Tag Archives: PhD

just write, damn it: the dissertation edition

One of my recent blog posts entitled “just write, damn it” got a lot of hits and positive feedback. All the feedback was for just writing and none was in favor of planning first. I was surprised that my methodological and analytical readers preferred to cannonball into writing without a lot of planning.

Someone told me about an approach to writing a dissertation that was somewhere in between just writing willy nilly and planning. It’s called the One Draft PhD Dissertation [pdf] by John Carlis, a professor of Computer Science and Engineering at Minnesota. His approach is to sketch a blueprint of a dissertation by planning what will be in each paragraph and then jump in and write each of those paragraphs. He writes:

What’s my story? While writing my dissertation and, at the same time, working as a professor (yes, it was stressful), several streams of thought happily converged. I read lots about writing, and was particularly struck by the forward to the John McPhee Reader, which described his disciplined, design-before-write way of work. I read Richard Mager’s Preparing Instructional Objectives, which has this message that translates to writing: teaching (writing) is about them, the students (readers), not the teacher (author). I taught software development using Yourdon and Constantine’s “Structured Development” (Addison-Wesley, 1975). Convergence began when the strong parallels in their content struck me. By then I already believed that software should be designed and not hacked, and, coming to the same conclusion about a dissertation, I decided to try to transfer software notions to writing. Since a paragraph is a unit of development, like a software procedure, I, following McPhee, chose to design everything down to the paragraph topic sentence level before writing. And it worked!

The one draft dissertation is a 5 step process:

  1. Believe you can do it.
  2. Understand that the purpose of a dissertation is to defend your claimed contributions to your field.
  3. Write for the right audience. Start by sketching a blueprint by writing for yourself then do the formal writing for your committee.
  4. Acquire draft writing skills (paragraph topic sentences, story telling, and logic to tie paragraphs together to tell your story)
  5. Design a dissertation.

What does “draft” mean? Here a draft is a completed unit, something that you give to others for review. I do not consider small scale revising, say editing within a paragraph while leaving the structure alone, the same as re-drafting. Do not misunderstand; “one draft” does not mean that you just start from a blank slate with final words of truth and beauty rolling off your fingertips. No, producing a dissertation is work, but it is merely work, not some mystical thing… You should believe that you can produce a one-draft dissertation, because focusing on contributions allows you to first choose vocabulary, craft figures, and grow a story tree down to paragraph topic sentences, and to then, for an audience of other experts, draft, one at a time, what you have the skill to draft, namely, paragraphs.

This approach has been tested on his students. It’s similar to one of my writer’s block antidotes. When I’ve struggled with writing something new, I often dive in and plan my story paragraph by paragraph by writing topic sentences, much like in the One Draft Dissertation. It’s planning, but it feels like another way to jump in and write. Just write topic sentences, damn it.

Finally, here is the average length of a dissertation by field, courtesy of beckmw.wordpress.com.

Thinking about getting a PhD? Here are some good resources.

Part of my job is to help students figure out if grad school is for them. Over the years, I’ve accumulated a few great resources for students thinking about a PhD. Here are some of my favorites.

Here is a link to the first of a series of ten blog posts by Jean Yang, who decided to leave Google to attend grad school in CS at MIT.

Philip Guo released an e-book on the PhD experience called The PhD Grind and has a 45 minute lecture on why to consider a PhD in CS. He has a bunch of other good posts about academic life as both a student and an assistant professor. He has a post specifically on applying to grad school.

Tim Hopper has a great series about whether one should consider a PhD in a technical field. I like this series because a broad range of people participated and not everyone encourages a PhD.

Chris Chamber’s blog post called “Tough Love: an insensitive guide to thriving in your PhD” has a lot of frank advice for determining whether a PhD is for you.

Below is a slideshare presentation I put together for the VCU math club a few years ago, since several seniors were planning to apply to graduate programs.

This is a woefully incomplete list, but I don’t want to delay posting it any longer. What other useful resources are out there?

thoughts on a PhD development course, part 1

I am teaching a 1 credit hour PhD development course for industrial and systems engineering students at the University of Wisconsin Madison. I am teaching the course with librarian Ryan Schryver, who is using the course to replace his office hours that students never came to. He found that students were not asking the questions that they needed to ask. Additionally, the department has a goal of exposing students to research and people across the department, but we have found that our students work in their labs with few interactions with students in operations research, manufacturing, human factors, and/or healthcare (our four department areas).

This course will fill these gaps. Our syllabus includes a variety of topics for students in their first 2-3 years, from understanding department and university policies to choosing a dissertation topic, technical writing, graphics, research organization, good programming habits, and prelims. We have a bunch of guest speakers plus me. I love hearing my colleagues’ take on things.

Student feedback thus far has been fantastic. I have urged the students to take ownership, and they seem interested in getting what they need from this course, not just in attending to get credit.

My favorite day thus far has been the PhD student panel, where PhD students asked questions to other PhD students. The questions came from all over the place, and the panelists were open to being quite honest about the process and the occasional struggles. Ryan and I knew it was successful when we had those unfiltered moments. Here are a few tweets from the panel.


Our guest speaker from the writing center, had some really good advice about the writing process:


Ryan talked about copyright and fair use, and he filled his talk with many pieces of useful information:


I’ll have a recap slide on this course at the end of the semester. In the mean time, please share advice and observations from similar courses you have taken so I can revise and improve the experience.

Related posts:

should you do a PhD?

Tim Hopper of the Stigler Diet has a wonderful series about whether to do a PhD in a STEM field [Link]. In each blog post in the series, somesone (usually a PhD) shells out some advice. There is a good mix of academia and industry in his series. My post is here. Mine is the only post with a picture in it.

I read through some of the posts and they are delightful. I highly recommend Tim’s blog series to college seniors who are thinking about grad school.

Below is a slideshare presentation I put together for the VCU math club for seniors who were planning to apply to graduate programs.

despite what you may have heard, getting a STEM PhD is still a good idea part II

This is an extension to my last post that argues that a PhD is still a good investment [Link].

I found an interesting NSF report [Link] on the science and engineering workforce that is useful for my argument that a PhD is still often worth it, at least in the US (HT: Gene Expressions blog). I’ll show evidence of low PhD unemployment rates and high PhD salaries. As I did in my last post, I will focus on recent PhD graduates, since trends in the past are not necessarily applicable to students currently considering a PhD.

ScreenHunter_04 Mar. 14 10.22

Unemployment for recent graduates in 2008 captures unemployment for those who earned their PhDs in 2003-2008. The overall unemployment rate for PhDs is 1.5%. It is 0.9% for computer science and math and 0.8% for engineering. Unemployment is 0.7% for those with an MS in computer science and math (slightly lower than for PhDs) and 2.6% for engineering (much higher than for PhDs).

ScreenHunter_05 Mar. 14 10.22

This post shows additional evidence of low unemployment among recent PhD recipients.

Unemployment rates according to years since degree. For recent graduates, look at "1-4 Years since degree" in 2003. This refers to PhDs who graduated in 1999-2002, which is a bit out of date, but it's relatively recent. Note that the unemployment is 3% for this group compared to 5% for BS degree holders.

Unemployment rates according to years since degree. For recent graduates, look at “1-4 Years since degree” in 2003. This refers to PhDs who graduated in 1999-2002, which is a bit out of date, but it’s relatively recent. Note that the unemployment is 3% for this group compared to 5% for BS degree holders.

ScreenHunter_02 Mar. 14 10.21

Salary distributions for BS, MS, and PhD degree holders. Again, this lumps together recent graduates with older graduates, so it doesn’t show that the PhD offers a salary premium for recent graduates. See the next figure for that.

ScreenHunter_03 Mar. 14 10.21

See “1-4 Years since degree” for recent graduates. This figure shows a modest salary premium for PhD degree holders as compared to MS degree holders. Getting a PhD 30+ years ago seems to have been quite lucrative in terms of salary, but I would not necessarily forecast those salaries for people getting PhDs now. Those with PhDs will, on average, earn more than those with MS and BS degrees, but I am not sure if it will make up for the ~5 years lost while in grad school.

despite what you may have heard, getting a STEM PhD is still a good idea

Unemployment among recent PhDs has been in the news quite a bit lately. I am going to make a case for getting a STEM PhD by showing you the good, the bad, and the ugly. My stance is that a PhD can be a great idea even though PhDs are not as lucrative as they have been in the past.

Jordan Weissmann wrote two posts in The Atlantic about unemployment among recent PhD graduates [Link and Link] and discussed these articles on NPR [Link]. The National Science Foundation collects the data at the time of graduation, and from what I can tell, their numbers are highly accurate based on excellent participation in their surveys. As far as operations research goes, IE degrees are in the “engineering” category and math and CS are in the “physical sciences” category. (I have done my own calculations on the NSF data that show the same trends. When I give seminars at universities, I like to share this data with students in a separate seminar just for students.) Here are two of Weissmann’s figures.

Unemployment at the time of graduation by year of graduation.

Unemployment at the time of graduation by the year of graduation.

Jordan Weissmann implies that post-docs are not employed. Sure, this may count as underemployment, since many of those with post-docs would prefer something more permanent, but it’s slightly disingenuous to consider post-docs to be unemployed. Still, I think it is important to separate our post-docs from other full-time jobs as Weissmann has done. He also notes that American-born PhDs do somewhat better than foreign-born PhDs in the job market – Americans have higher levels of full-time employment.

One reason that recent PhD unemployment levels are so high is because students may look for industry jobs after they graduate. Many find jobs immediately after graduation. This may be particularly true in fields like engineering, where more students than average get industry jobs (unlike academic jobs, industry jobs can be sought and found the summer after graduation). It’s worth noting that NSF does not track unemployment rates a year after graduation, so we don’t really know what happens. We can say that PhDs do not appear to have a problem with chronic unemployment. However, the data above show that it’s getting harder to find a job. It’s worth noting that between 2001 and 2011, the number of students graduating with PhDs increased from 40,744 to 49,010, an increase of more than 20%. This is a big increase in enrollment, and it seems to have affected employment opportunities. PhD graduation rates dipped slightly in 2001 (as compared to the late 1990s), and you can see in the figures above that 2001 was a good year for PhD graduates finding jobs.

Megan McArdle writes about how there may be too many STEM PhD students [Link].  She is concerned that the growth in science PhDs is driven by grant funding and the need for post-docs rather than by the supply of jobs that require a science PhD, which can lead to unemployment after graduation. And that is probably a legitimate concern in some fields. She cites NIH funding, which doesn’t grow the pool of operations research and mathematics PhD students too much. When my current department started a PhD program, we had to justify why our graduates would get employed with the state of Virginia – and that was back in 2007-2008 when PhD unemployment was not in the popular press so much. To be honest, I don’t know what the state does about PhD unemployment rates – presumably nothing since they are just starting to track students graduating with Bachelor’s degrees. I do think that as a nation, we need to not push harder to grow PhD programs without growing jobs as well.

These unemployment rates are depressing, but not getting a PhD may be worse. In another one of Jordan Weissmann’s articles, he writes that recent college graduates (BS/BA degree holders) have a 53% unemployment/underemployed rate [Link]. This seems horribly depressing. You can check the Bureau of Labor Statistics site [Link] for more statistics on unemployment/underemployed. Unfortunately, the BLS does not have many charts on education level. This one [Link] shows that unemployment for those with college degrees is now down to 3.8%, but that includes not-so-recent college graduates.

Here is another figure on BS/BA graduates.

2011 unemployment rates for recent college graduates. It shows a 22% unemployment rate and hints at a high level of underemployment. Importantly, it shows that major matters.

Several posts have argued with the premise that STEM PhDs have high levels of unemployment. Derek Lowe of “In the Pipeline” argues that unemployment among chemistry PhDs are way lower than unemployment in the general population and among those with Bachelor’s degrees [Link]. The American Chemical Society (ACS) cited a 4.6% unemployment level among its members in March 2011 (it’s lower now) [Link]. There are two problems with using these numbers to determine whether to get a PhD now:

  1.  This unemployment rates is from ACS members, and they have a biased sample. Unemployed chemists are less likely to pay ACS dues and are thus less likely to get asked to fill out ACS surveys. Plus, those who are unemployed are less likely to answer the unemployment question. This is the same issue that universities face with their graduating students – those with jobs tend to fill out outgoing student surveys, thus skewing employment levels. Therefore, the ACS survey likely underestimates the unemployment rate.
  2. The bigger problem is that the survey is from all ACS members. We are concerned with unemployment among recent PhD program graduates. The ACS survey does not give any insight into whether newly minted PhDs can expect to get a job.

However, a low PhD unemployment rate is consistent with all the statistics I can find. Here are a few other things to consider. First, the BLS shows a low unemployment for PhDs, which corroborates Lowe’s finding that PhDs can get jobs. But it doesn’t have enough information to show that recent PhDs can get jobs.

Unemployment rate by education level. This shows that the unemployment rate for PhDs is 2.5% compared to 3.5% and 4.5% for MS and BS/BA degrees, respectively. However, this doesn’t give any insight into unemployment rates among recent graduates.

This post [Link] by Catherine Rampell at the NY Times shows the unemployment rate for those aged 20-29 but doesn’t specifically look at PhD graduates:

Unemployment rates for recent graduates. Advanced degrees lumps together several important groups (MS, PhD, MBAs, etc), so it’s hard to tell what unemployment looks like for these groups. Given the ages considered (20-29), I’m sure that few PhDs were included in the “advanced degree” category, and therefore, their impact was diluted for better or worse.

These figures look at unemployment rate but not income. In an article that I cannot find at the moment, those with Bachelors degrees have seen their income just keep pace with inflation over the past 40 years. Those with advanced degrees have seen their incomes grow faster than inflation (see this link for the concept).

A PhD may not be as lucrative as it used to be, but I still think it’s a great idea to get a STEM PhD. Here are my reasons:

  1. Industry jobs continue to employ recent PhDs. And industry jobs pay very well.
  2. It’s a stochastic process: what was true in the past may not be true in the future. It’s hard to understand the employment rates among recent PhDs, but from what we know, unemployment rates for PhDs are lower than those for students with lower levels of education. Chronic unemployment does not seem to be a systemic issue for PhDs, and PhDs make more than their BS and MS counterparts. I see these trends continuing in the future albeit to a lesser degree.
  3. Yes, PhD students lose out on potential income for 5-6 years when they are in graduate school, but they get paid to get a STEM PhD, so it should not lead to extra loans. I started a Roth IRA as a graduate student. Getting a STEM PhD is different than going to law or medical school in this regard, and I think this is a huge plus.
  4. Many universities were not hiring in 2009-2010. This led to more post-docs than usual. Universities are hiring again, and all those  post-docs are ultimately getting full-time work. But recent PhDs often have to compete with post-docs for tenure track positions, which makes it harder for now. Eventually, this may work itself out. College is as popular as ever, and universities will need faculty to teach all of these students (although many are making do with instructors and adjuncts).
  5. And finally, field matters and the PhD program matters. The statistics above do not tease out those those who graduate from top programs (like top 10 or top 25). Students from top PhD programs are generally highly recruited – with some bias, I can recommend one stellar ISyE department for getting an operations research PhD. I recommend that students be very direct about employment options before they select a graduate school. Going to a top school makes a big difference for business school [Link] and law school [Link] – and it’s true for STEM PhDs, too. But if your employer is paying for your PhD, then this is less of an issue.

Do you think that PhDs are still worth it? My post focuses on getting a PhD in the US – are PhDs still worth it internationally? If you have a PhD, would you get one all over again if you were graduating from college now?

Update on 3/14/13: I found some additional unemployment and salary data for recent PhD recipients and posted it here: [Link]

so you’re thinking about graduate school in operations research

Today I gave a talk about applying to graduate school for the Math Club at VCU. My slides are below.

Here are a few links:

  • Advice on a career in operations research from INFORMS. I wish I had known about this link before–this FAQ answers many questions that students ask me.
  • A list of ORMS departments in the US from INFORMS.
  • By now, you’ve probably realized that you should become a member of INFORMS. It’s frugal ($37 per year) and valuable. Go for it! If you are not into OR, join AMS, ASA, ACM, SIAM, or whatever interests you.
  • Mike Trick made a plea for graduate students to sign up for twitter accounts. It’s a great idea, and I’ll let him explain why.
  • Students from state universities with strong STEM programs tend to be the most recruited in the nation. Those are great choices if you want a job when you graduate.
  • The students at the talk recommended forums, including the Xkcd forums and Grad café forums
  • One of the students at the talk recommended the book Getting what you came for: The Smart Student’s Guide to Earning an M.A. or a Ph.D.
  • I maintain some advice for student researchers. This advice illustrates some of the expectations for you in grad school. If it doesn’t frighten you, then graduate school might be a good option.

I want to thank my tweeps for their excellent advice (@tdhopper, @ksphil, @techstepper, and @dianam).