Technocracy Can Only Thrive When It Lies

Susan Dynarski of the University of Michigan made a stir with a New York Times Economic View piece, “Why Talented Black and Hispanic Students Can Go Undiscovered,” which hailed a short-lived change in how a large South Florida district, Broward County schools, identified intellectually gifted students:

In 2005, in an effort to reduce [racial disparities in the district’s gifted program], Broward County introduced a universal screening program, requiring that all second graders take a short nonverbal test, with high scorers referred for I.Q. testing. Under the previous system, the district had relied on teachers and parents to make those referrals.

The program was studied by David Card and Laura Giuliano, who found that replacing teacher recommendations with universal screening caused racial disparities to decrease:

The share of Hispanic children identified as gifted tripled, to 6 percent from 2 percent. The share of black children rose to 3 percent from 1 percent. For whites, the gain was more muted, to 8 percent from 6 percent.

As Steve Sailer pointed out, Dynarski’s column did not mention that the reduction in racial disparities was aided by a two-tier cutoff: students eligible for Free or Reduced Price Lunch (ie, low income students) and students designated as English Language Learners needed only score one standard deviation above the mean, while students not so designated needed to score two standard deviations above the mean.

A similar omission was made in a Quartz article about the same Broward County program, which makes Dynarksi’s soft-pedaled accusation of racism (“Teachers may have lower expectations for these children”) more explicit: “The ‘gifted’ system in US schools is broken, racist, and completely fixable.” A Wonkblog article about the same study (“These kids were geniuses — they were just too poor for anyone to discover them”) was more clear about the two tiers:

Broward is one of many Florida counties that maintains two sets of standards for gifted students. Most children have to score at least 130 on the state-mandated IQ test — roughly in the top five percent — to qualify. But children still learning English, and those from low-income households, only have to score 116.

In response to criticism from commenters and on Twitter for not including information about the affirmative action system, Dynarski responded that the cutoffs weren’t relevant:

Dynarksi

This is incorrect. The study does not provide evidence that teachers applied the previous referral system in a racially disparate way, i.e. that an IQ 116 lower-income white student was more likely to be referred than an IQ 116 lower-income black student. The study only provides evidence for disparate impact. If teachers were inconsistently applying the lower, FRL/ELL cutoff, and generally recommended kids in a race-blind manner who were above IQ 125 or so, then this would affect black and Hispanic students more often than whites, and produce results identical to what is observed. Inexpertly applying affirmative action is not the same as racism, no matter what Dynarski says.

I don’t have strong opinions about gifted programs. The newly enlarged elementary-school curriculum and testing regime is more than hard enough for most kids, even very smart ones. Common Core aside, there is more in heaven and earth than when you learn to factor quadratic polynomials, and better things to do as a kid than be academically challenged, no matter how smart you are. My childhood school district did away with gifted programs altogether the year I entered kindergarten, at the same time they instituted racially-motivated busing: I was certainly not challenged in elementary school, but I had good-natured teachers who were interested in talking with me and who would occasionally let me slip off into the reading corner for an hour or five while the rest of the class did their work. I was phenomenally happy, at least until I got to middle school and they expected you to actually copy down all the spelling lists and place value problems, at which time I talked the school into skipping me ahead a grade, in spite of the almost-all Fs on my report card.* My wife, who is demographically closer to the groups that public policy wonks wring their hands about not entering gifted programs in sufficient numbers, simply faked her mother’s signature to get into New York City’s old “Special Progress” program, which put you with the other smart kids and skipped you from 7th to 9th. So she skipped 8th grade, and later ended up teaching it five times, and I skipped 7th grade and ended up teaching it eight times; a lesson that those of us who do not learn the lessons of middle school are doomed to repeat them.

Back to Dynarski: why was she so loathe to put such a central fact about the Broward County system as the two tiers (almost every page in the NBER article refers to the tiers as Plan A and Plan B) into her article?

IQ Distribution

The simple answer is that technocracy requires an abrogation of values, at the very least an abrogation of the central American value of equality under the law. Algorithms will inevitably need to apply different standards to one group than another, lest they produce unpalatably one-sided results. (No one claims that a one-tier universal screening system would do anything but intensify racial and economic imbalances rather than redress them.) The informal systems of practice that teachers or police officers or local government officials apply will involve some accommodation to differences and multiple sets of standards, but these standards will not be systematized, nor will they be as tilted to equality as towards fairness. When technocracy intervenes– whether to institute stop-and-frisk programs for guns in New York or to increase representation of minorities in South Florida gifted programs or to judge teachers based on the “value added” of their students’ scores— it will be, in general, to increase racism, at least in the sense of different groups being held to different standards. The algorithms will happily be more racist than any teacher. No one likes to discuss that, even holding last year’s test score constant, black kids are expected to do worse this year by the value added model assessing their teacher’s skill. The writers hailing Broward’s program are reluctant to admit that the improvements are due to more systematic application of double standards; instead, they would like to suggest that the computer’s impersonal gaze will see genius where we mortals see only race.

This is the future that technocracy promises: racial disparities programmed into the algorithms that run public programs, even as the dominant voices of our culture become ever more strident in denouncing ordinary people’s racism; our most prominent academics and journalists angrily decrying discrimination in the personal and informal patterns of work and life while denying its institutionalization in the ever-watching electronic eyes that watch us, ready to save us from ourselves.

*I also faked being homesick on vacation, the last week of June, so that I could call my best friend to break into my never-locked house and steal my end-of-6th-grade report card from inside the mail slot, in fear that if my parents saw my atrocious grades they’d prevent me from skipping the grade; my beloved paternal grandfather died the day we returned and my parents never followed up with the school to find out how incredibly badly I had done. Some time after I graduated college I told them the truth. I’d like to think that my grandfather, who among all my relatives would probably have most shared my distaste for homework and desire to get back to a book (he spoke five modern languages as well as Latin and Greek but flunked out of medical school almost immediately after entering, moved to the States and graduated from Stanford in 18 months before joining the U.S. Army to go put his language skill to work against the Nazis; his well-thumbed copy of War and Peace sits at the head of my bed) does not hold it against me now.

13 thoughts on “Technocracy Can Only Thrive When It Lies

  1. The paper did not say that. It said:
    “all second graders in the District completed the Naglieri Non‐Verbal Ability Test (NNAT) (see Naglieri and Ford, 2003). The NNAT test ‐‐ which is similar to Raven’s Progressive Matrix Test ‐‐ takes
    less than an hour to complete and was administered by second grade teachers during
    regular class time. Non‐disadvantaged students scoring above 130 points on the test,
    and ELL/FRL students scoring above 115 points were eligible for referral for IQ testing”

    The final target was 130 on the IQ TEST. 100% of those above 130 on the IQ test were placed on the gifted program, and a smaller percentage from the 116-130 range of IQ were laced in order to attain an approximate 3.3-4% target of gifted. So they are talking about who were selected to participate in the IQ test.

    Like

    1. I’m not sure how that contradicts my statement. There were also two cutoffs for the IQ test, but the universal screening added an additional initial test administered to all students, which also had two cutoffs.

      Like

    2. Are you disagreeing with whether there were different cutoffs for the IQ test as well as the pre-selection test? There were, eg, from http://articles.sun-sentinel.com/2010-02-01/features/fl-gifted-students-20100201_1_gifted-testing-gifted-program-gifted-kids
      “Florida allows school districts an option to set their own criteria for free- and reduced-price lunch students as well as those with limited English speaking ability, both considered under-represented populations in the gifted program. In Broward, the minimum IQ is 116 and in Miami-Dade, the minimum IQ is 112 for these populations. The IQ scores are factored in with standardized test scores and other criteria to determine eligibility.”

      Like

      1. FOR CONSIDERATION, yes

        For placement in gifted programs, it was based on IQ scores + class scores + teacher recommendations.

        Resulted in IQ>130 => 100% chances of being placed in programs; between 116-130 >60% chance of being placed in the programs.

        I am not arguing that this is right or wrong, just that the paper does not lead to such conclusions. All the paper did was create a two-step process, which I can bet would have actually reduced the number of minorities and poor being placed in gifted programs, WHEN AVERAGED OVER MANY YEARS.

        In any reality (technocratic or otherwise), selection is based on multiple things. I think selecting all people > 130 itself is a wonderful thing; I would not do so. I would definitely check their class scores to weed out people who do not make the effort in class.

        Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s