# The Purposes and Problems of Labs

Over at Jim Henley’s place, Thoreau further justifies his status as an essential academic-physics blogger with a really good post about the problem of introductory labs:

In freshman labs, generally you?re trying to measure something (at least as it?s done at many schools). The measurement is never as clean as the stuff being taught in lecture (or interactive discussion-based peer-involved blah blah whatever). There is nothing wrong with the fact that lab measurements are not as clean as the stuff in lecture! However, it does mean that you aren?t spending those 3 hours thinking about the same things you were thinking about in lecture. I mean, yes, there?s probably a rotating object (or whatever) in the lab, and the professor?s diagrams in lecture probably also had a rotating object. But in lab you?re thinking about the best way to measure the tensi0n in that spring and trying to figure out whether the biggest factor dominating your uncertainty is the difficulty of keeping the system stable or maybe it?s the fickle nature of this meter you were using, and then you have to remember to subtract out this other force in your data analysis, and set up your data table correctly.

After two and a half hours, if all goes well, you get a graph that involves force and angular velocity or something, and that graph might even look like the graph that your professor produced in lecture when explaining some concept about rotational motion.* You are now ready to start your write-up and answer some questions before the 3 hours are up. However, you didn?t spend two and a half hours thinking about rotational motion. You spent two and a half hours trying to get this device to be stable and whatnot. Odds are that you didn?t learn anything new about rotational motion in that time. If things were done right, you probably did learn something about how to troubleshoot a tricky instrument.

He’s absolutely right about the problems here. In fact, this problem extends beyond the introductory level– no matter what the level of the lab class, labs present a significant problem. There are three things you might hope to accomplish in a lab class, and those three goals do not coexist peacefully.

The three goals of a lab class are, in no particular order:

• To teach students lab skills (how to use equipment, how to make measurements, how to record and analyze data)
• To complement the physics taught in lecture (either by providing a direct example of the physics discussed in class, or by exploring applications that aren’t covered in class)
• To teach students something about scientific writing (through writing lab reports)

These goals conflict with each other, in that trying to pursue one of them creates problems for the others.

If you want to teach students how to write, you necessarily need to have them do a lot of writing, which argues for doing as many labs as possible. On the other hand, if you want to teach lab skills, you need to give them time and freedom to experiment with the apparatus, which means doing a smaller number of longer labs. If you want to complement the lectures, you need labs that will reliably give something close to the theoretical prediction (as Thoreau notes), which argues for relatively “canned” labs, with minimal opportunities for students to mess up the apparatus.

You can sort-of satisfy two of these, by having students do lots of “canned” experiments and write them up, but the only lab skill that teaches is how to follow precise directions, which is not actually that useful in research (or anything else). There isn’t really a good way to do all three, though, and I’m not convinced that writing up “canned” experiments actually teaches that much about scientific writing (though this may be because I hate grading large numbers of reports whose procedure sections are basically copied from the lab hand-out).

This is a Hard Problem, and I don’t have a good solution to it. Anybody who does have a good solution, please leave a comment telling me what it is, because I would love to know.

1. #1 Jamie
September 18, 2009

Skip the labs all together, voila, problem solved.

You’re welcome.

2. #2 Laelaps
September 18, 2009

The strangest lab I ever participated in was the 2nd lab of a university Bio 101 course. I forget exactly what we were doing, but we were using the mass spec. to get results. At the end of the lab, though, everyone’s results were way off from what was expected. Rather than talk about this the TA said “Oh yeah, the same thing goes wrong every year in every lab. I’ll just give you the measurements you should have gotten.” Way to bring a sense of ethics into the lab…

3. #3 Uncle Al
September 18, 2009

Management is about process not product. Management is concerned with models, allocations, resources, doing lunch, and issuing orders. The real world is an unseemly lower echelon burden.

All labs must be reduced to software to control costs and eliminate inventory costs, storage, and depreciation. Risks from injuries will be ended. The few students who persist will become theorists whose torrents of pride are empirically untestable (or they will not be grant-funded). Studies must be conducted at the highest levels, replete with meetings and documentation. No labs! Management makes decisions, workers make mistakes.

Lab courses are a national security breach. Lab courses are each and all ITAR – International Trafficking in Arms. Lab courses will be overseen by Homeland Severity. Do you want your transportation hub destroyed?

4. #4 JohnV
September 18, 2009

You used a mass spec in intro biology?

5. #5 Anne Nonymous
September 18, 2009

Skip the canned labs. The students have opportunities on the homework to play around with the theory, and they can be assigned papers and long-answer homework questions to learn scientific writing. So the canned labs don’t provide any unique benefits, especially if the students are basically just copying the write-up from their lab manuals anyway. But they have nowhere else in the academic curriculum to learn how to just get down and dirty with the messy ol’ real world, in a scientific and systematic fashion. Do fewer labs, and make them worth doing.

6. #6 Sam K.
September 18, 2009

someone at Thoreau’s blog suggests: http://umdperg.pbworks.com/Scientific-Community-Labs. It definitely sounds interesting, and seems to at least attempt to cover all 3 goals you mentioned. You use lecture taught skills to design the experiment, you use your lab skills to carry it out, and since it’s pretty open ended, your scientific writing skills should be exercised too.

7. #7 Laelaps
September 18, 2009

John; That’s what they called it. It’s not something I have ever used since so I couldn’t say what it really was if they got it wrong.

8. #8 Thoreau
September 18, 2009

I suppose that if one wanted to go really far down the radical road of teaching concepts and writing, you could have them do simulations instead of labs (like some U. Colorado researchers suggest) and write up what they see. Even then, if you want them to explore a phenomenon, document their findings, draw conclusions, and justify those conclusions, it might take some time to really figure out what to simulate. Even if the simulation is totally plug and play and easy to generate output from, if they have several parameters to play with it takes time to explore parameter space, it takes time for a novice to figure out how finely to sample parameter space and over how large of a region in order to justify a general conclusion, and so forth. It takes time to figure out if a parameter should be sampled linearly or logarithmically (i.e. in steps of 1, 2, 3, 4, etc. or 10, 100, 1000, etc.).

So a good exploration of a simulation might take a couple of weeks.

And if they have to code it themselves, trouble-shoot it, validate it, and test the robustness against things like finite step size or whatever? No way are they doing one of these reports every week.

9. #9 Eric Goebelbecker
September 18, 2009

I think you should ask Emmy. She’d explain that Labs are for retrieving balls and birds.

10. #10 Jonathan Vos Post
September 18, 2009

My wife has taught Physics lab courses in 4 countries. I only taught one at college level. But I did receive superb Physics lab training, first at Stuyvesant High school, in the same class with America’s Undersecretary of Science Steve Koonin, where among other things I used Hall effect measurements on an Indium Antimonide sample I got from Watson Research Labs at IBM, to weight the charge carriers — holes. Also, the Freshman Physics Lab at Caltech. Since I’m a Theoretician, not an experimentalist, I worked extra hard. I would do 100 times more measurements than my classmates, and get an order of magnitude better results, after my thorough data analysis (using computers for the statsitics was unusual among Freshmen in 1968).

I think that such labs are absolutely essential in education. Hard to teach, but there is no substitute known.

11. #11 MRW
September 18, 2009

Perhaps a relevant quote:

“I can explain the conservation of momentum in 15 minutes, but three hours in the lab would only convince an honest student that the law is false.” David Griffiths

12. #12 CCPhysicist
September 18, 2009

You left out “design an experiment to test a specific idea”.

By the way, the “studio” approach is probably a better way to attack the second goal of complementing what is done in lecture.

13. #13 Kim Hannula
September 18, 2009

From the perspective of the other sciences that make students take intro physics, I would say that the first goal, to make measurements and play with the equipment, is the stuff that makes intro physics classes stand out in the intro science curriculum. The theory is important, too, but the number sense (and unit sense!) that comes with the kind of measurements people make in a Newtonian mechanics class are important.

Writing, on the other hand, can develop through time, and I’m happy teaching my students to write like geologists. But I’m never going to be as good at building a piece of equipment as the average experimental physicist (or the average physics major who becomes a geophysicist) is. So I would rather have my students learn that stuff from you guys.

14. #14 John Coleman
September 18, 2009

I hope this gets to the real point. I’m an old fart. But I had a live changing experience in a lab. This could be a long and potentially a very boring story, but I’ll make it short and get to the point. I was in an EE lab and we performed a really very simple experiment.Our results did not meet the very simple theoretical expectations.
We gave our results sheet to the professor. He threw it on the floor and said “This is shit.” He demanded we explain the results. We redid the experiment, resubmitted it with the same “shit” result. We eventually got the essence of the lab, not the experiment. The experiment was only (because of this brilliant professor) only a means to an
end. Obviously, it’s a lab I remember to this day. The experiment was performed 50 years ago.

15. #15 Anonymous Coward
September 18, 2009

In an intro lab, the first goal is the most important, and I would happily abandon the other two. I have rather extreme views: until the student learns the first item, I don’t see how labs could help with the second item. The students might as well watch a video demo.

To use an (admittedly unfair) analogy: we don’t start students in physics before they take algebra, so why do we use “canned experiments” that start them doing labwork before they understand the basic tools of labwork?

I would expand on the “how to record and analyze data”: for any future work in a quantitative science, it’s vital that the student understands what a graph means, what an error bar means, and what it means to fit data to a function. As far as I can tell, this is the only thing I learned in my first-semester freshman lab (other than “air tracks and gravity both work as expected”), but it was extremely helpful for all following labs and experimental work. It is also something that is sorely missing in student education at my current institution.

Once the student has mastered #1, then it’s time to look at some physics.

16. #16 Kaleberg
September 18, 2009

There is another reason for physics labs: to teach students where the stuff they learned in class comes from. That’s actually the most important reason when you come down to it. The laws of physics didn’t come from impeccable logical reasoning. They didn’t come from good writing. They didn’t come from computer simulations. They came from messy, confusing, conflicting, frustrating, annoying, tedious lab work.

In history, serious courses require students to read original documents. Contemporary observers rarely have a clue. You aren’t going to find a Declaration of Hundred Years War. Original documents were created to press grievances, announce events, record sentiments, urge action, inform superiors or record transactions.

Just as physicists make physics out of laboratory experiments, historians create history out of its raw materials. Just as serious history classes have to include some original sources, serious physics classes should include some lab work. It isn’t because it is good for one’s soul. It’s because it is at the heart of the subject.

17. #17 Sally Pardue
September 18, 2009

There is a missing goal from the list of three above. Promote critical thinking. This goal ties all the other three together. It is at the juncture of theoretical predictions and measured reality that students can experience significant learning. When we engage the student in the act of critical thinking (communicated through good writing as one means to measure their knowledge processing) we can open up meaningful dialog with them about the limitations of the theory (remember all those pesky assumptions) and the challenges of recognizing and then minimizing sources of uncertainty in our measurements.

18. #18 Tom English
September 19, 2009

One approach might be to have the students keep a laboratory logbook in which they document their experiences. Encourage them to record not only procedural information and observations, but also their frustrations and eureka moments. This log could supplement the more formal writeups and in the meantime provide a real “writing-to-learn” opportunity. I use something like this in my introductory observational astronomy course, where the observing log provides insight into the struggle to make sense of the workings of the sky. It makes for interesting reading, especially when you can see where students are making connections and when earlier confusions are resolved.

19. #19 Thoreau
September 19, 2009

There is another reason for physics labs: to teach students where the stuff they learned in class comes from. That’s actually the most important reason when you come down to it. The laws of physics didn’t come from impeccable logical reasoning. They didn’t come from good writing. They didn’t come from computer simulations. They came from messy, confusing, conflicting, frustrating, annoying, tedious lab work.

If the freshman physics labs were actually structured to help the students follow the path from messy, confusing results, on through refinement and testing, and all the way to some sort of synthesis where an experiment has been refined to get at a clean result that was hinted at in the earlier mess, I would agree 100%.

Instead, it’s 3 hours of mess, followed by a write-up that says “We did not verify Newton’s Laws, but, um, we know Newton’s Laws are true, so our data must be wrong” and then next week we’re on to some other mess.

September 19, 2009

There is another reason for physics labs: to teach students where the stuff they learned in class comes from. That’s actually the most important reason when you come down to it. The laws of physics didn’t come from impeccable logical reasoning. They didn’t come from good writing. They didn’t come from computer simulations. They came from messy, confusing, conflicting, frustrating, annoying, tedious lab work.

I count that as part of “lab skills.” Learning how to get good measurements from messy data is the ultimate lab skill.

I agree that this would be a wonderful thing to teach, but as I said above, this is something that requires a lot of time. You won’t get any meaningful progress on this in a single three-hour lab period. Hence, it is in conflict with the other goals of lab classes.

21. #21 Åka
September 19, 2009

Hmm, another benefit of labs is that they provide an important complement for those with a practical learning style. Some people learn better if they can connect the subject to anything they can touch. Also, I personally learned a lot from the informal discussions in the labs.

22. #22 Brian
September 19, 2009

One thing you might try is hitting concepts in lab before they appear in lecture. Give students projects that you know will result in the correct results, but make them think through how the measurements need to be made, and how to discuss the data. You can still have them write up the results, maybe after they’ve gained (hopefully) more insight after lecture. Seems like it might be more like how we really do science too.

September 19, 2009

Canned undergraduate lab courses are worse than a waste of time. They are actively harmful in creating a completely false impression of what scientific research is like, and turning students off of science. I say do away with them completely.

24. #24 Thoreau
September 19, 2009

Hmm, another benefit of labs is that they provide an important complement for those with a practical learning style. Some people learn better if they can connect the subject to anything they can touch.

That’s a great goal, and consistent with the goal of promoting conceptual understanding. However, the typical canned physics lab does not have much direct connection with the concepts. You aren’t handling objects that clearly display the phenomena discussed in class. You’re handling things that display all sorts of phenomena distinct from what’s seen in lecture.

If you learn best by applying your mind to things that you handle rather than things that you see on the board, you’ll spend those 3 hours thinking about trouble-shooting and errors and taming a mechanical device. These are important skills, and I think it’s great to teach them and engage students in learning them through a hands-on approach, but we need to stop pretending that in teaching these skills we’re also reinforcing what’s taught in lecture.

Now, if you want to complement the stuff taught in lecture with a hands-on approach, you need to have experiments that provide transparent, intuitive demonstrations of physical phenomena, not 3 hour measurement exercises that are forgotten the next week.

25. #25 agm
September 19, 2009

Hmm. Thoreau, you raise an interesting point: “If you learn best by applying your mind to things that you handle rather than things that you see on the board, you’ll spend those 3 hours thinking about trouble-shooting and errors and taming a mechanical device.”

In the environments I’ve been in (i.e., nothing like what Vos Post wrote about), these classes are basically for engineering majors. Full stop. The next largest contingent was premeds at my undergrad and grad institutions. And this is exactly what a lot of them will be doing as part of their jobs. Physicists need to learn how to teach that lesson better, “your job won’t be easy, and there will lots of messy things to resolve, and we’re starting to teach you how to get a result that is good enough to get the job done”

26. #26 CCPhysicist
September 19, 2009

The remark @22 is what underlies a new ? “discovery” approach to lab instruction that I have heard about from some friends in chemistry. Hence my suggestion @12.

However, what little I know about this approach suggests that it would require a radical change in how we go about teaching labs in physics. We would have to emphasize fewer skills and hope (try to ensure?) that they will retain a large fraction of that smaller skill set rather than almost nothing of the many different things we throw at them.

When writing that sentence, I realize that there is a similar philosophical discussion about how we approach the lecture course, which is but slowly evolving away from the mini-PhD structure it had when I was an undergrad, which was possibly before you were born.