New teachers are far less likely to leave the profession than previously thought, according to federal data released Thursday.

Ten percent of teachers who began their careers in 2007-2008 left teaching after their first year, according to the National Center for Education Statistics. But attrition then leveled off, and five years into their careers, 83 percent were still teaching.

That figure — indicating that just 17 percent of new teachers left their jobs in the first five years — stands in stark contrast to the attrition statistic that has been repeated (and lamented) for years: That between 40 percent and 50 percent of teachers leave the profession within their first five years.

The higher estimate, which has become a fixture in education debates, comes from the work of Richard Ingersoll, a professor at the University of Pennsylvania and a leading scholar on the nation’s teacher workforce.

But Ingersoll’s famous estimate was just that — an estimate. A “crude approximation,” he said in an interview Wednesday, made necessary by the fact that no one had tracked a cohort of new teachers over time to see how long they stayed in the classroom.

But now federal officials have done just that. They followed a representative sample of teachers who began their careers in the 2007-2008 school year in order to find out what happened to them. And the new findings present such a different picture that they have the potential to change the national conversation about new-teacher attrition, a problem that cascades across issues ranging from student achievement to school district budgets.

“I don’t think it’s that the earlier numbers that people like myself calculated were wrong,” Ingersoll said, explaining that his estimate eventually was buttressed by additional data that put new-teacher attrition in the same ballpark. “The hope is that there’s been an improvement and that teacher attrition has gone down.”

Ingersoll pointed out that comparing the new study to his estimate has an apples-to-oranges problem. His estimate included both public and private school teachers, for example, while the new data includes only public school teachers. His estimate looked at attrition that occurs after the fifth year of teaching, while the new data looks at attrition after the fourth year of teaching.

And he said that how you calculate attrition depends on how you define it. He believes 17 percent might be too low. As a member of the advisory panel for the database that gave rise to the new study, he has been doing his own calculations with the raw numbers, and he believes that a more accurate figure — which would include teachers who left the classroom and then returned within the first five years of their career — could be higher than 20 percent.

But even with those caveats, he said, the new figures are clearly lower than those that have been bandied about in education debates for years. And it’s too soon to know why, Ingersoll said, though there is likely to be no shortage of explanations.

Proponents of the Obama administration’s education policies might say that conditions for teachers have improved, Ingersoll said, while others might argue that his earlier estimates exaggerated the problem. And the recession also might have played a role: The new study began tracking teachers just as the economy went south, perhaps giving more teachers reason to stay put in their jobs.

“You can get different spins on this report,” Ingersoll said. “I certainly would hope that the reason the rates were lower is because so many of these reforms have hit pay dirt and we’re improving things. But the truth is, we do not know that.”

Among other key findings from the federal data released Thursday:

  • New teachers who are assigned mentors are more likely to continue teaching than those who are not assigned mentors. In 2008-2009, 92 percent of those who had first-year mentors were still teaching, compared to 84 percent of those without mentors. By 2011-12, 86 percent of those who had first-year mentors were teaching, compared to 71 percent who did not have mentors.
  • Teachers with higher starting salaries — above $40,000 — were more likely to continue teaching than those with lower salaries.
  • The proportion of teachers who leave the classroom involuntarily — for either budgetary or performance reasons — is not insignificant. Of the teachers who left after their first year, for example, 27 percent left involuntarily.
  • Older teachers who began their careers after age 30 were more likely to leave the profession within five years than younger teachers, and men were more likely to leave than women.
  • Teachers who entered the profession via an alternative certification program (such as Teach for America) were more likely to leave the profession than those who went through traditional training programs. In 2011-2012, for example, about 21 percent of teachers with alternative certification were not teaching anymore, compared with 16 percent of teachers with traditional training.
  • Teachers who spend their first year in higher-poverty schools (where more than 50 percent of students qualify for free or reduced-price lunch) are slightly more likely to leave the profession than those who spend their first year in lower-poverty schools. But the data do not say how many of those teachers who began in high-poverty schools then transferred into more affluent schools within their first five years. Such transfers have contributed to particularly high turnover in many of the nation’s neediest schools.

The new findings offer some information about teacher mobility — i.e., how often teachers who stay in the profession move among schools. But the study does not say how many first-year teachers are still teaching in the same building five years later. Isaiah O’Rear, who has managed the study for the National Center for Education Statistics, said a future report will take a closer look at how new teachers move from school to school.