When I meet other scientist types, this can be one of the most interesting questions to throw out there.
We can use mine as an example. I did my grad studies in Microbiology and Immunology, but basically I was doing biochemistry type work (cancer research with a lot of molecular stuff). It took me just over five years to finish this sucker which is pretty typical in North America. Of course, when I take a critical look at my thesis and calculate: "What if this thesis literally shows all of my work, because everything I did, worked? What if I had magic fingers throughout my research and never had a failed experiment!?"
Using this rubric, I calculate that my Ph.D. in biochemistry/molecular biology type work could've taken about, DUM-DUM-DUM...
Note that this figure also includes the 3 months needed to write the damn thesis itself! This means that technically my thesis is reflective of only 3 months of successful experiments: or as I like to think of it -- four and a half years of failed experiments!
This, of course, is how the scientific method works. It's slow, very slow. Most of the time, your marvelous ideas and experiments don't work out, and even if they do, there's almost always someone smarter than you telling you what's wrong with it. Even though it may not seem like it (especially when you read the headlines in newspapers and watch the stories on televisions), the reality is that scientists go about their discoveries in very very gradual increments.
Those doing the science know this already, but sometimes, I get the impression that this reality is lost to other folks. I'm assuming others in research have experienced this, but a classic example, are the many bright high school students coming to a lab, outlining a brilliant science fair project, and then telling you that they only have 6 weeks to do it, and oh-by-the-way "I can only come in after school on mondays, wednesdays, and thursdays!" Of course, this is when you tell them, that the experiments they're suggesting might actually take the better part of a decade, and probably a few dollars here and there.
Anyway, I'm also curious to see what other numbers people will come up with and especially in other fields, science or otherwise. So ask yourself: How long would your Ph.D. have taken if everything worked?