Teacher Retention, Teacher Experience, and Student Achievement

Urban teacher turnover is high, yet the first four years of teacher experience seem critical for student learning. Can more pay or training stop the exodus?

A review of several teacher salary schedules suggests the master's degree is perhaps as strongly correlated with teacher salary as classroom experience. This extra pay for the master's has been criticized by U.S. Education Secretary Arne Duncan as well as by Microsoft founder Bill Gates.

The first four years of teacher experience in fact may be more important for student achievement and success on standardized tests, particularly in mathematics, than the master's degree, according to a study of student achievement in Texas by Rivkin et. al (2000). A master's degree may however be coordinated with slightly less teacher transfer out of region(at least for New York teachers), according to other researchers (Hamilton Lankford, Susanna Loeb, and James Wyckoff, 2002), who studied urban and suburban teachers in New York State in the 90s.

Student Achievement and Teacher Experience

Rivkin et. al (2000) measured student progress among cohorts in schools with one and with more-than-one teacher per school per grade. Comparing student progress among cohorts who had all studied under the same teacher and cohorts who had been taught by different teachers, the researchers surmised, would enable them to sort out the amount of variation in student achievement within a school that could not be attributed to variation in teacher quality from the amount which could be attributed to variation in teacher quality. The researchers focused on comparing student achievement within a school rather than across schools, arguing that the role that teacher quality played in variation in student achievement in different schools could not be so easily measured, that it was likely to be confounded with the role played by other between-school factors, including demographic differences.

Rivkin et. al were able to observe considerable variation in what they believed to be student achievement within particular schools that might be linked to variation in teacher quality. They reported a link between this variation in teacher quality and the number of years of experience teachers had. Rivken et. al however reported little variation in student achievement and teacher quality when comparing teachers who had master's degrees with those who did not have master's degrees.

Teacher Experience More Critical for Mathematics?

Rivkin et. al looked at teacher quality for both reading and mathematics. They reported that the most pronounced improvement resulting from teacher experience was in mathematics achievement, a subject where many teachers are new or teaching "out of field." Rivkin et. al's data "supported the notion that beginning teachers and to a lesser extent second and third year teachers in mathematics perform[ed] significantly worse than more experienced teachers."

First Few Years of Experience Key?

The first few years of teacher experience seemed to be the most critical, according to Rivkin et. al. For teachers with four or five years experience in reading and mathematics, gains in teacher quality for each additional year of experience "were small," and "not statistically significant." And Rivken et. al did not report additional improvements with teacher quality for teachers with more than five years of experience.

Student Achievement and Teacher Certification

While teacher certification status like teacher experience, according to Lankford et. al, has a high correlation with student test scores (the "correlation coefficient" is 0.6, equaling perhaps about 36% cross-influence), higher than many other teacher quality measures (including teacher passing certification tests first time around, teacher degree from a competitive college., and teacher master's degree), no causal relationship has been shown. It may be simply that certified teachers who have passed their teaching exams, together with teachers who hold degrees from competitive colleges or master's degrees, are more likely to secure posts where student scores are higher.

Race and Measures of "Teacher Quality"

Researchers Lankford, Loeb, and Wyckoff, who as noted above, looked at New York state teachers in the 90s (1993-1998), observe that New York City attracted "less qualified teachers" than its suburbs. They argue however that measures of teacher quality other than teacher experience, including having attended a competitive undergraduate institution, holding a master's degree, or receiving passing scores on certification examinations, may be heavily confounded with race. Minorities who lack the more "competitive" degrees may be more likely than others to choose urban teaching posts. This may explain some of the differences between urban and suburban teacher "quality." However, suburban teachers tend also to have more experience than urban teachers, in part because of high urban turnover.

Teacher Flight?

At the same time that New York's urban teachers who received "poor" ratings were being transferred to schools in low-income areas within their original teaching districts, suburban teachers were transferring out of their original districts, according to Lankford et. al. These suburban transfers apparently transferred voluntarily, transferred to schools with fewer non-White students. The big draw the reason for the transfers seems to have been largely salary differences as teachers were more likely to transfer to schools where the base salary was higher than that of the original institutions.

Besides fewer non-White students in the schools teachers voluntarily transferred to, there were fewer Limited English Proficient (LEP) students in these schools. However Lankford et. al stated that the differences in the number of LEP students was not statistically significant.

New York pays its urban teachers less than its suburban ones. The same is not true for all metropolitan areas: Chicago pays its urban teachers more, as does Fort Worth, although in Fort Worth suburban class size is smaller, but urban teacher turnover remains high regardless. (The cost of commuting in urban areas of course helps to diminish any extra pay.)

Teachers' Unions and Teacher Pay

Teachers' unions continue to back higher pay scales for the master's. Nevertheless, If the first four years of teaching experience are the most important, as Rivkin et. al's study of Texas achievement and teacher quality suggests, perhaps greater salary increments during a teacher's first four years of teaching might help to encourage new urban teachers to keep teaching, gain needed expertise, reducing urban teacher turnover.

Hiring Practices

Balter and Duncombe (2005) reported similar hiring practices in New York state across districts, whether high needs or not, except that high needs districts tended to offer more bonuses for non-teaching experience or for supporting extra-curricular activities. Knowledge of subject matter and interviews were important in hiring in all cases. Fifty percent of districts asked candidates to teach sample lessons. A small percentage of districts looked at teaching portfolios. Shortages in high-needs districts in any case appeared not to be the result of hiring practices but of problems with teacher retention.

Other Options: Training, Mentoring New Teachers

Police officers must undergo a rigorous twelve-to-fourteen-week training program before they actually engage in police work. In some states, they must undergo more training. In New Jersey, twenty-five weeks of training is required.

Not so for "emergency-credentialed" teachers. They may be encouraged to observe in classrooms when they are not teaching (for example during the summer months or, in areas with year-round schools, during the time period their track is off), may meet occasionally or frequently with a "master" or "mentor" teacher, and may be required to complete a certain number of college credit hours toward a teaching credential each year they teach, but no prior training may be required (except of course a college degree where the School Board requires that).

Programs like California's SITE (Summer Institute for Teacher Education, described at the California Teachers Assocation web site) and "Teach for America" may help to move new teachers more quickly through the "hurdles" of teaching, and thus improve retention. Teachers' unions such as the National Education Association (NEA) have criticized "Teach for America" when it brings in people to teach in areas without need but otherwise have been supportive.

In addition of course teacher inducation programs, with mentoring (under a "master teacher") may help to reduce turnover. According to Ingersoll and Kralik (2004) the effectiveness of teacher induction and mentoring programs vary and is related in part to whether or not the master teacher teaches in the same field, to the amount of time spent with the master teacher, and finally to whether or not the new teacher is satisfied with the teacher induction program. Ingersoll and Kralik did note that teachers who expressed satisfaction with their teacher induction programs may have been more likely to remain in teaching to begin with.

Although it's associated with teacher retention, mentoring, like a master's degree, is not particularly associated with higher student achievement. Improving student achievement however has not been the focus of mentoring.

More Spending Per Pupil or More Community Support?

Once spending is adjusted to accomodate pupil needs, according to the a 2002 report by the United States General Accounting Office (GAO), nearly all of today's urban schools are underfunded when compared with suburban schools, with less playground equipment, fewer books per student, and a less chance of having a computer laboratory, among other differences.

According to the GAO, some inner city schools however performed as well or better than suburban ones. One of the characteristics of the high performing inner city schools was more parental involvement. Carey and Roza's research (2008) however that the most important difference between high-performing schools (with low teacher attrition) and low-performing schools, when both are located in low income areas, may simply be per-pupil spending.