Science Is Hard

Posted by Chad Orzel on July 27, 2006
(9)
More »

Jonah Lehrer at the Frontal Cortex asks an interesting question: Why is science so much work?

But I’m curious why science takes so long. I know this is an incredibly naive question, but why do post-docs have to work so hard? What is it about the scientific process that forces the average researcher to come in on Saturday (and sometimes Sunday)?

My own limited experience tells me that one of the main reasons science remains so labor-intensive is failure. Perhaps I was simply inept, but an astonishing amount of my time in the lab was spent repeating failed experiments, or repeating successful experiments that failed on the second try. After awhile, I became obsessed with knowing why certain PCR’s or Western’s or whatever didn’t work, but it was all to no avail. Even when everything was fresh and sterile, I would still find myself staring at a blank gel after 10 hours of hard work.

Does the cheap labor of post-docs (and the even cheaper labor of lab technicians) stifle the drive to reduce experimental failure? Or is experimental failure just a necessary inefficiency, a side-effect of doing science? If post-docs were paid more, would their be more interest in increasing scientific productivity?

The short answer is, it’s the nature of the business. The explanation is below the fold.

The key thing to remember here is that research, despite what you might think from the name, is the process of doing something new, something that’s never been done before. You’re measuring a new quantity, synthesizing a new molecule, blotting a new Westerner (help me out here…), boldly going where no human has gone before.

Not only do you not know what the result will be, you don’t even know if what you’re doing can work. There might be some effect you haven’t considered– either a fundamental principle, or a perverse technical quirk (“Porcupines are allergic to raisins.”)– that means your chosen method can’t possibly succeed. You won’t know until you try it. And in general, if these things were easy to do, somebody else would’ve done them already. At which point, it wouldn’t be worth doing that experiment at all, unless it was a step toward something else that nobody has done before.

Science is Hard, to paraphrase Barbie, and failure is an unavoidable part of the business. Most of what you try won’t work, but the only way to learn what does work is to try stuff and see.

A colleague at Williams once said something about introducing undergraduates to research that I have since stolen, and shamelessly repeat all the time: the hardest part about bringing a new student into a research setting is getting them to realize that it isn’t a three-hour lab class. New research students will generally keep at something for about the same length of time it would take an in-class activity to work, and then come say “This isn’t working.” At which point, you have to explain that not only do you not know if it should’ve worked by now, you don’t know if it will ever work, and they need to keep plugging away for several days at least.

Most scientists would kill to have the success rate of a lousy hitter in baseball– if I got one experiment in five to work on the first attempt, I’d have a publication list a mile long.

(9)
More »

Comments

  1. #1 Rob Knop
    July 27, 2006

    the hardest part about bringing a new student into a research setting is getting them to realize that it isn’t a three-hour lab class. New research students will generally keep at something for about the same length of time it would take an in-class activity to work, and then come say “This isn’t working.”

    Amen. I see this problem with graduate students, who haven’t figured out that research is different from a big class.

    There is another aspect to it. With a lab class, they’re always told just what to do. In the worst case scenario, the lab is a fill-in-the-blanks, follow the instructions and don’t think sort of affair. But even in the best case scenario, the lab is all prepackaged. It takes some time for research students to learn that in doing research, they don’t always have instructions– they have to be creative and figure out what to do, and that includes trying things that turn out to be stupid. What’s more, when you’re reading the literature, you have to explore — you can’t just read the papers your advisor tells you to read.

    I’ve seen grad students who were excellent classroom students but who never fully seemed to “get” the differences of research.

  2. #2 kstrna
    July 27, 2006

    It is true that the nature of science is to explore the unknown which means applying old techniques to new questions/situations or developing new techniques. You develop a hypothesis, you test it, you go back refine your hypothesis, so on and so forth. Once you get a good assay/technique then you can ask the questions you want to ask which also takes time. That doesn’t even entail the amount of effort to get like for a biochemist enough of your substrate, enzyme, etc. that is pure enough with all the quirks that entails (like for example: the E. coli under normal conditions just doesn’t want to overexpress your protein of interest).

    Now people do go off exploring why something doesn’t work a little too long or try way too many ideas. Figuring out when that is effective and when it isn’t is something to be learned in grad school & during a post-doc.

    There is a problem though in the system. The tendency I have seen is for professors to reward/like students that are like them, i.e. very passionate about research relative to other things in life. This is usually roughly “measured” by hours in lab working. When that is rewarded and the amount of data produced is about equal regardless of working 10 or 12 hours, then the tendency is for people not to work as efficiently as possible. It doesn’t encourage good time management skills. This increases the average hours/week. It also encourages those to come in on weekends when they otherwise might not need to.

    I am amazed by the faculty while in graduate school. They all are very busy & have a lot to do but then I watch and see how much time they waste by not being organized. Because they are soo passionate they are willing to put in super long days that masks organizational problems. Of course they are super-bright plays into that as well. They have a great ideas that pan out more frequently and can get the grants to do the experiments to begin with. Those who have other passions tend to leave grad school or leave academia. I have heard from students about faculty members who expected said students to work 80 hours a week. That just doesn’t encourage top notch organizational skills.

    The other passions include teaching. I TAed one extra semester and faculty members have been surprised I have been productive. The impression they usually impart is “wow, think how much more I could have accomplished if I hadn’t TAed”. In their minds any minute not spent in lab=time wasted. Little wonder that the quality of the teaching by many of the faculty in the department is lacking with that mentality.

  3. #3 Jamie Bowden
    July 28, 2006

    Can’t you just simulate it all on a computer using a model and call it done?

    :P

  4. #4 Mags
    July 28, 2006

    For the poor conused physicists:

    A western blot (more usually just called a Western) is just a (messy)way of detecting a known protein in a sample.

    You usually have a radiolabelled antibody to the known protein, and you run your sample on a gel. Then you blot it out of the gel onto a membrane and perform jiggery on that with you antibody so you can see where your protein migrated to, if present. It’s all very tedious. And phenominally sensitive to entirely random factors (not that I’m bitter about this or anything).

    A good run down is here – http://lifesciences.asu.edu/resources/mamajis/western/western.html

  5. #5 Doran
    July 28, 2006

    I was lucky to get a senior year physics lab class which did not consist completely of cookie-cutter work. The ethos of the class was “Here is your equipment, you have three weeks to get it running and analyze the results. Have fun.” This would involve much swearing at extremely fickle 10 year old electronics, and at times redesigning the experiments with new equipment. So one truly learns the ropes of doing experiments in a class setting, which I am at least grateful for.

    Chad, does your school have a Senior/Intro Grad – Advanced Lab class? How common is that across physics departments?

  6. #6 Adam C
    July 28, 2006

    Funny thing is by the sounds of it, my engineering physics program is training me more for what its like in research in physics than just doing physics research labs.

    Electronics labs take hours and hours of work, sometimes you never get them to work, sometimes you bash the power supply and it works. But mainly you spend good 8-10 hour sessions couple times in a week trying to drudge through the lab compared to the physics labs which you can blow through in a 4-5 hour session once a week.

  7. #7 revere
    July 30, 2006

    OK, let me be contrary and state a different point of view. Science isn’t hard, or at least it isn’t harder than many other things (NB: I am a scientist senior faculty who has published about 100 papers, so I’m not guessing about some of this; I’ve lived it). The discussion here seems to have confused “hard” with “putting in a lot of long hours at the bench.” If that is the definition, then being a post-doc isn’t much harder than running a restaurant or being a computer programmer who makes video games or someone who is building her own house with her own hands, etc. To say science is “hard” for this reason is a conceit that scientists work harder than other people and do it with their brains and not their hands. In all the other endeavors I just cited the main task is problem solving, not much different than what post-docs do. If you start a pizza place you run the risk of failure, you have problems to solve every minute, etc.

    The dirty secret we don’t teach our students is that most real research is tedious, time consuming and routine, just like any other kind of work. Whether you think it’s hard or the ride of a lifetime is primarily a matter of attitude. I left clinical medicine because even in the most advanced medical centers 95% of what a clinician does is routine, even if it’s the routine complicated or weird case. I didn’t like that kind of routine. i wanted the routine of my research area, and that’s what I got. For 30 years, now, I’ve had the luxury of having people lower down on the hill do a lot of the routine stuff for me because shit rolls down hill. But I substituted another kind of routine: committees, advisory work, teaching. I still get to do interesting research by harvesting the routine output of postdocs and graduate students but I get to do the “good stuff” even less than they do because of all the other stuff.

    I’m exaggerating a bit but only in the interests of provoking discussion and correcting a tendency I see in this thread.

  8. #8 Chad Orzel
    July 31, 2006

    Chad, does your school have a Senior/Intro Grad – Advanced Lab class? How common is that across physics departments?

    We do have a junior/senior level lab class, in which students do some more complicated lab experiments, in a more independent fashion than the typical cookie-cutter lab. I’ve only taught it once thus far (I’ll be doing it again this coming Winter term), and I wasn’t entirely happy with the way my module ran, but I have some ideas about how to improve things.

    As to how common it is, I really couldn’t say. Williams didn’t have a separate lab course, but they did have a lab associated with the junior-level modern physics class, which served much the same purpose. But that’s my entire sample right there.

    Revere’s comment is very interesting, but I think I’m going to promote that one to part of a forthcoming post.

  9. #9 assman
    August 2, 2006

    Actually I think research is also hard because professors tend to assign a lot of stupid projects to graduate students that require enormous effort and have very little payoff. This is because graduate students represent ridiculously underpriced labour and are subject to ridiculously high expectations. I also agree with what the professor above said. Most of academia is more tedious, mind-numbing and soul destroying than other jobs. It typically requires less problem solving and creativity than something like computer programming. There is only one real benefit to academia and that is that you have an excuse to learn about some really amazing things and if your are a genius do some amazing things.