A new Texas study punctures the commonly held notion that high levels of teacher turnover in poor, urban schools result from an exodus of the profession’s “best and brightest.”
The study, scheduled to be posted online this week by the Cambridge, Mass.-based National Bureau of Economic Research, a nonprofit research organization, draws on data on thousands of teachers and 4th through 8th grade students in an unidentified big-city Texas district that researchers call “Lone Star.”
See the accompanying item,
Rather than measure teachers’ quality by whether they had passed certification exams or had earned advanced degrees, the researchers looked at the test-score gains students made from year to year on state mathematics tests to determine which teachers were effective.
For the most part, they found, the teachers who left inner-city schools between the 1989-90 school year and the 2001-02 school year were no better at raising their students’ scores than those who stayed behind. In some cases, the analysis showed, the departing teachers may have even been worse.
The problem for urban schools, though, is that the resulting vacancies tended to be filled by brand-new teachers—a group the study shows to be less effective in producing student learning gains than many of the teachers who left. As a result, the researchers said, disadvantaged inner-city schools are still left with a disproportionate share of lower-quality teachers, even though most are novices who might one day turn out to be good at their jobs.
“This reinforces the idea that we ought to pay a lot more attention to retention issues and other decisions made after the point of hiring,” said Eric A. Hanushek, the lead author of the paper and a senior fellow at the Hoover Institution, a think tank based on Stanford University’s campus. “We haven’t pushed very hard on trying to find a way to keep teachers we know are good and helping poorer teachers find something else to do.”
Dueling Findings
Susanna Loeb, a Stanford researcher who has conducted similar studies using New York state data, said Mr. Hanushek’s study may be an “important first step” in understanding how teacher-mobility patterns contribute to student achievement in urban schools.
The report, “A Market for Teacher Quality,” is scheduled to be available from the National Bureau of Economic Research.
Her own research, however, suggests a pattern somewhat different from what Mr. Hanushek found. It suggests that the teachers who leave city schools for higher-achieving suburban schools tend to be more, not less, qualified than those who stay behind.
The difference is that Ms. Loeb and her colleagues measure teacher quality by looking at teachers’ general-knowledge scores from certification exams, whether they have a master’s or bachelor’s degree, and other background characteristics.
On the other hand, in his study, Mr. Hanushek said, “it turns out not many of those things are systematically related to what happens in the classroom.”
He found, for instance, that while new hires at higher-achieving schools and schools with larger minority enrollments tended to be teachers with master’s degrees, those teachers, in their previous, inner-city school assignments, had not been more effective than the colleagues they left behind.
Overall, Mr. Hanushek said, the departing teachers deemed to do a worse job than their colleagues tended to fall into two categories—those who moved to another school in the district and those who left the Texas public school system altogether.
But he noted an important finding: The teachers’ poorest classroom performance tended to come in the final year before they made their move.
“They either had a bad experience or, once they decided to leave, they didn’t work as hard,” he said.
The Value of Experience
As with similar “value added” studies, the Texas study also found that good teachers matter. Spending a year in a classroom with an experienced teacher who ranks at the 85th percentile in terms of effectiveness can translate to an average 9-percentile-point learning gain for students, according to the study.
On the other hand, having a brand-new teacher can negatively affect a student’s test scores. For instance, even the experienced teacher ranking at the 85th percentile would have produced only half as much average learning gain for students—around 5 percentile points—in the first year on the job.
Among teachers with four or fewer years on the job, Mr. Hanushek found, fourth-year teachers tended to be the most effective. Yet the statistics also show that many teachers leave the district before reaching their fourth year.
The analysis also indicates that students tended to learn more, as measured by their test scores, during years when they had teachers from the same racial backgrounds as themselves.
In addition, the report echoed his own previous findings that the draw for departing teachers did not appear to be the promise of making more money.
Teachers who switched districts boosted their salaries the following year by an average of $2,087, compared with the average $2,137 salary increase received by teachers who remained in the same schools. (“Study: Teachers Seek Better Working Conditions,” Jan. 9, 2002.)
