provement looks like. It should not be a semi-
random act, localized to a school that would
most likely not be sustainable once person-
nel changes occur. Improvement in educa-
tion should arise from a systematic approach
that was developed rigorously and iteratively,
leading to learning that not only results in
improvement at the school but lessons that
can support effective work of others.
Secondly, low-performing schools are
the places least likely to have the required
conditions for the kind of deep implemen-
tation and sustained focus that high quality
improvement science requires. In addition,
these schools are not likely to have the au-
tonomy and time required for effective im-
provement methods because state and dis-
trict policies often put more restrictions and
requirements on such schools.
Implementing any improvement meth-
odology scientifically requires the under-
standing that failure is a part of learning,
that learning what does not work is impor-
tant, and that a culture of continuous im-
provement requires a sense of professional
safety that our typical accountability struc-
tures confound.
Providing resources only to schools – and
not the district central office – would likely
be insufficient because the learnings needed
for the school to be more effective in the long
run are often needed at the districts to which
the schools are accountable. To get improve-
ment right requires a systemic approach that
acknowledges learning from improvement
requires more coherence and alignment
across related systems than is typical, and
offers local decision-makers the time to be
deliberate, to learn from failures, and to sup-
port staff as they manage the effort.
This often means 1) more autonomy, not
less, so that leaders can quickly approve
changes in practices and policies; 2) fewer
programmatic requirements rather than
extra ones, so that going deep on a few ef-
forts is possible; and 3) extra resources be-
cause deep, hands-on learning requires re-
sources often lacking in the neediest schools.
Recommendations to avoid the problems
caused by this misconception:
1.) Build broad capacity to do improve-
ment science: The power of improvement
science is the ability to help systems adap-
26
Leadership
tively apply a new intervention across a di-
verse set of contexts. Therefore, district lead-
ers should plan to implement improvement
science in a set of schools in diverse circum-
stances, rather than just the low-performing
ones. If only low-performing schools engage
in improvement science, the district will
learn how to change the direction of a low-
performing school, but not know how to
apply the design principles of effective inno-
vations in schools operating under different
conditions.
2.) Engage the district leaders in learning
too: Involve district leadership in the work
of an improvement science network from
the onset, with the expectation that they are
there to understand schools’ obstacles that
are rooted in the district office and to remove
those obstacles or provide schools with the
autonomy needed to navigate them. These
central office leaders are themselves devel-
oping as improvement scientists and will be
responsible for extending the capabilities to
other schools/departments. It would be best
if district leaders tackle a problem of practice
of their own, rather than just observing the
work of others.
Misconception #3: Improvement
science is just another set of data
tools and practices.
The tendency to see only the tools and
methodology of improvement science, and
not the underlying thinking, is a common
problem in adopting continuous improve-
ment methods, espe cially those that heavily
rely upon systematic improvement cycles.
The repetitive nature of the cycles can some-
times create a focus on the trees rather than
the forest.
Improvement science is more than just
the sum of its parts, and doing it in public
school systems requires a dramatic mindset
shift. In fact, improvement science when
done well is more of a movement than it is a
set of trainings.
Inspiring an entire district, or even an en-
tire school, to achieve this level of intention-
ality and coherence requires more than just
data protocols and training sessions. This
is one reason why the leaders at Carnegie
have been talking about the work in the lan-
guage of social movements. But generating
the type of collective passion that successful
movements have requires an attention to the
psychology of change – the intra- and inter-
personal aspects of getting people to work
differently.
Recommendations to avoid the problems
caused by this misconception:
1.) Focus first on the mindset of improve-
ment science: Improvement science has an
underlying curiosity and passion for results.
This requires teams to believe that meaning-
ful change is possible and can be a direct re-
sult of their own improvement efforts. To do
it well, one needs to think like Thomas Edi-
son, who said during his quest to invent the
light bulb, “I have not failed. I’ve just found
1,000 ways that won’t work.”
2.) Remember that you’re not just chang-
ing systems; you’re changing people: To get
individuals and teams to work more boldly
and more ref lectively, leaders must create
trustful, risk-tolerant workplaces. This re-
quires attention to community-building,
information flows, culture and identity.
Conclusion
A pioneer of improvement science in
health care, Don Berwick, offered a sugges-
tion for the education field based on what he
has learned in his own field: “Abandon reli-
ance on inspection to improve. It infects our
culture. It infects our world.”
We agree that educators should not focus
on inspecting for poor performers and cull-
ing them out of the system. Instead, we sup-
port improvement science, which looks at
every member of an organization as an im-
prover and a potential source of innovation.
We believe in improvement science’s
combination of tools, strategies, and atten-
tion to the interpersonal aspects of change.
Improvement science provides a systematic
means for getting better at getting better.
Derek Mitchell is the CEO of Partners
in School Innovation. Chris Thorn is the
director of knowledge management for the
Carnegie Math Pathways at WestEd. Brian
Edwards is a research writer with Partners
in School Innovation.