A decade of test-driven instruction has not raised SAT scores.
Please note the tight correlation between family income and SAT scores. Some people call the SAT the Family Income Index.
FairTest
National Center for Fair & Open Testing
SAT “COLLEGE READINESS” SCORES DECLINE AGAIN DEMONSTRATING FAILURE OF TEST-DRIVEN K-12 SCHOOLING;
MANY MORE COLLEGES IGNORE SAT/ACT EXAM RESULTS AS TEST-OPTIONAL ADMISSIONS MOVEMENT SURGES
SAT scores for high school seniors dropped again this year continuing a ten-year trend, according to data released today. SAT averages declined by 28 points since 2006 when the “No Child Left Behind” public school testing mandate went into effect. Score differences between racial groups increased, often significantly, over that period.
Bob Schaeffer, Public Education Director of the National Center for Fair & Open Testing (FairTest), said, “Test-and-punish policies, such as ‘No Child Left Behind’ have clearly failed to improve college readiness or narrow racial gaps, as measured by the SAT. Average SAT Scores declined since 2006 for every group except Asian-Americans. The ACT admissions exam and the National Assessment of Education Progress show similar trends.”
Schaeffer continued, “Fortunately, many more college and universities are recognizing that standardized exams fail to measure key factors for academic success. Over the last twelve months, 27 more schools have dropped ACT/SAT requirements for all or many applicants. New test-optional institutions include George Washington, Drake, Allegheny and Virginia Commonwealth.” A list of 850 institutions that de-emphasize admissions tests is posted at http://www.fairtest.org/university/optional.
| READING | MATH | WRITING | TOTAL | |
| ALL TEST-TAKERS | 495 (-8) | 511 (-7) | 484 (-13) | 1490 (-28) |
| Female | 493 (-9) | 496 (-6) | 490 (-12) | 1479 (-27) |
| Male | 497 (-8) | 527 (-9) | 478 (-13) | 1503 (-30) |
| Amer. Indian or Alaskan Native | 481 (-6) | 482 (-12) | 460 (-14) | 1423 (-32) |
| Asian, Asian Amer. or Pacific Islander | 525 (+15) | 598 (+20) | 531 (+19) | 1654 (+54) |
| Black or African American | 431 (-3) | 428 (-1) | 418 (-10) | 1277 (-14) |
| Mexican or Mexican American | 448 (-6) | 457 (-8) | 438 (-14) | 1343 (-28) |
| Puerto Rican | 456 (-3) | 449 (-7) | 442 (-6) | 1347 (-16) |
| Other Hispanic or Latino | 449 (-9) | 457 (-6) | 439 (-11) | 1345 (-26) |
| White | 529 (+2) | 534 (-2) | 513 (-6) | 1576 (-6) |
*High school graduates in the class of 2006 were the first to take the SAT “Writing” Test. The “No Child Left Behind” mandate to test every child in grades 3-8 and at least once in high school went into effect in the 2005-2006 academic year.
2015 COLLEGE-BOUND SENIORS SAT SCORES BY FAMILY INCOME
| READING | MATH | WRITING | TOTAL | |
| $0 – $20,000 | 433 | 455 | 426 | 1314 |
| $20,000 – $40,000 | 466 | 479 | 454 | 1399 |
| $40,000 – $60,000 | 488 | 497 | 473 | 1458 |
| $60,000 – $80,000 | 503 | 510 | 487 | 1500 |
| $80,000 – $100,000 | 517 | 526 | 501 | 1544 |
| $100,000 – $120,000 | 528 | 539 | 514 | 1581 |
| $120,000 – $140,000 | 531 | 542 | 518 | 1591 |
| $140,000 – $160,000 | 539 | 551 | 526 | 1616 |
| $160,000 – $200,000 | 545 | 557 | 534 | 1636 |
| More than $200,000 | 570 | 587 | 563 | 1720 |
Calculated by FairTest from: College Board, College-Bound Seniors 2015: Total Group Profile Report and College-Bound Seniors 2006: Total Group Profile Report
Diane Ravitch's blog 
I’m shocked, just shocked!
Could any of this be an effect of more students taking the test with the additional students being the less academically prepared students. Perhaps historically, only the top 20% took it because they were the only ones going to college. Now with the “everyone prepares for college” idea, more students are taking the test. And I believe in some states, ALL students take the test.
Either way, the NCLB testing mania clearly has not had the desired results.
Alice in PA,
Yes, changes in the population of students taking the SAT would explain the drop in scores. It might even be the case that NCLB did have the desired result but that was swamped by changes in the population of SAT test takers.
TE, the same arguments were made in the mid-1990s, in response to a continuing drop in SAT scores.
Before attributing the latest decline to changes in participants, the case must be demonstrated that there were meaningful changes in the test-taking pool.
Once again we agree, though I would emphasize the converse: before attributing the change of test scores to changes in high school instruction, the possibility of changes in the composition of test takers must be ruled out.
Don’t you agree?
TeachingEconomist, this is exactly why the “average” scores dropped. Review the reports and then determine for yourself why scores dropped. Folks, if you want to cheat, keep reading below.
– Group report for 2015
– Group report for 2014
And the takeaway:
1. White and asian students’ scores either stayed the same or increased from FY14 to FY15
2. Hispanic students’ scores dropped
3. The % of white students dropped while asians remained constant.
4. The % of hispanic students increased. Note that likely more poor hispanic studdents too the test (illegal immigrants) relative to the children of well-educated Hispanics thus causing the overall score drop here.
Basic critical thinking and math tell you this is a “mix” issue and not a “scoring” issue. It’s very revealing that our professional opt-out activists don’t have the skills to conduct this basic analysis.
The Washington Post piece says the lower the participation rate the higher the scores, so it sounds like you’re right.
If SAT scores really do track family income, maybe the scores should be flat. Maybe that’s entirely reasonable. Family income has been flat for more than a decade. There’s only so much schools can do. At some point the effects of a declining middle class will ripple and “interventions” in schools won’t be sufficient to stop the slide.
Yes, Alice, that is what’s happened, but it doesn’t fit FairTest’s narrative: http://www.vox.com/2015/9/3/9257121/sat-scores-2015
Even if it is a mix issue,the NCLB rheeformers have not been the magic bullet they claimed to be. Those gaps due to poverty still exist and are getting wider. The testing regime does not work.
Alice, if you don’t accept the theory that “mix” is responsible (the data provided by the College Board doesn’t tell us enough; all we know is that ~200,000 more kids took the test in 2015 than 2006, and that America’s public schools have higher proportions of black and Latino children than they did a decade ago), then how do you explain the improvement in Asian scores? Are Asians somehow avoiding the “testing regime,” or did many of their parents become wealthy between 2006-2015?
According to my hardcover first edition of “Reign of Error, if you go by measures like disaggregated NAEP scores and graduation rates, the state of K-12 U.S. public education has never been stronger.
Tim,
Test scores of US students have never been higher than they are today, as judged by NAEP. The period of greatest gains occurred before the enactment of NCLB. After NCLB, the rate of progress slowed.
Perhaps the high NAEP exam scores were caused by NCLB.
TE, the steadily rising NAEP scores over the past 40 years were NOT caused by NCLB. Nor were the NAEP scores increases in the 1990s. The biggest closing of achievement gaps was in the late 1970s. That wasn’t caused by NCLB either.
This is the kind of ill-informed, provocative remark that wastes my time and tempts me to ban you again.
The last time was for a personal insult.
But please find something else to occupy your time.
I seldom ban anyone.
You know that this site supports public education, and you haunt it just to taunt readers with pro-privatization, pro-testing comments.
Why?
Dr. Ravitch,
It seems to me that there needs to be an explanation for why we see detrimental effects of NCLB on SAT scores but we do not see these detrimental effects on NAEP scores. One obvious explanation for the lower SAT scores is that the mix of students taking the SAT exams has changed while there has been no change in the mix of students taking the NAEP score.
If all this reform has had a detrimental impact on education, shouldn’t we be seeing a drop in NAEP scores?
TE, Test prep produces higher test scores. Even on the SAT, the students whose families can afford to pay tutors get higher test scores.
Please, give it up. You are on the wrong blog. Very few people here–most teachers, retired teachers, and parents–believe in the sacred value of test scores.
Just you and Virginia.
We probably should take into account the amount of test prep that is done. Back when the idea that SAT was an aptitude test, test prep mostly meant familiarizing yourself with the test format and protocol. (One of my kids lost points because he did not read the SAT protocol that told him that he would be penalized for wrong answers but not for no answer. Being a diligent problem solver he attempted everything whether he was sure or not.) Now, we all know that intensive test prep can raise your score significantly which kind of puts a kabosh on the aptitude argument.
Even though I’ve explained this multiple times, 2old2teach shows she is too old to understand basic probability. This concept was my sample lesson when I interviewed with Kaplan back in the early 1990’s.
Trying to solve or guess a question you don’t know does NOT hurt your score when there are penalties for guessing. If there are 5 possible answers, then you receive a 1/4 pt penalty to wrong answers. So the expected value is:
80% you get guess wrong and get -1/4 pt
20% you get guess correct and get +1 pt
Net expected value = 0 pts or NO penalty. See my other responses to see incremental gains by eliminating 1+ answers.
Besides not thinking about the best strategy ahead of time, I submit this proves both you and your student failed the critical thinking portion of the SAT. This is simple logic. You not only need to be able to apply skills but to understand what information is important. In this example, you need to know the penalties for guessing. Your (and others) misunderstanding of this issue and the inability of the mass public to comprehend it caused the College Board to change this policy. Now, the tests are less fair because you get penalized if you don’t guess on every question you don’t know.
Frankly, I think this is also due to an overall demoralization and lack of confidence in the whole system, including awareness of the burden of student loans.
+1
Of course, much of that is courtesy of the deforms. Needless to say?
Another cause could be the narrowing of the curriculum. At one time the SAT was a broad knowledge exam. In addition, test-taking skills (knowing how tests are constructed) may not be understood by young teachers and so not taught.
Actually, SAT never was a broad knowledge exam. It has been about critical thinking, reasoning, problem solving.
But those are the very things that have been driven from curriculum by the low-level test driven classrooms.
There is a lot less requirement for kids to think as teachers are required to handout “study guides”.
The SAT decline reflects pretty well the dumbing down of schooling, public and private.
hmmm…could reported downward trend be a manipulated statistic to support the college board’s financial interest (a la Pearson) in prep courses, common core and additional education ‘reform’ measures?
Say it ain’t so, Sozo! Say it ain’t so!
P. S. Another possibility: A statement on my PSAT results claimed the Test was a reading test. Wide reading may be on a decline as STEM predominates.
Steven Krashen has written a great deal about the positive power of recreational reading on reading skills. Today’s students may be reading more STEM, or maybe they are spending more time on social media so they don’t take time to read for pleasure.
As the parent of high school sophomore, I have started receiving mail from SAT/ACT test prep centers. The recommendation from Huntington Learning Center (from their brochure) for freshman and sophomores for the 2015-16 school year: “Options: Redesigned SAT; ACT. Recommendation: Consider the ACT during your junior or senior year since the SAT will continually be changing.” So even the test prep tutoring businesses are steering students away from the SAT! I find that very interesting. I also found it very interesting that the College Board denied my son’s special education testing accommodations that have been part of his IEP for 10 years, and asked me to provide additional medical and psychological evidence of his disability, and a rationale for his use of assistive technology that he has used everyday in school since he was in the 2nd grade. After fighting them for a few weeks, I saw sample questions from the new SAT, and decided that he is not taking the test. Many colleges are dropping the requirement, his accommodations would mean the test would take 2 days, and the questions are the same type of CCSS nonsense that appear on the NYS standardized tests and new CCSS NYS regents exams, designed for high failure rates. Since we would have to pay for the SAT, I have decided to save my money, and save him from the torture and heartbreak these tests would bring, as well as the time consuming and expensive test prep (that now does not seem to exist for the new SAT). Thank you Fair Test! And thank you Diane!
That said, the entire testing system itself in question along with recent changes, due to its leadership.
Bad news only strengthens the hand of the privateers.
So much mental mathturbation* so little time!
*Gracias a SDP!
in one breath, Shaeffer says
“Test-and-punish policies, such as ‘No Child Left Behind’ have clearly failed to improve college readiness or narrow racial gaps, as measured by the SAT.”
And in the next he says “Fortunately, many more college and universities are recognizing that standardized exams fail to measure key factors for academic success.”
I think he needs to really think about the point he is trying to get across.
In talking, with Bob, a nice enough fellow, at the last two NPE conferences I couldn’t convince him that we need to reject the standards and testing regime which his group Fair Test tracks. It seems that he still believes we can “measure” the teaching and learning process. I’ve suggested to him to read and comprehend Wilson, but don’t know if he has done so.
Like so many people who struggle to throw off the thoroughly ingrained cultural habitus that is the supposed measurement of the teaching and learning process. I guess if everyone but a few nutcases like Swacker and Wilson believe in it, it must be okay. Ay ay ay ay ay!
Error in chart? $40,000 – $60,000 488 597 473 1458
$60,000 – $80,000 503 510 487 1500
Math scores for lower income students at 597 are higher than the next step up in family income, indeed they are higher than the 587 math score for kids who have family income of $200,000. I hope Fairtest will look again at their data entry.
LAURA, thank you for catching the error in the family income chart.
The correct math score for students in the $40,000-60,000 group is 497, not 597. FairTest corrected its press release.
Is student family income calculated when we evaluate teachers based on student test scores? I don’t think so.
Mack,
Family income was certainly included in the Chetty paper.
TE, Chetty’s paper used data collected when there was no high-stakes testing. Imagine: my third-grade teacher determined whether I might get pregnant 10 years later as well as my lifetime income. Who knew?
Dr. Ravitch,
I will try again to give a response that you will approve. Two points: 1) Mack Hick doubted that student SES status was taken into account when evaluating teachers by using changes in student standardized test scores. Student SES status was taken into account in Chetty’s work. This is easily verified by reading the paper. 2) one of the points of Chetty’s research is to see if good teaching has impacts outside the school. He found evidence that this was true. Why teachers want to discount and ridicule this evidence is a mystery to me.
TE, last response. I have things to do today that don’t include corresponding with you. No one on this blog says that teachers don’t matter. No one on this blog, including me, says that teachers don’t have impacts on children’s lives and on their future. What most people here say is that those impacts are hard to measure and that standardized tests don’t measure them. Period.
The Chetty study is an attempt to give evidence that having good teachers in school have a positive lifelong impact on students. Those positive impacts are hard to measure, but surely not being a teenage mother is evidence of that a teenage girl is making good life decisions.
TE, isn’t it amazing? What a third grade teacher did to raise standardized test scores “caused” a young girl of 17 to decide not to get pregnant. If you believe that, I have a bridge to sell you.
Dr. Ravitch,
Once again, we are looking for evidence that good teaching has an impact on students long after they leave the classroom. If you don’t think that lower teenage pregnancy rates for young women is evidence of the beneficial impact of good teaching on students lives, I really don’t know what you would consider evidence.
Perhaps it would be helpful if you could give the evidence that leads to your belief (or anyone else here) that teachers have positive lifelong impacts on students lives. Something more than anecdotes.
NY’s model takes into account whether students’ families receive economic assistance. A binary “yes/no” variable, not actual family income.
New research shows that poor kids have 9% less of (vital) grey matter in the brain. Dana Foundation
Maybe because it’s a common core based test and none of these kids grew up using the common core standards. They all got maybe 1.5 years of CCSS, just saying…
In the interest of not being called all kind of names by your readers, I will withhold my link on this one. But there is definitely a reason why the college board stopped showing “crosstab” results by race and income/education. In other words, you will see scores split by race. You will see scores split by income/education. But you will never, ever see scores split by both race and income in the same chart anymore. I have a link to the last time it was done.
Hint, when low-income students in one group score higher than the top-income students in another group, you realize Diane’s explanation above has absolutely nothing to do with anything. But it was a nice try Diane.
I do have a day job but you can’t possibly be putting out all these stories by yourself, can you? That would be unbelievably impressive.
Virginia, my day job is writing. I write everything on this blog except when I attribute to the author. What is your day job? Commenting here?
Diane, that is really impressive. You must literally generate 2-3x more content than anybody else out there. I would have sworn you had ghost writers. Oh well.
Commenting on here is hobby #3. Day job is consulting (not always done during 9-5 though) on performance metrics and cost accounting.
My current #1 hobby is holding schools accountable when they censor the public’s or teacher’s comments. Seriously, what kind of district implements this policy? Sounds similar to those districts who punish teachers for speaking out against CC, eh?
Hobby #2 is getting SGP data from the state of Virginia so we can make cool briefs. More briefs a’coming.
Golf has been a distant 4th this year. And the scores show it!
Alice in PA, poverty is a side effect, not the cause of the low AP scores. If you folks keep saying that, I’ll post some links that demonstrate that for certain.
retired teacher, you are correct that teachers shouldn’t be held accountable for SAT scores. You are also correct that kids of $200K income households will score higher than kids of $30K income households on average regardless of the teacher. You are incorrect in thinking that VAMs are based on aptitude tests like the SAT or that VAMs don’t take into account the income differences.
The income levels of the students determines the performance on the SAT as well as other standardized tests. We have been saying this all along. Somebody should start looking at facts for what they are, and not what they project to be. Do they think that the teachers of the students whose families earn over $200,000 per year are far superior to those of students whose families earn less? If so, they should send a group of teachers from the top tier scorers into the poorest areas. Let’s see how they make out, if they survive, that is.
“The income levels of the students determines the performance on the SAT as well as other standardized tests. ”
Income levels correlates with performance. Income does not determine performance. It is an important distinction. You do not have to come from a family whose income is $200,000 to score well, and you will not score well simply because of high family income. There are a lot of factors contributing to strong performance. I have a strong feeling that a family’s belief in the value of an education and whether they are able to successfully act on that belief have more to do with future success than family income.
You’re correct; I didn’t state that too well. My point is that the results confirm what we already know. There’s a danger in reading too much into it. ie it’s the teachers’ collective fault. Maybe the answer is as simple as the raw numbers of students taking the test are higher. The higher the percentage taking the test usually means the scores will be lower. They should post the numbers taking the test in successive years as well.
As one who has taught college 20+ years ago, as well as more recently, I can tell you that the quality and level of preparedness of incoming students seems to have dropped substantially. This for state college and/or junior college institutions that typically allow entry for most students who fulfil basic requirements. In particular, since I teach usually a basic Physics class, of the kind required for many biology, med., architecture, industrial engineering, kinesiology etc. majors (ie, non engineering/physical science majors), the level of math incoming students display on diagnostic tests would indicate the need for remedial teaching for about half of them. Instructors teaching for many more years than myself have confirmed a steady decline – especially among the typically white but non-privileged students who populate these schools. The stories i heard from high school teacher in Texas for example, are beyond distressing. And it’s not only that teachers/ professors have to fight the “competition” from ubiquitous electronic devices used by most students to ‘tune out” rather than opt-in. It’s the attitude that’s most distressing, as wel as the crazy emphasis on endless testing which is only good for passing the time, rather than improving knowledge base or skills or general competence.
The students I encountered in texas (as compared for example with California) gave more than a little cause for concern, These students probably represented the average SAT/ACT cross-section of the population rather than the top performers who are found at eg. Rice University and the like. What these students didn’t know – and I mean some pretty basic stuff – should give a pause to anyone worried about American competitiveness.
The recent SAT drop merely conforms what I already suspected. I can’t help but note the disproportionate increase among Asian Americans. that correlates with the ubiquity of these students in the physical science, computers and engineering departments of all the universities in the country. perhaps it is very much time to stop the silly arguments among people for whom education is a profession in and of itself and start listening to those who have to actually prepare students for jobs and occupations in the real world.
Sorry for the long comment. Alas, I didn’t even skim the tip of the iceberg, but better stop here.