Tag Archives: testing

Stagnant Test Scores, Ten Years Later

In spite of a decade of high stakes testing at our free public schools, our international test scores are stagnant and comfortably mediocre.


One reason is that high stakes testing is a terrible idea.

Even if it wasn’t, we use the wrong type of tests.

Even if we didn’t, we don’t have free public education.

No Child Left Behind, in spite of hopes, has not yet worked.

The plus side is our international test scores show that white and Asian students do pretty good. White Americans do as well as the Swiss. In school systems were whites and Asians or middle- and upper- classes are a large majority, we basically can experiment with ways of introducing more STEM (science, technology, engineering, and mathematics) courses, through things like programmatic learning.

In URM (under-represented minority) and low-SES (socio-econoimc status) schools, we do really badly. There parents are very risk conscious, and basically only want baby-sitters.

But the good news is because of the “great sort,” there are less and less high-achievement-capable students in URM and low-SES schools each year (apart from immigrant communities). In URM and low-SES schools, we should castrate the ranks of teachers and administrators as quickly as possible, and find some way of putting those schools under the controls of large local employers, such as light manufacture or semi-skilled clerical work.

Instead of leaving “no child behind,” we should have a system of trade education to leave some children behind — but not far enough behind that they hurt our society or economy.

Don’t ignore the poor

Every three years, the Organization for Economic Co-Operation and Development (OECD) conducts the Programme for International Student Assessment (PISA) to evaluate student achievement in many countries.

Because US schools are terrible, US students do badly on the tests. As shown in the below chart, the US is not in the top ten in either math, science, and reading.

One reason for this is that the US education system is not designed for the US population. As shown in this blog posts, US scores are acceptble if you ignore the poor — that is, “correct” for poverty. Of course, reality does not work that way — you can’t “correct” a population by ignoring it. You serve a population by writing curriculum and hiring teachers that are effective in addressing a population’s needs.

By writing off districts with poor students, US education policy does not only harm poor students — it harms all students in schools that are stressed by the presence of students from poor and low-performing populations.

Solving education does not have to wait till we win the “war on poverty.” The poor and the middle class deserve education too, even if they can’t afford to mortgage their futures to get it.

High-Stakes Testing is a Mistake

Too often, the education reform debate is split between two sides

1. Teachers, whose primary interest is diverting school funds from student welfare and to themselves, and
2. Everyone else, who are flabbergasted by the terrible US public education system,

I’m simplifying of course — there are three dimensions of fource and multiple stakeholders, but the American public school teachers as a class have been breathtakingly unconcerned with the needs of others for several generations. Still, understanding that teachers have done nothing to align their interests with other stakeholders (aside from one political party) and act like an abusive monopoly is important to understand the education reform debate.

Unfortunately, these two sides then break down as disagreeing on the issue of testing:

1. Testing should not be used
2. High stakes testing should be used

Teachers oppose testing because they do not want to be accountable for not doing their jobs. This is understandable, but of course dangerous to our nation.

Many education reformers support high stakes testing, because it is easy to politically & logistically easy implement. Unfortunately, testing is invalid to the extent that testing conditions different from desired recall conditions, and if we’re training students to only ‘know’ something in high-stakes pen-and-paper environments, we’re doing them a disservice.

It would be better to fully integrate testing into the curriculum. A personalized device (let’s call it a Skinner Machine, or an iPad) would work with the student to help him understand concepts, show him appropriate & challenging material, and of course continually assess his learning. Low-stakes, continuous, real-world.

Low-stakes and continuous testing would be a form of good testing that would be reliable, standard, valid, and practical. We don’t have them because writing good tests is hard, and teachers are opposed to testing for economic reasons. So as the hard work is getting the testing infrastructure set-up in the first place, high-stakes testing is better than no-testing in the context of the terrible status quo. But good testing — low-stakes and continuous — must be the next step.


That’s how testing should be done.

The Hard Way

Earlier I analyzed the political economy of education reform, and argued that supporting Publishers was the easy way to put better tests in front of students. This is because a testing regime flatters the money-seekingnature of Publishers, so their greed can be coopted. Nonetheless, it would be convenient to co-opt teachers as well.

I could see this happening in a number of ways, thought I don’t know if they are plausible


  • Teachers might be sufficiently educated through propaganda they actually become oriented around student outcomes
  • The solidarity in the teaching profession cracks, and teachers organize around different (though presumably still self-oriented) lines based on broader social and economic factors
  • Teachersaccept a smaller but more professional workforce organized around technology

The first and second of the above options assumes teachers stop acting as a rational actor. Teachers want money — they are politically active in order to divert welfare away from students to themselves — and because no one really knows if they are a ‘good’ teacher or not (as there is no good testing regime to tell), teacherse are naturally risk adverse. While no publishers see a risk of going out a business because of better tests, better tests would make it much easier to dismiss. low-than-great teachers

The third option — that teachers might agree to higher wages in exchange in exchange for integrating testing technology into the classroom – is possible. The West Coast Longshoremen struck an almost identical deal, and enjoy six-figure salaries as a result. This possibility requiers that teachers remain rational but have a hard-headed understanding of their operating environment. As the teaching profession has been lobotomized through generations of low wages, I doubt this degree of rationality will be forthcoming. Teachers have failed to use the federal-academic complex to pivot their front organizations into political success — what hope is there of teachers wising up now?

Even though working with publishers is the easy way to build a nation-wide testing infrastructure, and working with teachers is the hard way, both ways should be persued. Teachers still have some political power, and every iota of it that can be neutralized is an outa that can be used to improve the terms of trade with Publishers, a self-interested group that is not actively hostile to educating children through testing technology.

Of Fabrication and Testing

JR of Edgewise sends in “Counterinsurgents Should Consider A ‘Fabrication Cell‘” from SWJ Blog. Another idea, given to me by Isaac at the Boyd Conference, is that we should begin testing, in realistic environments, devices that can be used in counterinsurgency. Some devices to test would be obvious: everything from $100 laptops to anti-IED tools can be used to build connectivity and defense the military.

What else should be tested? What novel devices should we use in counterinsurgency?