It's no secret that our public schools are having problems attracting good teachers. The mass migration of women to more lucrative occupations is among the more obvious factors. Making matters worse, up to 60 percent of new teachers leave after only a few years. Unless these trends are reversed, teaching is in danger of becoming a refuge of leftovers.

Performance-based, or merit, pay has been advocated as one way to help turn these trends around. The idea is simply to link teachers' pay to performance, creating incentives to promote better teaching. Although theoretically sound, merit pay has problems in practice. And it is only one important aspect of a needed solution that involves more than just pay.

Merit pay is hardly a revolutionary concept. It works effectively in many professions and has been applied, on a limited basis, in public schools. It allows an employee and supervisor to define appropriate measures of performance by which an employee will be judged. Later, the results of the evaluation determine rewards.

Most compensation systems base teachers' pay on seniority, credentials and school-related activities. In Arlington County, about half the teachers (who haven't "topped out" on the scale) receive pay raises of up to 5 percent each year automatically through longevity step increases, independent of performance. A merit system would give increases only to those whose performances justified raises.

Determining valid measures is difficult. It is primarily because of perceived concerns in this area that teachers have strenuously resisted merit pay. This opposition is another obstacle --perhaps the most practical difficulty. Any system based on negotiation and mutual consent obviously is doomed if one of the principal parties does not support the process. But teachers' resistance is based largely on misunderstanding. Measures of performance are negotiated, not mandated. Defining them is difficult but not impossible. If teachers, as professionals, would join in improving their profession and its compensation system, negotiation and careful planning could work.

As a profession, teaching has relatively low entry requirements and is not highly selective. There is little differentiation among teaching roles and little opportunity for promotion. Evaluations, lacking any up-or- out standards, tend to be meaningless. They seem to count only when they are negative and tend to reveal only the most egregious incompetence. And these evaluations are not linked to rewards or training. Teachers rated outstanding receive almost no extra recognition, while those needing improvement seldom receive the right prescriptions. As a result, the profession lacks incentives to attract, motivate and retain the best teachers, improve performances of the less capable and root out the incompetent and ineffective.

A solution should include: 1) a definite career ladder, with new positions at the entry and senior levels (the "master teacher" concept, which establishes endowed chairs in certain fields or schools, is a promising variation of this approach); 2) a new salary structure, providing higher top salaries, better earning potential and adjustments for changes in skill demands; 3) a performance evaluation system, with appropriate negotiated measures at each step, based strictly on classroom effectiveness and contributions to school improvement; 4) comprehensive teacher training, tied to performance standards, with specific skills and training needs specified for each step; 5) long-term planning, involving teachers, parents, administrators and other local groups.

This is a subject that Arlington County should consider seriously. The debate should not be mired in limited warfare over merit pay, but should be expanded to include a farsighted analysis of all related administrative issues.