## Monday, November 23, 2009

### More on the Gender Gap for SAT Math Test Scores

A previous CD post showed that high school boys outperform high school girls for high SAT math test scores (see chart above, data here). For perfect 800 scores on the 2009 math SAT test, the ratio of boys to girls is 2.22 to 1 (6,928 to 3,124). What makes this outcome even more interesting are the following data from this SAT report from the College Board:

1. Girls are over-represented in the top 10% of high school students by GPA, which is 57% female and 43% male. Girl also outnumber boys in the second tenth of students by high school rank: 54% to 46%.

2. Girls outnumber boys for GPAs of A+ (60% vs. 40%), A (61% vs. 39%), A- (57% vs. 43%), etc.

3. The average number of years of math study is almost identical: 3.9 years for boys and 3.8 years for girls.

4. For students reporting more than four years of math study, the percentages are equal: 50% of boys and 50% of girls.

5. Both 50% of boys and 50% of girls report that calculus is the highest level of high school mathematics taken.

6. More girls than boys took AP Honors math courses, by a ratio of 117 girls for every 100 boys.

Therefore, it would seem that girls are equally prepared, if not more prepared (more AP math classes), than boys for the SAT math test, and yet boys outperform girls measured both by the difference in mean scores (35 point difference in favor of boys) and the over-representation of boys for scores on the high end (2.22 to 1 ratio for perfect scores), and these differences persist over time.

At 11/23/2009 10:25 AM,  W.E. Heasley, CLU, LUTCF said...

We can rest assured that the males and females that scored well in Math, past and present, have never and will never show up in the CBO.

At 11/23/2009 12:04 PM,  Anonymous said...

Still beating this dead horse but ignoring the GM news?

General Motors Co. on Monday asked European governments to help pay most of the euro3.3 billion (\$4.9 USD billion) it needs to restructure its struggling European operations.

Does the world really need to give so much more wealth to a company that should be able to make a go of it on its own or close the doors? Just imagine if the continued \$\$ being stolen from the people by governments and dumped in to save a shitty company could be otherwise diverted to companies that produce something that the world really wants and at the choice of consumers...

At 11/23/2009 12:15 PM,  American Delight said...

I have no idea how valid these findings are, but I know that even raising the issue can get you in deep dukey. Just ask Larry Summers.

At 11/23/2009 1:12 PM,  Size said...

Now, why is it you guys aren't smart enough to find the *&#@ hamper?

At 11/23/2009 1:46 PM,  Michael said...

I can't wait for the left to spin 2x=10 in to some kind of sexist discrimination cases.

"6. More girls than boys took AP Honors math courses, by a ratio of 117 girls for every 100 boys."

Just because the girls took the class doesn't mean they passed the class. It could be that girls are looking to date smart guys rather than jocks.

The data is meaningless.

At 11/23/2009 3:09 PM,  Size said...

It could be that girls are looking to date smart guys rather than jocks.

Um...it has been my experience that the nerds are rarely chased by girls in high school. More like the other way around.

I also know a lot of girls who took AP math, passed it with flying colours and didn't give a crap because they were going to waste their lives as a state employee to allow them time to have that all-important family.

I don't know why this disparity between the sexist. In the spirit of not allowing statistics to rule my life, I blew on the math portion of my SATs and then went on to excel in a math-heavy field after college.

At 11/23/2009 9:19 PM,  Milton Recht said...

1. High school grades are better predictors of college student performance than SAT scores. Female performance in math is consistently on par with males except for SAT scores. Maybe SAT scores are the problem and are measuring some unidentified variable.

2. Sat scores increase monotonically with family income. More divorces occur in couples with daughters than sons and single mothers have lower incomes. Possibly a known SAT income sociological effect that is manifesting itself by gender.

3. The issue is not SAT scores but why there are fewer females in Math and Sciences. Until college, females take as many math and science courses, including AP, as males and achieve comparable grades. The question becomes why in college do career paths and majors diverge so there are fewer female math and science majors and PhD candidates. Are the intended majors listed on college applications different by gender? Do they track actual college major choices equally or are there gender differences? Is the issue the gender major choice prior to college entrance or does the college experience create a negative effect on potential math and science female major candidates?

At 11/23/2009 9:55 PM,  Mark J. Perry said...

Milton: For being successful as a Ph.D. in the math, computer science or engineering departments at MIT, Univ. of Chicago, or Harvard, I think the math SAT score would be a much better predictor of success than high school GPA.

In the Larry Summers debate, the issue was underrepresentation of women in science and math at Harvard and MIT, etc.

At 11/23/2009 11:50 PM,  BMWright said...

Guys, no need to make excusses for the women about these little math scores. Is it really a problem if there are more male math wizards than females? I can assure you the baby-boomer women are still in charge of the world. Just ask my wife. As Prof. Perry pointed out the Women are exceeding the men in employment retention, a much more important score.

At 11/24/2009 1:01 AM,  Milton Recht said...

Mark,

In the mid-1990s, MIT admitted women with lower average SAT scores and two classes at MIT were tracked by gender in 1994-5.

"Part of the impetus for this study arose as college applicants and others questioned the policy of admitting women with on average, somewhat lower SAT scores than men....

Two classes of freshmen were followed from entry through graduation. In both classes, women had higher four-year completion rates (on-time performance) than men. The mean grade point average (GPA) was 4.3 for both men and women who completed their bachelor's degree requirements.

Some have argued that the equal level of GPA's may be partly due to the fact that women major in areas other than engineering or science (the former being considered easier majors than the latter) more often than men do. However, at the university studied, all students are required to pass a minimum of eight science courses and more than three-fourths of each class of women continued to major in science. When GPA's were compared by major, out of each class only one major had a statistically significant difference between men's and women's GPAs.

The authors conclude that for the majority of on-time undergraduates there is almost no difference in the GPAs of men and women, even within the same majors."

http://www.onlineethics.org/CMS/workplace/workplacediv/abstractsindex/perfmathsci.aspx

So "somewhat lower SAT scores" for women do not make a difference and if they are predicative, it is not in the ranges that MIT used. I do not have the data by math and verbal or for the average mean scores by gender, so "somewhat lower" is the best I can provide, but it appears that SAT predicative abilities do not improve for math and science majors at the undergrad level.

Currently, MIT undergrad is 45 percent women, with 32 percent of undergrad physics majors, 35 percent of undergrad math majors and 40 percent of undergrad engineering majors are female. At the graduate level, MIT math and physics is 23 and 17 percent female. Engineering is 24 percent female.

MIT PhD physics, math and engineering programs have fewer females as a percent as compared to their undergrad classes.

With the understanding that MIT PhD applicant pool is broader than MIT students, do MIT PhD programs accept a smaller percentage of qualified female PhD candidate applicants in favor of men or do they accept the same percentage for both genders but there are fewer female applicants than men? Is MIT (and other grad schools) relaying too much on GRE exam scores? Are undergrad professor recommendations used to evaluate applicants biased against females?

Studies have found that women also under-perform men on the GREs (like females do on SATs) in subjects such as physics and economics. When schools look at physics PhD grad school performance by gender on 2 year qualifying exams and on time to research thesis completion, females perform as well as males.

http://www.aps.org/publications/apsnews/199607/gender.cfm

GREs scores also correlate with family income and other socio-economic indicators. Again raising the issues of whether some university standardized admission tests are measuring a hidden socio-economic variable that is gender specific.

There are also changes that could be made to the SATs to improve female performance relative to men. Increasing the time allotted greatly improves female performance and significantly narrows the gap between men and women on SATs. Also, men do better on multiple choice questions because they are more willing to take educated guesses than women and eliminating or reducing the multiple choice part of the exams will narrow the gap between men and women.

At 11/24/2009 9:12 AM,  Mark J. Perry said...

Milton: I checked the website of the MIT Math Department and it was about 90% male, and it was the same for the Michigan Math Dept, maybe a little lower but still about 80-85%. So there might be a difference if we are talking about undergraduate math/science programs vs. Ph.D. math program vs. Ph.D professors at MIT.

There is also the issue of "selective concern on sex imbalances" that I have posted about before. Women outnumber men in 7 out of 10 graduate fields (no concern there about men being underrepresented) and men outnumber women in only 3 fields (business, engineering and physical sciences) - very strong concern there leading to proposed solutions, regulation, changes, funding, investigations, programs, etc.

At 11/24/2009 9:21 AM,  Lyle said...

Mark are the equivalent statistics on the Graduate Record Exam? It would be interesting to compare the results on the two exams to see the effect of undergrad school. Hiding in there are the socialization factors that have been mentioned. In any case at the time of HS graduation it is futile at best to make predictions of how one would do at a PHD program in the areas cited. People are still maturing and will change greatly in the 4 years of undergraduate school. (18-21 is when adult reason fully emerges, the personality stabilizes for the first adult period...)

At 1/17/2010 10:17 PM,  Anonymous said...

Isn't it interesting that there is a gender gap in the United States, but that gender gap is not worldwide. In fact, in Europe young women do just as well in math as men. If men are truly genetically superior in the area of math, then why is the gender gap not seen worldwide?

At 1/17/2010 11:32 PM,  Anonymous said...

Please review current information regarding gender differences in math:

Few Gender Differences in Math Abilities, Worldwide Study Finds,Science Daily (Jan. 6, 2010)
http://www.sciencedaily.com/releases/2010/01/100105112303.htm

At 1/22/2010 7:57 AM,  Pierre M. said...

So what...men are better at math than women in general...it's not like any of this data is new...these stats are almost identical as when I took the SAT's back in the late 70's...people have to realize the facts, get over it, and move on...women have other academic attributes that are on par and supersede those of men

At 3/06/2010 12:20 AM,  CK said...

Does the gender gap persist internationally or only in the US?

At 3/07/2010 5:41 PM,  Anonymous said...

People care about the differences in business, engineering and physical sciences because those are among the best paid majors/fields out there. Women outnumber men in the less well paid majors. In theory, since there are now more women than men in college women should outnumber men in every major.
I've read that 40% of math majors are now women, so obviously there could be a lot more qualified women in those math heavy majors like engineering. Part of the problem may be that those majors do not "sell" themselves in ways compatible with traditional gender roles. Not so many years ago there were few women majoring in biology. As soon as med schools opened up to women (medicine being a "caring" profession) women surged into biology.
If engineering sold itself as "doing good while doing well" (after all, designing buildings and bridges that don't fall down in earthquakes is certainly a way of caring), it might attract more women; as it is, women engineers are most often in areas such as bioengineering.
Or we could eliminate traditional gender roles. :) Then boys might not see doing well at school as being "feminine".