New Mexico’s three-tier teacher licensure system, which has cost the state more than $330 million, has little connection to whether teachers boost their students’ test scores, according to a new Legislative Finance Committee report.
The report backed the idea of basing teacher pay, in part, on student test scores, but urged caution because different calculation methods can yield very different results for the same teacher.
Sen. John Arthur Smith, D-Deming, who chairs the LFC, said Thursday he was “frustrated” by the findings on the three-tier system.
“We’ve rewarded, in many cases, mediocrity, and really haven’t accomplished what we set out to do,” Smith said.
Another key finding from the report is that New Mexico’s education funding formula doesn’t reward student achievement.
The formula gives more money to districts where teachers have advanced degrees and long careers, but it does not reward student achievement gains, or even reflect the three-tier pay scale.
The three-tier system pays tier-one teachers a minimum annual salary of $30,000, increasing to $40,000 for tier-two teachers and $50,000 for tier-three teachers. Teachers advance through the tiers primarily through experience, advanced degrees and a portfolio of lessons and student work samples they submit.
Three-tier has been partially funded through a 2003 constitutional amendment, which increased the amount of education money drawn from the state permanent fund. That funding has now begun to gradually decrease, through a sunset provision in the amendment.
Teacher evaluation in New Mexico is already in flux. After several failed attempts to revise the system through law, state education chief Hanna Skandera established a rule that requires school districts to evaluate teachers on a mix of test score growth, classroom observations and other measures of student learning. That system is being piloted in 68 schools statewide and will be adopted by all schools next school year.
However, the system is not currently linked to teacher pay. State law still requires teachers be paid based on the three-tier system.
Rep. Luciano “Lucky” Varela, D-Santa Fe, lamented what he called low starting salaries for teachers, but said student learning should drive their pay.
“This is inadequate in terms of what we’re paying our teachers,” he said. “But it’s got to be tied to performance.”
Researchers found students taught by top-tier teachers had slightly better test score growth than students with lower-tier teachers. But the gains were not significant, and in many cases the difference disappeared after researchers controlled for poverty. Top-tier teachers are more likely to teach more affluent students, who tend to have higher test scores.
LFC researchers also examined “value-added” measures of teacher quality. The idea of value-added is to control for variables that affect student achievement and isolate the amount of “value” added by the teacher. Researchers looked at two methods for the LFC report. The first is relatively simple and includes only how much a student’s test scores improved over time.
This method is thought to control for challenges like special needs and poverty, at least in part, because a student’s past performance on tests reflects their challenges. For example, a student who needed special education services one year is likely to need them during subsequent years, so every teacher who works with that student will face similar challenges. Some teachers, however, will help the student grow more than others.
The second method specifically controls for factors like poverty, special education needs, ethnicity and whether students are learning English.
The LFC researchers found these two methods had dramatically different results. One teacher, called “Mr. Wilson” in the report, teaches a gifted class in a large, urban school district.
Under the system that included only the student’s prior test score performance over two years, “Mr. Wilson” was rated “highly effective” in math and “meets expectations” in reading. However, when the second model was used, “Mr. Wilson” was rated “needs improvement” in both areas. The differences were even more stark when only one year of data was used.
The report concludes that the state should use three years of data whenever possible, and should use a combination of both value-added methods.
It also cautions that only about one-fifth of teachers in New Mexico teach subjects tested on the standardized test, leaving the majority of teachers ineligible for value-added calculations.
Skandera, who spoke at Thursday’s LFC meeting, said she agrees with the overall findings of the report, and believes it supports her efforts to overhaul teacher evaluations and to use student growth data in assigning school grades.
“I think the take-away that we’ve said all along, and was confirmed today, is value-added models aren’t a panacea,” Skandera said. “They’re not perfect, but boy, they are a step in the right direction, if done right and well.”
Skandera said Thursday no decision has been made on which value-added method will be used in next year’s evaluations. The pilot currently being done in schools is focused mainly on the classroom observations portion of the rule.
The simple method was used to calculate the “value-added” portion of A-F school grades, which have been criticized as volatile and hard to understand. With some exceptions, the A-F system on the whole assigned higher grades to more affluent schools.
— This article appeared on page A1 of the Albuquerque Journal