August 29, 2007

SAT scores continue to drop

SAT scores continue to drop.

Verbal scores are now down to 1994 levels. This year's mean score on the verbal portion was 502, dropping one point. Math scores also dropped three points to 515.

Did you know that the SAT was recentered back in 1995 to disguise the disturbing fact that scores had plummeted since the high water mark of 1963. Verbal scores had dropped so low that the College Board, which administers the test, boosted scores by 80 points. They boosted math scores by 20 points.

So this year's scores of V502 and M515 equate to a scores of 422 for the verbal portion and 495 for math. Both scores are below the traditional average of 500. Verbal scores are now below the 1994 level. Thank you whole language and balanced literacy. Math is still about 18 points above 1994 levels thanks to calculators whose use is now permitted during the test.

Here is a plot of mean SAT scores from 1951/52 to 1993/94 using the old scale.


Something very bad happened in education between 1963 and 1980. Something we haven't recovered from.

The Times reports some of the excuses being made for the lower scores:

The declines for the class of 2007 were not caused by a single factor, College Board officials said. But the increase in the number of traditionally underrepresented minority and low-income students taking the test played a role, they said. So did a new requirement in Maine that all high school seniors take the exam, including those who would not in the past have viewed themselves as college bound.


Here's the minority group breakdown courtesy of the Times:

13 comments:

Stirner said...

Simply appalling.

How nice that the Times is willing to hint the cause of the score decline is due to broadening the base for college admission. Considering that the bulk of the decline happened between 1963 and the early 80's, that thesis is obviously incorrect.

Perish the thought that the decline had anything to do with the watering down of instructional and curriculum standards.

KDeRosa said...

The base did broaden during the 1963-1970's period. This is frequently blamed for the nasty drop. But a closer analysis of the data reveals that the drop also occurred at the very top and among white students as well.

I do agree that the most likely explanation is a watering down of standards and curriculum.

dweir said...

Could the drop have been an unintended result of the National Defense Education Act of 1958:

Of even greater significance, however, the act opened the way for future legislation that redefined many of the relationships between the federal government and the education community.

The politics of this act are fascinating, in a jaw-dropping, maddening sort of way. Taken from an interview with Stewart McClure:

On education priorities prior to the act:
"RITCHIE: When you started working on educational issues with the committee back in the 1950s,
what in general did you consider to be the major educational problems? What was it that you and the members of the committee were trying to address?

MCCLURE: Well it seems to me that the principal focus was upon two things--I'm leaving aside the
special types of education, vocational, and GI, and all that sort of thing--but the biggest subjects were school construction, federal money for school construction, and teachers' salaries. That's what the NEA
[National Education Association] wanted, and they were the most vocal lobby, of course, but so did
the chief state school officers and most of the educational lobbys, higher, lower, middle, or whatever.
They were emphasizing a subsidy for the public school systems of the country, which basically needed
more buildings because we were in the baby boom of the postwar era, and higher teachers' salaries,
because at the time teachers were very poorly paid. Garbage collectors, policeman, and so on were getting much more out of
municipal and state budgets than were teachers, so they had a good case, from their own point of view
and from the public point of view, too. You can't educate children with dumb teachers, and you can't
hire teachers who are paid so low they either don't want to work or they could do something else and
do so. That was the main thrust."


On opposition to increasing federal funding:
"But beyond that they clung to some mythical constitutional
principle: the last thing that could happen in the United States was for the federal hand to be laid on local education, which belongs to the hands of the school boards and local council of education or
whatever they're called--which, of course, are all controlled by the Chamber of Commerce. In the real
world the business community dominates the school boards in every damn town in the country."


So, while there were some elements of the funding that went to math and science education, this was really just a political opportunity to open the federal funding floodgates:
"Yes, I think if there was one thing I ever did in my work on the Hill, my work for my whole career, it was to focus Lister Hill's attention on the opportunity which Sputnik, this Russian satellite, gave al I of us who were struggling, and had been for decades, to establish a federal program of monetary aid to
public education, and private, too, in some instances. And I'm really very proud of that. Of course,
someone else could have come along with the same idea, and probably did a week later, but I was first."

SusanS said...

Many gifted (and otherwise) middle schoolers take the SAT and ACT for talent searches, or even just practice. There is no ceiling for them like on the state tests, so one can get a clearer picture of their skills. I'm prepping my 12-year old for the ACT at the moment.

Have those kids been removed or are they included with the others?

Anonymous said...

---"But beyond that they clung to some mythical constitutional
principle: the last thing that could happen in the United States was for the federal hand to be laid on local education, which belongs to the hands of the school boards and local council of education or
whatever they're called--which, of course, are all controlled by the Chamber of Commerce. In the real
world the business community dominates the school boards in every damn town in the country."---

WOW. I learned something today.

Who thinks that JFK's promise to support legalizing the public employee unions (in return for key support in the Democratic primary) caused the decline in SAT scores?

Few people today would say the chamber of commerce "dominates" their local school board.

But somebody does dominate many school boards today. (Especially in areas where there are lots of low income students of color who perform poorly on the SAT.)

The teacher's union.

JFK feared at the time that he was helping to create a monster.

Is there a way to get the Chambers of Commerce back in charge? :)

CrypticLife said...

Hi Ken,

I'm a bit perplexed on this post juxtaposed against some of your other comments indicating that education in the US hasn't degraded, it's just always been bad. Can you clarify?

KDeRosa said...

Crypticlife, the nature of the SAT and it's self selceting population make it difficult to draw conclusions. Notice how most of the decline occured prior to the NAEP coming online.

le radical galoisien said...

So my 740 was 80 points higher than it should be?

Awww. I thought I was doing well. *mumbles*

I Reference My Stats said...

The statement "Verbal scores had dropped so low that the College Board which administers the test boosted scores by 80 points. They boosted math scores by 20 points" is incorrect. What they did was recenter, not just add points. See http://www.collegeboard.com/about/news_info/cbsenior/equiv/rt021021.html. So some people's math scores at the lower end are actually LOWER in the new system.

As for your claim of a drop in SAT scores between 1963 and 1980 meaning that something very bad happened that we have not recovered from, consider this. In 1963, 45 percent of high school graduates enrolled in college, in 2005, 68.6 percent did. (http://nces.ed.gov/programs/digest/d06/tables/dt06_186.asp?referrer=list) I call that a success.

KDeRosa said...

I did reference that the scores had been recentered which resulted the addition of points to the mean score that indicated.

The data is here.

I could have been clearer by pointing out specifically that the additional points occurred at the mean. But, this is a quick blog post not a thesis paper. But my general statement was more acurate than your misleading statement concerningthe left tail of one score.

For the verbal scores the amount additional points are the same along the scale. For math scores, the additional points come at three areas of teh curve, some scores (like mine) got no additional points, while some at the very bottom had points subtracted.

The recentering, however, was the result of a massive drop in scores that ocurred from 1963-1980. Some of that drop can be explained by demographic changes; however, not all of it can. (I alluded to this in a comment.) I've seen estimates that as much as 75% of the drop might be attributed to demographics. What about the other 25% an dwhat about the drop at the very top? Where did those kids go? Or what wheren't they taught?

I am not so impressed by the increase in college enrollments. Most of those marginal enrollees won't ever graduate. Many that do only graduate because there are majors, like education, in which the graduation requirements are close to non-existent.

Sherman Dorn said...

Something very bad happened in education between 1963 and 1980.

Since I'm not sure what the h*** the SAT has ever measured, other than testing well on the SAT, this statement is unwarranted. Even if you think that the SATs are decent evidence of how FTIC students will perform their first year, the SATs has never been validated as evidence of anything in K-12 education.

KDeRosa said...

Hi Sherman.

Smoking has never been validated as a cause of cancer either (at the last time I checked);however, the correlation data is sufficiently robust that we feel confident assuming that smoking is a cause of cancer until the validation data comes in.

Similarly, SAT scores correlate highly enough with IQ scores that I'm not worried about making that statistical leap until the causation data comes in.

And, unlike, say, Ravens Progressive Matrices, the SAT is testing a student's knowledge of academic content as its proxy for measuring IQ.

So, while non-academic content tests of IQ has shown IQ to be rising (Flynn effect), the SAT has shown the opposite. The difference is how well students are learning academic content which supposedly is being taught in schools.

A logical leap, yes, but no greater than the logical leap you guys in FEA are proposing for measuring academic performance.

Mark Montgomery said...

Great post, Ken.