So If we were to summarize some of the reasons for the score discrepancy between foreign and domestic students, I think this is a good list:
- Only the top foreign students have any chance of admission into US programs and thus only they will take the GRE. Or perhaps instead, only the top scorers will apply.
- There are a lot more smart foreign students than US students anyway because there are more foreign students period.
- Taking the GRE in a foreign country is often an expensive hassle that sometimes requires a whole day of traveling back and forth, and so the test takers are going to take it very seriously and be very prepared.
-Domestic students realize they can get in without a superhigh GRE.
On a side note:
Everyone keeps saying the GRE correlates poorly with grad school success and research. I don't think anyone should expect a big correlation with research. But what nobody addresses is the GRE is a good measure of one's knowledge and skill in undergraduate physics material (except for the 0177 test, that test sucked. This partly reflects the fact that the test is written by 5 profs who change from year to year). It's not supposed to measure our ability for research. However, there are still a few questions that require some type of data interpretation or familiarity with an instrument.
Most of the questions are not based on memorized facts and equations. Seriously, how many questions on the test required use of an unfamiliar equation? I'd say nearly all equations needed were ones we already knew, with a few exceptions (such as the velocity addition formula) which the ETS practice book told us we needed to know anyway. The problems require skill and creativity with the basics. Any responsible student should know he needs to memorize the double slit diffraction formula before taking the test, just like most exams we take in our advanced courses require memorization of some formulas.
I realize grad schools would rather have a student who will be useful in their labs than a student who has mastered basic problems but can't use an oscilloscope... But again, I think past research experience is not a fair metric to gauge candidates by... I mean seriously, seeing so many students walk away with sole authored or first authored publications, I don't think everyone has a chance for that...
If you're an experimentalist, you don't own a lab or any part of it. You don't control the grad students or anyone else affiliated with the lab. How are you going to sneak off into the corner, devise and run your own experiments, interpret the results with your undergrad background, and submit a paper without anyone else making you put his name on it?
Clearly it happens, but to marginalize an applicant because he hasn't accomplished these things is not fair. This isn't a reference to particular people here, as I know plenty of you know way more advanced physics than I do, but I see the craziest things happening all the time. Today a soph biochem major at my school who has barely passed her major classes and is taking calc 1 for her 3rd time announced on her facebook page that she's now working in a lab and doing cancer research. Maybe I'm bitter since she was one of the many disappointments in my life, but it remains true that everyone's doing research, everyone, and it's mostly random who gets the gold stars (pubs, awards) and who doesn't. Doing "research" is no longer a way to stand out but a requirement. I think the best way to stand out nowadays would be not to do research (hehe).
Think about how you got affiliated with your first lab... did you ask one of your instructors about his research? Did you get a grad student to hook you up with someone? How can you just walk in as a freshman stranger to a new lab and tell the professor, "Please put me on a project I can do myself."? If the prof doesn't laugh at you, the grad students are going to think you're trying to take over the lab.
If you're in theory, and somehow have the background to discover new things, you'll need to learn everything your prof has written over the past few years, and then go invent new things. Won't the prof want to be on the paper too since you used his work? And if you invented new things without consulting any prior work... let's just say you should not be in an undergraduate program.