February 22, 2013

How does Texas do well on the NAEP test?

Texas public school students usually score pretty well in the federal government's NAEP school achievement tests, at least when adjusted for ethnicity. I've always wondered how they do it. It would seem like the kind of thing worth checking into.

One way, it turns out, is by excluding more students from having to take the NAEP than other states do. Texas excuses 10% of its 4th graders versus 4% nationwide and only 3% in California. (See p. 5 of this new report on the NAEP performance of the 5 biggest states.) So, Texas has simply made a large fraction of Below Basic scorers vanish. That's a nice little running start for Texas.

If Texas has figured out how to fiddle with that parameter, I wonder what else they've figured out?

23 comments:

Who Set That Up? said...

steve i subscribe to your work on google reader - have been reading for years - was gonna say something disagreeable about your notes a few weeks back about car salesmen - because that's what i do - but - why bother....
what i would really like - separate from this article,
is an overall picture of america, crime, education stats, gdp, unemploy, etc/etc - you get the idea - based on data EXCLUDING african-americans (i use this term for lack of better...)
i think with your resources and savvy, you are up to the task.
i'll watch for it
thank you

Bob Arctor said...

Texas' reported median should be at the 55.75th actual percentile and California should be reporting the 51.5th percentile of all students.

Steve Sailer said...

That's assuming half the excluded students would score above the median. I think it's more likely that over 95% of the excused students would score below the median.

Anonymous said...

Sailer bait.

elvisd said...

When Bush was governor, he trumpeted graduation rates until someone pointed out the way Texas tracks students from 9th grade onward. A lot of dropouts weren't counted if they changed schools before dropping out.

James B. Shearer said...

That's assuming half the excluded students would score above the median. I think it's more likely that over 95% of the excused students would score below the median.

No, suppose you exclude the bottom 10%. Then the remaining 90% will range from 10-100 with median value 55%.

Douglas Knight said...

No, Bob Arctor is right. If you exclude the bottom 10%, the median of those who remain, the 10-100 precentiles is the the 55th percentile. In an extreme case, if you exclude the bottom 90%, the median of what remains is not the meaningless 90+50=140th percentile, but the 95th percentile.

countenance said...

I know a lot of urban districts in Texas artificially hide their dropout rate because they classify dropouts as being "home schooled."

Bob Arctor said...

countenance:

I know someone who worked at Houston ISD in the 90's-00's; not only did they recode dropouts as "homeschooled" but they also sent dropouts to $300 six week correspondence school diploma mills and then counted those students as "high school graduates" without even verifying if they even looked at the materials sent to them.

Another big difference between California and Texas that I've never seen anyone remark on is that TX Hispanics (especially before 2000) were overwhelmingly mestizos from northern Mexico whereas CA had a disproportionate number of Indios from southern Mexico and Central America. But as everyone enlightened knows that can't possibly have anything to do with why TX Latinos dramatically outperform CA ones, so I guess it will just have to remain a mystery...

Anonymous Rice Grad #6 said...

A couple of years back, probably with an eye toward trying to plant the seed for a move to Texas from the cranky, cold, socialist Northeast, I mentioned to my wife that I had read ("on that evil, racist blog I read sometimes, just to marvel at the depravity of its purveyor...") that white kids in Texas did better than those in the rest of the country on some standardized test.

Don't think I'll tell her about this little technicality...

Anonymous said...

From Frontline, take that as you will

It mentioned that anyone can for all practical purposes "buy" a high school diploma for about $200 that by law has to be accepted by post secondary schools in Texas. Not sure if employers have to count it.
It also showed the obfuscation built into the dropout numbers.
That they are pulling a Superintendent Chalmers and hiding the low scorers when test taking day arrives should shock no one.
Also in that frontline episode,
(Spoiler Alert) The single Asian student in the charter school they are filming in is the valedictorian.

Education Realist said...

I've been long skeptical of the "Texas kids just do better" meme, and am happy to see the exclusion rate brought up. But I thought that a few years back the states had been required to test all their kids. I must have misread.

Anonymous said...

"I think it's more likely that over 95% of the excused students would score below the median."

LOL

I worked as a teacher in Texas. I saw the lists of excluded students. I can assure you 100% of the excluded students would score below the median.

Anonymous said...

But I thought that a few years back the states had been required to test all their kids. I must have misread.

They have to test them all for Stanford or ITBS. And I mean all. If the kid is so disabled that he practically can't communicate they still fill out his name, age, etc., and proceed to have an aide read it to him. When no response can be elicited, then they can mark the student as unable to test due to whatever his condition was, but they submit a form for every warm body enrolled. That may be a requirement of the testing program in order to meet the reliability standard for the test. It would invalidate the whole thing if you could exempt students. NAEP is criterion referenced, so they can exempt away.

Anonymous said...


I know a lot of urban districts in Texas artificially hide their dropout rate because they classify dropouts as being "home schooled."


That actually makes a lot of sense. Homeschooling in Texas is totally unregulated. I homeschool in Texas. There is no requirement to prove anything to anyone ever. No notification has to be sent. No standardized tests have to be taken or reported. You just have to operate in a bona fide manner and teach good citizenship, math, reading, spelling and grammar. So, if a district wanted to hide dropouts, homeschool is a black hole into which no one will ever look because there is no way to check on home schoolers. I guess they could send a truant officer and make the parents say whether or not they are homeschooling. If the parents say yes. Then that is the end of it. If the parents say no, then they would be in violation of the compulsory education law and would be responsible for the truancy of the minor in their custody.

Anonymous said...

I know someone who worked at Houston ISD in the 90's-00's; not only did they recode dropouts as "homeschooled" but they also sent dropouts to $300 six week correspondence school diploma mills and then counted those students as "high school graduates" without even verifying if they even looked at the materials sent to them Well, Houston would do not as good as San Diego since it not only has a lot of Hisapanics but also blacks. San Diego has a lot of Hispanics low black population and more Asians.

Anonymous said...

Well, Texas largest cities have low white poverty between the 7 to 8 percent range. So, probably white scores are more correct than Black or Hispanic.

Anonymous said...

Probably, Texas is more dishonest than California, Nevada, Arizona or New Mexico or even Florida on their stats. As for moving there are many Red States in the middle of the US without the high Hispanic population. Try Ok,Neb, Kansas-some black areas and Missouri-some black and Mexican areas. North and South Dakota. Utah, Wyoming, or Id. Republicans don't have to go to the Republican verison of California. There are other states, when I present the negatives of Texas being as Hispanic as California and that populaion will grow in Texas Republicans don't listern that they are alternatives to Texas. By the way Texas is humid and sometimes gets cold but not as cold as some of the states mention above.

Anonymous said...

Well, Republicans use Texas as a model in some ways they do good, high property taxes and no progressive income tax. Tougher leanding laws, no housing bubble. Cheaper housing and overall taxes which attracts whites out of California and not always ones. In the few Republican counties in California many always talk of moving to Texas. But Texas cheats on its stats since it knows Hispanics with few exceptions like maybe lower drug use than whites or longer life usually do worst on most stats.

ATBOTL said...

The Texas house of cards will collapse soon. Let's hope it will be the death knell for aracial corporate conservatism.

Anonymous said...

The Texas house of cards will collapse soon. Let's hope it will be the death knell for aracial corporate conservatism I agree with you there on that issue but the Repubs for national office, Cruz, Smith and Poe are going against the amensty;however, Texas Republicans have allow illegals in their state to drive down the wages of construcation.

Anonymous said...

"'ve been long skeptical of the "Texas kids just do better" meme, and am happy to see the exclusion rate brought up. But I thought that a few years back the states had been required to test all their kids. I must have misread." - Texas is definitely cheating, though I'd probably be doing much the same if they hired me to "work" on such an intractible problem.

Gringo said...

elvisd
When Bush was governor, he trumpeted graduation rates until someone pointed out the way Texas tracks students from 9th grade onward. A lot of dropouts weren't counted if they changed schools before dropping out.

When Bush was governor, I researched class sizes versus dropout rates at a number of urban Texas high schools. There were MANY urban high schools that had senior enrollment at 50-60% of freshman enrollment. Yet these schools claimed dropout rates of ~ 2-6%.

There are obvious discrepancies in these numbers.

[These numbers are rough,as I am doing this from memory.]