The latest statistics from AIP is out on the graduate physics degree in the US. This latest data covers up to 2008.
The overall number of PhD being granted has increased since the local minimum in 2004, with the largest movement made by a significant increase in US citizens being awarded PhDs.
Several interesting facts can be seen from the data:
1. It still takes slightly more than 6 years to get a PhD (starting from a B.Sc degree) in the US (Fig. 6). The average is 6.2 years.
2. The largest number of PhDs still comes out of Condensed Matter Physics (Fig. 7). In fact, if one were to count "Material Science" and "Surface Physics" as part of this subject, the number is even larger.
3. The one that caught my eye was Table 6. It asks for a response to the question "“If You Had To Do It Over Again, Would You Still Get a PhD in Physics?” Most US citizens overwhelmingly said yes, either to doing it at the same institution, or at a different institution. However, non-US citizens seem to have a consistently larger percentage of either not wanting to do it all over again at the same institution, not doing a PhD in physics, or even not getting a PhD at all! It would be fascinating to dig a little bit deeper here to see what's going on.
These statistics are always intriguing to read. To me, it is still the most definitive survey of students and professionals in this field in the US.