April 27, 2010

Twin study of teaching quality

In reviewing Diane Ravitch's book a few weeks ago, I pointed out that it took decades for baseball statistics to mature, and that statistics measuring teaching effectiveness would likely take awhile, too. Here's the abstract of a clever study from Science:
Teacher Quality Moderates the Genetic Effects on Early Reading
J. Taylor,1,* A. D. Roehrig,2 B. Soden Hensler,1 C. M. Connor,1,3 C. Schatschneider1,3

Children’s reading achievement is influenced by genetics as well as by family and school environments. The importance of teacher quality as a specific school environmental influence on reading achievement is unknown. We studied first- and second-grade students in Florida from schools representing diverse environments. Comparison of monozygotic and dizygotic twins, differentiating genetic similarities of 100% and 50%, provided an estimate of genetic variance in reading achievement. Teacher quality was measured by how much reading gain the non-twin classmates achieved. The magnitude of genetic variance associated with twins’ oral reading fluency increased as the quality of their teacher increased. In circumstances where the teachers are all excellent, the variability in student reading achievement may appear to be largely due to genetics. However, poor teaching impedes the ability of children to reach their potential.

And here's the AP article: 
Study: Better teachers help children read faster

By DONNA GORDON BLANKINSHIP
Associated Press Writer

SEATTLE (AP) -- Genetics play the biggest role in determining how fast a child learns to read, but a good teacher can make a measurable difference as well, according to a study released Thursday.

Florida State University used twins assigned to different classrooms to develop the conclusions.

Researchers studied more than 550 first- and second-grade classrooms with at least one identical twin and more than 1,000 classes with at least one fraternal twin.

Among the identical twins, 42 pairs out of 280 pairs showed significant differences in reading improvement during the year studied, said lead researcher Jeanette Taylor, an associate professor of psychology at Florida State.

In each case, the teachers also had significantly different quality scores. Twins with similarly good teachers got similar scores.

"If you have identical twins, they should do very similarly in school," Taylor said.

Teachers whose students showed the greatest average one-year improvement in the number of words they could read out loud in one minute were considered the best teachers for the purpose of the study. ...

The study shows teacher quality is one reason for the differences between the achievements of twins, but it cannot explain the whole difference, Taylor said. Since these pairs were all in two different classrooms, other possible factors include how well behaved classmates are.

The researchers believe their results showed the best teachers made the biggest difference in learning achievement. Genetic differences between students seemed to disappear in classrooms taught by less effective teachers, because children don't reach their potential, the researchers found. ...

"Good classrooms are all alike; they maximize kids' potential. Poor classrooms are not only poor in one way; they are poor in multiple ways," Shanahan said.
  

14 comments:

Anonymous said...

OT: Here we go again.

http://www.nytimes.com/1999/09/30/business/fannie-mae-eases-credit-to-aid-mortgage-lending.html

Bill N said...

Not I have to look it up, "significantly" probably means statistically significant effect size rather than a practically significant effect size.

The strength of the correlation is pretty impressive, but it doesn't demonstrate direct causality. Considering how they measure, it could as easily be some other classroom factor, like time of day for reading lessons or some effect from the other students. For example, better classes can move faster. A couple slow students could drain the teacher's time.

In other words is the class better because the teacher is better or is the teacher measured better because he/she got a better draw on the class?

Before I look it up, I'll wager the authors of the study raised the cautions that the press didn't.

Cleanthes said...

In other twin news:

The Kaczyski twins first rose to public prominence as child actors in the 1962 film The Two Who Stole the Moon.

Jaroslaw Kaczinski, the identical twin brother to Poland’s late president Lech Kaczinski announced today he will run in the summer elections to replace his twin who died in an airplane crash on Saturday, April 10, 2010.

Anonymous said...

In circumstances where the teachers are all excellent, the variability in student reading achievement may appear to be largely due to genetics. However, poor teaching impedes the ability of children to reach their potential.

This might be even worse than the nerds don't get laid study.

If the taxpayers are having their money stolen to fund eggheads who need to prove that the sky is blue and the grass is green, then it's no wonder that this nation is staring at $100 TRILLION in un-funded government obligations.

PS: On the other hand, in this day and age, maybe we should just be thankful that they aren't out there proving that the sky is green and the grass is blue...

Anonymous said...

Perhaps there are other variables here that might a useful context?
A couple decades ago, a leading
homeschool advocate, Dr. Raymond Moore--himself a former public school superintendent--claimed there was no replicated research showing any advantage to early teaching of reading. In other words, boys especially starting to read at, say, age eight did not seem handicapped by the fully benign neglect of wedging them into reading instruction at, say, age 5 or 6 . This does not imply that endless delay is without serious compromise of reading potential but apparently readiness is heavily biologically "programmed" and can ripen as late as age 9 or so. Reading problems at early ages have a sex ratio of males having a vastly higher rate of problems. The proposal made by British psychologist Chris Brand in his (depublished ! but now online accessible ) book
"The 'g' Factor" was for a kind of smorgasbord of reading instruction methods and of smooth gradations along which smogasbord a child with any degree of motivation could smoothly interface by self selecting his/her own point of interface and thus escaping "iatrogenic" damage. It's important to focus upon the role of genetics and of teaching efficacy (it's a composite of art & science ). It's equally important to focus upon the damage of governmental coercion via mandatory schooling if, in fact, benign neglect belies a compelling state interest in requiring every 6 year old child to start marching in lock step. Dr. Moore pointed to the colonial history of portions of early America in which literacy rates were higher than we currently have without any governmental schooling coercion at all.

Anonymous said...

I no longer have a Science subscription so could not get the article online. How much variance was explained by these variables?

Toadal said...

I like the study since it gives two important and salient messages to the reader.

The first and obvious message is good teachers *do* contribute to children's improvement in oral reading. The second, not-so-obvious message is a young child's genetic potential contributes to half his or her success in oral reading.

The second, more subtle message needs public and political dissemination to show the importance our genes have to learning, education success, and national wealth. Hopefully the study and ongoing genetic research will help us overcome the class warfare our education system has inundated us with since the early sixties.

What we now need are astute educators and software designers to collaborate to create software learning applications 'renown for excellence' that can allow each child to fulfill his or her genetic potential.

keypusher said...

Proof that no what the studies say, the headlines will always be the same.

"Good classrooms are all alike; they maximize kids' potential. Poor classrooms are not only poor in one way; they are poor in multiple ways"

Tolstoy has evidently given up novels and gone into education research.

Anonymous said...

A background element of this topic, is the "iatrogenic" influence of coercive governmental mandatory schooling. There used to be little, if any, replicated research to indicate any reading potential being lost for children that delayed efforts to read until age 8 or even later. This is especially true for young males, who have a lot more problems with reading than do females.Homeschooling may not be the cup of tea for many people--something some of its practitioners don't seem to understand--but a good part of its success arises from a nurtured interface with reading readiness rather than from the inevitable degree of lockstep coercion that "public" schooling never gets far away from. The British psychologist,Chris Brand, in his 1996 (depublished! but now
accessible online for free download ) book, THE "g" FACTOR suggested for state schools an early primary level educational smogasbord of instructional methods and gradations along which a motivated child could freely choose his/her point of interface and rate of progress. The discussion that this notion deserved seems to have been buried thus far along with the depublished text. It's not just ineffective teachers that depress the influence of "g"--it is the structure of schooling itself.

kudzu bob said...

Genetic differences between students seemed to disappear in classrooms taught by less effective teachers, because children don't reach their potential, the researchers found.

The more good schools and teachers there are, the more manifest the black/white achievement gap must become.

Kylie said...

Anonymous said: "...maybe we should just be thankful that they aren't out there proving that the sky is green and the grass is blue..."

That's pretty much what they are doing when they ascribe that pesky persistent achievement gap to everything but the disparity in IQ.

Svigor said...

I'm with Bill N. It doesn't take long to get a graduate degree in detecting media bullshit. When they say "significant" a lot, but don't give you numbers, they're probably blowing smoke.

Roger said...

The study is just a simple nature/nurture twin study, but it has unsupported conclusions about teacher quality. It did not even measure teacher quality, except to attribute classroom score differences to teacher quality.

Anonymous said...

Looking at twins is superfluous for measuring teacher quality.

As long as students within a school are assigned randomly to a teacher (I'm not sure exactly how this works in practice), then on average the teachers will be working with students who have similar genetic aptitudes and home environments.

Hence, within a school and grade, differences between classrooms in terms of test scores should be attributed to either random noise or something in the classroom environment, rather than student aptitudes or home environments.

Using twins is analogous to matching subjects in clinical trials. It's sometimes done to increase efficiency (i.e. reduce drug company cost), but usually isn't if the sample sizes are large enough. The randomization already guarantees that the treatment and control arms are similar on factors other than the treatment, without any need for matching, and the same considerations apply here to twins.