Overall the number of incoming graduate students in physics and astronomy programs has risen by an average of 1.5% per year, which is only slightly faster than the overall population growth rate in the US of 1.2%. The growth rates for PhD's awarded in physics and astronomy were 3.9% and 3.2% respectively, while the growth rates for terminal master's degrees was 2.7% in physics and a whooping 4.8% in astronomy (although that corresponds to an increase of 1.2 terminal master's degrees in astronomy per year). In graphical form, here are the trends:
As you probably noticed, the most interesting feature in both plots isn't the trend, rather it's something consistent from year to year. For both physics and astronomy programs, the total number of degrees awarded is only about three-quarters of the number of incoming students. More precisely, if we assume that the grad student population is divergence-free and account for the overall trend in enrollment, the outcomes for physics and astronomy grad students are shown below.
For both types of programs, a little less than 30% of incoming graduate students will not earn a graduate degree. Since the AIP only tracks American programs, there may be people who transfer to a foreign program, but for the most part there is a clear and sizable attrition in our graduate programs. Interestingly while distribution between PhD's and Master's degrees varies considerably between physics and astronomy programs, the number leaving without a degree is nearly constant.
So what happens to the bright-eyed 1st year students? At least for some of them nothing in the way of academic degrees.
Update: The AIP data gives a break-down by school of the incoming grad students and graduate degrees awarded, but only for 2009. Here is a list of the program that had at least twice as many plus 1 (to account for noise) incoming students at degrees awarded in 2009. Of course with only 1 year of data there is a lot of noise here, but it might give an idea of the types of programs that contribute to the nearly 30% attrition rate. Program with more than 15 incoming grad students are in bold.
In order by state (alphabetically) and then alphabetical order, the physics programs are: U. of Alabama-Birmingham, UC-Davis, San Jose State, Yale, U. of Central Florida, Florida A&M, Idaho State, IUPUI, U. of Louisville, Louisiana State-Baton Rouge, U. of New Orleans, U. of Maine, Clark, UMass-Amherst, UMass-Dartmouth, Michigan Tech, Mississippi State, U. of Nebraska-Lincoln, UNLV, Syracuse, North Carolina Central, Wake Forest, Ohio U., Portland State, Carnegie Mellon, U. of Puerto Rico-Mayaguez, U. of Puerto Rico-Rio Piedras, South Dakota School of Mines, U. of North Texas, Southern Methodist, Texas Tech, U. of Texas-Arlington, U. of Texas-San Antonio, Brigham Young, George Mason, Hampton, Old Dominion, U. of Virginia, U. of Wisconsin-Madison, U, of Wyoming
In the same order but with bold for program admitting more than 7, the astronomy programs are: Yale, Florida Institute of Technology, U. of Virginia, U. of Washington
No immediate trends pop out of that list at me. Anybody else see anything?