All about the money

Recently, there was a discussion among faculty members and industry folks surrounding the role of training PhD students suitable for the workplace. The synopsis was something like this. Many students who get PhDs are interested in staying in academia but the realities are that for every 10-20 PhD students graduated, maybe 1 of them will get a faculty position. So what are the other PhD students supposed to do? Well, get jobs in industry. Except most of them have been trained to have an academic perspective on things and have not been trained to understand effective project management, deadlines, deliverables, etc. 

One of the ideas to solving this problem involves having the PhD student propose their thesis topic earlier in their training. The idea is that if the student identifies the scope of their PhD earlier, they can then focus more on getting those "deliverables" done. Also, from an advisor perspective, this is a contract of sorts where the PhD student can keep the advisor from moving the goal posts because there is an established stopping point to the student's work. 

While this is a grand idea and I whole-heartedly agree with it, it is simply not practical. And there is one simple reason for it. Money.

The most direct way that I see this relates to the size and duration of research funding awards. In my area, most research awards are 2-3 years with insufficient funding levels ($100k/yr) to establish instrumentation and equipment to complete a task. Thus, a student may start out on one project but because of my funding situation (for which the student has no control), that student may be forced to switch to multiple projects just so I can pay their stipend and satisfy multiple program objectives. However, these projects usually are not interlinked enough to draw a common thread for writing a thesis. This makes defining thesis objectives very nebulous when you don't know what task you will need to do to satisfy which program manager 6 months from now. I spend more of my time (probably 60%+) going after money than I do doing anything else, including actual research, teaching, service, mentoring, etc. As the PI, frankly I'm trying to solve the paycheck problem for my students first and then worry about the details like defining a thesis second. It is a sad state of affairs and I am disappointed in this but that's the practical realities that I face. 

The second way in which I see an issue with training PhD students for jobs in industry is that no funding agency actually awards funding for practical, applied research. (At least funding that is not ITAR controlled.) Virtually every response I get from review panels is "Wow, so-and-so company that wrote you a support letter should directly fund this.". I'm basically convinced at this point if you send anything practical to NSF it will get shot down for not being sciency enough. But they neglect that if they fund that research, it may turn into technology to spin off from the university, which may create jobs, which grows the tax base, and increases funding for entities like NSF. (BTW, that's the NSF standard lip service as to their motivation for funding projects). 

I've had a $2.1M NIH BioTech R01 turned down last week, an $875k fellowship turned down today, and two NSFs for about $800k turned down last month. All of them were doing some amount of science but all of them had practical applications which would be excellent training for students.That leaves 5 of my students with cloudy funding futures, several of which will have to put off their thesis proposals, because I can't secure funding. And that's not their fault. But the really infuriating thing about all of this is I have no shortage of companies that explicitly ask me to notify them when I have students graduating because they want students who have been trained in my area. But I fear I won't be in academia long enough for that to matter because the practical realities are that I need to bring in enough money to support the students and the funding agencies simply do not fund areas where students may be practically trained. 

Postdoc vs Research Engineer

In one of my last posts, I mentioned that my group has grown excessively large to 10 graduate students. I do not think I am writing anything that is not obvious to everyone in my group: it is very hard to manage that many people and write grants and write papers and perform the level of university and national service that I do. One of the things that I pride myself on is explaining the status of my group to the students. They know, e.g., the current financial situation (whether good or bad), know when I am hiring or not, and whether they can ask to attend conferences they would not normally attend. One of the recent things that has become apparent to me as a result of my participation in a faculty search committee is whether the candidate understands from where his/her funding comes. Now, most of my students are likely to end up in industry so this may be a moot point but I want them to at least understand where their funding comes from, even if they are not responsible for bringing it in.

Lately, I have been debating whether I should bring on a postdoc to help me advance my group. There are numerous reasons to do this: a postdoc can focus more on research instead of classes, can understand a higher level of applications, can write journal papers, and [hopefully] can manage graduate and undergraduate students. Now I know most STEM folks in the blogosphere are in the biomedical fields which have different expectations and funding mechanisms than those of us in engineering. When I talk to folks in the biomedical fields, they say that 40k is sufficient for a postdoc, maybe 42k if they are really good. But in engineering, you can easily make 10k-15k more than that with a BS, with stock options and a 401k so why would any engineer looking for a postdoc accept that salary??! The numbers I'e seen are more like 55k and when you factor in the 33% benefit rate and then 53% overhead rate, a postdoc costs me 112k/year. And that is for an average postdoc. To be honest, before I came to SnowU, I had a postdoc offer for ~73k/year. That's close to 150k/year if I wanted to hire someone at that salary. 

So this begs the question: is a postdoc worth the 112k-150k per year that I need to bring in to support them?

The basic (and more in-depth) answer is: NO. That same amount of money can get me ~2-3 students who will at least be here for until they get their degree, whereas a postdoc will jump at a TT position if offered. (And I'm not blaming them for that!) But I think the bigger question that needs to be asked is why would I hire a postdoc when (assuming) I could hire a research engineer. 

True, a research engineer would cost more than a postdoc but the big difference between the two is a research engineer (or research faculty depending on how you look at it) is able to be PI on their own grants. Other than NIH, which has its own shit-storm of a competition, there is no other mechanism where postdocs can submit their own proposals. However, a research engineer can submit their own proposals to NSF, DoD, DoE, NASA, etc. And if I, as a PI can land a big grant, I may be able to offset most of those costs for the research engineer and then have them help the finances of the group. 

But why would anyone choose to be a research engineer? Wouldn't they just jump at the first TT offer? I don't think that is the case if the right person is found. I know of several groups at SnowU where the lead PI has their research engineer helping them run their group, takes part in the management of students, helps write papers, and is content with trading the TT rat-race for no teaching duties, no service, and the aspects of soft money. Over the long term, especially with the current trends in academic funding, it is likely not sustainable unless you have NIH/DoD/DoE funding but they may be able to help me, as the PI, sustain a group and output level needed to maintain that level of support. 

I'm still not 100% sold on the idea but it has been something I've been thinking about lately. At this point, I think I would take all the help I can get but if I had my pick, I'll probably go with the research engineer rather than a postdoc at this point.

Research Group Dynamics

Today, over at Engineer Blogs, I discuss the dynamics of building a research group and put forth a few key ideas that I focused on in the beginning. If you have comments, they're greatly appreciated.

Freshman Advising

One of my service tasks that I am scheduled to do this year is freshman advising. At SnowU, they take freshman advising very seriously, whereas I'm not so crazy about it. I mean, apart from the social transition (and that can be a bitch for many students), the actual academic advising portion should be straight forward.

In engineering, unlike the Arts, students have to take 96% of their classes within their major (or general engineering). In other majors, you could have as little as 30% of your classes within that major (think Anthropology, History, or Communications). The one think about Mechanical Engineering that is so nice is you have this great little chart that says "Oh look, these are the 4-5 mandatory courses you have to take. No room for electives." Your decision is already made for you.

As incoming freshman, they don't know this, which is where I guess I come in. And I know things can get more difficult when you factor in skipping classes for AP Calc and Phys, etc. but even then it's not that hard.

So all in all, this shouldn't be so hard. But if you dig a little deeper, there's definitely a flaw with me advising freshman students. It's the blind leading the blind! How am I supposed to advise them on a graduation track that I don't fully understand yet?!? Yes, I should be able to get up to speed very quickly and this will help me get used to the curriculum. But still, I feel like these freshman will be like "Oh yeah, well my advisor doesn't know shit because he just started working here."

Any thoughts/advice/words'o'wisdom for me?

Abuse of Power

If you've been following my twitter account (@profgears), undoubtedly you've seen a flurry of tweets about student expectations, terrible advisors, and politics of publishing. This stemmed from GMP's post the other day over at Academic Jungle and DrugMonkey's post on the subject clearly in GMP's camp. DrugMonkey's post heading The Care and Feeding of your PI: a tip for grad students and postdocs should tell you all you need to know about their position on the matter.

I clearly disagree.

As a grad student, you're not there to serve a professor at their beck and call. As a professor, you're not there to wipe the nose of your students. Both should have expectations and duties that are clearly laid out. I've already discussed this a few weeks ago. The professor (or PI, or whatever you want to call them) is supposed to be the student's professional mentor. The prof is supposed to be the student's guidem helping transition from a student to being a autonomous researcher.

This clearly breaks down when there is a lack of communication between both parties. For instance, the prof expects the student to publish as many papers as they can before graduation. But you'd be surprised how many students don't know that and how many profs don't tell their students that. So you have a prof that's expecting paper drafts and a student wondering what to do with their data. The student tries to schedule meetings but the prof has little time to discuss it. Months go by. Still no paper. There is increasing frustration on both sides.

Inevitably, a downward spiral happens and things turn to shit. The prof did a bad job of mentoring their student, yet from the prof's perspective, their student is a slacker. The prof thinks their student isn't working enough and so the student gets even less time than the minuscule amount the prof already gives them. The student gets frustrated because their mentor and career guide doesn't respond back to them and now the student starts worrying about graduation deadlines and whether they are even going to get a recommendation.

At this point, both are in the wrong because they didn't communicate with each other. But the professor is always in the power position and it should never be abused. And it sure sounds like some Profs/PIs do abuse that power by holding graduation and recommendations over their students' heads. Those are the same Profs/PIs that are supposed to be mentoring their students. That's a terrible situation to be in if you're the student because the one person you're supposed to go to for advice is giving you shitty advice and guidance. In the odd chance that student makes it to the next level, how likely is it for them to mimic their advisor's behavior? I say that's highly likely and I sure as hell don't want to be that kind of prof.

-------

On the topic of publishing and who gets coauthorship, it's very hard to describe in 140 characters. My general feeling is this. Anyone who contributed to the writing, data gathering, interpretation, and oversight should be a coauthor. However, everyone one of those people should be capable of writing a significant chunk of the article. That doesn't mean only the person who wrote the article should be the sole author.

Here's how you test it. Ask each person if they could write the draft. If someone says they could probably get 60% of is done, that's fine (that's a significant chunk). If they said 10%, they don't belong as a coauthor. That's what the acknowledgements section is for.

In practice, most groups aren't that strict because everyone needs more papers to be considered successful. But if I adhered strictly to that, I'd be missing my prof on every paper. And for me, when I'm in that same situation, I want to be involved enough to always pass that test. Otherwise, I would and should expect my student to submit it without my name on it because I haven't contributed enough. But if it comes to that point, then there's obviously something else going on in our advisee-advisor relationship and there's a bigger problem at hand than coauthorship on a paper.

Templates, Because MS Word Makes it Hard Enough (Part 1)

Yesterday, Cherish the Scientist (@mareserinitatis) was kind enough to indulge my twitter (@profgears) rant about how I hate MS Word. It all started when I had to review a paper for a conference proceedings and the only template they offered was either *.doc or *.docx. NewPhD, just starting out, has a brand new laptop with Office 2007 on it. Meanwhile, I have the 4 year old tank laptop with lowly 2003 on it.

And MS, in their infinite wisdom, decided to change their proprietary format, so converting back between *.docx and *.doc destroys any formatting you thought you might have done. Now, I'm not going to keep ranting about this (at least for this post), but rather, I'm going to spell out how students writing their first publication should approach things like pictures, figures, and graphs.

A few days ago, I post about Advisor-Advisee Expectations, which included things like making sure to communicate, to respond rapidly to emails, etc etc. One of the things I forgot to add for students to do is Learn Compatible Software and one of the things I will teach them is How to Make Templates. I'll tackle the software stuff first.

One of the best things I've learned during my PhD studies is "if it looks professional, people will treat it professionally". This is very important for two aspects, namely written work (thesis and papers) and for presentations. If you want to make professional documents, the best way to go IMHO is with LaTeX. Latex is an open source typesetting system for preparing professional documents. I use the MiKTeX distribution. It has a high learning curve but once you get the hang of it, it makes everything easier. LaTeX is something that I would strongly advise any student starting out to learn. I know it seems crazy with all the other stuff you have to do, but once you start, and your peers start, you'll be able to help each other and rapidly climb the learning curve.

For large documents, the content and the format are largely independent. For instance, once you make your thesis template, you'll never have to change it. If you add more content, you just recompile the document and it's done. And, most journals and universities have LaTeX templates for their publications. In those cases, you don't even have to make the template, you just add content. I watched DrWife bitch and moan about MSWord changing fonts, section headings, chapter spacings, you name it while she was writing her thesis. Everything in MSWord is linked and you have no control over that. In LaTeX, everything is still linked, but you have full control over it. Fonts and Sections headings don't change because of the moon rise. Also, the files are a lot smaller, so you don't have to worry about it crashing.

Another reason why LaTeX is great is because it gets you making *.EPS (encapsulated postscript) pictures. Just google vector graphics or check out the wikipedia page. This is the real reason why I wanted to use Latex. I wanted my figures to look professional and be resizable with issues. That's really useful for transferring journal figures into presentations, which need to be scaled larger.

Over the next few days, I'll talk about how I set up templates and what programs I work with to make figures and drawings for publications.

Advisor-Advisee Expectations

A few days ago, I posted about Submitting it Anyway, which was a dig at advisors who take forever to give comments on papers. Or in my case, take forever to the paper and say "I have no idea what's going on but it looks good. Submit it".

If I take a step back, that's a somewhat terrible relationship to have with your advisor. That's really good way to sow dissension amongst the minions. Students don't want to work for you. You appear (as their advisor) to not give a damn because you're some really uber busy professor. That's why when I posted about my research group and number of students, I didn't think having more than 4 or 5 students was worth it. With the downward spiraling of academic funding, I don't think it's easier to maintain more than that. You're too focused on funding rather than your students, your teaching, and your research.

Later tonight, I have a presentation over the interwebs for first year grad students at SnowU who are looking for an advisor. In this presentation (and in the future), I lay out what I expect from students and what they should expect from me. If there's a discord in the future, then we should both be adults about it and discuss it to make things right. Miss MSE's comment on my post that she had to involve her department chair and graduate coordinator in her advisor's lack of involvement is very unfortunate. I'm speculating (and you can tell me if I'm wrong) that if Miss MSE's advisor was approachable and available, that they could have discussed this issue to iron out some kinks.

Below are the main points that I want to get across to first year students looking for an advisor. They should know what their advisor expects from them right from the start and they should know what to expect from their advisor.

As a Grad Student, I expect you to...

• Show up, be diligent, be willing to try new/odd experiments, be creative
• Treat this like a job. When you have your PhD, you'll have to anyway
• Get comfortable in the lab, you'll be there often
• Write journal papers and attend conferences (it helps both of us)
• Present your research regularly, both internally and externally.
• Support your colleagues as needed
• Not be afraid to ask for help from myself or any of your colleagues. We should be willing to say I don't know

As a tradeoff, you will learn a ton of things along the way, such as...

• Lab Skillz. When you walk into any lab in the future, you will feel comfortable
• Multi-disciplinary approach to problem solving. See that soldering iron? It works for mechanical engineers, optical engineers, physicists, just about anyone.
• How to write and review journal papers, critique your work, write tactful answers to dumb reviewers
• To give presentations where no one will be asleep
• To be confident about your research and your abilities
• Enough knowledge to give you a sizable base for a career in _______

But you should have expectations from me. As your advisor, I will strive to...

• Answer your emails in a timely fashion
• Show up in the lab (frequently!, daily?!)
• Help you transition to be an autonomous researcher
• Work with you to achieve your career goals, whether they're in industry, academia, or national labs
• Get you connected in our greater research community
• Be available, approachable, and reasonable.

If, for any reason, neither of us are living up to this, then we should be able to discuss it like adults and come to a suitable conclusion.

Submit it anyway

I’m tired of hearing about poor graduate students whose advisors don’t take the time look over their papers or don’t give timely comments on their work. If it’s about to be submitted for a journal, then it should be important for the advisor and they have an obligation to look at it and give extensive comments in a timely fashion. If your advisor is failing to do that, then they’re doing a bad job of advising in my opinion.

Contrary to anything you might hear otherwise, all authors on a paper are treated equally even though some think they only have "management/guidance" roles. All authors should know the research and should be willing to help. More importantly, all authors get the same amount of credit no matter how much work they individual contributed. So if you are working on a paper and feel like someone's not contributing their part, then say so. If you think you've got a good start, just say so. Publishing helps all authors equally so we, as co-authors, should be willing to write some of the work.

Let’s take a paper with 5 authors on it. The breakdown is usually as follows:

• Author 1: Main author, did most of the writing and research. Can explain the technical aspects, chug through math, models, data, etc, but has limited scope for the overall picture.
• Author 2: Generally the main author’s partner in the lab, immediate mentor, or fellow PhD student who helped gather/analyze data. Helped with some of the writing.
• Authors 3 & 4: These two Authors contributed some technical aspects and offered advice. If you’re ever looking for these two Authors, my money is on them not being in the lab.
• Author 5: Often the main advisor, grant getting, knows the overall plan for the project but (all too often) doesn’t know the technical details.

Officially though, everyone is equally counted as a co-author on the paper, and that’s the aspect that I disagree with. If Authors 3 and 4 only gave a little bit of advice, then thank them in the acknowledgements. If Authors 1 and 2 have kicked the paper around for a while, Author 5 should have the decency to comment within a few days. Otherwise remove their name! I think it’s fair for you (as the main author) to send the work and say I need comments within X number of days or a very good reason why you can’t do that, otherwise I’ll just remove your name and submit it anyway.

You may be a graduate student but this is why we have peer review. There’s no reason why you need a professor’s name on the paper to get it published. In certain instances it helps. But if your work is good enough, then go for it. If it passes peer review, then it’s good enough and then maybe your advisor will take notice that you’re not messing around and you’re not just a lab minion.

Quirky New Engineering Prof Seeks Passionate, Creative Student for a Professional Research Relationship

While I haven’t officially started at SnowU, I am essentially preparing when I do start. Aside from looking over class notes, preparing a new lecture, and lining up my proposal schedule, I need to find students. SnowU has been very good about identifying potential students for me and I will get a chance to pitch my research to new students before they generally make their decisions about a PhD advisor.

So this past weekend, I worked on a few advertisements about the first few projects I want to run. I initially started with the generic Background, Project Description, Responsibilities, etc. After writing a little bit, I took a break and then looked them over. There were two things that struck me upon rereading the advertisements: 1) these say nothing about my personality or about how I would run my group and 2) I certainly wouldn’t want to work for myself based on what I read.

I scrapped those and started over again. I started to go with ads that were essentially personals. One of the headlines for a laser project was Wanted: Cute Redheaded Laser seeks crafty Engineer for a Controlling Relationship”. I know it sounds cheesy but that’s certainly more eye catchy than Project Title: Control of Red HeNe Lasers. I thought those might get me into a little bit of trouble, so I started over again.

In the end, I settled on a format that shows a little bit of personality, a little bit of my quirkiness, and tells a little bit about what I expect from students, while still mentioning the basics like project description, etc. These are much more to my liking, while still be on the good side of professional.

When you were a first year student roaming the halls, what caught your eye on the bulletin board? If you didn’t have a specific advisor in mind, would you take a chance on “The New Prof”? I tend to think (and did so for my MS) students prefer to work with a well established Prof because they think that will help them in the long run. For my PhD, I worked for a new Prof, was still successful, but was almost in spite of my Prof. Is that still the current sentiment? Would you work for a Prof whose position advertisements start with Quirky New Engineering Prof Seeks Passionate, Creative Student for a Professional Research Relationship?

Dual Advisor Conflicts

Currently, I’m in that grey area as far as my job/research is concerned. Mainly, I’m trying to finish a few more papers before I officially move on. One of my other tasks is mentor the next PhD student to continue this current research trend. So with NewGrad, I’m trying to get a few more experiments completely, enough for maybe 1-2 papers and we collectively have a few conference proceedings submitted/accepted.

Research with NewGrad is going well. I expect a lot from my mentees and I give them a lot of tasks, but I also support my mentees much more than other PhD/Postdoc students do for their mentees in my group.

But as I’m slowly faded out of the picture, it’s time for the real Advisor (also technically my advisor) to take over. Since I’ve been pretty autonomous as a researcher for the past ~2 years, I’ve had to set my own path for the research, so I know where it should go. Not to toot my own horn, but I have been pretty successful at doing so. Advisor does not have nearly as in depth knowledge about the research topic as I do and thus, Advisor doesn’t exactly know where it should go. And Advisor wants NewGrad to go in a different direction than the direction I’ve set forth. These are also the same wonky decisions Advisor steered me toward.

Concerning my future, I am not so worried. Once I have moved to SnowU, I will be guiding my own research, unless I get myself as a grad student. In general, I will be able to proceed down the path I want on this topic.

I am more concerned about NewGrad and how this will change NewGrad’s situation. There is clearly a disconnect between myself and Advisor and NewGrad is aware of that. Also, NewGrad knows that I’m leaving shortly and Advisor will be the main person to turn to for the rest of the PhD project. And NewGrad does not want to run the risk of alienating Advisor so early in the project.

I think it is a weird situation to be in for the both of us. What would you do if you were in NewGrad’s shoes? Have you been part of a supervising tandem where you had conflicts with the other supervisor? How did this affect your underlings? Thoughts? Comments?