subsided, with the courts ruling that
who taught reading in the 2009–10
not to the current year or a future year
the individually identifiable ratings
school year. These teachers were spread in which your child might be in her
could be released to the media, the
across 703 elementary schools, with
classroom. Given what we know about
department released ratings from
an average of about four teachers per
how unstable value-added measures are
2007–08, 2008–09, and 2009–10. The school; 33 schools had just one 4th
from one year to the next, perhaps the
local media publicized these ratings
grade teacher who taught reading, and
fine print should mimic that in auto-
widely in the winter of 2012, with the
three schools had 10 or more.
mobile ads: “Your mileage may vary.”
tabloids publishing front-page stories,
As researchers and scholars alike have
And finally, there’s the peculiar
replete with ambush-style photos, about noted, these value-added scores are not notion that parents are actually allowed
the teachers identified as the “worst”
precise. I found that for more than half
to choose their child’s teacher within
teachers in New York City (Chapman,
of the teachers, the range of plausible
a school. In most schools, the mecha-
Lesser, & Fanelli, 2012; Macintosh,
scores was so wide that it’s unclear
nisms for assigning students to teachers’
2012; Roberts, 2012).
whether they were in the top 20 percent classrooms don’t provide much room
There’s little question that the pub-
of teachers or the bottom 20 percent.
for parental choice because allowing
lished ratings subjected a great many
New York City teachers to shame and
ridicule. One of my own students, a
teacher in Brooklyn, reported over-
hearing a student say to one of her
colleagues, “I don’t have to listen to
Who wouldn’t prefer
their child to have a
parents to pick the most popular teacher
in the school for their child would be
neither equitable nor efficient.
With these caveats in mind, how
might a parent go about using value-
added rankings to pick a teacher for his
you! You got a D!” And at least one out-
standing middle school math teacher
or her child? Of course, in the 44 ele- teacher judged “highly
mentary schools in New York City with
decided to leave teaching at least partly
due to a Teacher Data Report that
effective” over one
ranked her as the worst 8th grade math judged “ineffective”?
teacher in New York City (Pallas, 2012).
only a single 4th grade reading teacher,
there would be no choice involved. In
2009–10, there were 146 elementary
schools with two 4th grade reading
The political leaders who engineered
teachers, and one of the teachers had
the public release of the Teacher Data
a higher value-added score than the
Reports may have misjudged the
To take this one step further, I
other in just 26 of them ( 18 percent).
political fallout. The annual ratings were compared the scores of the 4th grade
The percentage was similar for the 522
shown to be highly imprecise, with an
reading teachers who were teaching in
schools with three or more 4th grade
average confidence interval of more
the same school. This comparison is
than 50 percentiles (Corcoran, 2010).
presumably what many proponents of
Overall, if we tabulate the nearly
And many parents, as well as others
the public release of teacher evaluations 4,800 comparisons among pairs of 4th
in the school community, saw that
have in mind: empowering parents to
grade reading teachers teaching in the
the ratings didn’t align with their own
choose the best teacher in the school for same New York City elementary school,
impressions of their children’s teachers. their child.
one teacher in a pair was demonstrably
As in Los Angeles, there is little evidence
First, let’s admit that there are several more successful in promoting perfor-
so far that New York’s publication of
problematic assumptions weighing
mance on the state English language arts
unofficial teacher rankings has resulted
down the free-market concept here.
test in only 12 percent of the comparisons.
in new pressures from parents to place
For example, a value-added measure
And we would expect teachers to differ
their children in the classrooms of par-
might purport to describe Ms. Walters’s 5 percent of the time by chance alone.
contribution to the achievement of her
Of course, many of these 4th grade
average student, but if your child isn’t
teachers who were teaching reading
How to Choose?
average, the value-added measure may
were also teaching mathematics, and
To illustrate the difficulty parents
not indicate how your child would fare their value-added scores on the state
would have in identifying the “best”
in her classroom.
mathematics test could be subjected to
teachers on the basis of public ratings,
Second, a value-added score for Ms.
a similar analysis. When teachers teach
I examined the value-added scores of
Walters will almost certainly pertain
multiple subjects in self-contained
2,656 New York City 4th grade teachers to her performance in a prior year,
classrooms, parents don’t have the