In short, the survey found that performing research increased undergraduates' interest in science and technology fields (so-annoyingly-called "STEM" disciplines, for Science Technology Engineering Mathematics). Such undergraduates were also more likely to go on to advanced degrees, although here the causality isn't necessarily clear. The survey did find that students with higher grades tended to be more likely to get involved in research; this raises at least the possibility that "getting involved in research" and "going on to an advanced degree" are affected by a common cause, and that the former doesn't necessarily increase the probability of the latter.
Of great importance was the fact that undergraduate research seemed to improve the confidence and future success of underrepresented minorities and women. I'm not sure I can tell you what is particularly "white male patriarchy" about classroom performance, but if this is a way to help people realize their true abilities in science regardless of their ethnic background, then it could be an important component in the continuing problem of minorities and women in science. (Indeed, the title of the Science article is "The Pipeline: Benefits of Undergraduate Research Experiences.")
I had very good undergraduate research experiences myself. I was very lucky to spend all four summers after each of my four college years working in nuclear physics at the 88" cyclotron at Lawrence Berkeley Lab. I think that the people there were happy with me— and that's why I was hired back on subsequent summers. But I cannot deny that at least that first year, it was family connections that gave me the opportunity to get in the door. I also did astronomical research my senior year at Harvey Mudd, working with Shane Burns on measuring the distance to galaxies using the Surface Brightness Fluctuations method. Nothing publishable came out of that, and the result of the research program was our own understanding of how to use the method— but the experience itself was invaluable to me.
I'm going to stick to Physics and Astronomy, since that's my own field, even though the Science article is more general. If you think about it, there are (at least) two things that go by the name "Physics." The first is a body of knowledge. Call it the current state of our understanding of Physics. The second is a field of study, a potential career. The way college education is set up, we do a far better job teaching the former than the latter. Almost all of our classes, even the lab classes, teach what is known and understood in Physics, what we have learned. There is, traditionally, very little or no education in how one goes about being a physicist.
Undergraduate research is the best place to learn something about the latter. It may be that you love the physics, but don't love doing the physics. Wouldn't it be nice to figure that out from your undergraduate education? Or, vice versa, it may be that you struggle with some of the concepts of physics, but you have a true talent in contributing to the body of knowledge of physics. Wouldn't it be nice to figure that out before you write off the very notion of going on to graduate school?
I would even like to see undergraduates in Physics who don't go in in Physics to get involved in research. Again, lab classes are often useful or even essential, but are very different things. It would help if people who are "out there" as senators and lawyers and teachers and so forth had some idea what it was like to be an actual practicing physicist.
I've advised a number of undergraduate research projects myself, and have had four students in the last four years graduate with honors theses; all four are now in graduate school somewhere. I'm also on a committee at Vanderbilt that gives out a prize for undergraduate research each year. Let me tell you, reading some of those applications is intimidating. Last year's winner was the first author of a Nature paper (the cover story, even), and was a paper I was already familiar with as a result of it having been cited by a number of the telescope time requests I'd read the previous week. For some of these applications, I'm wondering: should we give these students a prize, or should we give them a Master's degree?
It is true that undergraduate research is like graduate research in the first few years: often, perhaps most of the time, it takes more time for the professor than the professor gets back. Certainly a lot of what the students are doing we could in less time than it takes for us to teach them how to do it. I've heard this from other sources as well, so I know I'm not just blowing smoke here. However, I view it as an investment. Some of the students, after a while, do get to the point where they are making real contributions. But, also, I view it as a key part of my mission as a physics professor to getting undergraduates, as well as graduate students, involved in research. As I said, it should just be a standard part of an undergraduate physics education.