Ohio has just started reporting value-added indicators. Schools and districts get an averaged rating of expected, below expected or above expected growth for the year. I don't know how much I trust the averaged rank, but the same ranks are given by grade/content, which a more helpful way of looking.

But I believe that these support Downey's assertion that actual "growth" is much more equitably spread across income levels than is achievement. And I suppose this makes some folks feel better (and others feel worse). And it is certainly helpful to look at where in the 3-8 grade span kids are getting lost and where they are picking up speed.

But NCLB really poses a different challenge than seeing that every kid gets the same amount of growth. It suggests that there is a minimum amount of knowledge or ability that we can shoot for providing to every kid. This doesn't say that some cannot get more (as we all know that they do). But we really have to focus PUBLIC education on ensuring that we have a public that leaves school at a certain level (or above). This implies that certain schools will need more resources to achieve this minimum level (if their environment isn't providing as much as some others). We seem to stall at this point--and all the arm-waving from teachers who feel blamed isn't helping us to look at a redistribution of resources.

I think that the Japanese get the point about ensuring a level of achievement for all. They don't seem to be troubled by "teaching to the middle," in order to assure the highest common good. This is a value that they hold. I don't know that we do. They rely on the joku to provide either tutoring or enrichment, and seem to feel that this is appropriate.

Measuring the impact of a school doesn't move me, if the impact is still insufficient to meet a minimum grade-level of achievement for all students. What this says to me is that some students need more. Why are we so loath to see that they get it?

Hi Margo,

I wholeheartedly agree with the social goals you set out around ensuring a basic level of achievement, and ensuring that kids, irrespective of their background, get there.

But we need to separate the issue of social goals from the issue of measurement, especially if our primary objective is achieving the societal goals we both agree upon. If a school is making exemplary progress based on a growth or impact measure but is not making proficiency targets, we don't serve the students well by sanctioning the school and its teachers.

Why not measure impact or growth to establish whether the school is performing worse than it should be first *and* then provide more support and additional interventions to students/schools if the school is not "low-performing" according to these more sophisticated ways of isolating school performance?

OK--so we measure impact or growth first and what do we do with the results? If the goal is the get to point A, and we suspect that we've got kids starting from B, C, D and F, what good does it do to know how fast they are moving?

But, I think there is something else that is crucial about Downey's methodology, which is that it doubles the amount of testing--with the only value-add that we can separate out the school contribution from the summer contribution (or loss). Which, again, seems to be a red herring, unless we are going to be able to tackle the environment (or the summer).

I get really nervous when people start talking about the "sanctioning" of schools. The "sanctions" in NCLB all relate to making improvements, beginning with enlarging the options available to students and providing tutoring, and moving forward with planning and implementing changes. While the press likes to focus on the drama of reconstituting personnel, this is one possibility mentioned along with improving curriculum, and extending the school day and year. I hear very little about districts who have gone this direction. In fact my own district has gone in the opposite direction (shortened school day--across the board. No one even mentioned the possibility of protecting any of the schools that were already doing poorly).

The other day, skoolboy mentioned an Ohio study that looked at funding disparities within districts. Just for fun, if someone has time, it would be interesting to see whether those disparities relate at all to the growth disparities.

Margo/Mom,

The study by Condron & Roscigno of Columbus, OH does in fact show that within-district disparities in expenditures are associated with school-level achievement when controlling for prior achievement, which is essentially a growth model. But I have to caution that the effects of spending differences on achievement growth generally are neither large nor consistent, for a number of reasons: the unit of analysis is most commonly the district, which means inequalities among schools within districts are not taken into account; spending is a coarse proxy for exposure to resources that might matter for classroom teaching and learning; and how money is spent may be as important, and perhaps more important, than how much money is spent.

To respond to Margo/mom's earlier point about why don't we see that kids get what they need to end at the same minimum level, even if they are starting at positions of varying advantage of disadvantage?

It's a good question. Part of the problem may be that truly leveling the playing field may need to take place before a child hits kindergarten, with high-quality preK and very early childhood programs that include parent education. For an example of what might be possible, I suggest reading Paul Tough's new book, Whatever It Takes.

I am interested in collecting research testing conventional wisdom such as the childhood obesity study you cite. To the public it just seems obvious that schools with their vending machines, their terrible lunches, the reduction of PE and recess encourage obesity. Policy decisions are often made on the basis of faulty conventional wisdom.

I would like to make a collection of other such unexamined assumptions related to education and the pertinent research, or note the lack of research.

As Margo pointed out, if we were to universally apply Doug Downey's approach, we'd have to double the amount of standardized testing - not a desirable goal.

Further, using the test scores as the measure runs the risk of strengthening the illusion that test scores can tell us a lot about learning, especially in an era of intense teaching to the test. This is certainly not Downey's fault - it is mostly what we have for usable large-scale data - but it ignores untested subjects and vital aspects of the tested subjects themselves.

Questions: Do schools serving more affluent students provide more of these unmeasured attributes, such as ability to solve problems, engage in synthesis or evaluation? Or are they distributed more evenly, as Downey finds with school-year gains on reading and math tests? Or even more common in low-income schools? Which attributes would be want to know more about? How would be find out about them?

Years ago U Wisconsin researchers such as Fred Newmann reported studies in Chicago finding that student acquisition of more complex knowledge and skills was distributed across Chicago schools - but rare in most schools. They concluded this by looking at classroom-based work, also finding that less didactic, more interactive, instruction was found in the classrooms with greater evidence of students acquiring the higher order skills. NCLB has certainly tended to reinforce the dominant highly didactic approaches. That may explain the declining rate of improvement on NAEP - but even NAEP is not an adequate measure of complex, higher-order thinking.

Downey's work is indeed useful, but we have to get beyond the test scores in research as well as educational practice.

maritza: I'll probably get around to reading Paul Tough's book eventually, since I am reading so much ABOUT it. But I have some cautions about over-reliance on early childhood programs to level the playing field. I believe that some of early longitudinal studies of Head Start showed that it was effective in delivering students to school at a point closer to their upper income peers. But they lost ground again by second grade. Sorry--I cannot cite specific studies, perhaps someone else recalls what I am remembering.

Margo,

It's true that many early childhood interventions, including Head Start, have had what are known as "fadeout effects," in which short-term gains quickly fade out over a few years. The conventional wisdom, as I understand it, is that (a) the quality and intensity of the early childhood intervention matter a great deal, and some programs are much better than others on these dimensions, and (b) the children who are the targets of early childhood interventions may need ongoing supports after the intervention is concluded and the children move into elementary schools which vary in quality and curricular alignment. The existing evidence on the potential of early childhood interventions, when coupled with ongoing support in elementary and high school, is what leads Nobel-laureate economist James Heckman to be such a staunch supporter of high-quality early childhood interventions.

Margo:
I was not sure what your comment about Japanese schools was trying to get at. Japanese schools definitely do not teach to the middle. They do ensure that every student receives the delivery of the same educational content. For a primer, on Japanese schools, testing and the role of the juku in Japanese education, I suggest my article in the Oct 1993 issue of The Kappan, entitled The Secret of Japanese Education.

Skoolboy--I don't disagree a bit. It takes both appropriate early childhood care/education and high quality and appropriate elementary and secondary education. To me the value of the research was that it pointed out that it is insufficient to look to head-start-like programs to level the playing field. We have some real equity and quality issues to address within elementary schools as well.