It's a matter of degree. A _typical_ undergrad thesis is more spoon-fed from the perspective of "here's the idea, go do it". A typical masters project is still

It's a matter of degree. A _typical_ undergrad thesis is more spoon-fed from the perspective of "here's the idea, go do it". A typical masters project is still more advisor-generated than student-generated. A typical Ph.D. project is more student-generated than advisor-generated in many cases. Please add copious "not all"s to this paragraph as the variance is extremely high and everyone's case is unique. And the rigor expected increases substantially. Again, generally. The major thing of a Ph.D. isn't really how much it moves the state of the art forward, it's how much it moves you forward. You will almost certainly approach research and your area fundamentally differently after. That's true for me as well as a professor with respect to research: I'm pretty convinced that my largest contribution is the Ph.D. students I've mentored more than the chunks of research I've done. I'm proud of the research. But I'm more proud of the students. They've gone on to do wonderful things. The thing it hangs on is "justifies" - financially, a Ph.D. in CS is hard to "justify." Intellectually, it's easy. Depends a lot on your individual utility function! I could have dropped out of my Ph.D. program to go be a pre-IPO employee at Google or Akamai and, ah, let's say that my financial picture would have a few extra zeros on it compared to where I am now. :-) But I'm really glad I didn't. It's been a very fun journey and I've never felt like I'm just doing the routine. That's worth a lot for me.