From TED TALKS:
'. . . How do you make a teacher great? It seems like the kind of question that people would spend a lot of time on, and we'd understand very well. And the answer is, really, that we don't. Let's start with why this is important. Well, all of us here, I'll bet, had some great teachers. We all had a wonderful education. That's part of the reason we're here today, part of the reason we're successful. I can say that, even though I'm a college drop-out. I had great teachers.
. . . When I first learned the statistics I was pretty stunned at how bad things are. Over 30 percent of kids never finish high school. And that had been covered up for a long time because they always took the dropout rate as the number who started in senior year and compared it to the number who finished senior year. Because they weren't tracking where the kids were before that. But most of the dropouts had taken place before that. They had to raise the stated dropout rate as soon as that tracking was done to over 30 percent. For minority kids, it's over 50 percent. And even if you graduate from high school, if you're low-income, you have less than a 25 percent chance of ever completing a college degree. If you're low-income in the United States, you have a higher chance of going to jail than you do of getting a four-year degree. And that doesn't seem entirely fair.
So, how do you make education better?
. . . the more we looked at it, the more we realized that having great teachers was the very key thing. And we hooked up with some people studying (***study excerpts and link below) how much variation is there between teachers, between, say, the top quartile -- the very best -- and the bottom quartile. How much variation is there within a school or between schools? And the answer is that these variations are absolutely unbelievable. A top quartile teacher will increase the performance of their class -- based on test scores -- by over 10 percent in a single year. What does that mean? That means that if the entire U.S., for two years, had top quartile teachers, the entire difference between us and Asia would go away. Within four years we would be blowing everyone in the world away.
So, it's simple. All you need are those top quartile teachers. And so you'd say, "Wow, we should reward those people. We should retain those people. We should find out what they're doing and transfer that skill to other people." But I can tell you that absolutely is not happening today.
What are the characteristics of this top quartile? What do they look like? You might think these must be very senior teachers. And the answer is no. Once somebody has taught for three years their teaching quality does not change thereafter. The variation is very, very small. You might think these are people with master's degrees. They've gone back and they've gotten their Master's of Education. This chart takes four different factors and says how much do they explain teaching quality. That bottom thing, which says there's no effect at all, is a master's degree.
Now, the way the pay system works is there's two things that are rewarded. One is seniority. Because your pay goes up and you vest into your pension. The second is giving extra money to people who get their master's degree. But it in no way is associated with being a better teacher. Teach for America: slight effect. For math teachers majoring in math there's a measurable effect. But, overwhelmingly, it's your past performance. There are some people who are very good at this. And we've done almost nothing to study what that is and to draw it in and to replicate it, to raise the average capability -- or to encourage the people with it to stay in the system.
You might say, "Do the good teachers stay and the bad teacher's leave?" The answer is, on average, the slightly better teachers leave the system. And it's a system with very high turnover.
Now, there are a few places -- very few -- where great teachers are being made. . . And the whole spirit and attitude in those schools is very different than in the normal public schools. They're team teaching. They're constantly improving their teachers. They're taking data, the test scores, and saying to a teacher, "Hey, you caused this amount of increase." They're deeply engaged in making teaching better. . .
http://www.ted.com/talks/bill_gates_unplugged.html******
The study behind the changes sought for teachers.... for those who want the facts and the research behind the policy. You may not like it, you may find it personally offensive, but the statistics seem to speak for themselves. I think it speaks to the fact that some teachers are GREAT - those who have a "calling" as it were. Those who aspire to be TEACHERS and not those who want a "job"...
I'm not posting this to be drawn into a debate. I already know that some are going to be highly incensed, but this is for those people of the public who do not understand the science behind the policy. I think some will be very surprised.
Identifying Effective Teachers Using Performance on the Job
Education, Teachers
Douglas O. Staiger, Professor of Economics, Dartmouth College
Robert Gordon, Senior Vice President for Economic Policy, Center for American Progress
Thomas J. Kane, Professor, Harvard Graduate School of Education
". . . Recent evidence demonstrates that teacher certification
is a poor predictor of teacher effectiveness.
Figure 1 plots the distribution of teacher
impacts on average student math performance in grades
three through five in Los Angeles Unified School District.
The figure is based on the performance of roughly
150,000 students in 9,400 classrooms each year from
2000 through 2003. Figure 1 shows the distribution
of teacher impacts for three different groups of teachers—
those who were certified when hired, those who
were uncertified when hired but participating in an
alternative certification program, and those who were
uncertified and not participating in an alternative certification
program.1 Controlling for baseline characteristics
of students and comparing classrooms within
schools, there is no statistically significant difference
in achievement for students assigned to certified and
uncertified teachers (Kane and Staiger 2005).2
While the differences between the three groups are small,
the differences within the three groups are quite dramatic.
In other words, there is not much difference between
certified and uncertified teachers overall. But effectiveness
varies substantially among certified teachers and
also among uncertified teach The difference between the 75th percentile teacher and
the 50th percentile teacher for all three groups of teachers
was roughly five times as large as the difference between
the average certified teacher and the average uncertified
teacher. The difference between the 25th percentile
teacher and the 50th percentile teacher is also about five
times as large. And those larger differences are evident
even after adjusting for the obvious socioeconomic and
educational factors that affect student performance. A
similar analysis for distributions of reading scores yielded
similar results: that is, certification does not seem to
affect classroom performance much, but there is wide
variation across teacher effectiveness even after adjusting
for many other factors that affect student performance.
To put it simply, teachers vary considerably in the extent
to which they promote student learning, but whether a
teacher is certified or not is largely irrelevant to predicting
his or her effectiveness. But could school district leaders
learn anything useful about a teacher’s likely future impacts
by measuring that teacher’s impact on student test
scores in the past? How long would it take to make reliable
distinctions between more and less effective teachers? To
test how well a district could predict future effectiveness
using performance during the first couple of years on the
job, we focused on a sample of teachers whom we observed
in their first, second, and third year of teaching. We measured
their students’ performance during each of those
three years, controlling for students’ previous test scores
and demographics. We then ranked teachers based on
their estimated impact on their students during their first
two years of teaching, sorting them into quartiles. Figure
2 reports the distribution of estimated impacts of teachers
during their third year, using four separate curves, with
each one representing the quarter of the distribution of
effectiveness in which the teacher was categorized during
the first two years of teaching.
While certification status was not very helpful in predicting
teacher impacts on student performance, teachers’
rankings during their first two years of teaching does provide a lot of information about their likely impact
during their third year. The average student assigned to
a teacher who was in the bottom quartile during his or
her first two years lost on average 5 percentile points
relative to students with similar baseline scores and demographics.
In contrast, the average student assigned to
a top-quartile teacher gained 5 percentile points relative
to students with similar baseline scores and demographics.
Therefore, the average difference between being assigned
a top-quartile or a bottom-quartile teacher is 10
percentile points.
. . . In related research, Hanushek and Rivkin (2004) summarize
the research on the predictive power of master’s
degree completion and find little consistent evidence
that graduate degree attainment can identify effective
teachers. Similar results are reported in Murnane (1975),
Summers and Wolfe (1977), Ehrenberg and Brewer
(1994), and Aaronson, Barrow, and Sander (2003).
. . . III. Recommendations
Recommendation 1: Reduce the Barriers
to Entry into Teaching for Those Without
Traditional Teacher Certification
(there is significant information to explain these recommendations - for those interested in knowledge and understanding, the link is below)
Recommendation 2: Make It Harder to
Promote the Least Effective Teachers to
Tenured Positions
Recommendation 3: Provide Bonuses to
Highly Effective Teachers Willing to Teach
in Schools with a High Proportion of Low-
Income Students
Recommendation 4: Evaluate Individual
Teachers Using Various Measures of
Teacher Performance on the Job
Recommendation 5: Develop Data
Systems to Link Student Performance
with the Effectiveness of Individual
Teachers over Time
How reliable are quantitative measures
of teacher effectiveness?
One concern is that quantitative measures of teacher
effectiveness will be unreliable because of statistical
noise. Even if a teacher’s skills and effort remain largely
the same from one year to the next (after the teacher
has a few years of experience, at least), the average performance
of students in the classroom will differ from
year to year. In a typical fifth-grade classroom, with
only about twenty-five students taking the test each
year, a few particularly bright or particularly rowdy
pupils can substantially affect the average performance
of the class. A teacher may look good either because
the students in that year did unusually poorly on the
baseline test or unusually well on the follow-up test.
Some years a construction crew may be working loudly
across the street on the day students take their evaluation
exam, driving down the test scores for the class, or
perhaps two of the low-scoring students in a class will
come down with flu on the day of the test, bringing
up the class average.
These extraneous sources of variation mean that evidence
on teacher effectiveness mixes together true differences
between teachers and other, potentially random,
factors. As described in the technical appendix, we have
attempted to adjust downward the variation in teacher
effectiveness reported in figures 1 and 2, using our best
estimates of the proportion of the variation that is due
to nonpersistent or random factors. While we may still
be making mistakes due to random errors in categorizing
individual teachers, the total variation depicted in these
figures does not reflect these other factors (that is, the
total variation has been “shrunk” to match our estimate
of the persistent variation).
We also find that the teachers who seem to have a positive
impact on math achievement also seem to be able to raise
reading achievement. On a scale where a correlation of
0 means that those teachers who have a positive impact
on math are completely random in how they perform in
teaching reading, and a correlation of 1 means that all
the best teachers in math are also all the best teachers
in reading, the correlation between teacher effectiveness
math and reading achievement was roughly 0.6.
While there is a degree of randomness in evaluating
teachers, the evaluation does capture actual aspects of
performance. Our proposal does not attempt to use measures
of teacher performance to make fine gradations,
but instead focuses on who will look either quite effective
or quite ineffective largely regardless of the evaluation
system that is used.
Why not focus on improving teacher
quality by investing in training for
existing teachers?
Many school districts currently invest heavily in professional
development for existing teachers. However, we
believe that efforts to selectively retain the most effective
teachers are more likely to generate large increases in
average teacher effectiveness than additional training of
the existing teaching force.
Figure 4 illustrates three interesting facts. First, there
are large gains in teacher effectiveness between the first
and second year of teaching, but much smaller gains
between the second and third year. The difference
in mean math impacts is approximately 3 percentile
points between the first and second year of teaching
and roughly 1 percentile point between the second and
third year of teaching.
Second, the distribution of teacher effectiveness does
not seem to become more narrow by the third year: the
curve for teachers in their third year is just about as wide
as the curve for teachers in their first year. (In fact, it is
slightly wider.) In other words, as teachers gain experience
on the job, their effectiveness does not seem to converge.
This has potentially important implications. For
example, suppose that some teachers started out effective
and remained so and other teachers started out ineffective,
but got better. We would expect the distribution
of teacher impacts to become narrower with each year
of experience. This does not happen. In other work, we
have shown that the reverse is true: those who start out
effective in their first years of teaching tend to get better
faster than those who start out ineffective (Kane and Staiger 2005; Kane, Rockoff and Staiger, 2005). In other
words, the teachers to start out more effective seem to
improve at a slightly faster rate than those who start out
less effective.
Third, the magnitude of the payoff to experience—
about 4 percentile points over the first three years of
teaching—is small relative to the difference in effectiveness
between those identified in the top and bottom
quartile. Remember from a previous section that
the difference in teacher effectiveness, as measured by
impact on the math score between a teacher identified
as having been in the top and bottom quartiles in their
first two years, is 10 percentile points. That is, the
return to moving from one to three years of experience
is less than half as large as the difference between
teachers identified to have been in the top and bottom
quartile in their first two years.
. . . All this said, changes to tenure policies should be complements
to, not substitutes for, teacher training efforts.
One possible use of the evaluation systems described in
this paper would be to identify the highest-achieving
teachers and single them out to provide mentoring to
teachers who are struggling. Our point is that, rather
than simply invest in professional development in the
hope of solving the problem of ineffective teaching, districts
should place greater emphasis on selectively retaining
effective teachers and then invest in professional
development for them.
V.CONCLUSION
Although it can be difficult to know with much
certainty who is likely to be an effective teacher
during a job interview, we have shown that
school districts can learn a lot about teachers’ future
effectiveness simply by scrutinizing their record during
their first few years on the job. Currently, such information
is not being used. Indeed, it is usually not
even assembled—since most districts now cannot link
individual student test scores to teachers.
Over many years, American schools have experimented
with various reform strategies, from increasing accountability
to reducing class sizes. Given that history, we are
unlikely to get dramatic new results from pushing a little
harder on these familiar levers for reform. For instance,
in school systems that already have good accountability
systems, further ratcheting up the pressure is not likely
to produce sudden improvements. Moreover, raising
the hurdles for entry into the teaching profession a little
higher is not likely to generate a watershed improvement
in teacher quality. But partially because most districts
have never assembled the data required to calculate the
“value-added” by individual teachers, the payoff to beginning
to do so could be enormous.
Traditionally, policymakers have tried to raise teacher
quality by raising the hurdles for those entering teaching.
But our results suggest that those hurdles are often
not related to teacher effectiveness. Rather than continuing
to focus on teacher credentials, our proposal would
build the infrastructure to measure teacher effectiveness
on the job and to encourage states and districts to use
that information.
http://www.brookings.edu/~/media/Files/rc/papers/2006/04education_gordon/200604hamilton_1.pdf