An interesting experiment (for a company that actually has a job and is not leading people on) would be to distinguish a job posting that says
"Hiring for this position will be based in large part on a work-sample test during a half day at our office"
versus
"Previous experience . . . preferred, but will train the right candidate."
We can tell that the United States economy is in recession (as is the case in many other countries with Hacker News participants) because we keep seeing new stories submitted every day or so about the hiring procedures of companies, with multiple comments. In a long FAQ post I've posted recently
http://news.ycombinator.com/item?id=4270768
here on Hacker News, based on many helpful comments from other participants, I summarize a LOT of research on company hiring procedures. If you want to hire someone good in the United States, make a work-sample test part of your hiring procedure. Work-sample tests are much better than biographical reviews of resumes for finding good workers. If you want to get a good job in a well managed company, develop the skills to get past a realistic work-sample test for the position you seek. Many more details appear in the FAQ,
http://news.ycombinator.com/item?id=4270768
which is quite long but well worth a read if you are looking for a job or if you are a business manager trying to hire someone who will do a good job.
P.S. I still hear of young people who are gaining full-time, full-benefits jobs in today's economy. In the usual case, they are getting those jobs by showing what they can actually do as part of the hiring process, degree or no degree.
After edit: I recently read the VERY interesting book The Folly of Fools: The Logic of Deceit and Self-Deception in Human Life
http://www.amazon.com/Folly-Fools-Logic-Deceit-Self-Deceptio...
(I wonder if this book has been discussed here on HN yet?) and that reminded me of one of the main reasons that hiring by screening resumes is demonstrably less effective than hiring by giving work sample tests: many people lie on their resumes. The author of the interesting submitted blog post was of course posting a fake job ad, and several comments here on HN point out that many job ads may just be fronts for recruiters rather than postings by actual employers. Participants here gave examples (NUMEROUS examples) of job ads getting responses that don't appear at all to fit the job, but an employer also has to worry about "false positives," applicants who look like they fit the job but who are inflating their educational credentials or multiplying their years of actual work experience. The only way to know what an applicant can do is to test. Perhaps announcing up front that the hiring process includes actual testing of job-related skills MAY screen out some of the poseurs from even sending in their fake resumes (although the many shotgun applicants who don't even read the job ad closely will still be sending resumes all over the place only to waste the time of anyone who receives the resumes).
One way the author of the blog post could have demonstrated statistical acumen is by labeling his data presentations "Self-Reported Experience" rather than "True Experience" and "Self-Reported Credentials" rather than "Education." He has NO idea what the actual educational credentials or work experience (or other aspects of biography) of any of his applicants really are. He would have been aware of this point if he had taken a good statistics course in college, but alas good statistics classes are very rare in the United States.
http://statland.org/MAAFIXED.PDF
http://escholarship.org/uc/item/6hb3k0nz