(As first appeared in Education Week, December 8, 2010. Reprinted with permission from the authors.)

There is a business training film in which a man in a gorilla suit walks slowly into a room, waves, and slowly exits. But most people don't see the gorilla—they are watching what they have been told to watch, people tossing a basketball. In education, we likewise have a "gorilla" staring us in the face—how we fail to prepare teachers to succeed—and keep missing the obvious.

For an enterprise that spends almost all of its money on teachers, public education expends far too little effort to get the human-capital chain right. In the recently released "Teacher Quality Blueprint," the University of Pennsylvania Center for High Impact Philanthropy argues that "the issue is poor management of human capital in education—from recruitment through to retention. ... settling for any teaching candidate who shows interest, training them poorly, deploying them unevenly, failing to support them adequately once they arrive at a school, providing a work environment that is not conducive to teacher or student learning, and evaluating and compensating them in a way that fails to recognize weak or strong performers."

Public opinion (and moviemakers) can shift the conversation to bemoaning union due-process protections, or we can restate the problem so that we also look for solutions from the entities that can make a difference: the 50 states. States need to take responsibility by not sending school districts so many underprepared teachers who too easily attain certification and tenure. Unfortunately, when it comes to teachers, we set the bar too low, and we train too little. We must improve state policies that influence how we recruit, educate/train, certify, hire, develop, and retain teachers before we work with the unions on how to usher out the poorest performers.

There are too many teacher-training slots. U.S. Secretary of Education Arne Duncan put it plainly last year when he said that "many, if not most, of the nation's 1,450 schools, colleges, and departments of education are doing a mediocre job of preparing teachers for the realities of the 21st-century classroom." Though training teachers may be a cash cow for some schools of education, states must lead the way by reducing the number of available slots, allowing those schools to make education a more highly valued major. Making entry into teacher-training programs more competitive will help attract better candidates, as it has done in Japan and Finland—and in Teach For America.

Upgrade clinical training. We underinvest in training teachers, as is most evident in a critical semester of academic training: student-teaching. We place student-teachers with no metrics to decide which classrooms they should enter. How is it decided that a particular teacher should be the powerfully influential "cooperating teacher"? Few programs review that teacher's instructional skills or student outcomes, or train the cooperating teacher to be an effective coach and mentor. We need state-certified cooperating teachers who earn extra pay for this certification. They should meet the highest levels of student-outcome measures and provide evidence of excellent instruction, such as through Robert C. Pianta's Classroom Assessment Scoring System.

Any reform of student-teaching must deal with its role as a "cash cow" for schools of education. Students pay full tuition for their student-teaching semester, despite its low cost to the school of education. Field supervisors are paid very little; likewise, cooperating teachers receive limited financial rewards or training. Clinical training could be the best place in the training pipeline to develop crucial teaching skills and to weed out would-be teachers who can't succeed in the classroom. Yet teacher-trainees too often receive a mediocre student-teaching experience and an almost automatic passing grade. The obvious contrast is the medical profession, where clinical training is about the very best practices possible.

Teacher certification should provide a meaningful bar to entry before ineffective teachers become permanently entrenched in their jobs. Not everybody with an education degree should be certified; look at the pass rates for bar exams, medical boards, and recognition in other professions. Certification should be based on rigorous subject tests and field evidence of effective instruction. Effective certification practices should also include on-the-job evaluations of instruction for two or three years before certification.

Connecticut is one of the states leading the way: Only 70 percent of prospective elementary and preschool teachers passed its subject test on teaching reading. State Commissioner of Education Mark K. McQuillan defended the test, saying: "You can't really effectively teach what you don't know."

Train principals in how to hire high-quality teachers. How is it that we do not adequately train principals for arguably their most important job: deciding whom to hire? Principals need expert training and coaching in how to recruit. They need time to devote to recruiting and training, using lesson-observation protocols during the hiring process. Districts and unions need to leave hiring decisions to principals and local site committees so that large urban districts have the flexibility to hire early and for specific grades, subjects, and schools. We believe that hiring practices are a fertile area for experimentation and research—for example, testing the use of short probationary periods to assess their impact on teacher quality.

Tenure shouldn't be automatic. Tenure is too often the "default" decision, and few school districts have any metrics for evaluating tenure or candidates, or spend much time and energy on the process. A recent Los Angeles Times series ("Failure Gets a Pass") noted that 98 percent of teachers in the Los Angeles Unified School District receive tenure after two years—based on a half-hour review. This is all too typical of lax state regulation where it counts.

Dramatically improve professional development: focusing on teaching practice, not the individual teacher's personality strengths and weaknesses. Research tells us that in-class, one-on-one coaching, and the Japanese lesson-study model are the most effective practices, giving teachers ownership over their own development.

The Children's Literacy Initiative's federally supported Investing in Innovation, or i3, project develops grade-level "model teacher" leadership and expertise, using in-class, content-focused coaching by its own staff. Teachers need to learn how to plan lessons, not read them from a script. They need expert help in learning how to deliver lessons while closely observing student learning, not just in assigning pages from a textbook.

We have to get over the simplistic notion that exemplary teachers are just identified and then given bonuses. They must be developed and given leadership opportunities.

Good teaching is built practice by practice, not simply discovered in the raw. In the "model classroom" framework of the Children's Literacy Initiative, exemplary teaching is measured, replicated, and recognized with opportunities for "model teachers" to become grade-level leaders. The best teachers can play a vital coaching role inside their own schools by fostering continuous learning among teachers.

There are 3.6 million K-12 public school teachers in the United States, and within this enormous workforce lies much of the challenge and promise of public education. Of course bad teachers should go, but as Secretary Duncan has said, "we can't fire our way out of this problem."

We cannot expect to replace millions of teachers with those somehow blessed with innate charisma or a superhuman work ethic. Mary Poppins is not waiting in the wings to save us. Only state policies can prevent so many inadequately screened and trained teachers from receiving tenure and provide a sufficient pool of quality teaching candidates to fix this human-capital train wreck.

(Linda Katz is the executive director of the Children's Literacy Initiative, in Philadelphia. Andrew Belton is a partner in the New York City office of the Bridgespan Group.)