SAN FRANCISCO — Which of the following is the best way to solve our nation’s educational crisis?

A) Reduce expectations.

B) Disregard the worst-off students.

C) Make sure that every child learns.

Obviously, the correct answer is C. And making sure that every child gets a high-quality education is exactly the logic behind President Bush’s education initiative, the No Child Left Behind Act.

By the 2013-2014 school year, under NCLB, schools across America are required to bring every student up to proficiency on state reading and math tests. Unfortunately, in their quest to meet that goal, school districts are showing a tendency to bypass the students who need the most help.

Consider Billy, who lives in an affluent suburb and attends a highly regarded public elementary school. He’s an average-performing pupil whose reading and math test results are just below the state’s definition of proficiency in those subjects.

Now consider Maria, who lives in a low-income neighborhood and attends a problem-plagued public elementary school. She performs very poorly on state tests and is nowhere near proficiency level.

Which student is likely to get more attention?

The truth is that many state accountability systems would focus on Billy. The reason for this counterintuitive choice is that state bureaucracies are trying to meet NCLB goals by imposing simplistic annual growth targets that create perverse incentives.

California’s plan, for instance, calls for 13.6 percent of students to be proficient in English in the 2003-2004 school year, and increases the required percentage by just over 10 percent every year for the next decade.

But this system motivates states to focus on Billy first.

That’s because improving Billy’s performance a little will push him over the proficiency bar and have an immediate impact on the state meeting its annual proficiency goals.

Because the system focuses on groups or percentages of students hitting the proficiency bar rather than individual students, the natural tendency is first to target the easiest students — those performing just below proficiency.

It’s a classic case of the low-hanging-fruit syndrome. Only later will the state get to Maria, who by then will no doubt have fallen even farther behind.

If you think this sounds crazy, it could get even worse.

Chances are that as the federal deadline approaches, the next perverse incentive will be to dumb down the definition of proficiency so that Maria can simply be reclassified as “proficient,” even if she shows little improvement.

Fortunately, there’s an alternative. If states stop focusing on group performance and percentages, we can get out of these statistical games and make sure that every child really learns.

A new measurement model, detailed in a study by the Pacific Research Institute, calculates a rate of expected academic change — or REACH — by using an individual student’s test scores to come up with an annual improvement target for that student.

In other words, given a student’s current place on the ability scale, the REACH model tells teachers, principals, parents and officials how much Billy and Maria need to improve each year in order for them both to reach proficiency by the time they leave school.

Other measurement models report student gains in terms of simple point increases or through comparisons with local, state and national averages. The REACH scores, in contrast, focus not on comparison to average scores, but on each student’s progress toward subject-matter proficiency, which is the goal of NCLB.

Using the REACH model, teachers can identify students who need remedial help and require greater attention. Also, by tracking each student’s performance, the REACH model can be used to evaluate whether a particular educational program has helped or hurt.

And it can be used to find out which teachers are doing the best job of moving students forward, inform the public about best instructional practices, and individualize teacher professional development and training to address teacher weaknesses.

Colorado already is moving toward a REACH-based testing system, and legislators in other states have begun to take notice too.

With the NCLB clock ticking, let’s make sure that we really meet its goals.