It may also be possible to get an undergrad degree through some combination of on-line classes, proctored tests, and evaluation of knowledge and skills gained through nonacademic experience. Thomas Edison College is an example of a college that awards such degrees, as discussed in the NY Times. Some grad schools might have specific requirements for GPA etc. that could create problems for students who are primarily self-taught, though. And you still need to convince a prospective major professor that your unconventional education has prepared you well for research in her lab.
Interesting discussion this week in NY Times, Slashdot, and UnCollege.org about alternatives to a spending an expensive 4 years in college, now that free courses and lots of other educational content are so widely available on the web.
College grads tend to make more money, with some well-known individual exceptions, though the reasons aren't clear:
1) knowledge and skills learned in classes are a good match for high-skill jobs?
2) contacts met while getting drunk?
3) employers assume that people with degrees are smarter or work harder, even if most skills needed will actually be learned on the job?
Note that reasons #2 and #3 would make individuals with college degrees more competitive for the best jobs but wouldn't necessarily imply any benefit to society from greater investment in education, as I've discussed before. But that's not today's topic.
What if your goal isn't to get rich from writing an ap, but to do science? Depending on what kind of science you want to do, you might need access to an electron microscope, a gene sequencer, and dozens or hundreds of journals costing hundreds or thousands of dollars each per year. Having a wide range of experts to talk with would help, too. In other words, you need access to a major research university.
How can you get such access, without running up a lot of debt? Easy. Go to grad school. Grad students in science don't usually need to take on much additional debt, unless they have kids or an expensive life-style. 20 hours a week as a teaching assistant and your tuition is covered, plus (barely) enough to live on. Or, if you're lucky, you get a fellowship or a research assistantship that pays you to work on your thesis research.
But can you go to grad school without getting an undergrad degree first? It's not easy, but is it impossible? Suppose you spent 2-3 years reading lots of scientific papers, working part-time in a lab, and asking interesting questions in department seminars. Then you ace the GRE and get a great letter from your boss. Is that enough to get into a good grad school?
Absolutely, if you did the above while also getting an undergrad degree from the least-expensive accredited school you can find. (Later in life, nobody will care that your undergrad degree is from Southern Nowhere State, if your PhD is from Big-Name University.) Without any undergrad degree, alternative credentials like great GRE scores and a published paper would convince many professors to take you as a grad student. (That's how I got into Cornell, on probation. I did have a degree from Evergreen, but they weren't accredited yet and didn't give grades.) But would the professor's grad school agree? Most would not, at least today. But could that change, if there were significant numbers of such nontraditional students applying?
I think it would be very risky to assume that you can skip college and still go to grad school, even if you can manage to learn more science "on your own" than you would have in college. But on-line learning is starting to shake things up. Who knows what might be possible in the future?