## Friday, October 14, 2011

### Huge Gender Differences Persist on SAT Math Test

Math SAT Percentile Ranks: Males and Females, 2011
Score   Percentile    Males Females   M-F Ratio
800999,1204,6831.95
790997714791.61
780993,1141,4722.12
770986,8244,0631.68
760985,6483,2631.73
750974,5762,4401.88
740967,0114,7141.49
730965,1713,2371.60
720956,1344,0341.52
7109410,0376,8331.47
7009310,8097,4601.45
6909212,5288,7231.44
6809013,88610,0671.38
6708914,65710,6201.38
6608714,55411,4701.27
6508615,03111,9601.26
6408416,87813,8051.22
6308216,39614,0331.17
6208017,22314,7281.17
6107727,25124,8571.10
6007518,34217,0281.08
5907317,59117,1721.02
5807021,62221,0381.03

The College Board recently released detailed data on percentile ranks by gender for the 2011 SAT tests, and data for the 2011 SAT math test are presented in the table and chart above.

1. For all math SAT scores of 580 and above (70th percentile), male students outnumbered female students (see table above).

2. As the scores increased by 10-point intervals from 580 to 800, the male-female ratio increased in almost all cases, reaching a peak of 2.12-t0-1 for scores of 780.  For perfect scores of 800, the male-female ratio was 1.95-t0-1.

3. More women (873,896) than men (768,554) took the SAT test in 2010, and to adjust for those differences in sample sizes, we can calculate that 1.19% of males had perfect 800 scores compared to 0.54% of females, for an adjusted male-female ratio of 2.21-to-1 (vs. the 1.95 unadjusted ratio).

4. For top math scores between 750 and 800, there were 183.25 males earning those high test scores for every 100 females (see chart above).

These results are especially significant because female high school students are generally better students overall than males, and equally or better prepared for the SAT Math test than male students:

a. Females outnumbered males in the top 10% of their 2011 graduating classes - there were 127 female students in the top 10% of high school classes for every 100 male students (56% female to 44% male).

b. Nationwide, there were 144 female high school students with GPAs of A+ for every 100 males (59% female vs. 41% male).

c. Females had a higher average GPA of 3.40 compared to 3.27 on average for male students in 2011.

d. More high school girls than boys took advanced AP/honors math classes in high school (117 females per 100 males) and science classes (122 females per 100 males).

e. More high school girls (54%) than boys (46%) studied mathematics for four years in high school, and those percentages were the same for students studying four years of high school science.

Bottom Line: Despite being better prepared academically by many different measures, both overall and for mathematics specifically, female high school students score significantly lower on the SAT math test, and the 30-point differences in test scores favoring males has persisted since Richard Nixon was president.  At the high end of math performance, high school males significantly outperformed their female counterparts on the 2011 SAT math test - for test scores of 750-800 there were more than 183 males for every 100 females achieving at that level.

Despite the huge differences in math performance by gender (both on average and at the high end of performance), we frequently hear statements like this: "There just aren't gender differences anymore in math performance," says Universityof Wisconsin-Madison psychology professor Janet Hyde, "So parents and teachers need to revise their thoughts about this.  Stereotypes are very, very resistant to change, but as a scientist I have to challenge them with data."

Given the significant and persistent gender differences in SAT math test scores that have persisted over many generations, the scientific data about gender differences in math performance would seem to present a serious challenge to Professor Hyde's claims of no gender differences in math performance.

At 10/14/2011 8:21 AM,  Reliapundit said...

TELL LARRY SUMMMERS.

At 10/14/2011 10:52 AM,  Mike said...

Oddly, I had a conversation with the parent of 2 girls about this last night....after discussing SAT scores with my wife last weekend.

Is there any way to see and compare practice SAT scores? Or non-timed tests of the same caliber?

Obviously, this is purely anecdotal, but the one thing that was brought up (completely coincidentally) was that both are very good at math...unless they are timed. They seem to have a tendency to stick on one problem that has them stumped, rather than move on and come back. My first thought was that this is sports related.

At 10/14/2011 11:44 AM,  AIG said...

She is a psychology professor...and she thinks she's a "scientist". How cute!

It would be interesting to see similar comparisons from other countries. It may be some cultural trait in the US (I've noticed that girls here are brought up to be "cute" like the psych professor above a lot more than in some other countries)

At 10/14/2011 12:25 PM,  \$9,000,000,000 Write Off said...

You could've said the same thing about medicine, law, politics, etc at earlier points in U.S. history. A 30 year SAT profile in this country is a challenge, but the serious challenge to your assertion is that other countries don't have the math or the moron/genius distribution gender profiles as the U.S.

At 10/14/2011 12:26 PM,  Anonymous said...

"Despite being better prepared academically by many different measures, both overall and for mathematics specifically, female high school students score significantly lower on the SAT math test, and the 30-point differences in test scores favoring males has persisted since Richard Nixon was president."

If the high school females were better prepared academically by many different measurements, either the measurements were inaccurate or invalid for the SAT. Are the academic measurements designed to measure how well a student will do on the SAT? If so, the test preparation is not working. If not, why is anyone using them to measure something they were not designed to measure? It looks like it is back to the drawing board either way to me.

Of course, the SATs should show a high correlation between those scores and college mathematic achievement to make any difference in the first place. Do the makers of the SAT explicitly state and prove that claim with longitudinal data?

At 10/14/2011 12:40 PM,  Buddy R Pacifico said...

From the Science Daily article:

"Again, the effort uncovered little difference, as did a comparison of how well boys and girls did on questions requiring complex problem solving. What the researchers did find, though, was a disturbing lack of questions that tested this ability. In fact, they found none whatsoever on the state assessments for NCLB, requiring them to turn to another data source for this part of the study."

Questions involving complex problem solving would seem to be a test of reading comprehension - not math problem solving. So, the longer the narrative of the question, the more it skews towards females.

Thus, the answer to STEM inequality is longer (drawn out) question narratives for females on tests.

At 10/14/2011 12:49 PM,  \$9,000,000,000 Write Off said...

By the way, you are ignoring the larger evidence by focusing on SAT.

People solve math problems differently, deploying spatial reasoning, venn diagrams, geometry, etc. At a general level, the genders use different methods and the female methods are slower while the SAT privileges speed. (Also, note that the SAT disparity is declining since Nixon.)

Better evidence of math aptitude comes from presenting math problems without time pressure. That's exactly how high school and college mathematics are tested. And what are the results? Girls take as many math classes as boys in HS, but get superior grades. Girls get as many bachelor degrees in mathematics, with equal grades.

Faced with this evidence, the "girls aren't cut out for science" crowd, like all of their predecessors, have moved on to a new scientistic argument, the one made by Larry Summers, the distribution: Girls may be equal to boys at the general science/math level, but girls aren't geniuses like boys. Those gender distribution disparities don't hold in a number of European countries, so its hard to call that an intrinsic quality in gender.

I'm not math aptitude and gender is definitive, and counters can be made to everything I've written. But we know enough about historical gender bias and its pseudo-scientific claims, to be skeptical of "the science is settled" arguments like yours.

I write this a reader who is generally sympathetic to your posts.

At 10/14/2011 1:00 PM,  Anonymous said...

Mike,

Our teamwork training classes show women usually need collaboration when they are unsure of an answer to a question while men will usually try to go it alone even if they actually help.

In the real world, which would you rather have, someone who knows she is not positive about the answer and seeks help or someone who thinks he is right but is not? Maybe men are just luckier gamblers who are willing to inherently assume more risk that pays off in timed tests given to them individually.

Yeah, men and women are different, but aren't you glad we are?

At 10/14/2011 1:12 PM,  AIG said...

Write Off, I think you are drawing from 3 different distributions in your arguments. The first is the distribution of kids who take the SAT. The second is the distribution of all HS kids. The third, is the distribution of kids who graduate with bachelor degrees in math.

All three are pretty different. The first and the second aren't the same. A larger % of boys either drop out of HS, or don't take the SAT. The third, of course, is unrelated.

I do agree that the SAT isn't the best measure of gender abilities in math. But its as close as we can get.

PS: The time argument doesn't seem convincing to me. Given enough time, a lot of us can solve even the most complex problem on the SAT, given that its a multiple-choice exam, and thus getting any of the answers on the paper...even through trial and error...will improve your odds. But someone getting the answer in 10 seconds isn't equivalent to getting the answer in 10 minutes.

At 10/14/2011 1:14 PM,  AIG said...

Plus, isn't getting the right answer faster...a measure of competency in math?

At 10/14/2011 1:21 PM,  uclalien said...

"Females had a higher average GPA of 3.40 compared to 3.27 on average for male students in 2011."

I'm a little surprised that no one else has commented on this quote. It implies that the average GPA for all high schools students is in the ball park of 3.34.

I've heard a lot about grade inflation, but this is pretty damning evidence.

At 10/14/2011 2:40 PM,  Larry G said...

I would not worry about US males verses females...

look at how we compare internationally:

Rank Countries Amount
#1 Japan: 557
#2 Korea, South: 547
#3 New Zealand: 537
#4 Finland: 536
=5 Australia: 533
=7 United Kingdom: 529
=7 Switzerland: 529
#9 Belgium: 520
#10 France: 517
#11 Austria: 515
=12 Denmark: 514
=12 Iceland: 514
#14 Sweden: 510
#15 Ireland: 503
#16 Norway: 499
#17 Czech Republic: 498
#18 United States: 493

http://www.nationmaster.com/graph/edu_mat_lit-education-mathematical-literacy

At 10/14/2011 5:58 PM,  Anonymous said...

"I've heard a lot about grade inflation, but this is pretty damning evidence."

Grades can't be inflated unless you have a reference point to measure from or grade on a curve with a stated distribution. Any increase from the past could just mean the prior numbers were too low or other factors have changed.

At 10/15/2011 1:28 AM,  uclalien said...

"Any increase from the past could just mean the prior numbers were too low or other factors have changed."

Of courses "other factors" have changed. Many schools/courses now operate on a 4.5 or 5.0 grade point scale.

At 10/15/2011 10:18 AM,  Anonymous said...

AP courses would have to be taken out of the newer data because they did not exist in the old data as well as a multitude of other changes over time because that is what is causing the GPA scale change.

Additionally, targets would need to be defined to what a "normal" GPA should be.

Next, the mix of everyone receiving a GPA but not everyone taking the SAT would have to be addressed.

And finally, the same students who graduate with a mathematics degree would have to be compared to their SAT scores in math.

Grade inflation? Maybe. But the evidence I have seen so far does not convince me of a GPA process variance over time that is inflated for students’ benefit that I could put my name on in a Six Sigma report.

At 10/17/2011 3:09 PM,  yoyodyne said...

How many women have won the Fields medal? Or Nobels in math-heavy sciences?

There are more men at the extremes of math, just like in Chess, just like in Music. There's no female Mozart. There's no female Bobby Fischer or Kasparov or Tal or Anand -- in fact there's only been one female GM *ever* who could come close to those guys.