Over at The Blog of Doom, the Cap’n makes some good points, namely

Labs are not ways to reinforce your teaching. Labs are ways to teach.
…
The goal of labs should be to let students figure things out for themselves. Let’s not tell students to verify Schmoe’s Law. Let’s tell them, “there might be a relationship between these two variables. Find out what it is and if there’s an equation that can describe the relationship. You get to design your own experiment to do so.”

That would be real learning.

(emphasis in original, though it doesn’t seem to show up in the block quote)

Yes! I agree. One problem I see, though, is that it’s difficult to recreate the atmosphere of having the science be an unknown, since the students can simply read ahead and know what the answer is supposed to be. One possibility is with simulations. You could, for example, have a computer simulate a CRT with a new and different magnetic and electric field — instead of having the Lorentz force law hold (F = qE + qv X B) you would make things nonlinear and have the students deduce the law. The source might be limited to just a few energies (you could pretend they are monoenergetic radioactive samples), and maybe the source particles are “mysterions” that don’t have an integral charge. Let them have a taste of what it’s like to do an investigation where you don’t know what the answer is supposed to be.

I think simulations fail in some circumstances, though. I recall a computer program that simulated single-particle interference building up over time to give the familiar pattern. But it didn’t actually prove anything, since the pattern will be whatever the programmer wants it to be. New and different science probably needs to be more involved with the actual apparatus. When I was a TA for a “modern physics” class I was a little surprised at how neat the students found the labs to be, even those not majoring in physics, even though it wasn’t as “hands-on” as the introductory/general physics labs were; you were relying on some measurement apparatus to show electron diffraction, or radioactive decay but that didn’t matter. The results weren’t necessarily expected — even though you saw the Bragg equation in class, seeing the rings actually show up and change when you changed the accelerating potential was far more satisfying than confirming the value of g.

Another example — a colleague of mine was describing a lab one of his kids had recently. It was a black box, with some items inside that were taken from a list of possibilities, and the students had various investigative tools at their disposal (perhaps a scale and a magnet, among other things) and they had to determine what was in their box. Nobody knew the answer ahead of time, and the students had to go through and explain their reasoning for why a test confirmed or excluded a potential item on the list as being in the box. I wish I had had labs like that in school.