measures don’t) then poverty has declined substantially in America. If these figures don’t include transfer income than matters would be even more lopsided in favor of the United States vis-a-vis say a Mexico. If it does, it is still lopsided.
So, now, someone please explain to me again why American Hispanic and African-American kids score so close to the median PISA score for Mexico. Oh yes, there is poverty in the United States, and students of color suffer from it in a disproportionate fashion. Nevertheless, the incidence of childhood poverty in Mexico is far more severe than in the United States. Moreover, public schools in America enjoy lavish levels of per-student funding when compared to their counterparts in Mexico.
Put another way- how is that schools in Mexico have such a greater bang for the buck in overcoming student poverty when compared to urban schools in the United States? Perhaps we should redirect our international education crowd from Finland to Mexico.
Jay P. Greene's Blog 
look here:
http://www.oecd-ilibrary.org/sites/factbook-2010-en/11/02/02/index.html?itemId=/content/chapter/factbook-2010-89-en
Using 50% of median for Poverty Line (not 60% as above) and including Gov. transfers USA has a poverty rate around 18%
Significantly less than the 30%+ (as above) … so I would guess the chart you posted does not include Gov. transfers.
I think the high rate of low performance among groups of low income USA children has a lot to do with:
1 .. culture
2 .. too much following of invalid Ed Philosophical beliefs rather than (doing what works) using data to improve instruction for educationally disadvantages learners.
As a confirmation that perhaps culture is a factor … look at American Indian test performance gap in AZ between students in Low Density schools (less than 25% American Indian enrollment) and the much lower performance of those in High Density Schools (more than 25% American Indian enrollment)
In WA state the gap is even larger when Low Income American Indian students state wide average is compared with performance with low income on the Rez.