Research & Insights

Some
students are more apt to be assigned better or more experienced teachers than
other students. That’s not news. Past studies have found that lower-income and minority students tend to be assigned to
teachers with less experience than their peers.
A new study by Rebecca Wolf of SRI
International plays this pattern out but goes a step further to see whether
some schools or, more intriguingly, grades within schools get a larger share of
novice teachers.

Wolf
finds that the biggest apparent driver of differences in who gets the newest
teachers within a school was the student’s grade level—not whether students were high or low performing. While the level
of student achievement played some role, the effect size was relatively quite
small (students who scored basic on the state math test were about only one
percent more likely to be taught by a new teacher than a higher-achieving
student was). However, a 9^th grade student was 10 percent more
likely than a 12^th grade student, regardless of her academic
standing, to be taught by a novice math teacher. Sixth grade was the exception
to this finding, with student achievement having a bigger impact than grade
level. A low-achieving 6^th grade student (the first grade of middle
school for most schools in the district) was much more likely to have a novice
teacher than other 6^th grade students.

So
why does assigning the newest teachers to the lowest grades in a school,
especially to the lower-performing students in those grades,
matter? The problem is that success in 6^th and 9^th
grades, the years referred to as “transition grades,” has significant
implications for students’ long-term educational attainment and engagement.
Studies show that student experiences during these years have a relationship to
drop-out and student achievement rates that can persist for years into
students’ academic careers.
Wolf’s study may mean that
principals should think twice about where they’re placing their newest
teachers.

This
year, Clark County School District in Las Vegas started the year in
a pinch: district-wide, there were over 600 teaching vacancies and student
enrollment continued to grow. In response, the district pulled out all the
stops to recruit new teachers (see ads in
airline magazines and a zip-lining superintendent) and is now rethinking how to deploy
existing staff.
Taking
lessons from Public Impact’s Opportunity Culture program as well as other
school districts and charter management organizations, Clark County is piloting
two staffing models in which effective teachers take responsibility for an
expanded group of students. They’ve launched a pilot program that includes a
blended learning model and a “teaching and learning model” in which
excellent teachers are responsible for their own classrooms as well as leading
other teachers. Clark County School District is applying the same principle to
their principals. Next year, two excellent principals will be leading two
schools each in an effort to “franchise” their approach to
leadership.
We expect that there will
be tweaks to the models for both teachers and principals, but we applaud the
Clark County School District’s effort to take risks, try new ideas and learn
from other systems. Watch this space in the coming months to see what we can
learn from them.

Teacher quality happenings this week: a new American Federation of Teachers’ survery find that teacher’s job satisfaction has dropped, James G. Cibulka is out as president of CAEP….

This week: closing poor performing schools doesn’t academically hurt displaced students, San Francisco is struggling to hire teachers for next fall, and more!

April’s Trendline is all about transfers and excessing, highlighting how districts prioritize transfers and how they decide where to place teachers when excessing occurs. Plus, we take a detailed look at mutual consent policies across the country.

Chicago teachers are going back into contract negotiations, CAP has a new report, Missouri’s new and tougher teacher licensure test is angering some preps programs and teacher candidates, and more.

TQ News for April 22, 2015

The myth: U.S. teachers spend upwards of 50 percent more time in front of their students than teachers in
other countries (see more examples of this myth here,
here
and here).

The
reality: They don’t. A new studydigs into why survey data incorrectly suggest that our
teachers spend so much more time teaching a classroom of students (as opposed
to other kinds of activities such as 
planning, classroom duty and tutoring students, which are counted
separately) than teachers in other countries.
The graph above depicts teachers’ responses to the survey question used
to gather U.S. data on teaching time–notice a pattern?
The question intends for teachers to add up how much time they spend
teaching each week and then round to the nearest hour. The responses spike at
intervals of five, suggesting that teachers estimated the amount of time they
teach each day to the nearest hour and multiplied by five (for the five
instructional days in the week), leading to an inflated amount of teaching
time.
For example, teachers who teach five 45-minute classes a day should have
reported 19 hours of teaching per week, but instead reported 25 hours, rounding
a 45-minute period to an hour.

This error of rounding seems to explain why, year after year, the OECD’s
Education at a Glance reports that U.S.
teachers spend significantly more time in front of their classrooms than do
teachers in other countries.
Another reason to
suspect the reported amount of teaching time is inflated? About a quarter of
the survey respondents reported teaching their students more hours than the
length of the school day itself.

The
lack of a major new teacher-related initiative in the Alexander-Murray ESEA draft doesn’t have us especially worked up. It
shouldn’t be a surprise to anyone that NCLB’s highly qualified teacher (HQT)
provisions are nowhere to be found in the Senate draft, which does include some
nods to—but no requirements around—teacher effectiveness.
HQT
has been on a slow path to ignominy, with many viewing its focus on teacher
qualifications over performance to be an anachronism. But HQT deserves much
credit for shining a spotlight on teachers’ need for subject-matter knowledge.
As a result, content tests are now a standard part of
teacher licensure and out-of-field teaching is much less common. 
All
but four states now require all elementary teachers to pass a content test as a
condition of initial licensure. Forty-three states require content tests for
secondary teachers, and every state now requires secondary teachers to have a
major in their subject area. Just 24 states required majors for secondary
teachers in 2001 before the passage of NCLB, and only 13 did in 1991.
Call
us optimists, but we believe content testing is now so fully ingrained in the
certification process that states don’t need federal carrots or sticks to make
sure teachers know the subjects they are licensed to teach. As for the very few
holdouts, if the feds didn’t force compliance in the last dozen years, why
would anything change now? 
On
another TQ front, we weren’t expecting to see a teacher evaluation mandate, but
we’re also not shedding any tears over its exclusion. This is another place
where states have come a long way over the life of NCLB. Actual state policy
that requires annual evaluations and evidence of student learning is preferable to vague promises to the feds
any day.
We
were pleasantly surprised to see a little gem that limits the use of Title II
Part A funds for class-size reduction to “evidence-based levels.” With so much
of this funding currently supporting small reductions that, according to
research, make no difference in student outcomes, this nugget could
dramatically change how districts use these dollars.
With
the House not sharing the Senate’s bipartisan approach to reauthorization, the
path forward for ESEA is far from clear. But another bill may have a greater
chance at moving soon, and this one—FERPA—has
us truly worried.
FERPA
is the Family Educational Rights and Privacy Act that gives parents protections
over their children’s educational records. With so much focus on data leaks and
privacy rights, it’s not surprising that a FERPA overhaul would be on Congress’
agenda.
Data
privacy is not really in NCTQ’s wheelhouse, and parents are certainly entitled
to know that information about their children is safeguarded. We’ll leave it to
those
with privacy expertise to
talk more about that.
What
has us so concerned are provisions in the draft currently circulating that
would allow parent opt-out of any use of student data by a third party. By our
read, this could not only devastate educational researchers’ access to student
information, but could also wreak havoc on the use of performance data for
accountability purposes, if the state or district contracts any of that
analysis to a third party.
How
will we put stock in any research study when we won’t know if the results are
skewed due to the exclusion of students whose parents opted out? How will we
know that school accountability or teacher evaluation results are
representative and accurate?
Surely,
we can protect parents and children without sacrificing our accountability
mechanisms and important research.

Are you worried, too? Let your friends on the
Hill know.

As I
think about my career plans for next year when I will begin work as a first-year
high school math teacher, I am both excited and nervous about taking on
teaching’s many demands. Managing a classroom full of twenty students will be a
challenge which I would like to prepare for as much as possible. So when I was
asked to research a classroom management strategy for NCTQ, I jumped at the
opportunity.
During
my research, I came to learn of a classroom management strategy called the Good
Behavior Game (GBG); it was developed in 1967 by a novice fourth grade teacher
named Muriel Saunders. The GBG creates a functional, engaged classroom by
rewarding groups that best exhibit defined behaviors, such as following
classroom rules. 
Education
is a field in which only one in 1,000 studies is replicated even once (Makel and Plucker). In contrast, the effects
of the GBG have been replicated in over 50 studies between 1969 and 2015. Each
study has proven the GBG’s effectiveness in reducing disruptive behaviors in
pre-K through 12th grade and in producing longer-term positive effects on
academic performance.
After
doing my own research on the GBG, I set off to determine how many of my future
colleagues will likely have learned about the strategy as they enter classrooms
around the country. The results were surprising. Of the 12 classroom management
textbooks evaluated in NCTQ’s classroom
management report,
including popular texts such as Marzano’s Classroom
Management that Works and Wong’s The
First Days of School, not a single one mentioned the Good Behavior Game.
While
I didn’t have ready access to a larger sample of textbooks, I was able to
estimate that approximately two percent of teacher candidates are exposed to
the GBG in their preparation courses. To arrive at this estimate, I looked
through all of the citations of Barrish (the seminal study of the
GBG) to find 16 classroom management textbooks published in the last 10 years
which reference the GBG. I then developed an inventory of the textbooks used in
90 classroom management courses. Of those 90 courses, only two use one of those
16 textbooks.
Why is
the GBG ignored in teacher prep when classroom management is so critically
important for novice teachers? NCTQ’s classroom
management reportprovides a full discussion, but in short, teacher educators promote
engagement as the be-all and end-all of classroom management. Approaches that
rely on behaviorist principles—such as the positive reinforcement at the heart
of the GBG—are belittled or ignored.
How wrong is it to deprive new teachers of an
approach that can allow them to be more effective instructors and their
students to be more successful learners? The Good Behavior Game is a useful, research-driven
strategy that I am fortunate to have learned before beginning my future
teaching career. The fact that so many of my peers will not have had the same
opportunity is an embarrassment for the field of teacher prep. Perhaps more
programs will benefit by adding this useful strategy in their preparation
curriculum.

We were heartened to find not one–but two–education studies this month using data gleaned from twins to answer research questions about teacher quality and classroom performance. This is a novel approach for education research, although it’s quite common in other kinds of social science research. The working theory is plausible (though not foolproof): by comparing how twins perform under different teachers, researchers can pull out real differences attributable to schools and not students’ backgrounds.
A Netherlands studytakes advantage of a policy in Dutch schools where twins are always assigned to different classrooms to examine the impact of teacher experience on student achievement. Researchers were able to compare the academic outcomes for 495 twin sets in grades 2, 4, 6 and 8. 
Early in their schooling, twins’ scores in grade 2 show meaningful (and statistically significant) differences based on the experience of their teachers. For each additional year of experience one twin’s teacher had over the other’s, there was about 1.5 percent of a standard deviation’s growth; that may  sound small, but over time that difference adds up. 
The researchers think that an examination of twins’ performance is an especially robust idea for early grades because there’s little chance that students would have been sorted into classes based on their academic abilities at so young an age.
That might explain why in higher grades, the impact of teacher experience decreases: the effect on students in grades 4 and 6 is only found in reading and all gains due to teachers’ years of experience disappear in grade 8.
The most interesting finding here may be that there was no evidence of the ‘plateau effect’ from experience that American studies frequently find at about year five in a teacher’s career. Rather, the effect of experience found in this study is linear, meaning that a 20-year veteran is consistently more effective than a 15-year veteran who is consistently more effective than a 10-year veteran, and so on.
Back in the U.S., another twin study also looks at the relationship between teacher characteristics and student outcomes. The research considers master’s degrees (where no relationship with student outcomes has ever been soundly documented), national board certification (where a relationship has been found that correlates with increased student outcomes) and teacher experience. Of the three factors, teacher experience seemed to be the teacher characteristic most strongly related to student outcomes. Unlike the Netherlands study, though, the effect of a teacher working longer tapers off, as almost every American study has shown, after about a five-year career.

Given the
general direction of federal and state policy over the past few years, one
could easily think most large districts include, or have set plans to include,
some measure of student growth in teacher evaluations. But, as our latest Teacher Trendline
shows, 28 percent of districts in NCTQ’s Teacher Contract Database still do not
include such measures.
When it
comes to evaluation frequency, the majority of districts in our database do evaluate all their teachers annually,
both tenured and non-tenured. Nearly 77 percent of districts evaluate
non-tenured teachers once a year. Clark
County (NV) stands out,
formally evaluating non-tenured teachers three times a year. Far fewer
districts (54 percent) evaluate tenured teachers annually. Nearly a quarter
only do so once every three years. Worse, in five California districts—Fresno, Long Beach, Los Angeles, Sacramento and San Diego, teachers are only evaluated only once every
five years after they’ve taught ten years.
Read more about teacher evaluation trends across
the districts in our Teacher Contract database, here.

For as long
as NCTQ has been tracking teacher policy, states have overwhelmingly set a
lower bar for licensing new special education teachers. This is especially troubling now, as most
special education students are expected to meet new college- and
career-readiness standards.
Consider
this: 35 states still offer or exclusively grant K-12 special education teacher
licenses, sending an unequivocal message that the content knowledge and
pedagogy needed for an elementary and secondary special education classroom are
interchangeable. Just 15 states require special education teachers to even pass
a content knowledge test.
Of the 18
states that require general ed elementary teacher candidates to pass an
adequate test of the science of reading, only 11 also require special education
teachers at the elementary level to pass the same test. Considering that
reading difficulties are the primary reason for referrals to special education,
the absence of a requirement should trip some alarms. The remaining seven
states, Alabama, Florida, Minnesota, Mississippi, New Hampshire, New Mexico and Ohio, all need to close this loophole.
There are notable exceptions. New York, in particular, stands out. In
addition to requiring special education teachers to pass content tests that
appear to be as rigorous as what is required of other teachers, New York is the only state that
requires secondary special education teachers to demonstrate content knowledge
in all subjects they are licensed to teach.

The
surprising news about Russ Whitehurst’s departure from Brookings Institution’s
Brown Center on Education Policy, a center he had ably led since 2009, has us
scratching our heads as much as anyone else in the DC policy community.

As an
organization, NCTQ is very grateful to Whitehurst for his continuing work on
NCTQ’s Audit Panel, where he’s helped make sure the ratings processes for the Teacher
Prep Reviewmeet
the highest standards. But of course his justly earned claim to fame is his
having founded the U.S. Department of Education’s Institute for Education
Sciences. And it was as director there that he helped lay the intellectual
foundation for far stronger teacher preparation and classroom instruction than
we have today.
One of the
main obstacles to improving teacher effectiveness has long been the meager
research base on what teachers should
be trainedto be able to do. The
relative paucity of research in what works in teaching as compared to, for
example, medicine, has paved the way for the field’s abdication of training
altogether.
But while
there remains much we do not know about what constitutes good teaching, there
in fact is a core set of strategies, identified through high-quality research,
that every teacher should master. Russ had the IES assemble and pressure-test
this research, and then put out highly readable practice guides highlighting the teaching strategies that have the
greatest demonstrated impact. Thanks to him, teaching and teacher prep now have
some of the blocks around which the profession could be built.
Unfortunately, as we’ve found in our own past
research on training in classroom management and our upcoming report on the fundamentals of
instruction, these practice guides have gone largely ignored by teacher
educators. But by publicly rating programs on how well they train teachers to
use the strategies in the practice guides, we aim to draw the field’s attention
back to the strong research contained in the practice guides and firmly
ensconce Whitehurst’s legacy in the training of new teachers.

As scarce as
international studies on teaching generally are, we were glad to see two new
studies looking at what it takes to be a successful math teacher. Both provide
more evidence that there are two distinct types of knowledge relevant to
teaching mathematics: content knowledge and pedagogical content knowledge. The
former is self-evident; the latter involves the capacity to understand possible
student misconceptions, to recognize alternative problem-solving approaches,
and so on. 

One study looks at teachers in Belgium (Teachers’ content
and pedagogical content knowledge on rational numbers: A comparison of
prospective elementary and lower secondary school teachers) while the other focuses on teachers in Taiwan and Germany
(Content knowledge and pedagogical content knowledge in Taiwanese and German
mathematics teachers). The studies share a few common and not wholly surprising
findings:
The two types of knowledge are definitely
different and, fortunately, both can be measured.
The fact that secondary teachers have greater
content knowledge than lower level teachers doesn’t always imply that they have
greater pedagogical content knowledge.
Frankly, the
most interesting piece of information in the two articles is how poorly
Belgium’s elementary and lower secondary teachers performed on a test of
pedagogical content knowledge involving fraction problems (even though Belgian
students outperform ours both as 9-year olds and
15- year olds). It’s well established that American
elementary and middle school teachers are relatively weak in math, but it’s an
eye opener to see that on a math problem involving division by a fraction, 86
percent of teachers in higher-ranked Belgium were incorrect.
In any case,
this revelation of the difficulty elementary and middle school teachers have
with fractions no matter what side of the pond they’re on poses an excellent
opportunity to show very graphically the difference between two elementary math
textbooks we’ve evaluated and how each takes a very different approach to
developing both content knowledge and pedagogical content knowledge.
First a bit of math on the topic of division by
a fraction: Teachers should most definitely not reinforce the idea that the basis for this operation is
a mystery. (“Ours is not to wonder why, just invert and
multiply!”)  A good teacher can demonstrate the two types of situations
modeled by this equation:
In the first
situation, we’re finding out how many quantities of 1/4th are in 1/2 (there are
2);  in the second situation, we’re
finding out that if we did 1/4th of a job with 1/2 of a quantity, it will take
2 of the quantity to complete the job.
(Keep
reading.  No one said this was easy!)
Here’s how a
textbook to which we gave a low score (Mathematics
for Elementary School Teachers, Bassarear, 5th ed.) presents this topic on p. 280:
There’s half a page of discussion that draws on the use of the multiplicative
inverse in an abstract way and provides no explanatory graphic. In fact, the
author explains, “Unlike most of the other algorithms we have examined,
[the invert-and-multiply algorithm] does not lend itself to a diagrammatic
representation.”
Here’s the final part of the explanation
offered:

Good thing
that Sybilla Beckmann (author of a textbook to which we gave a high score, Mathematics for Elementary Teachers with
Activities, 4th ed.) wasn’t paying attention to that conclusion. She
provides three pages of discussion demonstrating both types of situation
described above, with no fewer than four explanatory graphics—exactly the types
of graphics elementary teachers can use in their own instruction.
Here’s one
of the graphics (and
the full discussion can be found here): 

The quality
of textbooks in both elementary math and early reading prep vary dramatically.
Since our first reports on the preparation of elementary teachers in math
and reading, we’ve invested a lot in textbook reviews (analyzing the several
dozen elementary math textbooks and close to 1,000 early reading textbooks).
Ratings are based on extensive textbook reviews done by experts; in the case of
math textbooks, reviews are done by mathematicians well versed in the art of
teaching mathematically-skittish elementary teacher candidates. We’ve posted
both the math
and reading textbook reviews; doing more to ensure that teacher prep
instructors hear about these evaluations the next time they choose a textbook
for their course is critical.

This month’s Catching up on Contracts reports on contract changes in five districts in our Teacher Contract Database: San Francisco Unified School District, Albuquerque Public Schools, Brevard County School District, Elgin School District U-46 and Tulsa Public Schools.  

For this month’s Trendline, we closely examine teacher evaluation policies in the largest districts across the country. Specifically, we take a look at the major components within evaluations, frequency and teachers’ ability to grieve their evaluation ratings.

This month’s Catching up on Contracts focuses on salaries and benefits in the five districts we reviewed. In four of the featured districts, Oklahoma City Public Schools, Omaha Public Schools, Long Beach Unified School District and Hartford Public Schools, the salary changes are nothing out of the ordinary. However, one district, Mesa Public Schools, introduced a complete overhaul of its previous salary structure this school year.