Affirmative Action in Computer Science, A Hiring Perspective

I am now in the midst of hiring the next class of Stanford CS106 section leaders for the fall quarter. Section leaders (SLs) here at Stanford are typically undergraduates (and occasionally graduate students) who teach a small section of a larger lecture class. These classes often also have a head TA who will help coordinate the high-level academic aspects of the class.

CS106 is Stanford's introductory computer science sequence, and includes CS106A, CS106B, CS106X, and CS106L. The enrollment is so high that we now maintain an active pool of around 60-80 SLs who will teach every quarter. The entire teaching program even gets its own course designation: CS198. You can read more about the semi-complex structure of the teaching program on the CS198 web site.

I am one of the co-coordinators of the CS198 program, and as such am responsible for hiring SLs to try to maintain an appropriate number of staff every quarter. Most SLs return quarter after quarter, so we only typically hire between 10-15 now per quarter out of a pool of around 80 applicants.

This puts me in an interesting position. Stanford's CS department has around 4 men for every woman. [Please correct me if this statistic is incorrect or outdated.] Many of the students who go into Computer Science, especially girls, cite the CS106 courses, and especially their amazing section leaders, as their initial impetus to pursue the subject. Many women especially see female SLs as role models in their pursuit in a field so dominated by men, and with a culture so defined by men. So, an obvious question is, should I pursue a policy of affirmative action?

I have always been completely supportive of efforts to fill in the gender gap in engineering, but have always been against affirmative action in hiring, admissions, and other similar decisions. The main problem is that the affirmative action can cause the fallacy that women are bad at computer science to become a self-fulfilling prophecy much too easily. Imagine an admissions scenario. A woman is accepted to fill some gender ratio quota, whether implicit or explicit, but otherwise would not have been accepted. She may do well, but statistically, she will likely do worse than her male peers. Then, the proportion of women in the program doing badly will be greater than the proportion of men doing badly, and this gets noticed by people both within and outside the program. The prophecy is fulfilled.

Yes, I know the whole argument of affirmative action is that, if you believe it, if we give many women a chance, some will overcome the barriers and encourage other women to pursue the field. But realistically, unqualified women will just reinforce the gender fallacy, and successful women will inspire, but successful men will inspire as well, and overall, affirmative action does more harm than good. What we really need is something that is much more difficult to implement than a quota. We need to change the culture surrounding engineering, as well as to encourage those women with high potential to pursue engineering.

This is what I am doing now in hiring. The final hiring standard for both sexes are exactly the same, but during the application period, I try especially hard to seek out women who would make great section leaders and encourage them to apply through personal emails, and even allow slightly longer extensions on the application. But, once past the application stage, the entire interview and decision process is gender-blind.

Using this strategy, the ratio at the final round of interviews last quarter was almost 50%, and the final pool of SLs hired was 30% female. We hope that this quarter, we can push even closer to breaking that 50% mark.

The REMIX Culture

This is one of Lawrence Lessig's classic speeches. I'm sure most people who read this already hold the viewpoint that copyright law needs a major overhaul, so this video will just be a friendly and clear reminder of why that is true. For those of you who believe that copyright law needs no change, hopefully this video will explain why so many people believe firmly that the current system is broken.



The highly ironic part of this is that Warner Music has issued a DMCA take-down notice on the presentation.

Really?

William Perry on his Confirmation

The following is a hilarious segment of one of former Secretary of Defense William Perry's lectures in my Stanford MS&E 293 class on national security:

"The Washington Post responded to the announcement by running an article observing that I was soft-spoken and humble, and questioning whether those were the right personality traits for the job of Secretary of Defense. They were simply reflecting the well-known fact that those in Washington that traveled the high road of humility are never bothered by heavy traffic. As a result of the Post article, my confirmation hearing was a surreal experience. Senator Byrd, instead of asking me questions about my views on defense issues, took off on the Washington Post article. He observed that George Washington was humble. Robert E. Lee was humble. And yes, that Jesus was humble. This is true. So he concluded that my humility ought to be an asset. He did concede, however, that in his 40 years in the Senate, he had never before seen humility in a Cabinet officer. After that ringing endorsement, the other Senators fell in line and confirmed me as Secretary by a vote of 98 to 0. After all, they did not want to get on the wrong sides of any of their constituents, who might be admirers of Washington, Lee, or Jesus. That experience was the highlight of my experience with Congress. It was all downhill from there."

Private Data

My private information may be public now. Shit.

If you get offers to buy my social security number from some sketchy source, please let me know. kthxbai.

College Endowments not Coffers of Gold To Be Plundered

Articles like this one in Trees and Things rile my blood. Not only is this activity blatantly predatory, but shows both a lack of concern for many of the nation's pressing issues and a fundamental lack of understanding of (or desire to understand) how endowments work.

Yes, the thing known as the Endowment in most top private universities across the United States is increasing, and yes, there is a deficit, but coveting the money of educational institutions to help offset the national debt is an atrocious proposition. Many congresspeople complain that the major Universities with the $10+ billion endowments should allocate more money to financial aid, and use the money to expand programs and admit more students. However, things are far more complicated than than. Stanford's $17-something billion endowment is not the giant vault that people imagine it to be. It is a conglomerate of tens of thousands of donations, each earmarked for a specific purpose, with only a small fraction going to the "general fund," the part of the endowment where the University has discretion for its use, and another fraction going to financial aid, mostly earmarked for certain groups of students.

Forcing Universities such as Stanford to use a certain percent of their endowment for financial aid may simply be illegal, or, if not, would force schools to allocate more money to financial aid from the general fund than otherwise prudent, diminishing funding for research and other academic activities , the very activities that make the top institutions so successful and attract the world's top talent. Leading from this, there is the conception that the best tuition is no tuition. While this is an honorable goal, it is economically flawed. The schools with the top endowments currently offer the best financial aid in the nation, excepting those few like Olin College that offer a full tuition scholarship for all accepted. These top endowment schools typically offer over 95% of a student's financial need in terms of grants, work study, and low-interest loans. Free tuition sounds attractive, but economically, having a cost equal to the value of the benefit from the education is another way to attract only those students who will value the education. As a staff member of a Stanford dorm, I know that for large events, there should preferably be a small charge, even if there are funds to cover everyone's cost, just to make sure those who sign up will not balk, and that the people signed up at least value it enough to pay x dollars.

So the next question is, why not admit more students? That is perhaps easiest to answer. The value of a Stanford education is that students have the opportunity to take classes from eminent professors and first-hand contact with them, as well as to do research with them. There simply aren't more top professors; they are continually stolen back and forth in a friendly war between top institutions. Admitting more students would have a definitely dilution effect on the quality of education. Ask any student in a Big Ten school, and you will find that contact with a professor is rare and valued. Schools like Stanford, MIT, Harvard, and others, are slowly working to make education material from its classes and seminars available freely online through programs such as MIT's OpenCourseWare.

Finally, the large endowments serve as an important rainy day fund in case of catastrophes. A large earthquake is set to strike the San Francisco Bay Area within the next thirty years, and while all of Stanford's buildings are earthquake-reinforced, there are large collections of multi-million dollar scientific equipment that may be damaged during the Earthquake. The fund serves as a security policy against natural disasters, and equally damaging disasters such as policy changes that result in dramatic funding cuts from the federal and state governments.

There is no doubt that the United States is slowly losing its edge in science and technology. The last thing we need is to punish successful schools for their high donation rates from their successful and happy alumni. I would be less angry if the proposed "endowment tax" concepts would give money back to education, but the billions of dollars the federal government would essentially steal from the top research institutions by modifying some laws to their advantage would would mostly be used for offsetting a few days of fighting in Iraq.

Wake. Up.

Play Video Games. Beat Molecular Biologists. Feel Good.

Harnessing the millions of easily distractable minds worldwide, groundbreaking science is being done. Rosetta@Home, the software that is modeled after Stanford's Folding@Home and designed for distributed calculation and prediction for 3D protein shapes on millions of computers worldwide, now has a new feature. As the Economist article describes, the new Rosetta@Home software contains a game in which users (players) can manipulate the 3D protein structure according to basic laws of chemistry and physics in order to minimize energy. This was created due to the fact that 3D protein alignments and folding are still hugely computationally intensive, and often, the best 3D structures are found by molecular biologists working by hand rather than a computer using a heuristic algorithm. Moreover, humans often find these optimal structures faster than would the Rosetta@Home distributed computing network. There are some friendly competitions going on already.