It’s job-hunting season in academia, which also means it’s talking-about-the-job-market season. After writing the previous post, I noticed a post on the same topic by Steve Hsu, who was interviewed for a Chronicle of Higher Education article (temporary free link, look quickly!) about the lousy job market in science. Steve has most of the relevant bits in his post, including this graph about the job situation in physics:
Of course, being a contrary sort, I’d like to quibble a bit with the graph and its presentation in the article.
The article makes a point of noting that:
In physics nearly 70 percent of newly minted Ph.D.’s go into temporary postdoctoral positions, whereas only 43 percent did so in 2000.
Well, yes, it looks really bad if you choose to compare the situation now with the second-best year of the last two decades. I went on the market in the fall of 2000, and I can confirm that it was a great time to be looking for a job– every college and university in America had made a bundle in the stock market during the dotcom boom, and they were all hiring new faculty. But you’ve got to go back to 1985 before you find a similar fraction of new Ph.D.’s taking potentially permanent positions as during the 1996-2001 boom. Looking over the full thirty-year span of the graph, the current situation is not all that far out of line. Yes, it’s at the high end of the historical range, but this isn’t some IPCC hockey-stick kind of graph, where we’re orders of magntiude above the previous average.
The other thing that bugs me about this graph is that the dramatic swoops and dives create something of a false impression. Looking at this, you would think that the number of permanent jobs doubled between 1993 and 1995, and then was cut in half between 2000 and 2004.
What you’re looking at isn’t the number of jobs in a given category, though, it’s the percentage of people who got Ph.D.’s in a given year who took jobs in a given category. That means it’s a ratio of two numbers, and neither of those numbers is terribly large. And, in fact, if you look at the number of degrees awarded over time, you’ll notice an interesting correlation: that high point in permanent jobs from 1996-2000 corresponds to the recent minimum in the number of Ph.D.’s produced, while the historical low point in 1993 lines up with the all-time high in the number of Ph.D.’s produced.
In other words, I think that you could reproduce most of the variation in the graph from the Chronicle article by keeping the total number of permanent positions constant, and letting the number of Ph.D.’s vary. What the graph shows isn’t so much a change in the supply of jobs, as it is a change in the demand for those jobs.
Which is not to say that the job market doesn’t suck. But I don’t think that graph is an accurate representation of the actual suckitude of the job market in physics.