Wednesday, April 22, 2009

McKinsey Goes Skin Deep

This McKinsey report The Economic Impact of the Achievement Gap in America's Schools is odious in the way it shifts between comparing US schools to foreign schools, and comparing the achievement gaps within the US, without comparing the amount of inequality in the US to that in high-achieving countries, or looking at measures of child welfare in the US vs. high-achieving countries. If you point out that the amount of variance in achievement based on income in Iceland, Finland and Canada is much less than that in the US, without pointing out that the reason for that is that there is less variation in income, you are peddling propaganda.

Of course, these comparisons are difficult to make, because there isn't another advanced country that does not provide health care to all its children, so we don't know if they do a better job of education sick kids than we do. You can't check to see if another country does a better job of educating a population that was brought to the continent in bondage, enslaved until 150 years ago and not granted full voting rights until 45 years ago, because we're the only ones who have done that. You can't see if other countries have higher achievement with similar incarceration rates for low-income populations and minorities because nobody else matches us. You can't compare how we do with other countries that rely primarily on local property taxes to fund education because there aren't any.

On the other hand, the Czech Republic beats us on PISA scores, but as of 1995 "In the Czech Republic, 46.4% (other estimates place this at around 75%) of Roma children are in special schools for the mentally handicapped" and "The German education system fails to provide adequate language training for children who speak non-native mother languages and shows a strong tendency to reproduce social inequality." These comparisons are much more complex than McKinsey would lead you to believe.

I could go on and on of course. You can say:

By comparing several neighboring-state pairs with similar demographics, we can see how dramatic this disconnect can be between overall achievement and the racial gap. New Hampshire and Connecticut, for example, have similar overall fourth grade reading scores; yet the gap between white and black scores in Connecticut is more than twice what it is in New Hampshire.... State variations in the racial achievement gap cannot be explained by the proportion of blacks and Latinos in a state’s educational system, furthermore, although school-level segregation may play a role in influencing outcomes.

Well... yes, but what are you supposed to do with that. Connecticut is an extremely inequitable and segregated state (I lived there for a while):

Between 1991 and 2002, the increase in the average annual income of Connecticut’s wealthiest 20% of families ($35,093, i.e., from $109,867 to $144,960) was nearly double the average total annual income of Connecticut’s poorest 20% of families ($21,003). Indeed, the growth in the gap in income between the top 20% of Connecticut families and the bottom 20% of families (as measured by the change in top-to-bottom ratio) was greater in Connecticut than in every state except Tennessee.

Even Connecticut’s middle income families (those in the middle 20% of families) have experienced income growth that pales beside the growth experienced by the top quintile. Over the 1980s, middle-income families saw their incomes increase by 15%, compared to an increase of 44% for very high-income families. By comparison, between 1991 and 2002, middle-income families again saw their incomes increase by 15% while the incomes of very high income families increased by more than twice this amount (a 32% increase).

And that's before the Connecticut hedge funds started their pillage of the past seven years. So you can scratch your head and imply that Connecticut's schools are to blame, but presumably McKinsey are paid to do deeper thinking than this.

Last point -- the premise here is that the achievement gap(s) cost us money -- which I'm sure they do. But if the effect was as strong as this report argues, wouldn't there be a bigger economic upside to individual states that seem to be doing better on closing the gap in test scores? Why are DC and Connecticut at the top in per capita income, with Florida and Texas sitting at 18 and 32?

Ultimately, the bottom line here is that McKinsey sees educational attainment, or lack thereof, as simply the cause of social problems -- lack of education causes health problems. The real world is considerably more complex.


Bill Kerr said...

Since one important function of the education system is social sorting along class lines then the need for reform would be less noticed if the overall average put USA higher up the ladder

eg. in Australia we rank much higher on PISA scores (top 10) but also have a long tail of underachievement

I've never quite understand what the real purpose of NCLB is ...maybe just a way to control teachers more

john thompson said...

Great post.

Tom, do you think McKinsey will offer you a job so they can clean up their act?

What's the point of McKinsey, that we should spend more on education or that we should spend more on education policies of Texas, NYC, and Joel Klein et. al?

When Districts outside NYC do not close gaps as kids advance in grades, McKinsey condemns them. However, when their NYC data shows that 3rd graders in the lowest quartile have a 13% chance of moving into the top 1/2, they say it shows that low performing 3rd graders have "varied" outcomes as they advance.

And where does that data appear? Not in the section on student performance, but in the area that discusses the social costs of educational failure.

The McKinsey report looks like a brief for Klein, albeit with amazing production values.