U.S. Department of Education

A recent federal study found that a much smaller percentage of beginning teachers leave the field in their first five years on the job than the widely quoted figure of 50 percent. It’s 17 percent, according to the new research.

The U.S. Department of Education’s National Center for Educational Statistics made the new finding in a study released in April. Not extensively reported, the study conflicts with the widely held perception that new teachers experience a high turnover rate. This may be true in some districts, in some regions, but it wasn’t a nationwide trend in the five years studied, 2007-08 through 2011-12.

The longitudinal study, “Public School Teacher Attrition and Mobility in the First Five Years,” found that 10 percent of new teachers in 2007-08 didn’t return the following year, increasing cumulatively to 12 percent in year three, 15 percent in year four and 17 percent in the fifth year. The totals include teachers who were let go and subsequently didn’t find a job teaching in another district.

The most exhaustive study of teacher attrition to date, the study followed 1,900 teachers, with follow-up paper questionnaires and contacts by phone.

The frequently cited statistic that “half of new teachers leave after five years” stems from a 2003 study, also using federal data, by Richard Ingersoll, a sociology professor at the University of Pennsylvania Graduate School of Education, who concluded that between 40 and 50 percent of teachers didn’t return for a sixth year of teaching – one year longer than the new study. His finding was based on yearly approximations. Unlike the current study, the federal report that Ingersoll used didn’t track what happened to individual teachers after the first year. His data included private school teachers and excluded the 3 percent of teachers who left (perhaps on maternity leave) and returned to teach within the five-year period

“Two important findings support what NEA has advocated for a long time. That high-quality mentors and competitive salaries make a difference in keeping teachers,” said Segun Eubanks, director for Teacher Quality at the National Education Association.

Another reason for the differences in findings could be the time periods in which the studies were done. Ingersoll used data on first-year teachers collected in four years between 1988 and 2000. The recent study followed one group of teachers from 2007-08 to 2011-12, during an economic recession that could have affected job mobility, said Isaiah O’Rear, the project officer from the National Center for Education Statistics. The study didn’t delve into that issue, he said. The center is not planning another in-depth study at this point, he said.

The new study didn’t break down the findings by state and region, but it did analyze teachers’ earnings, ages, education and school locations and cited a number of findings:

  • Money – 97 percent of teachers who earned more than $40,000 their first year returned the next year, compared with 87 percent who earned less than $40,000. By the fifth year, 89 percent of those earning $40,000 or more were still on the job, compared with 80 percent earning less than $40,000.
  • Guidance – 92 percent of teachers assigned a mentor their first year returned the next year, and 86 percent were on the job by the fifth year. Only 84 percent of teachers without mentors returned in the second year, declining to 71 percent in the fifth year.
  • Degrees – There was no statistically important difference in attrition between teachers who began teaching holding bachelor’s degrees and those with master’s degrees.
  • Mobility – By the second year, 16 percent of teachers had moved to another school or district. One-fifth of the “movers” moved involuntarily or didn’t have their contracts renewed.

Reacting to the study, Segun Eubanks, director for Teacher Quality at the National Education Association, said on the NEA website, “Two important findings support what NEA has advocated for a long time. That high-quality mentors and competitive salaries make a difference in keeping teachers.”

Eubanks also said, “Not surprisingly, the study found that teachers who spend their first year in higher-poverty schools are more likely to leave the profession than those who spend their first year in lower-poverty schools.” But O’Rear said the difference – an attrition rate 3 percentage points higher in year five – was not statistically significant.

The study, however, didn’t note which teachers moved to lower-poverty schools after the first year. O’Rear said that a follow-up report, which will delve into more details, will be out later this summer.


Comments

Leave a Comment

Your email address will not be published. Required fields are marked *


Comments Policy

The goal of the comments section on EdSource is to facilitate thoughtful conversation about content published on our website. Click here for EdSource's Comments Policy.

Expand Comments
Collapse Comments
  1. Lucas F 1 week ago1 week ago

    I think I would lean with the department of education numbers on this subject and not with a biased report that is trying to mask an epidemic rather find the root cause and a solution. It's very dishonest to say that the notion that half of teachers quit within their first five years is based on research from 2003 when in fact there have been many studies since that show numbers in between 40% and … Read More

    I think I would lean with the department of education numbers on this subject and not with a biased report that is trying to mask an epidemic rather find the root cause and a solution. It’s very dishonest to say that the notion that half of teachers quit within their first five years is based on research from 2003 when in fact there have been many studies since that show numbers in between 40% and 50%, including surveys by departments of education.

  2. Dr. Charles Bickenheuser 1 month ago1 month ago

    If you use the percentage of teachers who leave each year for the first five years, then the cumulative rate (para. 3) is 55.87 percent remain for years six. In an average expression, you could then say that about 44 percent of teachers leave before year six.

  3. No 9 months ago9 months ago

    The study was started in 2008 when a recession hit. Teachers were not leaving because they were afraid that they would not get another job. For example, in NYC, there was a hiring freeze for two years! Wondering what was the motive for the person who worked the numbers to disprove that teachers quit at a higher rate. They do. I have seen it in NYC.

  4. Karl Meer 9 months ago9 months ago

    I think schools pay low salaries in the first 15 years knowing a large percentage of teachers will not make it to the top salary because of the pay, having a family, burn out, and such. It would be better to increase the first 15 years and decrease the top years some, so you have money to start a family.

  5. Miseducation Nation (@MisEduNation) 1 year ago1 year ago

    I don’t believe the study. Actually, I don’t believe anything posted on this site.

  6. Don 1 year ago1 year ago

    We are told over and over of the difficulty in retaining teachers at "hard-to-staff" schools and we know that low-performing schools have much higher layoffs. This 3% figure seems questionable in that light. The study (or the reporting) doesn't address how many teachers leave voluntarily or are laid off. Every time we lay off new teachers and give them their walking papers after a long, hard and expensive journey through college, we condemning … Read More

    We are told over and over of the difficulty in retaining teachers at “hard-to-staff” schools and we know that low-performing schools have much higher layoffs. This 3% figure seems questionable in that light. The study (or the reporting) doesn’t address how many teachers leave voluntarily or are laid off. Every time we lay off new teachers and give them their walking papers after a long, hard and expensive journey through college, we condemning students to a diminishing pool of talent and put another nail in the coffin of public education.

    Replies

    • Concerned Parent Reporter 1 year ago1 year ago

      I concur with your important well thought out and balanced views Mr. Don.

      I say the studies need to,desperately find out and analyze if interns are the ones that leave.

      age discrimination, and the need to,pay a bit more for non intern teachers have driven the quality of,teachers down.

      cheap,intern teachers view teaching as a way to jump to another job.

      The CDE needs to not allow interns to teach.

      • Paul 1 year ago1 year ago

        Contrary to your assumption, many Intern Credential holders are second-career teachers or teachers with prior public school substitute or private school experience, who desperately want to be in the classroom. Highly-motivated, they meet exactly the same requirements as traditionally-prepared teachers, except that they teach for 1 to 2 years WHILE completing the certification coursework. (Traditionally-prepared teachers perform a variable but inevitably lower number of hours of student teaching while taking classes.) Regulations mandate that interns … Read More

        Contrary to your assumption, many Intern Credential holders are second-career teachers or teachers with prior public school substitute or private school experience, who desperately want to be in the classroom. Highly-motivated, they meet exactly the same requirements as traditionally-prepared teachers, except that they teach for 1 to 2 years WHILE completing the certification coursework. (Traditionally-prepared teachers perform a variable but inevitably lower number of hours of student teaching while taking classes.) Regulations mandate that interns receive support from both a school district teacher and a university-appointed supervisor.

        Whether certified traditionally or through an internship program, all new California teachers complete the same final assessment (TPA) and receive the same preliminary credential. At that stage, would you rather have a teacher with 1 to years’ experience running his own classroom, or a few hundred hours helping someone else run a classroom?

        The Intern Credentisl has existed since 1967.

        This information, easy to find on the Web site of the California Commission on Teacher Credentialing, is often ignored, misunderstood, or not reported. Most school principals do not know it (there are 2,000 to 3,000 interns in a given year in California, out of 300,000 teachers, so most principals have never worked with one). Parents certainly wouldn’t know the information.

        Interns do not seem to meet the sampling criteria for this particular research study, and even if they did, no conclusions could be drawn for such a small subgroup.

        • FloydThursby1941 1 year ago1 year ago

          I see the union people attack these professionals, but it seems more knee jerk to me than scientific. For instance, I'm 100% sure Caroline would oppose these people, and Gary, before it comes up. The thing is, maybe they're right, maybe they're wrong, why don't we just compare test score results of their students, students of teachers in each group. Then we will see factually whether or not they are right. … Read More

          I see the union people attack these professionals, but it seems more knee jerk to me than scientific. For instance, I’m 100% sure Caroline would oppose these people, and Gary, before it comes up. The thing is, maybe they’re right, maybe they’re wrong, why don’t we just compare test score results of their students, students of teachers in each group. Then we will see factually whether or not they are right. Let’s not be automatic. Let’s be scientific. Let’s study this strictly by the numbers.

        • Parent Concern 1 year ago1 year ago

          No, say 95% , say 3,000 to say 8,000 , are people with no previous credential.

  7. Bill Younglove 1 year ago1 year ago

    And how, I wonder, does this teacher-leaving-rate compare/contrast with other professionals who have also spent years and years and a small fortune to become those who serve the public?

  8. navigio 1 year ago1 year ago

    If one were to choose a time period to measure, this would probably be one of the worst possible ones. 07-08 was peak funding, and subsequent years saw large decreases in funding along with associated drops in staffing levels and apparent increases in emergency credential use. Our district pink slipped almost 15% of our teachers in one of those years. And if the enrollment in teacher prep programs is accurate, instead of 50% of teachers … Read More

    If one were to choose a time period to measure, this would probably be one of the worst possible ones. 07-08 was peak funding, and subsequent years saw large decreases in funding along with associated drops in staffing levels and apparent increases in emergency credential use. Our district pink slipped almost 15% of our teachers in one of those years. And if the enrollment in teacher prep programs is accurate, instead of 50% of teachers leaving the profession after becoming credentialed, we now have 50% of the expected applicants ‘leaving the profession’ by not bothering to even start it.

    Replies

    • Paul Muench 1 year ago1 year ago

      Interesting points. Did the study only measure voluntary departures? Maybe the poor economy also influenced more teachers to stay on due to the lack of other jobs.

  9. Paul Muench 1 year ago1 year ago

    Keeping teachers in the classroom is a low bar. Paying teachers significantly more to attract a much different group of people into teaching is a better approach. I’m willing to gamble that fixing teacher pay would mobilize public support to work through the remaining issues such as pensions etc..

    Replies

    • Gary Ravani 1 year ago1 year ago

      Pauls:

      So you want more avaricious people in the classroom?

      • Paul Muench 1 year ago1 year ago

        Do you agree that lack of money can make people feel taken advantage of? That’s what I’m looking to prevent.

        • FloydThursby1941 1 year ago1 year ago

          The pay should be higher at the beginning but not go as high at the end. It should be based on contribution, not loyalty to the union. People don't see that paying so much after 20 years reduces how much you can pay from 1-10 years. I'm not against paying more if there is more effectiveness but from my observations and including days missed from work and replaced with a sub, I … Read More

          The pay should be higher at the beginning but not go as high at the end. It should be based on contribution, not loyalty to the union. People don’t see that paying so much after 20 years reduces how much you can pay from 1-10 years. I’m not against paying more if there is more effectiveness but from my observations and including days missed from work and replaced with a sub, I don’t believe there is as much of a difference in contribution as their is in pay scale, according to the union salary scales. I believe part of the reason it is as it is now is not based on what’s ideal for children but based on rewarding those in the union with the most power who have been there the longest and ensuring loyalty. I believe flattening the curve somewhat would increase retention of young teachers?

          I’d like to hear from teachers and union members and experts in this area. Without considering politics, in your opinion, A. is the pay schedule fair based on contribution to children and B. Would flattening it bring more young people into the profession? Or what about merit pay so young teachers, if they stay every day and really push and tutor their students after school and do better than their peers, however they do it, after accounting for demographics and previous year’s test scores, they could earn as much as a lower performing older teacher, or more. I do believe there is room to attract more teachers and drive teachers to put in that extra level of focus which can help raise test scores across the board.

  10. Gary Ravani 1 year ago1 year ago

    From the above story: "Another reason for the differences in findings could be the time periods in which the studies were done. Ingersoll used data on first-year teachers collected in four years between 1988 and 2000. The recent study followed one group of teachers from 2007-08 to 2011-12, during an economic recession that could have affected job mobility, said Isaiah O’Rear, the project officer from the National Center for Education Statistics." And: "Reacting to the study, Segun Eubanks, … Read More

    From the above story:

    “Another reason for the differences in findings could be the time periods in which the studies were done. Ingersoll used data on first-year teachers collected in four years between 1988 and 2000. The recent study followed one group of teachers from 2007-08 to 2011-12, during an economic recession that could have affected job mobility, said Isaiah O’Rear, the project officer from the National Center for Education Statistics.”

    And:

    “Reacting to the study, Segun Eubanks, director for Teacher Quality at the National Education Association, said on the NEA website, ‘Two important findings support what NEA has advocated for a long time. That high-quality mentors and competitive salaries make a difference in keeping teachers.'”

    Well, yes.

    Replies

    • Paul 1 year ago1 year ago

      It’s so funny how the NEA grasps at straws. A box checked on a questionnaire does not mean that the respondent had a mentor, much less an involved and effective one. The mentor/no mentor distinction is one of flimsiest elements of this research project.

      • Gary Ravani 1 year ago1 year ago

        Maybe for this project. But there is a lot of support for the idea that keeping new teachers in the classroom improves when they have a well trained mentor in their early years.

  11. Zeev Wurman 1 year ago1 year ago

    Very interesting. Thanks, John!

    I’d like to point out that even the discredited “half in 5 years” number was not as frightening as people made it to be.

    For example, “Only about 38% of college graduates whose highest degree is in an S&E field work in S&E occupations” http://www.nsf.gov/statistics/seind12/c3/fig03-07.gif

  12. Melissa V Rentchler, MLISc, M.Ed., CA State credentials Teacher and Teacher Librarian 1 year ago1 year ago

    I just examined the “study”. It is not very “longitudinal”, but rather less than a decade of data examined at best.

    Replies

    • John Fensterwald 1 year ago1 year ago

      Melissa: You obviously have found the data, starting on page 12. The study is what is says: a look at the first five years of employment, starting in 2007-08. My assumption is these types of studies are expensive, and this is the most thorough look at nationwide teacher attrition to date. It would be great if a new study, starting in 2015-16, under different economic conditions, could be done. Congress, at least the majority, … Read More

      Melissa: You obviously have found the data, starting on page 12. The study is what is says: a look at the first five years of employment, starting in 2007-08. My assumption is these types of studies are expensive, and this is the most thorough look at nationwide teacher attrition to date. It would be great if a new study, starting in 2015-16, under different economic conditions, could be done. Congress, at least the majority, would like to cut the Department of Education budget, which would affect research.

  13. Melissa V Rentchler, MLISc, M.Ed., CA State credentials Teacher and Teacher Librarian 1 year ago1 year ago

    Where is the data for the year/s this article is representing? I hypothesize that there might have been/is more perseverence and staying power during tight employment times ie., the recession era 2008+ and the slow “recovery”.

  14. Paul 1 year ago1 year ago

    I am a big believer in statistics, and knowledgeable about math. Even I have a hard time accepting statistical inferences made from a tiny sample. If all 2,000 teachers in the sample were from California, the sample would comprise 0.67% or 67 out of 10,000 in our credential pool of 300,000 people. California has roughly 12% of the nation's population, and if 12% of the teachers in the sample were from California, the sample would comprise 0.08% … Read More

    I am a big believer in statistics, and knowledgeable about math. Even I have a hard time accepting statistical inferences made from a tiny sample.

    If all 2,000 teachers in the sample were from California, the sample would comprise 0.67% or 67 out of 10,000 in our credential pool of 300,000 people.

    California has roughly 12% of the nation’s population, and if 12% of the teachers in the sample were from California, the sample would comprise 0.08% or 8 out of 10,000 people in our credential pool.

    Can the habits of 240 teachers really predict the habits of 300,000?

    Some will argue that the denominator should be new credentialholders, not all credentialholders, but that would still yield a sampling rate of 1.1% or 111 out of 10,000 people (versus the 0.08% or 8 out of 10,000 mentioned above).

    To the researchers’ credit, they avoid many of the frauds that the CTC commits in compiling its retention report. The CTC relies on a sample of convenience of BTSA completers, ignoring teachers who exit before finishing BTSA. Since BTSA is a two-year program, no one who stops teaching before the end of Year 2 can be counted in CTC attrition figures. Also uncounted are teachers ineligible for BTSA, due to part-time, part-year or substitute employment. Last but not least, teachers who leave after starting BTSA late, interrupting it (e.g., due to a layoff), or never starting, don’t count as having left.

    Both these researchers and the CTC ignore people who earn teaching credentials but never find a teaching job.

    These researchers make much of whether a new teacher had a mentor, but the notion is ill-defined. Anyone might have checked the box.

    Unless and until the CalTIDES tracking system is funded, we have no data to support optimistic teacher retention claims in California.

    Replies

    • John Fensterwald 1 year ago1 year ago

      Paul, as you know, this was a national study, not a California study, and I am confident that the NCES researchers had a valid and reliable sample of the 152,000 new teachers nationwide in 2007-2008. As my article noted, The new study didn’t break down the findings by state and region, … And it would be incorrect to infer California trends from the national study.

      • Manuel 1 year ago1 year ago

        John, that's a nice way of making sure that what you reported is factually correct without giving an impression that it is more than what it is. Not saying that you did anything wrong, mind you, as you are just being the messenger while navigating between Scylla and Charybdis. (Why is it getting attention just now if it was released in April?) But if what Paul states in his comment is also factually correct (no, I … Read More

        John, that’s a nice way of making sure that what you reported is factually correct without giving an impression that it is more than what it is. Not saying that you did anything wrong, mind you, as you are just being the messenger while navigating between Scylla and Charybdis. (Why is it getting attention just now if it was released in April?)

        But if what Paul states in his comment is also factually correct (no, I don’t want to do a deep dive into the study, at least not yet), then it is fair to ask how significant this study is to get an idea of what really goes on in the totality of the teacher pipeline.

        Regardless of how one feels about this issue, the undeniable fact is that if teachers are not retained in sufficient numbers there will never be a sufficiently large pool of highly qualified teachers who are willing to put up with student scores in standardized tests as true measures of their effectiveness. True, many argue over what exactly is a highly qualified teacher, but if there are not enough teachers with experience, who is going to staff our schools? After all, the districts that you and Ms. Ellison mentioned in your other article have felt the decrease in supply.

        In regards to a better study, my guess is that we get what we pay for: a very limited study done during a period of very unfavorable economic circumstances. In my opinion, it is not in the interest of many of the Usual Suspects to get an accurate and current look at this problem.

    • Wendy 1 year ago1 year ago

      I wonder how many new teachers stay after the first five years because their loan debt is so high that they don’t feel comfortable moving to another career? I find the data extremely limited as well. Too bad.

      • John Fensterwald 1 year ago1 year ago

        Wendy: The project manager said that a follow-up report on the study will be coming out later this summer. I don’t know what it will explore, but I will link to it, if not report on it.

      • Paul 1 year ago1 year ago

        Interesting point, Wendy. Given low starting salaries and precarious employment, teaching is a particularly bad way to pay off student loans. State assumption of up to $12,000 of a teacher's loans (A.P.L.E.) runs its course after four years, and a $5,000 federal assumption payment arrives after Year 5. Apparently, new A.P.L.E. slots were de-funded during the recession, as part of the Cal Grant cutback. (Can someone confirm?) The state and federal loan assumption programs require consecutive years of … Read More

        Interesting point, Wendy. Given low starting salaries and precarious employment, teaching is a particularly bad way to pay off student loans.

        State assumption of up to $12,000 of a teacher’s loans (A.P.L.E.) runs its course after four years, and a $5,000 federal assumption payment arrives after Year 5.

        Apparently, new A.P.L.E. slots were de-funded during the recession, as part of the Cal Grant cutback. (Can someone confirm?)

        The state and federal loan assumption programs require consecutive years of full-time teaching in the same economically-disadvantaged or low-performing school, a bar which is nearly impossible for a new teacher to reach today, what with temporary status, “no cause” non-reelection, seniority-based layoff, or seniority-based bumping/involuntary transfer.

        What’s more, the $17,000 total did not rise with the drastic CSU and UC tuition hikes of the last decade. Today, the whole sum covers less than one year’s tuition and living expenses, against the five years of university study required of teachers.

Template last modified: