June 18, 2008

My syllogism for improving student achievement

To avoid any further confusion going forward, here is my hastily written syllogism with respect to improving the academic achievement of low-SES students:

Many Low-SES kids are different than their middle-class peers -- different in ways that make them more difficult to educate.

Most schools are designed to educate middle-class kids, as a result many low-SES kids are unable to access the education provided in these middle-class schools.

Many Low-SES kids require compensatory education which includes improved instruction that compensates for their skill deficits and the utilization of improved classroom management practices.

Compensatory education can eliminate or substantially reduce many of the academic achievement gaps that exist between many , but not all, low-SES students and their middle-class peers.

Compensatory education can be provided at today's funding levels.

With proper training, many teachers would be capable of providing a compensatory education to low-SES students; but most currently lack these skills.

Most schools do not provide compensatory education to their low-SES students and as a result the low-SES students who could have benefitted from compensatory education students fail academically.

Academic failure, coupled with present mainstreaming policies, exacerbates in-school behavioral problems, further decreasing the school's ability to educate the low-SES students who have access to the standard education provided.

Though many correlations can be found between factors thought to be associated with low SES and student achievement, there is little reliable evidence that improving one or more of these factors leads to increased student achievement. The weight of the evidence actually suggests the opposite -- that improving SES and the factors associated therewith will have either no effect on student achievement or a small educationally insignificant effect.

Nor is there reliable evidence that improving these SES factors will improve the academic benefits conferred by providing compensatory education.

There may exist other benefits to be gained by improving SES of low-SES people, but those benefits do not affect academic achievement and so should not be included under the mantle of education.


Parry Graham said...

You say: "Though many correlations can be found between factors thought to be associated with low SES and student achievement, there is little reliable evidence that improving one or more of these factors leads to increased student achievement. The weight of the evidence actually suggests the opposite -- that improving SES and the factors associated therewith will have either no effect on student achievement or a small educationally insignificant effect." (italics added)

Would you mind pointing to some of that evidence? Meta-analyses, in particular, would be helpful.


KDeRosa said...

I'd start with

The Minnesota Transracial Adoption Study

The Minnesota Twin Family Study and Brouchard's Reanalysis

The Minnesota Texas Adoption Research Project and Willerman's various papers.

Also see THIRTY YEARS OF RESEARCH ON RACE DIFFERENCES IN COGNITIVE ABILITY (Rushton and Jensen) which discusses much of the extant research in section 12.

Parry Graham said...

Some additional research that I would recommend that appears to come to, if not different conclusions, at least more complex conclusions:




(Has a nice research/literature summary at the end) -- http://www.ncchild.org/action/images/stories/Poverty_Brief_final.pdf

KDeRosa said...

The Turkheimer study has a restriction of range problem. They separated out, and studied separately, the cohort of low IQ subjects and determined that nature played a smaller role in determining IQ than it did when compared to the larger group from which these subjects were drawn.

The ETS study is a bunch of correlation studies that mostly agrees with my point.

The Dahl study is new for me, but it appears to be another correlation study.

Parry Graham said...

I guess I’m still having a difficult time connecting the assertion that you made (“the weight of the evidence actually suggests the opposite—that improving SES and the factors associated therewith will have either no effect on student achievement or a small educationally insignificant effect”) with the research reports that you cite.

Please correct me if I am reading the studies inaccurately, but the studies you cite seem to focus on several points:
• IQ and personality/psychological traits have a strong genetic component that resists environmental factors
• There are genetically-based, immutable differences between the IQs of whites and blacks (a point with which I disagree, but that’s a topic for a different discussion)

Again, please correct me if I am wrong, but with the exception of one reference I found to a Willerman study, none of the studies you cite examine student achievement, which is the factor you consider in your assertion. In addition, I believe at least one of the studies suggested that improved environmental conditions did result in raised IQ scores during childhood (i.e., when a child would be of school-age years), but those increases faded once the children studied reached adulthood.

Whether one agrees or disagrees with the studies you cite, I think it is probably fair to say that the weight of research evidence does suggest a strong connection between genetics and IQ (the Turkheimer report suggests that the strength of this connection may be masked by factors associated with poverty, but I don’t think it challenges the general connection). But from my reading (again, please tell me where I am incorrect), none of the studies you cite (with the possibility of the one Willerman example) address the research question, “Do improvements in SES—or, probably more accurately, do improvements in many of the factors associated with low SES—have an effect on student achievement?”, which was the heart of your point.

KDeRosa said...


My reading of the twin studies is that they provide some evidence that cognitive ability has a genetic component whose proportion is increasingly expressed as a child reaches adulthood. The APA has stated that the genetic component is about 0.75 for white adults. That leaves about 0.25 attributable to environmental effects.

The Minnesota Transracial Adoption study looked at both IQ and academic achievement. See here under results and the Willerman study (The aptitude-achievement test distinction: A study of unrelated children reared together. Behavior Genetics, 7, 465-470.)found that the IQ results were generalizable to academic achievement.

I think those are the links you are looking for.

I am not claiming that environmental effects don't matter. They do. Severe malnutrition and lead poisoning, for example, can depress IQ. But these kinds of severe environemental affects are not widespread problems in the U.S. Poor kids may be hungrier than their middle-class peers using very loose definitions of that term, but they are not malnourished.

The fadeout effect starts well before adulthood. By age 7 the coefficient is down to 0.07 and it is at 0 by age 12. See here.

So I do think there is more support for, rather than against, the assertion that improving factors associated with SES will have little or no effect on student achievement.

I think that the largest SES factor is the low language environment that many low SES children grow up in as an artifact of the gentic component of IQ. Since their parents are more likely to have low IQs, which results in low education and levels and language usage among low-SES children. So I'd say that providing a high-language pre-K experience for these kids will provide the biggest bang for the buck if instruction is improved at the k-12 levels to capitalize on those gains. But no one has suggested this as part of any of the recent initiatives.

Parry Graham said...

I think I buy what I believe is at least part of your argument: “raising SES” as a policy proposal to address the under-performance of impoverished children is non-sensical. SES really serves as a proxy for a host of factors that may or may not relate to student achievement, and it would only be through successful attempts to improve on those factors that do impact student achievement that policy proposals might be successful.

For example, in the Barton study that I mentioned, student mobility and birthweight are mentioned as correlates to student achievement. Policy efforts to reduce mobility among impoverished families, or to improve the healthcare available to poor, pregnant women, might be examples of ways to support improved student achievement, in a round-about way.

Research suggests, and my own experience as an educator convinces me, that students that come to school on time, well-fed, well-rested, physically comfortable (e.g., not suffering from the pain of unaddressed cavities in their teeth), wearing glasses to help them see (if needed), and feeling safe (e.g., not worried about the threat of violence in their home or neighborhood) come to school more ready to learn than their peers for whom those conditions are not in place. To the extent that public policies or services can help to ensure those conditions are in place, I say “Good job”.

Now, if those students then show up at a school that provides ineffective instruction, all bets are off. And effective schools figure out ways to adjust when those environmental factors are not in place. Schools are responsible for controlling the factors that they control—which are appropriate and effective curriculum, assessment, and instruction—and their mission is to ensure that students are successful despite those environmental variables. But it does make the job easier when the kids show up ready to learn.

I know you are a big DI fan, and I am not opposed to the DI model. In my opinion, successful learning occurs when students are taught an explicit and appropriate curriculum, they are assessed on an ongoing basis, and instruction is tailored to meet students’ needs as evidenced by assessment results, and from what I can tell DI incorporates those elements. I am not a fan of what I see as DI’s inflexibility—my background in organizational behavior convinces me that rigid, overly-scripted approaches can obviate the opportunity for innovations and continuous organizational improvement—but there is no arguing that DI has achieved some impressive results, especially with impoverished and certain special education populations.

I remain unconvinced that improving SES “should not be included under the mantel of education”, but I agree that the real imperative is providing our children with the schools that they require and deserve.

Anonymous said...


Have you read Meaningful Differences? It sounds like you may have.

I agree with Parry that some students who attend effective schools could benefit from certain improvements in SES. Mobility is the biggest one -- they will be much better off if they do not move away from the effective instruction.

While vocabulary development for pre-k students would have measurable educational effects if followed up with effective instruction in elementary school, I do not think it is nearly the same priority as effective instruction for school age children.