Epilogue—The meaning of the gaps: Yesterday, we said some data concerning the PISA aren’t available yet.
As it turns out, they are. If you want to disaggregate the new PISA scores by race/ethnicity, that can be done for all three subjects tested—reading, math and science.
Yesterday, we showed you those data for math. If you want to see them for science and reading, click here, scroll to pages 34 and 46.
Those data help you contemplate the size of our nation’s gaps. What’s the overall meaning of our nation’s very large gaps?
Once again, let’s review the horrible data in which scores on the NAEP are broken down by race/ethnicity and income.
In this chart, “lower income” refers to kids who quality for free or reduced price lunch because of family income. This is not a measure of poverty, despite what you’re constantly told by the rising number of Fox-like hacks of the left.
“Higher income” refers to kids who don’t qualify:
Average scores, Grade 8 math, 2011 NAEPHow large are our nation’s achievement gaps? Today, we’ll articulate a distinction we left unsaid yesterday.
All students: 283
White students, higher income: 299
White students, lower income: 278
Hispanic students, higher income: 280
Hispanic students, lower income: 266
Black students, higher income: 273
Black students, lower income: 258
Asian-American students, higher income: 313
Asian-American students, lower income: 286
On the NAEP, lower income white kids produce higher average scores than higher income black kids! In our view, this is part of what that means:
Obviously, many black kids are doing great in school. Many white kids are struggling. Beyond that, black kids have been scoring substantially better on average than was the case in the past.
(It’s astounding to see the way the liberal world refuses to state that last fact.)
On average, black kids have been scoring much better. But white kids have been scoring better too. This means that our gaps remain, even though they have gotten smaller.
Our gaps remain—and in various ways, our gaps are very large. Here’s the way that grabs us:
In a nation with very large gaps, what does it means when we’re told that we should eschew so-called “tracking?” When we’re told that we should adopt a set of higher “standards” to which everyone should be held?
Those prescriptions may make perfect sense in a society with small gaps. In this society, the gaps are very large. Do those prescriptions make sense for us?
Our gaps continue to be very large. Have you ever seen an “educational expert” come to terms with that fact?
We taught in the Baltimore City schools from 1969 through 1982. Mostly, we taught fifth grade.
Happily, NAEP data say that our nation’s tenth percentile scorers—our kids who are struggling in school—are now scoring substantially better than was the case in the 1970s. But the gaps are still extremely large between our lowest and highest scorers.
Have you ever seen an “educational expert” attempt to address this fact? Ever, by which we mean even once?
When we read Amanda Ripley’s new book, The Smartest Kids in the World, we were struck by the way she argued against various practices she referred to as “tracking.”
To appearances, Ripley argued this way because Andreas Schleicher had issued a decree to the world on the subject of “tracking.” For the most part, Ripley is an echo for Schleicher, the slightly cultish figure who devised and runs the slightly cultish PISA.
The implication of Ripley’s presentation is fairly obvious. All our kids should be taught the same challenging things in school.
This sounds like a great idea, if you’ve never set foot inside an American classroom. If you don’t have any real sense of the size of the gaps.
Here at The Howler, we’re wondering what our schools should do for the kids who are lagging behind. And make no mistake: in our society, many of those deserving kids are lagging far behind.
What do we do for children like them? Our gaps are quite large, and those children are suffering. It’s no fun to go to a school where you struggle and fail every day.
What should we do for the kids who are lagging behind? In our country, those kids are often way behind. What do we do for them?
We’ll examine these questions the week after next. Meanwhile, we marvel at the lack of concern for those kids routinely displayed by those of us on “the left.”
On the left, it’s becoming hip to sneer at news of our country’s mid-level average scores. “What, us worry?” we lefties now say. Our country has always had mid-level average scores. And our economy is still leading the world!
We’re amazed by that approach, which we’ll examine in detail next week. Those mid-level average scores are produced in large part by the large numbers of kids who are lagging way behind. When we sneer at our average scores, we’re sneering at those suffering children.
The gaps in this country are very large. In our view, you can’t begin to discuss our schools if you don’t understand that fact.
The outcry against tracking has occurred because: (1) minority children are disproportionately tracked into lower performing groups; (2) the tests used to determine assignment to tracks were shown to be racially biased; (3) worse teachers who did a poorer job of teaching were routinely assigned to teach the lower tracks; (4) there was little mobility between tracks once a child was assigned, resulting in stigmatizing and labeling.
ReplyDeleteIt should be possible to go back and correct these problems with tracking but the problems listed were so closely associated with it, that no one has tried to do that. It is easier to just call for the elimination of tracking based on test scores.
Once again, the reasons for eliminating tracking were benign. Are you now going to criticize educators for responding to complaints of the parents of minority children and doing what was asked of them? Claiming that those who eliminated tracking don't care about minority children is grossly unfair.
off topic,
Deleteeducation wonk alert:
m. night shyamalan discusses ``I Got Schooled: The Unlikely Story of How a Moonlighting Movie Maker Learned the Five Keys to Closing America's Education Gap.''
cspan2 at 6:45 or so
i wont be watching.
7;45 eastern time
Delete"We’re amazed by that approach, which we’ll examine in detail next week. Those mid-level average scores are produced in large part by the large numbers of kids who are lagging way behind. When we sneer at our average scores, we’re sneering at those suffering children."
ReplyDeleteTerrific insight.
This is a very very good post.
ReplyDeleteHowever, I think by tracking what is meant is deciding by means of a high stakes test given at an early age, say 11 years old, the entire future of a child based on the result.
In practice, this means writing off whole groups of children. Those who score well will be given an education that could lead to college, those who score poorly will be given custodial care until they drop out (or work sheets-- if they are lucky) and expected to do low-paid labor (or jail), effectively being labeled failures the rest of their lives.
Amanda Ripley is correct about this. That is what happens when kids are tracked. They are not helped to catch up. They are written off and resources go to the highly achieving ones instead.
The humane thing would be to reward everyone for their progress, no matter what their "level". Focussing on the "level" is exactly the wrong approach. It is a kind of crazy puritanical essentialism.
Imagine if physical therapy were done this way. In fact, there are very few cases where people cannot progress.
Not "tracking" doesn't mean that kids should be given work that they are not ready to do. In order to learn people need have the feeling of success that comes from mastery. In good teaching methods (as is the case in music or language, probably also math), they should be doing mostly manageable tasks combined judiciously with just enough amounts of newer more advanced work that will help them proceed smoothy to the next level without being discouraged. They should not be made to feel like failures at any stage, because this interferes with learning. -E
Fine comment.
DeleteEllen, isn't this a straw man? I don't know of this kind of tracking being done in the US (formally, as opposed to de facto), or any proposals that it be done.
Delete"This is not a measure of poverty, despite what you’re constantly told by the rising number of Fox-like hacks of the left."
ReplyDeleteThis is kind of a stupid comment because it's based on a technical error that actually weakens the argument of those who say it. Yes, it actually includes many kids whose family incomes are above the poverty level. Effectively, if about 50% of students qualify for free or reduced lunch, it's roughly equivalent to comparing the the 25th percentile with the 75th percentile. But it is completely consistent with the observation that wide disparities in income (which also happens to correlate fairly highly with race) result in wide differences in average test score performance. Identifying it as "lower income" instead of "poverty" corrects the mistake.
If you compared the top 10% in income with the bottom 10%, the gap would be far wider. Accordingly, while we might believe all reporters should know that qualifying for the lunch program does not mean living in poverty as technically defined by the Census Bureau, it's clearly a no-harm, no-foul mistake.
The gaps do correlate with ethnic group and wealth to a degree, but IMHO that's not the key gap. The real gap is the one between good students and bad students, regardless of ethnicity. Presumably if one compared, say, the top 20% with the bottom 20%, the gap would be even larger than the gap between various ethnic groups. Because of this very wide gap, some kind of tracking is needed.
ReplyDeleteThis one had me chuckling for various reasons.
DeleteYes, there's a gap between good students and bad students. By definition. This unsurprising gap tells us nothing, however, also by definition.
No, this very wide presumed gap wouldn't tell us that some kind of tracking is needed. That would depend on the number of low scorers. Assuming there were a sufficient number of low scorers to target, the very wide presumed gap would tell us nothing about how to track or whom to track.
Yes, the very wide gap is a presumed one because DAinCA didn't check on the numbers. Because he never checks. TDH tells us that in grade-8 math in 2011, higher-income white students outscore lower-income black students by 299 to 258, a gap of 41 points. The NAEP reports that the 25th percentile on this test is 260 and the 75th percentile is 309, a gap of 49 points. Larger than the gap based on race and income, but not much larger.
(The NAEP gives only quartile scores, not quintile.)
Thanks for verifying my guess, deadrat. I was talking about the gap between thetop fifth and the bottom fifth. If you look at this quintiles based on their midpoints, that's something like the 90th percentile vs. the 10th percentile. No doubt that difference is more than the 49 point difference between the 25th and 75th percentiles.
DeleteIt seems obvious to me that these two groups can't take the same classes. The 90th percentile could study calculus in high school, while the 10th percentile would be at a much lower level.
I repeat my main point. The ethnic compositions of the top and bottom groups are irrelevant to the question of whether tracking is needed. Whether they're white, black, purple, or chartreuse, the top 20% can learn calculus and the bottom 20% can't.
Sure it's a "verification" if by that word you mean an evisceration. And yeah, yours is a guess. Which is only one of two modes of investigation you have. And in this case, not a very good guess. I have no idea what it means to look at quintiles "based on their midpoints" or what that has to do with deciles.
DeleteThe NAEP data gives only quartiles, so the comparison we have is the one between the 25th and 75th. That gap is larger than the largest gap based on both race and income, but not by much.
Lots of things seem obvious to you. Which turns out to be a constant problem in your getting things right. In this case, though you've pretty much missed the obvious. The score gap between the 10th and the 90th percentiles will almost have to be larger than the gap between the 20th and 80th percentiles. Almost by definition. So this fact tells you nothing interesting and certainly gives no one a clue about how to proceed.
"Whether they're white, black, purple, or chartreuse," eh? Seems an odd thing for such a fan of Charles Murray to say, but maybe that's good news.
deadrat -- In my original comment, I guessed that gap between the top 20% and the bottom 20% was larger than any of the ethnic group gaps. How does one measure this gap? I'm comparing percentiles 80-99 vs. percentiles 0-19. A convenient way to do this is to compare the midpoints of each of these two sets of students. That's why a comparison of the 90th vs the 10th percentiles would be a way to measure what I said.
DeleteBTW deadrat, I don't know if you actually read The Bell Curve. Most of that book discusses IQ differences of human beings, regardless of ethnicity. One of the book's conclusions is that much of the difference in people's success in school and in life is due to IQ differences. I don't fully buy this conclusion, but it wasn't about race or ethnicity. Put crudely, it was about smart people and dumb people.
Now, one chapter of The Bell Curve does address the lower average IQ scores earned by blacks. The book states repeatedly that these lower test scores can not be attributed to inherent differences. Rather, the book says that blacks' lower tested IQ scores may be a temporary effect of their unique struggles in this country. It's unfortunate that many critics reported that the book said just the opposite.
People took issue with the book because of Murray's arguments in the second half, not so much Herrnstein's graphs. Murray tried to make a case that programs such as Head Start were useless and that society should give up on low-scoring kids (e.g., blacks) because it is too difficult to change their scores. There was a strong assumption that because scores are correlated with success in life, scores are determiners of life outcomes, something that is not true at all.
DeleteIn order to buy the argument that difficulties raising scores means difficulty in later life, you have to assume causality between scores and life outcomes. These are correlational studies. Scores don't cause success in life. Opportunities and preparation do (with motivation and encouragement). That's why Murray's arguments were crap, regardless of the validity of Herrnstein's analyses (against which minor, highly technical criticisms were made). Because one cannot set aside the effects of racism in our culture, the rest of the exercise in that book is pretty meaningless, except to people like you who wish to believe everyone starts out on a level playing field, ignoring the quite obvious fact that they do not. Critics of the book quite rightly grasped Murray's intention, even when they failed to understand the book's content. The book was providing cover to conservatives who wanted to defund social programs on the grounds that they were an exercise in futility, condemning black children to the underclass. It was ugly and appropriately rejected, since America was not yet a place whose values supported a two-tier system in which the haves live in comfort off the sweat of people who struggle to make ends meet. Getting there now though.
A convenient way to do this is to compare the midpoints of each of these two sets of students.
DeleteA comparison of the 90th and 10th percentiles is a way to measure the difference between the 90th and the 10th percentiles. If you want to compare the top 20% and the bottom 20%, you look at the 20th and 80th percentiles. The relationship between the 10th and 20th and the relationship between the 80th and 90th depend on the distribution of the scores.
It seems to me that I read both The Bell Curve and The Mismeasure of Man back when those books came out, but it's been almost two decades since the former was published, and I'll have to admit I don't remember much of that egregious piece of pseudoscience. I do remember two things -- that intelligence can be captured in a single number and (contrary to your claim) genetics is the major determining factor.
A brief time with the google reveals that the number is called g or the g-factor, and according to TBC, that's the factor attributable to native intelligence that explains the correlations across different tests. The claim is that inheritance explains 40% to 60% of the g-factor.
I believe that the TBC attributes societal differences among ethnic groups to the differing g-factors of these groups. That was the whole point of the book and the cause of much of the uproar, turning around the -- dare I say "liberal"? -- point of view that socioeconomic factors explained differing test scores.
I could be convinced by evidence that my recollection is faulty, but I really don't expect you to provide any.
Anon: Murray tried to make a case that programs such as Head Start were useless and that society should give up on low-scoring kids (e.g., blacks)
DeleteAs written, this is a racist statement. Being black isn't the same as having a low IQ. I'm quite sure that David Blackwell's IQ was a whole lot higher than yours or mine.
If you equated "low-scoring kids" with " some blacks" or with "blacks disproportionately", that would have been OK. However, there are plenty of blacks with high PISA scores and plenty of non-blacks with low PISA scores. The absolute number of non-blacks with low PISA scores is probably larger than the absolute number of blacks with low PISA scores, just as the number of non-blacks on welfare is larger than the number of blacks on welfare.
The book was providing cover to conservatives who wanted to defund social programs on the grounds that they were an exercise in futility, condemning black children to the underclass.
No, if anything, the book wanted to the underclass to remain the underclass, regardless of ethnicity. However, The Bell Curve argued that these people will remain as the underclass, regardless of the social programs. IMHO there's a lot of evidence that many social welfare programs are not effective. However, these programs have powerful vested interests. That's why opposition to TBC was so strong.
deadrat -- If you want to compare the top 20% and the bottom 20%, you look at the 20th and 80th percentiles.
deadrat -- Here's an example that I hope will make things clearer. Suppose a high school calculus class consists of all the students with PISA scores in the top quintile. The students' PISA percentiles range from 80 to 99. The average PISA percentile of this class is the 90th percentile. Now suppose you added all students in the lowest quintile to this class. The average student added would be in the 10th percentile.
This is simple stuff. Perhaps you and I are looking at different questions.
David, you have no idea what my IQ might be. On the internet you don't know who you are talking to. Were he not deceased, I could even be David Blackwell for all you know.
DeleteI am pleased to hear you call Murray racist. I think he was too. You also admit that you agree with Murray's belief in the ineffectiveness of social programs. That is an empirical question -- by that I mean, one that can be answered with data. Head Start and similar social programs have not been shown to be ineffective. They have been shown to have a more immediate than long-term effect, especially in the absence of continuing intervention. That makes them just like a whole bunch of other programs. Do you imagine your cholesterol would stay low if you stopped taking your statin and then tested it a few years down the line? Probably not. Does that mean your statin drug was ineffective?
Social programs also have effects that are not necessarily reflected in test scores. For example, gifted programs show no effect on educational attainment but do show a strong effect on ambitiousness of educational goals and willingness to seek out greater challenge in grad school and later life. So, are they worth spending any money on? They do nothing to improve test scores, so lets get rid of them. If preschool Head Start keeps low-performing kids out of jail later on, but does little to give them higher grades by fourth grade, should it be abandoned? Murray says yes. He doesn't care about any social benefits. He also doesn't ask the obvious question -- if Head Start produces short term benefits, why should it not be extended to the higher grades too so that struggling kids can continue to have interventions needed to keep them going as work becomes more difficult for them? Why are we not willing to invest along the line when our early intervention programs do produce benefits -- to the extent that there is continuity in attendance and an attempt to address family problems concurrent with schooling? Why don't conservatives ask these questions?
A while back there was a very controversial study published showing that by the time sexually abused children were adults, there was little lasting impact of their abuse on their success or their lives. Some people thought this study was saying that child abuse was not a horrible thing because of the finding that it need not result in long-term disorder. I see a parallel with Head Start. If an intervention has an impact in the short term (this time for the good), but that impact is miniscule compared to the other influences on a child's life so that it is not measurable after many years have passed, does that mean it is not important in the short term and does it mean it is not worth doing at the time? For the record, so there is no misunderstanding -- I think child sexual abuse is a horrible thing that must be prevented and if it occurs, treated. I also think that providing positive experiences that benefit a child in early childhood are good and we should do them, even if the long-term benefit is overcome later by other influences on a child's life.
Do you disagree David?
Just to be clear, Anon, I didn't mean to call Murray a racist.
DeleteFor all I know, you might be smarter than Blackwell, but IMHO it's unlikely, because he was so darn smart. Among other achievements, he was the co-creator of a field of mathematics called Dynamic Programming.
I agree that the effectiveness of social programs is an empirical question. My impression is that programs like Head Start and School Lunch have not produced academic improvement commensurate with their cost. YMMV.
That's not to say that they don't do any good at all. OTOH it's conceivable that the money for these programs might be spent in ways that are of greater benefit to the poor. Ideally, Head Start should be able to demonstrate not only that it does some good, but that it does more good than alternative ways that money could be spent.
Unfortunately, cause and effect are really hard to measure. I don't think anyone can say with certainly what would have happened to the kids in HS if that program hadn't been available. And, it's even harder to say what would have happened to these kids in some hypothetical program that doesn't exist. If the government operated like a well-run business, they'd be trying different alternatives in different locations to get actual data on relative effectiveness. However, the government is mostly about giving money to special interests. These special interests won't let them discontinue HS in selected areas.
David, you cannot measure cause and effect with respect to children's outcomes because it is unethical to put any child into a control group where he or she is denied attention, much less deliberately assign them to a condition of poverty when those are not their original circumstances.
DeleteYou are the one proposing hypothetical programs that do not exist, when you suggest that Head Start shouldn't be funded because there might be a more effective alternative, without saying what that alternative might be. When you start talking about effectiveness in relation to cost, instead of effectiveness as measured by improvement in outcomes, you are no longer being empirical but are bringing "value" into the situation. My impression is that conservatives do not place sufficient value on educational outcomes for children and wouldn't fund any public education for anyone if they could avoid it. Home-schooling for all, and too bad for you if your mom cannot do it. Your idea that no one is working on improving interventions is totally wrong. You also do not understand that even within programs such as Head Start, new ideas are being introduced regularly in an effort to improve early childhood education. In fact there are journals devoted to it.
I also find it ironic when conservatives complain that Head Start shouldn't be funded because it is "ineffective" while clamoring to fund abstinence only sex education, which has been demonstrated not to work and yet keeps getting added to Legislation because it is a special interest of conservatives.
It does strike me as very unfair to criticize programs for not doing what they (1) are doing already, (2) have done all along, (3) and you would know it if you knew more about such programs than just what conservative propagandists tell you.
IQ scores do not predict specific accomplishments in a field, especially breakthroughs. They do not measure creativity or motivation (IQ is uncorrelated with creativity). IQ scores do not measure persistence or critical thinking or willingness to break barriers or take a new path. Terman's longitudinal study of gifted adults showed that they became accomplished in their various professional fields, but they did not show genius on the level of Mozart, Jefferson, or Einstein. You have confused potential with performance in your comment about Blackwell. Again, you have no idea what I have or have not done in my field. I'm glad Blackwell did well. When a child overcomes multiple disadvantages and risk factors to obtain an advanced degree and do work at the same level as others, perhaps that is a display of genius akin to what prodigies like Mozart do when they grow up under more favorable circumstances. People who overcome adversity tend not to talk about it much, so others assume they had the same upbringing as everyone else. I suspect that is a reason why various interventions do not get the credit they deserve as they support children's struggles to do well in later life.
Anon -- Instead of guessing what conservatives want, just look at what they say. Conservatives want more extensive use of vouchers. BTW a substantial majority of black parents want them too. Unfortunately for black students, Dems have bowed to the teachers' unions to kill voucher programs wherever they could.
DeleteI think we conservatives tend to think of individuals, while liberals tend to think of groups. My wife and I were big supporters of public schools, yet we sent one daughter to private school for 3 years when her individual school situation was intolerable. Conservatives want poor parents to have the same options for their child.
Differences in aptitude are likely between groups and emphasis on math as an accepted determinant of potential is misguided especially in diverse societies
Delete"We’ll examine these questions the week after next.."
ReplyDeleteYeah. Longtime readers are on the edge of their chairs.
OMB (Don't leap off when BOB does)
Delete"On the left, it’s becoming hip to sneer at news of our country’s mid-level average scores. “What, us worry?” we lefties now say."
Name one BOB. Name names. Name one such sneering leftie or forever quit attacking people in the media when they don't "name names."
We know BOB reads only comments on blogs other than his own, and mentions comments when he wants to sneer or they fit his meme. So he won't name names.
So I'll challenge BOBfans worldwide. Cecelia Mc., deadrat, Anonytrollbanner, come on down. Name someone to whom BOB's quote applies.
KZ
KZ,
DeleteYou're laboring under a couple of misapprehensions. The first is that I'm a "BOBfan," by which I take it you mean that I adore TDH's every word. Or maybe it's just that I think your trollish side is adorable yet still trollish. My main purpose here seems to have become correcting DAinCA's ever more ignorant comments. Even I'm not sure why that is.
How again do we know that TDH "reads only comments on blogs other than his own"? What do his reading habits have to do with his naming or not naming names?
To be fair, TDH says that next week he'll examine "in detail" the sneer approach. Let's see if he names names then. For now, his sentences about mid-level scores make an apt and accurate paraphrase of Diane Ravitch as she was quoted by The Washington Post in their the "Education" section of 12/2. Does she count as a "leftie"? Is her comment a sneer about students scores? Eye of the beholder, perhaps.
I think you have the right suspect deadrat. I am simply not
Deletesure how a Bushie becomes a leftie.
KZ
Bushie? Only if you mean she was an appointee of George Herbert Hoover^H^H^H^H^H^HWalker Bush. Also of Bill Clinton. On the left hand, the educationist establishment loves her; on the right, she's a fellow at the Hoover Institution.
Delete