"Continuous Improvement" Doesn't Work

This from Richard Rothstein in the National Journal:

[M]any supporters of No Child Left Behind are complaining that Senator Harkin’s proposal weakens the NCLB benchmarks by substituting “continuous improvement” for “adequate yearly progress” towards a fixed proficiency goal.

But as I wrote in my Economic Policy Institute blog last week, there is less here than meets the eye. “Continuous improvement” continues the poorly conceived NCLB approach, seeking an unrealizable national whip that will somehow transform American schools for the better. The effort ignores both evidence and common sense.

The proposal developed by Senate Democrats will relieve states of having to meet federally specified achievement goals in math and reading. Instead of requiring all students to be “proficient” in these basic skills by 2014 (as NCLB demands), or to be “college ready” by 2020 (as the Obama Administration proposes), the Harkin bill will require only that schools show "continuous improvement" for all students, and for students from low-income families, those who don’t speak English, minority students, and students with disabilities (see page 52 of the draft bill).

According to a report in Education Week, “state and local officials likely will be exchanging high-fives, since that would give them much of the flexibility they're looking for.”

They are in for a shock. “Continuous improvement” is no more reasonable or achievable than “proficiency for all,” or universal college readiness.

Of course, citizens should expect every public school to strive for its peak level of performance, but some schools have much farther to go to reach this level than others. Unlike present policy, a well-designed accountability system could judge how far each school can and should go, and whether it is on the right track to get there for the several populations it may serve. In each case, this is a difficult judgment to make, and a slogan is no substitute. In this regard, a single one-size-fits-all metric such as “continuous improvement” is no better than “proficiency for all.” The Broader, Bolder, Approach to Education campaign has described the outlines of a more reasonable accountability system, and a book, Grading Education, goes into more detail.

NCLB’s attempt to require all students to be proficient at a challenging level led to the absurd result that nearly every school in the nation was on a path to be deemed failing by the 2014 deadline. The demand ignored an obvious reality of human nature - there is a distribution of ability among children regardless of background, and no single standard can be challenging for children at all points in that distribution.

Expecting all children to be college-ready suffers from the same problem, and more. In a nation where 32% of all young adults now earn bachelor’s degrees, and where the Bureau of Labor Statistics projects that only 30% of job openings by 2018, even in a healthy economy, would require a bachelor’s degree or more, the notion that 100% of students would be able to succeed in an academic college by 2020 is even more fanciful.

So why not “continuous improvement” instead? It’s a nice slogan, borrowed from a management fad promoted by W. Edwards Deming and others who thought this was the key to Japanese auto manufacturing success. But while consistent attention to small improvements makes sense as a management tool, no company has ever continuously improved, overall, indefinitely. There are spurts of improvement, and plateaus, and then the most successful companies fade, to be overtaken by others. No management expert would recommend that firms be dismantled if they are consistently profitable, but just not more profitable year after year after year.

But continuous improvement will now, if Senate Democrats have their way, be the trajectory for every school in the country, by law.

Today, some students and schools already perform at their peak capacities. Many high-performing students and schools should be praised for maintaining high levels of performance, not condemned for not improving further. Imagine a parent with two children of high natural ability, one of whom is an “A” and the other a “C” student – should the parent have equal expectations of both for improvement? I’m certain that Sen. Harkin and his staff have not made a study of schools and students nationwide to determine that there nowhere exist schools for whom maintenance of quality, rather than improvement, should be our expectation. If there are such schools, they will be the first to be labeled failing under a continuous improvement standard.

More importantly, many “C”-performing schools and students are now achieving at their peak capacities, in view of the external constraints over which public education has no control. Many disadvantaged students, and schools serving them, are low-performing not because of inadequate school efforts, but because children come to school unable to focus because they are hungry or suffer from untreated minor illnesses such as asthma or tooth decay, or with inadequate early childhood literacy preparation because their parents are poorly educated, or with interrupted instruction because of homelessness or displacement from housing instability, or stressed because of parental unemployment shocks, home foreclosure, or neighborhood crime.

It is easy for Senators to issue fiats that schools serving such children should “continuously improve” their math and reading test scores, but many of these schools are already performing gargantuan feats, given their social and economic contexts. Indeed, many (not all) schools with low test scores serving disadvantaged children are having more positive impacts on their students than typical “high-performing” schools serving affluent children who would test well even with the poorest instruction. A Congressional proposal that both types of schools should “continuously improve” is based on no empirical examination of what is possible in different types of schools, and how what is possible might vary.

A new Russell Sage Foundation book edited by economists Greg Duncan and Richard Murnane, Whither Opportunity? Rising Inequality, Schools and Children's Life Chances, notes that growing income inequality in America has been accompanied by a widening gap between the achievement of students from rich and poor families. In a Chicago Tribune op-ed, Duncan and Murnane explain why they (and one of their co-authors, Stanford economist Sean Reardon) think this has occurred:

Growing economic inequality contributes in a multitude of ways to a widening gulf between the educational outcomes of rich and poor children. In the early 1970s, the gap between what parents in the top and bottom quintiles spent on enrichment activities such as music lessons, travel and summer camps was approximately $2,700 per year (in 2008 dollars). By 2005-2006, the difference had increased to $7,500. Between birth and age 6, children from high-income families spend an average of 1,300 more hours than children from low-income families in "novel" places — other than at home or school, or in the care of another parent or a day care facility. This matters, because when children are asked to read science and social studies texts in the upper elementary school grades, background knowledge is critical to comprehension and academic success.

With unemployment now stuck at unacceptably high levels, especially for the parents of minority children, the ranks of students without family-provided enrichment activities can only grow, and the isolation and concentration of such students in schools serving peers without such activities can also only grow. National legislative demands that schools show “continuous improvement,” without any amelioration of the economic hardship faced by parents, would be laughable were it not so tragic.

And then there are the reductions in resources available to schools, a result both of the faltering economy and Republican attacks in many states on public education. Last month alone, 24,000 jobs were eliminated from public schools. Heidi Shierholz of the Economic Policy Institute has calculated that since the recession began in December 2007, nearly 278,000 public school jobs have been eliminated; another 48,000 necessary to keep up with growing enrollment were not created, for a total loss of educators and other public school workers of 326,000. Such reductions will continue until the economy stabilizes.

Some who are hostile to public education cheer this development, seeing it as the shedding of inefficiency. But I’d guess that the Senate Democrats who are drafting the ESEA re-authorization are not among them. Sen. Harkin and his colleagues probably think that counselors, reading specialists, librarians, classroom aides, parent coordinators and teachers whose jobs are being cut make a positive contribution to student achievement. Yet if so, how can Senate Democrats demand that student achievement show “continuous improvement” when these Senators can do nothing to prevent the resources on which that improvement depends from being eliminated?

Some schools, of course, should improve, and dramatically so. Some schools serving low-income students and some schools serving affluent students operate at far below their potential. But those that are doing as much as we should expect cannot be distinguished from those that are doing far less, simply by looking at student test scores or how those scores grow or don’t grow over time. Making such distinctions requires holistic evaluation of school curriculum, leadership, and teacher quality that is expensive, requires trained experts, and is nested in broader community social and economic contexts. States now have no resources or capacity to implement such accountability systems, and proposals for a new Elementary and Secondary Education Act are noticeably silent when it comes to addressing this shortcoming.

If ESEA is re-authorized with a simple requirement that all schools and students show “continuous improvement,” we’ll be back for the next re-authorization in five years with new demands for wholesale waivers from Congressional expectations that had no basis in reality.