Test-based evaluation has demoralized teachers because they know it is unfair to judge them by student scores. Many believe it has contributed to a growing national teacher shortage and declining enrollments in teacher education programs.

A major problem with test-based evaluation is that students are not randomly assigned. Teachers in affluent suburbs may get higher scores year after year, while teachers in urban districts enrolling many high-need students will not see big test score gains. Teachers of English-language learners, teachers of students with cognitive disabilities, and teachers of children who live in poverty are unlikely to see big test score gains, even though they are as good or even better than their peers in the suburbs. Even teachers of the gifted are unlikely to see big test score gains, because their students already have such high scores. Test scores are a measure of class composition, not teacher quality.

Seventy percent of teachers do not teach subjects that have annual tests. Schools could develop standardized tests for every subject, including the arts and physical education. But most have chosen to rate these teachers by the scores of students they don’t know and subjects they never taught.

Scholarly groups like the American Educational Research Association and the American Statistical Association have warned against using test scores to rate individual teachers. There are too many uncontrolled variables, as well as individual differences among students to make these ratings valid. The biggest source of variation in test scores is not the teacher, but students’ family income and home environment.

The American Statistical Association said that teachers affect 1 percent to 14 percent of test score variation. The ASA is an impeccable nonpartisan, authoritative source, not influenced by the teachers’ unions.

The Gates Foundation gave a grant of $100 million to the schools of Hillsborough County, Florida (Tampa), to evaluate their teachers by gains and losses in student test scores. It was an abject failure. The district drained its reserve funds, spending nearly $200 million to implement the foundation’s ideas. Gates refused to pay the last $20 million on its $100 million pledge. The superintendent who led the effort was fired and replaced by one who promised a different direction.

Should Massachusetts cling to a costly, failed, and demoralizing way to evaluate teachers? Should it ignore evidence and experience?

Common sense and logic say no.

Should teachers be judged “subjectively”? Of course. That is called human judgment. Is it perfect? No. Can it be corrected? Yes. Most professionals are judged subjectively by their supervisors and bosses. Standardized tests are flawed instruments. They are normed on a bell curve, guaranteeing winners and losers. They often contain errors — statistical errors, human errors, random errors, scoring errors, poorly worded questions, two right answers, no right answers. No one’s professional career should hinge on the answers to standardized test questions.

Massachusetts is widely considered the best state school system in the nation. The hunt for bad teachers who were somehow undetected by their supervisors is fruitless. The Legislature is right to return the decision about which teachers are effective and which are not to the professionals who see their work every day.