The certification pathway that New York City teachers took to their classrooms seemed to have little relationship to how effective they were in raising students’ scores, concludes a study that matched some 10,000 teachers with six years of test results.
The research also found that the variation in effectiveness among teachers with the same certification status—traditionally certified, alternatively certified, members of the Teach for American program, and uncertified—was much greater than the variation between groups.
The study is the second within three months to look for links between certification type and teacher effectiveness in the 1.1 million-student district, and to find that the connections are weak or nonexistent.
“What Does Certification Tell Us About Teacher Effectiveness: Evidence From New York City” is available from the Harvard Graduate School of Education.
“If you look at Teach for America and traditionally certified folks, they’d seem dramatically different on many observable characteristics,” said Thomas J. Kane, one of the authors of the report and a professor of education and economics at Harvard University’s graduate school of education. “What’s striking is that … the [effectiveness] differences are small, and the big story is that the distribution of differences looks so similar.”
The studies also reflect the heightened interest among researchers in mining year-to-year increases in individual students’ test scores for clues to what makes a good teacher. In that effort, the nation’s largest school district stands out as a laboratory because the number of alternatively certified teachers there has exploded over the past six years as the district worked to replace uncertified ones.
When to Be Selective
Of the 50,000 teachers hired in New York City from 1999 through the first half of 2005, the period of the study, 46 percent were traditionally certified, 34 percent were uncertified, and 20 percent were alternatively certified, a proportion that grew to 28 percent over the last three years of the period.
Traditionally certified teachers were prepared in university programs before being licensed to teach. Most of the city’s alternatively certified teachers had been named New York City Teaching Fellows and received the bulk of their training while on the job. That was also true of a smaller number of recruits to Teach for America, a privately organized national program that places recent graduates of selective colleges in hard-to-staff schools.
Mr. Kane said the study’s results suggest that initial certification status is not the best way of selecting teachers, though it has been a standard one. One instance of its important role was the order to New York City schools from Richard P. Mills, the New York state commissioner of education, that only certified teachers be hired, starting in September 2003.
“If you are going to try to get selective, get selective at the end of two years, when you’ve got a lot more information to be selective on, as opposed to before hiring,” Mr. Kane said.
Under most union contracts, teachers can be more easily fired in the first two or three years on the job—before the protections associated with tenure kick in.
The study does not endorse any particular way of weeding out teachers, but possibilities Mr. Kane mentioned include a combination of their students’ test-score growth, peer and principal evaluations, and parent ratings.
A paper released in December by the Teacher Pathways Project, a partnership between researchers at the State University of New York at Albany and Stanford University, also found the relationship between certification status and effectiveness to be small. But researcher Susanna Loeb of Stanford wrote in an e-mail that her team drew “somewhat different” conclusions from those of Mr. Kane and his co-authors, Jonah E. Rockoff of Columbia University’s business school and Douglas O. Staiger of Dartmouth College. The complications of the data could make a significant relationship hard to see, she asserted. (“Big Changes Found in Teachers’ Paths to N.Y.C. Schools,” Jan. 4, 2006.)
One problem, Ms. Loeb wrote, is that the different certification routes are not as distinctive as their names suggest. Teachers’ preparation before entering a classroom actually varies and overlaps more than than the broad categories of “traditionally certified” and “alternatively certified” suggest, she wrote.
Teach for America Weighed
Other observers rejected taking any approach that de-emphasizes standards for teacher education and certification, arguing that those tools offer the best guarantee that all students will get decent teachers
“Let’s distill what makes [teachers] effective, and build licensing and screening around those characteristics,” argued Thomas Blanford, the associate director of the teacher-quality department of the National Education Association, the 2.8-million-member teachers’ union.
Mr. Kane said he agrees that selection at the time of first hiring could be an important part of raising teacher quality, but he warned academic background—used extensively in choosing both New York City Teaching Fellows and Teach for America participants—also seems to be a poor predictor of effective teaching.
Mr. Kane’s paper used calculations of the returns teachers get on their effectiveness from their first two years of classroom experience to address a long-standing debate about the value of Teach for America.
Echoing other research, the study found that teachers, no matter their certification status, made significant effectiveness gains during their first two years in the classroom. But those gains are lost more frequently when positions are filled by Teach for America recruits, because their turnover rate at the end of four years is more than twice that of other rookie teachers’.
The researchers concluded, however, that the job the TFA teachers do in raising student test scores is just enough better than their peers’ to largely compensate for the loss of experience to the district in any given year.
