Pedagogical Poll: Good Results or Historical Accuracy?

As a less-than-fascinated physics student doing the transparent version, I'd probably be more annoyed by the results than swept up by the transparency; as the same student doing the other demonstration, I probably would not be aware of what I was missing. So as a theoretical teaching moment, the transparent version seems better: the student sees what's going on *and* gets an introduction to "other things we don't understand can have an effect on our results" life of science. But I don't know how that would work in application. (I'm guessing there's some physics reason why the older version is so inaccurate, but it's something that's too complicated explain to this level of student, yes? )

If the purpose of the lab is to teach something about the photoelectric effect, then use the "black box" version. The other version would, in my mind, be a separate lab that is teaching about instrument design, not physics as such.

Old school.

I vote for the old school version for 2 reasons. Granted, I've not taught at a SLAC, so my experience is oriented towards future engineers.

First, fighting with equipment yet still working toward a goal. The real world is messy, and you are not making anyone's education better by hiding that from them. The practice is much more important than the number they turn in at the end. (Caveat: one does have to explain that it's not surprising for other processes/effects/broken equipment/etc to interfere with measurements, and that this is being done anyways. There won't be any black boxes available on the job.)

Second, oscilloscopes. Damned if I can recall how to really use them. Forces me to be extra clear and careful in assessing what students are doing or in aiding them as they work through lab exercises, which in the end is better for the students.

What's the intent of the lab and the level of the students?

If they're undergrads, and you're trying to teach them what the results are, and demonstrate that the results are what you say, then for the love of God, let them do it the accurate way.

If they're higher level undergrads or, for some reason, grad students, or if you're explicitly trying to teach them how to do indirect measurements then make them do it the hard way.

That Germanium essay could well have been written by my lab partner and I when we were sophomores(!!) in college, because we did a similar experiment, with similarly catastrophic results, probably because the damned diode was older than we were, and PN junctions in dime store diodes might diffuse over the course of thirty years. Yes, I'm still really bitter about it, too. because that was just an inappropriate lab for sophomores who need to know what the physics are, as opposed to (say) seniors who need to know how to dream up indirect measurements. (God, I hated that class.)

(Scans up, reads comments.) So, um, yeah, I agree with Tim. I'm just still bitter and hostile, eighteen years later....

Also, realistically speaking, you aren't really contributing anything to education by assuming that these are anything but exercises. We might be using the word "lab", but this isn't research. We shouldn't pretend that it is, or that anything groundbreaking will come out of freshmen fiddling with equipment.

And if something novel (in the research sense) does show up, you hire them...

I say old school. The point of labs isn't to get the right answer, its the process of doing them. The way I see it (and the way its done at the school I am at - I am TAing sophomore undergraduate labs at the moment) is usually the experiments are done with fairly old equipment, and then one black box/plug and play apparatus is there to show the students how wonderful modern technology is. Sometimes the labs are painful, but there will be a lot more learning going on, and learning how to troubleshoot is a very valuable skill that is hard to learn with black boxes that you can't tinker with.

Like others have said - it depends on the purpose of the course.

If the purpose is to train students in getting plancks' constant (you know, because girls like guys with skills - planck's constant skills) - then go with the blackbox version.

If the old-school version is too tedious, then just don't do that lab (is it an option).

For me, blackboxes don't really do much. You could just make a blackbox that when a student pushes a button, gives planck's constant +/- 1% (with a little random number generator inside).

That link to the Electron Band Structure lab report was The Awesome. Thanks.

Since this is a physics lab, and not an instrumentation lab, I would say use the new equipment.

However, since you have the time, it may be worth setting up at least one set of the old gear for demonstration purposes. You might even be able to do a single run through to get data for a 'compare and contrast' bit in the report, (e.g. having them come up with reasons for how bad the results from the old school gear are).

I agree that it depends on the level of the students in the class, and the purpose of the course.

In fact, during my senior year as an undergraduate physics student, we had an upper level lab class. One group in this class actually took that Pasco experiment you mentioned, and used it as a guide to built their own photoelectric effect experiment. As upper level physics students, we did not need to be convinced that the photoelectric effect works the way the textbooks say it does, and simply seeing the right answer pop out would have been a bit useless to us. Building the experiment ourselves however, was a valuable experience. Indeed, it didn't work very well, but in exploring the ways in which it failed we probably learned more about the practice of experimental science than we otherwise would have.

In terms of actually doing research however, I'd rather have the commercially produced, high quality apparatus, whenever possible. With the sole caveat that the manufacturer must provide adequate explanations of how it arrives at its results, and tests showing that they are indeed valid.

Unrelated, but I mention this because it will be familiar to you, Chad. In this lab class, I was in a different group than the one that did the photoelectric effect. I built an external cavity stabilized diode laser, and used it to examine the hyperfine structure of Rb around 780nm. This laser was then donated to the research of the professor who taught the lab class. He uses it (and about a dozen like it) for atom trapping experiments. Its been a while since I was there, but I believe he is working on dual-species traps, MOT-on-a-chip experiments, and producing arrays of vortexes in Bose-Einstein condensates using beams with spatial modes containing angular momentum. It was an interesting place to work. Quantum optics is sweet.

I suppose there's no way they could do both, or at least a scaled-down version or demo of the old school experiment? There's definitely something to be said for seeing how things are supposed to work and getting experience with o-scopes. In the present economy, you never know when people may have to revert to actually building their own equipment (gasp!) instead of spending big money on hi-tech gadgets. They'll realize that physics can be done on stuff that isn't so expensive (isn't that the tradition of Russian physicists?). On the other hand, it's nice to see that technology has progressed in the last half century, too, and there's nothing more frustrating than doing a lab that doesn't seem to work.

See, my feeling is that the purpose of the lab is to show that experimental measurements of the photoelectric effect agree well with the Einstein model. The more complicated version doesn't really add to that understanding, and in fact, the complication tends to obscure the physics. Students spend so much time fretting over the experimental details that they lose track of what it's supposed to show.

You can argue that they're learning lab skills in the process, but I'm not all that impressed. The only really useful thing they get out of it is how to use an oscilloscope, and there are other ways to teach that. There's some fuzzy data-selection heuristic stuff going on in deciding exactly what to use as the stopping potential for any given point, but it's hard to explain that in such a way that they don't leave the lab thinking "it's ok to fiddle with the data to get something closer to the target value." That's not only not what we'd like them to learn, but is actively harmful.

My feelings on this are also influenced by the fact that I tend to follow the Six Ideas That shaped Physics method of talking about the photoelectric effect, by speaking of it in terms of simpler idealized experiments. The transition from the idea of the stopping potential as something that occurs naturally to the idea of applying something to stop the current is a little awkward, and again, tends to obscure the important physics.

Do you want them to get the right answer or to appreciate that good science is not an easy thing to achieve? Doing all the Official statistical crap on data infuriates students, and well it should when applied to manicured data. Go pull a legitimate signal out of real world noise.

Uncle Al saved a research project that could not for the life of the group make a phosphatase assay work. "Don't use phosphate pH buffer." Curiously, they weren't effusively grateful for having received the answer.

>the purpose of the lab is to show that experimental
>measurements of the photoelectric effect agree
>well with the Einstein model.

That is easy to accomplish. "And it was shown that the Einstein Model agreed well with the experimental measurements. See you all next week."

(I have almost never met a lab I felt was worth the pedagogical trouble.)

I lean toward good equipment. There is certainly a lot to be said for struggling with difficult instrumentation, but if you're going to give people bad instrumentation to fix up then you need to give them enough time to succeed at it (i.e. probably more than a single session of 3 hours), and some background on the instrumentation, not just on the phenomenon being measured. Without that context and time, there isn't much that they'll learn.

Yeah, yeah, when I was their age we walked uphill through 10 feet of snow to use broken equipment and learn how to struggle with shitty equipment and survive, but still. A bit of frustration is useful, but if the result of all that frustration is a number that's an order of magnitude off, and no chance to improve, what was the point? There should be a reward for that effort, not just a lesson that life sucks.

Besides, even with a fairly reliable apparatus, there can still be data analysis steps or calibrations, and if you want to teach some technique you can even have them check the effects of different things (adding a filter here, playing with a calibration parameter there, whatever). Unless, of course, the machine is so black box that you turn the key and the number comes out and that's that. But as long as the more reliable instrument still offers some way of checking process and analyzing an indirect step, there's a lot that can be learned in a well-designed exercise.

See, my feeling is that the purpose of the lab is to show that experimental measurements of the photoelectric effect agree well with the Einstein model.

If that's all you want to do couldn't you print out some relevant papers from Physical Review or Annals or whatever journal it was and show them data? I'd argue that reading and making sense of a relevant paper would almost certainly have more merit than pressing buttons and watching a box that they know next to nothing about doing its thing.

The older setup sounds like it offers some benefits. There seems to be some thinking in designing the experiment to do what they want. And the error sounds more like a feature than a bug - it allows for more questioning about whether the model makes sense, and for where these errors are coming from. I personally see very little pedagogical value in the black box approach, at least as I understand it here.

Given your stated purpose, then I have to advocate for the good equipment. There is a time and a place for teaching difficult, even creative, metrology techniques, but it doesn't sound like your class is at that location.

And there is a real risk-- take it from a former student who has a real grudge against that other professor-- that you'll end up just annoying the hell out of a student who would be otherwise enthusiastic about the subject. Also, this might be obvious, but if you do end up teaching labs that are all about difficult and creative test measurements with big error bars... tell the students that that is the point, so they don't freak.

That other professor may have been trying to achieve the same thing, but he didn't help himself by hiding it.

See, my feeling is that the purpose of the lab is to show that experimental measurements of the photoelectric effect agree well with the Einstein model.

If that's your goal, then by all means use the modern equipment. I've always been a fan of the "try to teach students about the process of science, not just the results", though, and understanding that it isn't all simple and certainly isn't a black box is really important to that. Presumably there are other lab sessions in the course, though, where "process" and "results" aren't quite so violently opposed.

I should also note that, at least the way I did it this term, the lab is not totally a black box. I provided light sources (mercury lamp and HeNe laser),diffraction gratings, lenses, and the detectors, along with various mounting hardware. The students had to decide for themselves how to set things up so as to be able to make the necessary measurements.

The detector is on the black-box side, but not much more so than most of the rest of the equipment we use (scintillation counters for the Compton effect and positron annihilation experiments, for example).

This made me recall a lab I did when taking freshman chemistry, years ago. It's a well-known but wildly inaccurate method for finding Avogadro's number. You create an _allegedly_ one-molecule-thick layer of some floating liquid and measure its area with a dusting of lycopodium powder that sticks to it. There was _considerable_ variance from student to student in the values obtained. I remember joking that my percentage error was closer to Avogadro's number than my value for Avogadro's number.

But you know what - All these years later, this is the _only_ freshman chemistry lab that I can still clearly remember. The _lesson_ stuck, quite independent of the gross inaccuracy of the method.

::Lab B: Box oâ Parts Method:

I did this lab in a senior level âIntermediate Laboratory Physicsâ course. My professor gave us a phototube and plopped us in a room with power supplies, âscopes, a big monochromator in the back corner (hand-cranked! We had to calibrate the scanning range ourselves because the dial was broken), and a mercury vapor lamp (and a plethora of other equipment acting as red herrings. )

The prof. gave 2 weeks to set up an experiment and present findings. We were thrilled after 1.5 weeks of crazy hard work to get values accurate to 5%. But more importantly we were expected to justify our measurements and account for every bit of our error.

This was the real value of that class: not demonstrating the effects that the labs were specifically about, but learning to take careful measurements and identify your sources of error.

So actually I guess I'm changing my vote to = it depends.
If the goal is to demonstrate an effect then maybe go with the Black Box.
If the goal is to teach the dirty Physics of the Photoelectric effect then monochromator + tubes all the way.

P.S. -- This was one of my favorite labs of that semester, thanks for the good memories.

Physics high school teacher from Sweden writing.

What is the point of the lab? When I do it the point is to make the students understand the physics not to actually measure a constant they can find in physics handbook. Black box labs normally donât do a very good job of making students think about what is going on.

When I plan a lab I try to be clear about the physics concepts the lab is about, and how the equipment work but not how they should use it. Getting numerical results +20% in education setting is good enough in my opinion, the result of the lab is how many students have a better understanding of the photo electric effect after the lab.

Pedagogy must be informed by the level of the students.

Your "modern" course is only one step removed from the 2nd semester physics class I teach, where most of the students are going to be engineers. Nearby University gets a lot of engineers in their "modern" course because it leaves only a short step to a physics minor for them and is relevant for EE majors, but you might not. In any case, we can't forget that they are, for the most part, still concrete learners - and that some might want to be theorists.

Formal and informal surveys indicate that my students like labs that are short (1 hour beats 2 hours) and that give correct answers. They will trade time for "really cool" observable phenomena, and don't mind the time quite as much when the results work out.

I like the suggestion about having one version of the setup for demonstration purposes. Maybe even operate it yourself. (That will entertain them no end!) I do something like that for our oscilloscope lab, pulling out a really old one to show what is inside an analog scope and that it actually works.

I also second the suggestion of designing an advanced lab around the "how this works" concept. If you know you will have control over that lab course at some point in the near future, you could even tell them that your plan is to explore what is going on inside the PASCO box so they can begin to learn how to earn money by designing physical apparatus.

I should first state that I think the purpose of undergrad labs shouldn't be to let students validate models themselves, but to teach them experimental methodology. From that point of view, I'd vote for the old school version.

That aside, though, I don't see how doing the experiment with the PASCO black-box detectors would "show that experimental measurements of the photoelectric effect agree well with the Einstein model." Suppose I was skeptical of that Einstein's model has been validated. Would I be convinced that it has by doing the PASCO experiment? No, because I don't know that the apparatus accurately converts the energies it measures into stopping potentials, and that the values output by the apparatus actually are those of the stopping potentials. It's natural to suspect that what the black box is doing isn't what my instructor claims it's doing, since the instructor has a vested interest in telling lies-to-children about what the box does. (I will ignore the question of whether I should trust my instructor. But we can see that inferring the model's goodness from the black box experiment involves an extra epistemic step, so a skeptic of the model has a good reason for being less convinced by the black box version than by the old school version.)

I'd say it depends on (a) the level of the students and (b) the amount of time you have. It's a great idea to get students to understand the full process by using the old-school method, but only if you have time to do it right and the students have the background to understand what's going on.

I've seen some real-life biological equivalents of "Electron Band Structure of Germanium, My Ass", and they were the results of trying to teach both a method and a concept to a group of rank beginners during the same introductory lab where we had to completely use up the first hour with bureaucratic stuff.

If you do have time and are not certain of the students' prior preparation: Why not do the lab with the black box and then repeat it later in the semester with the other equipment? That would be a lesson in scientific practice that very few students ever get.

There was a photoelectric effect problem on the Physics GRE which involved the old-style lab setup, so it'd be good for students to understand the principles behind it. Since the PASCO-aided lab takes less than an hour, why not spend the rest of the lab section talking about the old-school method? Just talking about it abstractly won't be as "fun" as actually doing it, but you won't have to deal with o'scopes and the vagaries of equipment.

Hey, I stumbled onto this blog at random, but as an undergraduate physics student, I figured I might as well throw in my two cents.

I would LOVE for once to do a lab for determining the value of a constant that produced results that weren't off by an order of magnitude . Good data makes labs a pleasure to write, and for once I'm proud to show off my margins of error. Otherwise I find myself just spending half the report trying to find reasons for failure other than "The equipment was a pile of crap".

I think that you can remove some of the problems of lack of transparency by requiring a beefed up introduction where students are required to explain what the equipment does. I don't know about physics programs at other universities, but I had plenty of opportunity to use an osciloscope in electricty and magnetism labs in second year. Since this is meant to be done in a single lab period, where students will not have an opportuinty to actually coax anything reasonable out of the old equipment, or attempt to improve their technique, I feel that the lab loses value on the write up side, and doesn't gain much on the practical side. They aren't learning how to get good results out of oldschool equipment, they're just getting bad results.

... I've gotten plenty of bad data, in my undergraduate labs. It's more the norm than the exception. As long as every experiment that students do is not a 'black box', then they will still learn to deal with bad data. At least here they can deal with good data too.

I agree with Ursula. There's a time and a place for dealing with errors, but there's also a time and a place for having something work. Besides, in an earlier incarnation I did experimental physics, and I can't think of any difficult piece of equipment that I got good results from on the first try. Yet I was willing to do those initial failed sessions because I knew there'd be a pay-off in later sessions. When we have students wrestle with bad equipment for one session and then take it away, we deny them the satisfying part of struggling with a bad apparatus: Finally mastering it and making us of that mastery. If we're going to give them crappy equipment, let's let them take 2 weeks on the experiment.

Bonuses for us:

1) We only have to buy half as much equipment because we only do half as many labs. (That sound you hear is my Dean shouting "Yes! Yes! Yes!")
2) If students actually have the time to get something down, we can expect lab lab reports that say something other than "The error was caused by human error" or "The equipment must be broken" or "The error was caused by standard uncertainty in random experimental error." (Yes, that last piece of word salad has shown up in reports.) If they have time to figure this out, they can actually figure out which piece of the apparatus does what, and whether a malfunction in that piece would give a result that's too big or too small.

3) If they do half as many experiments, we grade half as many lab reports.

This is a debate that has been going on for years. I have taught labs in both variants: complex equipment with good results vs. not-so-good equipment with not-so-good results. I am torn. The good results are nice, and students are frustrated when the experiment doesn't work. If you are not careful, they get to thinking that physics is inexact. On the other hand, if the experiment is too much of a black box, or if components of the experiment are way beyond the level of the student to understand (just exactly what does the phase lock amplifier do?), then they don't really understand how the experiment is being performed. The idea of the lab is to reinforce the concepts of the lecture.

For my majors' course, then the more accurate results are better. Those students can better understand what the black box does. For the non-majors' course, though, the black box is just magic, and they are better off with the low-tech equipment and lesser quality results. One thing that I sometimes do, though, is to combine exercises. They do a low-tech version to see how the physics works, and then a high-tech version to see the good results. They often don't understand everything in the high-tech version, but it verifies the concepts, and the low-tech version, with its poor results, shows them qualitatively how the physics is working in a way that they can understand.

The old one you can see inside. We KNOW hbar. Thats NOT the point of these labs!!!! It's to teach technique and error analysis. Now after that one to show them the Pasco one so they know you CAN measure better is great, or do that in a lecture demo.

hbar is one anyway ferchrissakes......

Labs are to learn, not confirm lecture. When they bitch "we haven't talked about this in lecture" they must be told, or just told at the very beginning that IT'S OK TO DO THE LAB FIRST.

I would have the students assemble and run the experiment on the old-school setup but have the PASCO experiment set up at the head of the class and have each see the demonstration of how it is built and used. Also a handout that explains the difference in the experiments and why one is more direct and intuitively obvious but yields vague results whereas the other is more a conceptual leap of faith as to what is going on but nails the results.

Best of both worlds and it gives a deeper understanding of both the underlying concepts your covering and the useful understanding that science is often a matter of sifting the information from raw data. Seeing through the noise and artifacts produced by imperfect apparatus and conditions and seeing the connections.

Lower level undergrad: Black box.
Upper level undergrad: Old school.

In my undergraduate degree, all of the labs were old school. It took quite a while to stop writing lab reports as "Electron Band Structure In Germanium, My Ass", but gradually I got it: it wasn't that I was rubbish at physics, or that the accepted results were wrong. The problem was that doing physics experiments is hard, and you need to understand what you're trying to do and be very careful. I'm very grateful I learned that lesson, and it's served me well in more than just work.

Oops, forgot to add: When I did my masters degree, I started at the same time as a guy who had done his undergrad at the same time in a different college with black box labs. He was seriously discouraged for the first year, because he had no experience of things taking weeks to work not because they're broken or you're an idiot, but because it's hard.