I was very poor at applying my stats knowledge at the time, a sample size of one is hard to draw conclusions from. One thing that is very true is there is not a lot of control over research. As a Phd you are bound to your advisor (which you can choose) and as a non-tenured faculty you are bounded by what can get funded. As an undergrad I thought, "Being a faculty member is great, you get to create whatever you want in Computer Science." Once I realized that it was indeed a real job, my enthusiasm wrenched very (too?) hard in the other direction.