Last year, The Atlantic's Rachel Brown offered an idea that is as relevant as ever today: get rid of teacher licenses.

When it comes to hiring teachers, much of what we thought we knew turns out to be wrong. Performance on teacher-certification tests is a poor indicator of success in the classroom, and research shows that teachers who come from traditional schools of education aren't necessarily more effective than those from alternative certification programs, such as Teach for America. It's time we scrapped our outdated and inefficient system for recruiting and training teachers, and allow anyone with a college degree and a background check to teach.

As Malcolm Gladwell recently elucidated in The New Yorker, we are terrible at predicting who will turn out to be a great teacher and who will be lousy. The traditional logic has held that people who pass a credentials test, earn a license, and have a degree from a school of education will make the best teachers, and consequently, we should set the bar high for entering the profession. But that isn't the case. Much of what teachers do is unrelated to the kinds of skills that are measured on tests like the Praxis, the standard teacher-licensing exam. A talent for, say, facilitating student interaction, or classroom management, can't always be taught, and may be exhibited only in practice.

A master's degree -- long considered the mark of a qualified educator, and used in many school districts as a criterion for increased pay -- has actually been shown to have little or no effect on student learning. What's more, master's degrees are expensive, both for the teachers and for the districts that pay their raises, without corresponding gains for students. One recent study even concluded that teachers who earn a master's degree after they've started teaching are less effective at their job than teachers who don't have a master's at all. As Education Sector's Kevin Carey puts it, it appears that "master's degrees either reward teachers who were already worse, or they make them worse."

So what do we do? Gladwell suggests an apprenticeship program that trains teachers and evaluates them while they're teaching, not before they begin. This isn't dissimilar to what many alternative certification programs do with their recruits, focusing on character and past accomplishments during selection, and then putting the novices through a grueling summer training program, with instruction and feedback that continues through the school year. Last year, an Urban Institute study found that high-school students taught by Teach for America educators in North Carolina actually performed significantly better on end-of-course assessments than those taught by educators with three or more years of experience. Yet even Teach for America participants are required to attend graduate courses at night and eventually do earn their certifications. By that time -- years into their teaching careers -- isn't a license something of a moot point?

Lowering the entry barriers to the teaching profession would certainly make waves, and could result in controversial shifts in teacher pay and tenure, union agreements, and existing schools of education. However, it's widely acknowledged that public schools are largely under-performing, and with more than a third of America's 2.3 million teachers approaching retirement, it's time to start actively getting bright new teachers into our classrooms. Eliminating the bureaucratic and largely irrelevant teacher-licensing process will make it easier to recruit teachers who majored in or worked in sparsely filled subject areas like math and science and prepare them for teaching, without requiring investment in an expensive master's-degree program or certification course before they can work. We have little to lose and much to gain by broadening the teacher applicant pool. The result will be more talent for the classroom, and more benefits for America's students.

Objections? E-mail conor.friedersdorf@gmail.com and I'll publish the best.