Previous month:
December 2010
Next month:
February 2011

January 2011

Industry or Academia? On an ex-Harvard prof turned Googler

A Harvard professor who had recently received tenure and writes a popular blog announced a few months ago that he would not return to the academic world after his sabbatical at Google - instead, he would join Google for good.

While his recent post on the contrast between his current life as a Googler and his past life as an academic strikes a wrong note (a commenter even felt the need to remind him of the "honeymoon" period at the beginning of each job, which might be clouding his judgment; the blog author later defended himself saying he had written the post "in jest", although he seemed to be the only one finding his post jest-like), I think his posts about academia and his decision to leave do vividly illustrate the dilemmas faced by professors in engineering, who have job opportunities outside the ivory tower. Since they can provide valuable information for doctoral candidates trying to pick the best career path, I want to point them out to my readers.

The earliest post I'm going to link to is dated May 2010 and entitled "The Secret Lives of Professors". In it, the author shares his doubts on the professorial life; in retrospect, it is not surprising that he chose to leave. His comments focus on the pressure to find funding and on time management. He writes: "The biggest surprise is how much time I have to spend getting funding for my research. Although it varies a lot, I guess that I spent about 40% of my time chasing after funding." He did have a huge group, and acknowledges he was not terribly good at grant-writing, although he does not explain what, if anything, he did to improve.

As for time management, in his words: "Another lesson is that a prof's job is never done. It's hard to ever call it a day and enjoy your "free time," since you can always be working on another paper, another proposal, sitting on another program committee, whatever." I am not quite sure in what respects this differs from life in industry; the biggest issue with academia is that it is easy to work from home, and thus to never take a real break and relax. The ability to stop work at the end of the day and forget about it until the next morning, no matter the number of unfinished tasks that await you when you get back, is certainly very helpful to avoid burnout.

The last point surprised me - "Most of my days are spent in an endless string of meetings." You would expect a good department chairman to protect untenured faculty (which the author was until he was granted tenure and took his sabbatical) so that they can get their research program off the ground.

In his "Why I'm Leaving Harvard", the new Googler explains: "And of course the amount of overhead and red tape (grant proposals, teaching, committee work, etc.) you have to do apart from the interesting technical work severely limits your ability to actually get to that point [where someone at Google or Microsoft or Facebook implements the ideas he has read in a paper of yours]." Obviously, if you view even teaching as distracting from your core purpose - which you happen to define as doing technical work - and chafe at advising graduate students rather than doing the work yourself (a point the author mentions elsewhere in his posts), joining industry is the best decision you could take for yourself.

Teaching and advising make an immediate and lasting impact on society by training tomorrow's workforce. Many doctoral students find jobs based almost entirely on the strength of their PhD dissertation - that means that at some level, they would not have gotten the job without the skills their adviser taught them. Of course you don't have your name next to theirs the way you do on research papers, and you don't receive any clear recognition for your mentoring, but the indirect impact you have will be more lasting than a research paper, as it often affects the whole direction of a student's career.

As the author himself admits, academia is not for everybody - "I also admire the professors who flourish in an academic setting, writing books, giving talks, mentoring students, sitting on government advisory boards, all that. I never found most of those things very satisfying, and all of that extra work only takes away from time spent building systems, which is what I really want to be doing." Working at Google will certainly provide him with many gratifying opportunities to see his work implemented and contribute to one of the most famous companies out there. I think the decision to leave an institution such as Harvard, especially now that he had achieved tenure, shows integrity and a desire to follow his dreams that deserves respect, although it is a pity he had to waste seven years in academia when this seems to have been such a poor fit for him. But at least it allowed him to see for himself and have no regret about the academic life.

To his credit, the author let his area dean write a guest post ("Why I'm Staying at Harvard") in response to his announcement that he would not return to Harvard; taken together, these posts give a balanced account of the pros and cons of the academic career path.

I'll only quote one paragraph of the dean's post (abridged below), which captures his overall themes quite nicely - please refer to the whole post for more details: "I enjoy the freedom of working on whatever I find interesting... having the opportunity to work with a whole variety of interesting and smart people, from undergraduates to graduate students to CS colleagues all over the globe to math and biology professors a few buildings down; the ample opportunity to do consulting work... the schedule that lets me walk my kids to school most every day and be home for dinner most every night; and the security that, as long as I keep enjoying it, I can keep doing this job for the next 30+ years."

The author did the research community a great service in launching a discussion of these issues; hopefully, he and his dean will help doctoral students avoid mistakes in their own career path by giving them a more accurate picture of the professorial life.

Analytics in the Lehigh Valley

Continuing in my series of blog posts about analytics, I'll write about a study that I heard about today, at an event called: "State of the Lehigh Valley: Community Trends at a Glance" which was organized by RenewLV in partnership with the Lehigh Valley Research Consortium and held at Lehigh University's Iacocca Hall. (LVRC conducted the study.) The event consisted in a presentation of the report's findings followed by a panel discussion with a Q&A section.

The report is supposed to be online soon, so I'll just touch upon some points that I found particularly interesting.

Social scientists use the percentage of students eligible for free and reduced-price lunch as a proxy metric to measure poverty. While a staggering 75% of the student population in Allentown (City) was eligible for lunch subsidies in 2010, and 51% in Bethlehem, the presenter made the case that (a) this percentage increased in every school district between 2004 and 2010, not just the inner-city ones, and (b) the rate of increase of the metric has been greater in the suburbs than in the inner cities. In other words, contrary to what some believe, poverty also exists in the suburbs.

The recession had been expected to result in an increase in crime; however, that has not been the case. I found that intriguing because of the underlying assumption (held in other parts of the country too) that financial hardship would make people commit crimes. What the data shows here is that the crime rates have either slightly decreased or remained constant. In addition, "concern with becoming a victim of crime reached its lowest level since 2006 with 58% of Lehigh Valley residents reporting that they were either very or somewhat concerned with the possibility of crime victimhood" (report, p.16). The report also states that most people rate positively law enforcement in the area.

Someone commented that the current composition of the Top 15 employers (in terms of number of employees) in Lehigh and Northampton Counties is heavily tilted toward the public sector. Here is the list for 2010 (State of the Lehigh Valley 2010 report, p.5): 1. LV Hospital Center, 2. St Luke's Hospital, 3. Air Products & Chemicals, 4. Allentown School District (SD), 5. Lehigh County Government Center, 6. Bethlehem Area SD, 7. Giant Food Stores, 8. State Government, 9. Lehigh University, 10. Northampton County, 11. Federal Government, 12. LV Physician Group, 13. Wal-Mart Associates, 14. Easton Area SD, 15. Guardian Life Insurance Company.

Health care and education are expected to be two key sectors of the US economy in the coming years (see for instance Chart 5 in this report from the Bureau of Labor Statistics); the employment situation in the Lehigh Valley already reflects that trend.

There was a concern that companies such as Giant Food Stores and Wal-Mart provide mostly entry-level jobs with very few opportunities for advancement. I would have wanted more information about college graduates - among the students graduating each year from a college in the Lehigh Valley Association of Independent Colleges, how many find jobs here and which local companies have the most employees with a college degree?

Young professionals interested in small businesses, innovation and entrepreneurship are the most likely to find the LV appealing enough to stay after graduation, but their contributions won't be captured by looking at the biggest local employers. Someone in the audience suggested that the companies below the Top 15 should be studied. A better idea would probably be to study small businesses instead.

It's worth noting that the Lehigh Valley area was ranked #4 in the creativity rankings of medium-sized cities presented by Richard Florida in his book "The Rise of the Creative Class" - you can find the rankings here, where the Lehigh Valley is referred to as the "Allentown, PA" metro area. The rankings were based on the 2000 census and of course much has changed since then, but back then 28.7% of workers were labeled creative and the area was ranked #3 among medium cities in terms of innovation. So it would be interesting to add measures of local creativity and innovation to the report.

The director of Community Health for St Luke's Hospital made a great speech that contained many fascinating insights. For instance, she pointed out that PA is in the bottom half of states in terms of health status and gave as her source America's Health Rankings 2010 - indeed, PA is ranked 27th. According to this detailed profile of PA, the strengths of the state are: low rate of uninsured population, high rate of high school graduation, and low percentage of children in poverty. The challenges are: high levels of air pollution, low per-capita public-health funding and low immunization coverage. (Shocking stat: "Since 1990, the prevalence of obesity increased from 12.5 percent to 28.0 percent of the population.") You can find a list of PA's rankings across many health metrics here.

The director also said that on top of that, the LV isn't doing terribly well compared to other areas in PA (I didn't find the data source for that statement online); she explained that the health system in the LV is very good in restoring the health of sick people, but preventive care, i.e., keeping healthy people healthy, needs to be improved. Other concerns she had were the local air quality and the high incidence of low-weight births in the area. The report mentions positive statistics about the poverty rate, which at 7.4% is about half the rate in the US as a whole (14.3%) and is also significantly lower than the average rate in PA (11%). Improving health care in the 3rd biggest metropolitan area in PA is an important but also exciting challenge that resonated with many people in the room.

Some people at the meeting were very enthusiastic about regionalization of public services but agreed they needed to do a better job explaining to the community how this would improve overall operations. A lunch attendee mentioned the case of the Jacksonville, FL, area as a "poster child" for successful regionalization. There was also a discussion about business opportunities in the Lehigh Valley, the impact of the recession on companies that had recently moved here from New York and New Jersey (the LV has lagged the US as a whole in its recovery), the fact that some sort of permits take a year to be granted because of the caseload, and the recent developments trying to make New Jersey more of a competitor in attracting or retaining businesses.

This was the first-ever public release of the report, and I hope this becomes an annual event. In future years, I'd love it if there could be a discussion of the next steps to be taken based on this data.

HBR on Robert McNamara and Analytics

The December issue of Harvard Business Review had an interesting feature on Robert McNamara's business legacy - "Robust McNamara and the Evolution of Modern Management" - which I thought would be a great topic for an analytics-themed blog post. (As a reminder, I have created a new category called Analytics for my blog, and the first post in this category is here.)

While McNamara is primarily known for serving as Secretary of Defense under Presidents John F. Kennedy and Lyndon B. Johnson, he was also President of the World Bank from 1968 to 1981. Most importantly for this article, he taught analytical approaches in the Army Air Forces, serving most of World War II with the AAF's Office of Statistical Control and analyzing the efficiency of US bombers' missions; later he joined the Ford Company and, as one of the "whiz kids", helped engineer Ford's turnaround through quantitative techniques. His legacy, though, is irremediably linked to the Vietnam war, as his New York Times obituary made clear.

HBR writes: "Because of his role in [the Vietnam tragedy], he tends to be caricatured as smart but not wise, obsessed with narrow quantitative measures but lacking in human understanding... In [McNamara's] legacy we see the triumphs of modern management as well as its most troubling limitations." The article retraces McNamara's early professional steps, including his learning business statistics at Harvard Business School in the late 1930s; it also mentions him as a champion of passenger safety who gave up a $410,000 a year in salary and bonuses at Ford, and chose not to exercise stock options to avoid the appearance of a conflict of interest, to become a cabinet secretary (paid $25,000) under JFK.

After praising McNamara in his early days, the article also does a good job when it highlights the pitfalls of overrelying on analytics during the Vietnam war. "Data that were hard to quantify tended to be overlooked, and there was no way to measure intangibles... Much of the data about Vietnam were flawed from the start... The Pentagon depended on sources whose information could not be verified and was in fact biased. Many officers in the South Vietnamese army reported what they thought the Americans wanted to hear, and the Americans in turn engaged in wishful thinking."

(The best-selling novel Matterhorn by Karl Marlantes, about the author's experience as a Marine during the Vietnam war, touches upon similar themes; it seems that numbers of enemy casualties on the ground were repeatedly inflated throughout the chain of command, so that decision-makers in Washington had a very distorted picture of what was happening. Data is only as good as its easiest-to-manipulate part.)

The article also has an insert about "What the Whiz Kids Missed", i.e., aspects of human behavior that have to be taken into account in decision-making processes: bounded rationality, escalation of commitment (the incentive of taking additional steps in a bad direction because so much has been invested in it already and the marginal cost is small), emotional intelligence, the wisdom of crowds (as opposed to central decision-making), disruptive innovation ("incumbent firms often fail precisely because they are well managed"), and dispersed networks (importance of local and grassroots efforts).

Getting a PhD: Much Bitterness at The Economist

Most articles in the print edition of The Economist are very good - which is why it was so surprising to recently find a bitter rant mixed in with the high-quality articles the magazine has accustomed its readers to. The article, entitled "The disposable academic", has for subtitle: "Why doing a PhD is often a waste of time."

The following paragraph will give you an idea of the article's perspective: "One thing many PhD students have in common is dissatisfaction. Some describe their work as “slave labour”. Seven-day weeks, ten-hour days, low pay and uncertain prospects are widespread... Meanwhile, business leaders complain about shortages of high-level skills, suggesting PhDs are not teaching the right things. The fiercest critics compare research doctorates to Ponzi or pyramid schemes."

Ponzi or pyramid schemes? Seriously? As a commenter wrote, "It seems to me our correspondent had a sour experience during his/her PhD"; indeed the author is spewing bitterness at a furious rate. Far down in the article, she admits that she got "a largely pointless PhD in theoretical ecology", whatever that means. The author's own disappointing experience certainly explains a lot in an article that is supposed to represent unbiased journalism rather than an opinion piece.

The article hints at important questions and it is a pity they were not better addressed. For instance, the issue of sometimes misaligned interests of employees and employers is relevant in any company, but the challenges are greater in universities where graduate students cannot transfer without losing years of hard work. Who benefits from the PhD process the most - the student or the university?

Whatever the answer is, it is not the advisor, who invests a lot of time in training a student to do work the advisor already knows how to do. In humanities departments such as English, I believe that the advisor does not even co-author papers with his students; advising is viewed more as service than research since the advisor's name will not appear on publications. Therefore the student's doctorate is often financed by positions as Teaching Assistants or even instructors; most (all?) sections of English 1, a course required of all undergraduates at my institution, are taught by graduate students.

While teaching represents an important experience for doctoral students serious about a career in academia, it also allows universities to staff basic-level courses cheaply and free professors to teach more advanced topics. If a graduate student does not make progress in his dissertation because he has to teach semester after semester, he might begin to blame the system, in which he brings immediate value to the university through positions that do not benefit from him doing research. At the same time, people who intend to devote their career to academe - where they will have to teach no matter what institution they join after graduation, from community colleges to research universities - should know what they are getting into. Teaching in graduate school is part of the training.

The author also mentions the increase in unionization, "though Yale and Cornell, where university administrators and some faculty argue that PhD students who teach are not workers but apprentices, have resisted union drives." The fact that doctoral students are paid a stipend rather than a salary gives weight to the PhD-student-as-apprentice vision far beyond the Ivy League. Indeed, if international PhD students were primarily workers rather than apprentices, they should be on a different type of visa.

Further, comparing the stipend (not salary) of graduate students with that of full professors (who have received a doctorate, tenure and promotion in the academic hierarchy by demonstrating excellence in their field, including their research area) does not make any sense at all and yet the author blissfully quotes numbers as if they applied to similar individuals: "A graduate assistant at Yale might earn $20,000 a year for nine months of teaching. The average pay of full professors in America was $109,000 in 2009—higher than the average for judges and magistrates."

Full professors have many more years of experience, many more responsibilities, and of course an additional degree; their average age is somewhere around 55. The average age of a Yale grad student is closer to 25. (Time for a prop for this old post of mine, "There is age in average.") The only conclusion of these numbers is that, for the lucky few grad students who make it to full professors many years later, yes, a PhD is worth it.

Another issue, pointed out by the article's author but not explored in any depth, is that "There is an oversupply of PhDs. Although a doctorate is designed as training for a job in academia, the number of PhD positions is unrelated to the number of job openings." But no one knows how many job openings there will be four, five, six, seven years in advance. Also, some students might prefer to join a teaching institution while others want to continue to do research; there is no way to predict how many incoming students will end up in each category. Others will drop out before they finish.

The regulation mechanism rather occurs through the level of available funding, in areas where students are mostly funded through research grants. Many departments figure out how much money their faculty members can spend on research assistantships in the coming year(s) before deciding how many students to admit. It is of course an imperfect science - some faculty members will leave an sabbatical or choose not to take any new student if their research group is already well-staffed. I view a system where students are funded through a mix of Teaching Assistantships and either Research Assistantships or Fellowships as healthier than the all-TAs model, but some fields, especially humanities, receive few grants and will not have funding for RAs, no matter how much they would like to.

The author correctly notices the "rise of the postdoc" as a novel trend, but cannot resist painting it in ominous terms. "The rise of the postdoc has created another obstacle on the way to an academic post. In some areas five years as a postdoc is now a prerequisite for landing a secure full-time job." Rather, the postdoc has become a risk mitigation technique for universities, which do not want to give employment for six or seven years (it is rare to be forced out before tenure, although one can receive signals that it is time to update one's resume) to someone who will prove incapable of developing and managing his own research program. If there is a "secure full-time job" down the road, isn't it preferable to give it to the better-suited candidate?

The article goes on and on in giving a very negative view of getting a PhD, in any field (already a consterning over-simplification across humanities, sciences, business and engineering). I could keep writing about it but I'll try to wrap up, having other things to do.

The article could have started an important discussion on today's PhD students and some universities' maybe ill-founded eagerness to be "research institutions" (with PhD programs) rather than the less prestigious but very important "teaching institutions" (without), and it is a pity that the author did not give a more balanced view of the situation. There is a piece of good news hidden in all this: The Economist might be looking for a competent writer for its higher education section (or its Christmas specials section, which is where the article originally appeared.) Send resumes to or