Expertise in Teaching

24 Expertise and Expert Performance
in Teaching
James W. Stigler
Department of Psychology, University of California at Los Angeles,
California
Kevin F. Miller
School of Education, University of Michigan, Michigan
Introduction domains of expertise. But in education, teachers

are in a very real sense at the mercy of students,
There are approximately 3.5 million full-time
and policy-makers, and curriculum designers,
elementary and secondary teachers working in
and so on. This is something we will need to
US public and private school classrooms (US
address as we consider what it means to be an
Department of Education, 2016). Given their
expert teacher.
potential to impact the lives of young people, it
Another obstacle to understanding expertise in
would seem to be of great importance to under-
teaching is the “pseudo-expertise” one develops
stand what makes one teacher more effective than
as a student. Before anyone starts their formal
another, what expertise in teaching looks like, and
training as a teacher, they already have experi-
how it develops. Despite the fact that teaching is
enced well over 10,000 hours as students in class-
one of the oldest and largest of human profes-
rooms, making teaching the profession with the
sions, however, we still lack a clear conception of
most intensive and lengthy apprenticeship of any.
what it means to be an expert teacher. This is not
One consequence of this experience is that every-
because researchers have not tried to pin this
one in our society, including teachers, thinks they
down. They have, and we will try in this chapter
already know what an expert teacher is, without
to make a contribution to this effort. But it is
any serious consideration of the research. This
worth discussing, at the outset, why these ques-
leads to bias in research on teaching, with a dis-
tions have proven so difficult to answer.
tinct lack of research designed to investigate
Perhaps the biggest challenge to understanding
theories that “everyone knows aren’t true.” It
expertise in teaching is that teaching is not an
also may reduce variation in teaching methods
individual endeavor in which the teacher him-
within our culture, which makes it more difficult
or herself is the only actor. Teaching is a complex
to explore alternatives.
system of interacting elements, and effective
Finally, our work is hampered by a lack of a
teaching requires that all of these elements work
consensus on the aims of education. We have a
together to produce the desired outcomes. Later
wide array of desired educational outcomes, yet
we will consider the nature of this system. But
teaching practices that prove effective for one
for now, consider simply that teachers cannot
outcome might be ineffective, or worse, for
achieve their goals without the cooperation of
others. Similarly, practices that yield impressive
students. Chess experts don’t require the coopera-
short-term results (e.g. high scores on year-end
tion of the chessboard. The chessboard, the musi-
standardized tests) may have negative effects on
cal instrument, and so on are invariants in many
432 p ar t v . i d om ai n s o f ex p er ti s e : p r of e ss i on s
long-term outcomes such as career advancement The Nature of Teaching

and satisfaction decades later. Work by Jackson
and colleagues (Jackson, 2012; Jackson, Rockoff, A Definition of Teaching
and Staiger, 2014), for example, has shown that
Coach John Wooden famously said, “Everyone’s a
teachers who produce the strongest gains on
teacher, to someone” (Gallimore & Tharp, 2004,
achievement tests are not the ones who succeed
p. 119). The ubiquity of teaching means that we
at reducing absences and suspensions, variables
need to restrict the domain we seek to describe. In
shown to predict future educational attainment
this chapter we will narrow the focus somewhat to
and adult earnings. This complicates the straight-
include only classroom teaching. Despite this
forward strategy of studying teaching expertise
narrow focus, classroom teachers make up the
by studying acknowledged experts.
largest segment of public sector employment,
Although it is tempting to rally round a single
with 3.5 million teachers in the United States
measure of student outcomes (e.g. the currently
(US Department of Education, 2016) and more
popular “value added” models of student achieve-
than 29 million in the world (UNESCO Institute
ment), we must be careful not to get too focused
for Statistics, 2015).
on the metric instead of on the underlying process
We base our definition of teaching on one
the metric is intended to reflect. Campbell (1979)
offered by Lampert (2003; see also Ball &
warned of this in what came to be known as
Forzani, 2007). Lampert sees teaching as “work-
“Campbell’s law”:
ing in relationships,” specifically, the relationships
The more any quantitative social indicator is used among the three core elements of a classroom
for social decision making, the more subject it will lesson: the teacher, the students, and the content
be to corruption pressures and the more apt it will be that is being taught. Teachers must manage the
to distort and corrupt the social processes it is content and the students, but most importantly
intended to monitor. (Campbell, 1979, p. 85)
they must manage the relationship between stu-
These concerns don’t invalidate the effort to dents and content, over time. These relationships
understand and measure student educational define the “problem spaces” in which teachers
outcomes or to understand how teachers develop work. So, for example, teachers must relate to
the skill to promote these outcomes, but they do students, and must collaborate with students to
help explain why understanding teaching exper- get work accomplished, which generally means
tise is not a simple matter of identifying and getting students engaged with studying the con-
studying acknowledged experts. tent. Problems arise in the management of each of
In this chapter we try to take a broader these relationships, and teachers, over time,
approach to understanding the nature and develop routines for handling the problems that
development of expertise and expert perfor- recur.
mance in teaching. Because the literature on Although we find this model to be a useful
expertise in teaching is limited, we instead try starting place for a full definition of teaching,
to integrate a number of ideas and findings from we will broaden it in three ways. First, teaching
literatures as diverse as cross-cultural compar- always implies some goal for students (e.g. a
isons of teaching, cognitive psychology, and learning goal) and some sequence of events
systems improvement, among others. At a designed to achieve that goal. Although some
minimum, we hope to start a new conversation educators may prefer a more democratic view of
about what expertise looks like in the highly relationships within the classroom, we believe it
complex domain of classroom teaching. is important to see the teacher as an actor with a
Expertise and Expert Performance in Teaching 433
specific agenda for what she wants students to of time. And there are only a finite number of
learn. This goal-directed nature of teaching was lessons in the school year to cover the prescribed
cited explicitly in some early definitions of teach- curriculum. Thus, in some sense, planning of
ing. For example, Thorndike (1906) defined individual lessons and sequences of lessons is a
teaching as the methods used to help students zero-sum game: as new activities are added,
achieve the learning goals valued by society. others must be deleted.
The importance of making this agenda explicit The centrality of time in teaching becomes
has grown over the past 10 or 20 years as states even clearer when we consider students.
and districts have focused more on creating and Berliner (1990) noted that the amount of time
implementing clear academic content and perfor- students spend actively engaged in learning is
mance standards for what students should know often a small fraction of the time they spend in
and be able to do at each grade level. Despite the school. The amount of such engaged time can
recent influence of national and local standards, vary greatly across classrooms, and this can be a
however, the teacher must still decide what learn- strong predictor of student learning (Fisher et al.,
ing goals the students in her class can and should 1980). Teachers’ skillful management of time is a
achieve – at the end of the lesson, at the end of the key factor in determining student learning. That
unit, or at the end of the whole school year. management, and the planning and decisions that
A second modification we make to Lampert’s undergird it, are an important aspect of teaching.
model of teaching is to broaden “teaching” to Thus, we offer this definition of teaching:
include the planning and reflection that go on Teaching consists of the interactions of teachers,
before and after the lesson. It is common in the students, and content, in classrooms, that are
United States to overemphasize the importance of intended to achieve some goal or goals for stu-
what happens during the lesson – the classroom dents, within a specified period of time (e.g. a
performance of the teacher – and de-emphasize classroom lesson), together with the planning
the intellectual work outside the classroom – that takes place before, and the analysis that
what happens during planning and reflection. In takes place after, the lesson.
fact, the effectiveness of a classroom lesson can
be determined as much or more by the plan as by
Teaching Is a System
the on-the-fly decisions made by the teacher
during the lesson. Planning is a teacher’s single More than many domains of expertise, teaching is
most powerful leverage point for improving the a complex system with many moving parts. An
quality of what happens during a classroom expert pianist, for example, can have near total
lesson. Likewise, the teacher’s later reflection control over the product of her expertise, reliably
on, and analysis of, problems that occur as the performing beautiful music time after time. She
lesson unfolds can lead to revisions in the plan for can depend on a great deal of constancy in the
next time, revisions that yield improvements in environment. Unlike students, the keys on the
instruction. Reflection, or analysis, as we will call piano on which she performs can be counted on
it, can be thought of as planning for the future to stay in place and respond consistently; the
informed by evidence gathered during the lesson. Mozart concerto will be the same each time she
A final element in our definition of teaching is performs it; and the audience will apply a com-
time. Classroom teaching is highly constrained mon set of standards as they appreciate and eval-
by time. Teaching is often implemented as a ser- uate her work.
ies of lessons, and these lessons are separated in Teaching, in contrast, is constrained by a num-
time. Each lesson is often limited to a set amount ber of variables, many beyond the teacher’s
control. First there is the matter of students: a teaching and its development require institutional
teacher can succeed only if students cooperate supports and only make sense in a larger cultural
and engage in the tasks and activities the teacher context.
assigns. According to Cohen, a leading education
researcher, this is one of the key predicaments
Teaching Is a Cultural Activity
that define the teaching profession (Cohen, 2011).
But it is not just the students that co-determine the Not only is teaching a complex system, but it is also
results of teachers’ work: curriculum, textbooks, a cultural system, what some have called a cultural
assessments, policies, parents, peers, and so on, activity (Gallimore, 1996; Stigler & Hiebert, 1999).
all interact and contribute to the nature and out- Our own thinking on this emerged from the
comes of the system we call teaching. Third International Mathematics and Science
The fact that teaching is a complex activity that Study (TIMSS) video studies, conducted during
must be considered from a systems theory per- the 1990s (Stigler & Hiebert, 1999, 2004). In
spective presents challenges to those who would these studies, videotapes of national samples of
describe and study expertise in teaching. A tea- eighth-grade mathematics and science lessons
cher may appear to be an expert in one environ- were collected in seven countries: Japan, Hong
ment, yet seem more like a journeyman when Kong, the Netherlands, Switzerland, Czech
observed in a different environment. The success Republic, Australia, and the United States. The
of a lesson depends in part on a teacher’s prepara- study had two goals: to document what “average”
tion and planning and her ability to assess stu- teaching looks like at a national level, and to
dents’ understanding and respond to their needs. compare teaching in the United States with teach-
But it also depends on factors that are largely out ing in countries that are high achieving when
of her hands, such as the previous experiences of compared to the United States based on assess-
those students as well as institutional culture, ments of mathematics and science achievement.
routines, and support for learning. Findings from these studies indicated a striking
The question of whether expert performance homogeneity of teaching practices within coun-
“belongs” to the expert or to the organizational tries, but marked differences in practices across
context is a matter of continuing controversy. countries. Even within a country as diverse as the
Mueller and Dyerson (1999) have argued that United States – racially, ethnically, linguistically,
expertise in complex institutions requires organi- and socio-economically – a national sample of
zations that can develop and take advantage of eighth-grade mathematics teachers appeared to
individual skills; absent that, human experts may be following a common script, despite the fact
fail to make a contribution commensurate with that teachers are given high levels of autonomy
their abilities. Others, such as Hoffman (1998), and control over the methods they use. The US
have argued that the question of whether knowl- lesson script appears designed to produce what
edge resides in institutions or individuals is Skemp (1987) called an “instrumental” under-
largely an artificial one that relates to how exper- standing of mathematics. Teachers walk through
tise is used rather than whether or not individuals example problems, then supervise students as
can be experts. In this sense, an expert teacher in a they practice solving similar problems, the goal
dysfunctional school system would not be some being for students to remember the steps used by
kind of oxymoron, but rather a waste of human the teacher, and then to be able to execute the
resources. Because larger cultural factors are cen- steps without errors.
tral to both the development and practice of In Japan, by contrast, the lessons follow a
expert teachers, we would argue that expert different cultural pattern. Japanese teachers
typically begin by presenting students with a Cultural activities can be hard to see simply
difficult problem to work on, one they have not because they are so widely shared within a cul-
seen before. The teacher does not instruct tural group. The teaching methods that are preva-
students how to solve the problem, but simply lent in US schools today appear natural to those
lets the students struggle to find a way to solve of us who grew up here, but may appear strange to
it on their own. Because students have not those who experienced a different tradition of
been told how to solve the problem they schooling. Breaking free of our cultural lenses is
usually come up with a variety of different solu- one of the main benefits of cross-cultural compar-
tion methods – some correct, some incorrect – ison. When Japanese students are asked to solve
which they then discuss in class. Through these problems they have never seen before, they strug-
discussions, teachers focus on what Skemp would gle and often appear confused, yet teachers do not
call a “relational” understanding of mathematics, intervene to simplify the problems or resolve the
working to deepen connections with core under- confusion. When we observe this in Japan it sud-
lying mathematical ideas (see also Schwartz, denly draws attention to how uncomfortable US
Chase, Oppezzo, & Chin, 2011). students and teachers are with the experience of
Cultural routines such as these are not created confusion. These different cultural routines are
deliberately. Instead, they evolve slowly over supported by wider cultural beliefs about the
time as cultures adapt to an ever-changing envir- role of confusion in learning.
onment. Modern schooling is itself a cultural Finally, cultural activities are hard to change
invention, perhaps one of the most successful because they are multiply determined. Many fac-
cultural inventions of our modern era. As the tors conspire to keep things as they are. We know
world economy shifted from agriculture to indus- from cognitive science research that confusion is
trial production, schools developed to prepare actually a critical part of deeper learning (D’Mello,
workers who could fit into the jobs and living Lehman, Pekrun, & Graesser, 2014). Yet if a US
arrangements that were emerging. Schools, in a teacher tried to induce confusion in their students
sense, went viral, and now they are everywhere. (let’s call it “productive confusion”) using teach-
And significantly, the cultural routines of teaching routines similar to those used in Japan, she
ing that developed in the United States more than would no doubt get a lot of pushback. Students
a century ago appear to be mostly unchanged would complain (“We haven’t had that!”), parents
(Cuban, 1990; Hoetker & Ahlbrand, 1969). would complain (“It’s not fair to expect students to
Cultural activities are learned implicitly, which work on problems you haven’t taught them how to
makes teaching quite different from most other solve”), textbooks would not introduce material in
domains of expertise. As a cultural activity, teach- the right order, and so on. The cultural nature of
ing is more like dinner-time conversation than it is teaching raises problems for educational change,
like flying an airplane. The routines of dinner-time but it also has implications for the nature and
conversation are learned from growing up in a development of teaching expertise.
family and observing how others behave at meal
times. People do not take a course or read a manual
Expertise and Cultural Activities
to learn this. They learn to participate in cultural
activities by observing and imitating others. There Cultural activities are implicitly learned and often
is evidence that people learn to teach in just this operate outside of our awareness. This raises
way – by observing their teachers during their 13 some interesting issues in terms of expertise.
years of schooling before entering college and by Most participants in cultural activities are neither
imitating what they remember (Lortie, 1975). experts nor novices; they are simply operating
within the normal limits of variation for their Expertise in Teaching

cultural system. In Ericsson’s terms (Ericsson,
The first step in studies of expertise typically
2008), they have reached a (perhaps premature)
involves identifying a set of experts and analyz-
level of automaticity, a level at which they are
ing how they differ from novices. In the case of
“good enough.”
teaching, three main approaches have been used,
An important consequence of the cultural nature
each with important limitations. These are: (1)
of teaching is that one can be an expert in imple-
comparing beginning and experienced teachers,
menting teaching routines that are not themselves
(2) studying teachers who have been identified as
optimal. Hatano and Inagaki (1986) distinguished
experts through a process of nomination or certi-
between two courses of expertise. Routine experts
fication, and (3) looking at student outcomes to
become efficient at implementing relatively set
identify expert teachers.
routines (an example might be someone working
in a fast food restaurant) without necessarily
understanding why they work or being able to Experience as a Proxy for Expertise
reproduce them in a different setting. Adaptive A large body of literature supports the idea that
experts (such as a sushi chef) deal with constantly expertise requires practice over long periods of
changing problems and need to develop both an time, but experience alone does not guarantee the
understanding of why things work as they do and development of expertise. Studies of the effect of
an ability to alter their approach as circumstances teacher experience on student achievement gener-
change. ally find positive but small relationships between a
It seems obvious that teaching involves adap- teacher’s years of experience and their students’
tive expertise, because students present con- learning (e.g. Nye, Konstantopoulos, & Hedges,
stantly evolving challenges to teachers. Yet to 2004; Rockoff, 2004); some studies show benefits
the extent teaching is nested within a set of cul- during the first several years, but none after that
tural beliefs, the range of exploration of possible (e.g. Hanushek, 2003; Rivkin, Hanushek, & Kain,
strategies will be limited. Those cultural beliefs 2005). The problem with conflating experience and
vary across cultures but also change over time. expertise has long been recognized (e.g. Berliner,
Resnick and Resnick (1977) showed that societal 1986). Despite this, experience still has been the
expectations on literacy have changed dramati- main variable used to indicate expertise in
cally over time. Strategies for teaching that work teaching, at least until recently. Given that most
when only a low level of literacy is expected for practitioners in a domain will plateau before
most students will not be as successful when we reaching the highest levels of expertise, these mod-
expect much more from all students. This in turn est relationships probably underestimate the effects
has implications for what constitutes expertise in of expert teachers on their students’ learning.
teaching literacy.
Whether we should expect expert teachers to
Studying What Recognized Experts Do
be able to transcend the limits of their cultural
teaching routines is an open and challenging The second approach to understanding expertise
question. We have now discussed some of the involves studying people who have been recog-
obstacles to a straightforward analysis of exper- nized as experts. One obstacle to this approach is
tise in teaching and described a model of what that, at least in the United States, teaching is an
teaching entails. In the remainder of this chapter activity that is often observed only by students.
we will discuss what this means for studying and Berliner (1986) noted that judging of skill and
developing teaching expertise. outcomes in athletics, livestock, dogs, crops, and
other fields takes years of systematic training and that students of NBPTS certified elementary
practice, but we lack anything similar in the school teachers in North Carolina did better than
evaluation of teaching. those of uncertified teachers, with effects on the
Efforts to systematically observe elementary and order of 0.05 standard deviations. Goldhaber and
secondary school teaching have a long pedigree, Anthony reported a larger effect, of 0.11 standard
going back at least to a program in the 1880s and deviation for students eligible for free or reduced
1890s that John Dewey participated in at the cost lunch.
University of Michigan (Williams, 1998). More A much smaller effect was reported by Harris
recently, the National Board for Professional and Sass (2009), who looked at data from Florida.
Teaching Standards (NBPTS) has developed a pro- They also reported larger effects for teachers who
cess for certifying teachers. To date more than had received NBPTS certification when it first
112,000 teachers have completed their certification became available. They argued that this suggests
process, which involves submitting a multimedia that the certification process initially identified a
teaching portfolio and taking a three-hour assess- set of committed and effective teachers, but that
ment examination. as it became more widespread the certification
An early evaluation of the program (Bond, process was less successful in identifying more
Smith, Baker, & Hattie, 2000) looked at a set of effective teachers.
65 experienced teachers divided into groups The NBPTS has provided a systematic way of
based on their performance on the assessment, designating some teachers as at least relatively
with approximately equal numbers passing and expert. There is some evidence that the teachers
failing the NBPTS certification assessment. so designated differ from their peers in ways
Interviews and classroom observations showed consistent with current ideas about effective
consistent differences favoring the designated teaching, but less consistent and convincing evi-
expert teachers in, among other areas, the depth dence that their students do better than do those of
and challenge of problems set for students, the other teachers. This suggests that we might do
ability of teachers to anticipate and plan for class- better by simply identifying expert teachers based
room problems, and the depth of their representa- on how their students perform.
tion of classroom situations. Student writing
samples from the NBPTS certified teachers
Identifying Experts Based on Student
demonstrated higher understanding than did
Achievement
those from the comparison group. A later study
by Hogan and Rabinowitz (2009) compared An inductive approach to identifying experts
NBPTS certified and novice teachers and found involves looking at results and designating as
similar differences in the depth of their represen- experts those teachers whose students develop
tation of classroom problems. faster than expected. As discussed at the start of
Efforts to show that NBPTS certified teachers this chapter, this approach is at the heart of the
have better student outcomes than do their non- “value added” approach to teacher assessment.
designated peers have shown more mixed results. Finding ways to connect teacher expertise to
Cavalluzzo (2004) compared student achieve- student outcomes will be critical to improving
ment in mathematics in a large urban school dis- education, but efforts to do this to date have
trict in classes taught by Board-certified and other shown how complicated this seemingly straight-
teachers, finding consistent effects favoring the forward approach is.
certified teachers. Goldhaber and Anthony (2007) One recent attempt to identify practices asso-
and Clotfelter, Ladd, and Vigdor (2007) found ciated with student learning was the colossal
Measures of Effective Teaching (MET) project In the case of teaching, we can imagine several
sponsored by the Bill and Melinda Gates additional reasons such correlations would be
Foundation. The study was the most ambitious low. One might be the alignment between tea-
study of teaching ever undertaken: more than chers’ goals and the measures used to indicate
20,000 videotaped lessons were collected from student outcomes. If standardized tests aren’t
3,000 elementary and middle-school teachers’ measuring the student learning outcomes of high-
classrooms in seven urban school districts across est priority to teachers, then you would not expect
the United States. Students were surveyed, and a correlation between teachers’ behaviors in the
students’ test scores – both on state standardized classroom and students’ end of year test scores.
tests and on supplemental tests of higher-order But perhaps even more important is the contex-
thinking in both mathematics and literacy – were tual nature of teaching. A move that might be best
collected and matched to the video data. Even in one situation might be precisely the wrong
more impressive, in the second year of the study move to make in a different situation. For exam-
a subsample of 800 teachers were randomly ple, a critical remark to one student may be just
assigned, within schools, to different classrooms what he needs to engage him in digging deeper on
of students, and again, students’ learning at the a problem. But the same remark may come as a
end of the year was measured. crushing blow to a different student, who needs
This was the first large-scale study to identify more encouragement.
teacher effects on students’ learning using In this sense, teaching is more like driving to
random assignment. The measures of teacher work than like shooting a rocket ship to the
effectiveness based on students’ learning in moon. Most of the work involved in shooting a
year one (non-random assignment) were highly rocket ship can be done in advance. The trajec-
predictive of the randomly assigned students’ tory can be calculated with near-perfect accu-
learning at the end of year two (Kane, racy, such that when the button is pushed,
McCaffrey, Miller, & Staiger, 2013), and impor- everything unfolds as predicted. Driving to
tantly, the size of the teacher effect in year one work, on the other hand, requires constant
was the same as that in year two. Clearly, some adjustment to the expected and unexpected var-
teachers are more proficient than others at pro- iations that occur. Taking a left at a certain
ducing student gains on standardized state tests. intersection may usually be best, but not
But disappointingly, the observational measures always. Unexpected obstacles or weather con-
applied to the videos of classroom teaching ditions require a response from the driver. The
yielded very little of note, predicting almost expert commuter knows when to adjust, when to
none of the variance in student learning at the go around. Teaching, thus, is more like driving
end of the year (Kane & Staiger, 2012). to work. With goals in mind, teachers must
These findings are in some ways reminiscent of constantly read the situation, monitor progress,
the early work on chess (de Groot, 1965, 1966). and make necessary adjustments.
Despite numerous attempts to find differences in This analysis leads us to reject the idea that
the way chess masters and weaker players play expertise in teaching can be defined in terms of
chess (e.g. number of moves considered, search decontextualized “best practices.” Our view is
heuristics, depth of search, etc.), the only reliable that correlations between teacher actions and stu-
difference turned out to be in the quality of the dent learning are low not because we haven’t yet
move: chess masters make better moves in any identified the right set of best practices, but
given situation, and thus win more games, than because teaching itself is contextual, meaning
the weaker players. that such correlations will always be low.
Further support for this view comes from the The Construct of Learning Opportunities
TIMSS video studies. If expertise in teaching is
If teacher expertise cannot be equated with a set
defined by a set of best practices, then one would
of best practices, what does explain the consider-
expect the practices used by teachers to be similar
able teacher-level variance in student learning
across the highest achieving countries. For exam-
outcomes? We believe a clue may be found by
ple, because Japan is a high-achieving country in
digging deeper in the TIMSS videos of class-
mathematics, we might expect that Japanese
rooms in the high-achieving countries. In fact,
teaching routines (as described earlier) would be
we do see commonalities, not at the level of what
similar to those used in other high-achieving
teachers do but in the kinds of learning opportu-
countries.
nities they manage to create and sustain for
In fact, however, this proved not to be the case.
students. Teaching routines differ among the
Although students in the Netherlands, Czech
high-achieving countries. But although they use
Republic, Switzerland, and Hong Kong – as
different routines, and the actions of teachers
well as Japan – all score relatively high on inter-
differ, all appear to have found ways to create a
national mathematics tests, teaching methods
common set of learning opportunities for students.
across these countries vary markedly from one
Based on results from the TIMSS video studies,
another. Features of teaching seen as desirable by
and on our reading of research on teaching and
education reformers in the United States – for
learning more broadly, we propose three distinct
example, the use of manipulatives, real-world
types of learning opportunities that are necessary
problem scenarios, and group work – were
to produce high levels of learning in mathematics
found in some, but not all, of the high-achieving
and, we believe, in other subjects as well:
countries. Almost everything coded in the TIMSS
video studies varied among the high-achieving • Productive struggle – This can be simplified to
countries (Stigler & Hiebert, 2004). the aphorism no pain no gain: deep learning
We interpret this finding, again, as further evi- requires some element of struggle. Despite the
dence for the cultural and contextual nature of romantic view of learning as ideally fun and
teaching. In Japan, a teacher can pose a difficult enjoyable, students learn more when they are
problem at the start of a lesson, and students will engaged in hard intellectual work which, though
immediately set to work on it, even if they find it rewarding in the end, is not necessarily enjoy-
frustrating and uncomfortable. If an American able in the moment (see, for example, Bjork &
teacher adopted this same practice the results Bjork, 2011).
might be quite different. In a project aimed at • Explicit connections – Evidence from the dis-
helping American teachers to increase student covery learning literature suggests that just
discussion in mathematics lessons (Wang, struggling with core content will not necessa-
Miller, & Cortina, 2013), we found that helping rily lead to learning (Kirschner, Sweller, &
teachers develop skills at leading mathematical Clark, 2006). It is also important to help stu-
discussions and giving them daily feedback was dents make explicit connections between the
not sufficient to promote change. We had to problems they are working on and the concepts
include some “professional development” for stu- they need to understand. The source of these
dents as well, because they did not know how to connections (individual discovery, listening to
listen to and engage with each other’s mathema- peers, or teacher summaries) may matter less
tical ideas. Teaching expertise exists within a than whether or not they are explicitly made.
cultural matrix, which means that expert teachers • Deliberate practice – Finally, struggle and con-
in different cultures may act very differently. nections need to be sustained over time,
through practice; and not by repetitive practice learning have proven less than fruitful, there is
but by deliberate practice, as demonstrated in another reason: learning, especially the kind of
the literature on expertise, with the opportunity deeper learning called for in current education
to vary strategies and get informative feedback. standards, takes time. Unlike chess, where the
final outcome can be known in the space of an
Whereas the most common paradigm for
hour or so, student learning of core content stan-
researching the effects of teaching on learning
dards often takes place over months and years.
looks for the simple effect of teacher actions/
Experts need to orient their actions simulta-
behaviors on student outcomes, we propose that a
neously toward long-term learning outcomes
slightly more complex model will be required:
and to more immediate indicators of learning.
teacher actions and behaviors create learning
These more immediate indicators, we believe,
opportunities for students, which, in turn, produce
are best found in a theoretically motivated con-
student outcomes. Under this model, expert tea-
struct such as learning opportunities – similar to
chers are not defined as those who employ a set of
what Lipsey (1993) has called “small theories.”
best practices, but instead those who (1) have the
ability to assess students’ current knowledge state
Expertise in the Classroom
both prior to and during instruction, (2) formulate
clear learning goals, (3) consider a large number of What does it take for teachers to create these
strategies and routines in their repertoire, (4) make learning opportunities in the classroom, espe-
good judgments about which strategies are most cially given the complex nature of teaching?
appropriate in any given situation, and (5) are able First, teachers must have clearly defined goals
to implement the strategies effectively to create for what they want students to learn, and the
learning opportunities for students. This model is ability to assess the gap between students’ current
represented in Figure 24.1. knowledge and where they are trying to help
Are these additional layers really necessary? students go next. Expert teaching is not just
We believe they are. Besides the fact that simple performing the acts of teaching. It is a highly
correlations of teaching actions with student contextualized endeavor in which teachers must
create the precise learning opportunities that will
move students to the next level of learning and
development. Doing this requires great skill at
formative assessment, which is one of the core
elements of expertise in teaching (Black &
Wiliam, 1998).
Based on these assessments, which must be
continually updated during instruction, teachers
then must create learning opportunities targeted
toward students’ current needs. Doing this will
require knowledge, skill, and judgment.
Knowledge. Much research has focused on what

expert teachers need to know in order to teach
effectively. Clearly they need to know the subject
that they are teaching. But this begs the question of
what we mean by “know.” A great teacher of high
Figure 24.1 A model of expertise in teaching. school physics does not need to know physics as
well as a professional physicist might know it. Many researchers have studied the relationship
In fact, the professional physicist might be a between pedagogical content knowledge and stu-
below-average teacher of physics to high school dent learning, but as with research on teacher
students, or even to university students (see practice, the correlations have been disappoint-
Feldon, 2007). Knowledge for teaching is not the ingly low (Hill, Rowan, & Ball, 2005). One reason
same as knowledge for some other purpose. for these low correlations must certainly be the
What do teachers need to know? Shulman complex systemic nature of teaching; with so
(1987) coined the term pedagogical content many variables in the mix, the correlation of any
knowledge to refer to the special knowledge tea- single one would be expected to be low. Another
chers need in order to teach effectively. However, reason may be that the most common measures of
pedagogical content knowledge is just the tip of pedagogical content knowledge consist of paper
the iceberg. Shulman proposed a taxonomy of and pencil multiple-choice items, which may be
teacher knowledge that still proves useful today measuring inert knowledge (Bransford, Goldman,
(Shulman, 1987, p. 8). It includes: & Vye, 1991; Whitehead, 1929). As with the case
of teaching practices, the key point may not be
• content knowledge;
whether or not you know something, but whether
• general pedagogical knowledge, with special
you are able to access and apply the knowledge
reference to those broad principles and strat-
when you need it to improve students’ learning
egies of classroom management and organiza-
opportunities.
tion that appear to transcend subject matter;
Kersting and colleagues have developed an
• curriculum knowledge, with particular grasp of
innovative measure of teacher knowledge that
the materials and programs that serve as “tools
they view as a measure of “usable” knowledge
of the trade” for teachers;
for teaching. This measure, called the Classroom
• pedagogical content knowledge, that special
Video Analysis (or CVA) assessment, requires
amalgam of content and pedagogy that is
teachers to view video clips of authentic classroom
uniquely the province of teachers, their own
episodes and then comment on what they see in the
special form of professional understanding;
video in terms of interactions among the teacher,
• knowledge of learners and their characteristics;
students, and content being taught. Coding of tea-
• knowledge of educational contexts, ranging
chers’ open responses yields a measure of teacher
from the workings of the group or classroom,
knowledge that correlates with Hill and Ball’s
the governance and financing of school districts,
MKT measures (Kersting, Sherin, & Stigler,
to the character of communities and cultures; and
2014). More important, however: the CVA mea-
• knowledge of educational ends, purposes, and
sures have been shown to be far better predictors
values, and their philosophical and historical
of student learning than have the paper and pencil
grounds.
MKT measures (Kersting, Givvin, Thompson,
Much of the research that has been inspired by Santagata, & Stigler, 2012; Kersting et al., 2016).
Shulman’s taxonomy has focused on pedagogical
content knowledge. Hill, Ball, and colleagues Skill. Despite the importance of teacher knowledge,
(Ball et al., 2006; Ball, Thames, & Phelps, there is a performance component as well to teach-
2008; Hill & Ball, 2004) have invested consider- ing. It’s not enough to just know what to do, but you
able effort in further subdividing, and measuring, also need to be able to do it well in a variety of
pedagogical content knowledge, developing mea- situations. Most teacher education and professional
sures referred to as the Mathematics Knowledge development programs for teachers have focused on
for Teaching (MKT) scales. making teachers more knowledgeable; few actually
give teachers opportunity to practice the skills of rewarding, but having a teacher one does not
teaching. But some recent work has recognized the want to let down can keep students engaged
importance of skilled implementation for creating until they develop enough skill for the activity
learning opportunities in classrooms. itself to be rewarding.
One such project is the one led by Lampert, This is an aspect of teaching that is often
Franke, and Kazemi (Lampert et al., 2013). These downplayed in US discussions. We interviewed
researchers have identified a set of instructional elementary school teachers in China and the
routines that are within the capabilities of pre- United States about their ideas of what contrib-
service teachers, yet still complex enough to uted to their students’ success in learning mathe-
represent the work of teaching. Such routines matics (Correa, Perry, Sims, Miller, & Fang,
(which they refer to as activities) include Choral 2008). Chinese teachers were much more likely
Counting, Launching and Using Word Problems, than their US counterparts to emphasize the
and so on. Participants in the program engage importance of their relationship with their
in “rehearsal” of these instructional routines students (as one teacher put it, “First of all . . .
through a process that very much resembles I think I’ll let the students love me first. The
deliberate practice. students need to love their teacher before they
Participants in the program begin by planning love the subject, so I should develop a good
mini-lessons that incorporate one of the instruc- relationship with them”).
tional routines, and then implement the lesson
with their peers. After getting feedback from Judgment. Finally, just knowing that and know-
peers, they implement the lesson in a real class- ing how are not enough. Because teaching is
room and record the result on video. They later highly contextual and complex, teachers must
take these videos back to the university and ana- also have the ability to decide which of many
lyze their implementation, getting feedback from possible strategies they should pursue in any
experts and peers. By using such routines as a site given time and place. They need to be able to
for deliberate practice, teachers in training are size up a situation, decide which strategy to
able to integrate the skills of teaching with the employ, and then adapt it to achieve their specific
concepts and knowledge they are being taught in instructional goal. In other words, teachers need
their teacher education program. The goal is to judgment. They can be highly knowledgeable in
develop adaptive knowledge of the routines – the all the ways Shulman (1987) describes, and
ability to implement them effectively to achieve highly skilled at implementing a wide variety of
instructional goals in a range of contexts. instructional strategies. But unless they make
Instructional routines are not the only aspects good decisions about when and how to employ
of teaching that require skilled implementation. their knowledge and skill, their knowledge may
The skills of managing emotional connections not serve to support students’ learning.
with students are perhaps equally critical, espe- Expert teachers’ judgments must be based, in
cially if we expect students to engage in the hard part, on an expert reading of the situation, some-
work of productive struggle. Csikszentmihalyi, thing that has been studied across a wide range of
Rathunde, and Whalen (1997), in their influential domains of expertise. Experts in fields from chess
study of talented teenagers, identified the rela- to radiology to electronic circuit design all appear
tionship that students had with their early tea- highly skilled at extracting structure from the
chers or coaches as an important theme. world as it relates to their domains of expertise
Activities such as playing a violin or running (Hoffman, 1998; Kellman & Garrigan, 2009).
long distances are not initially intrinsically Sometimes referred to as “situation awareness”
(Endsley, 1995), experts are able to quickly How Teachers Become Experts
recognize the important features of complex
If there is a lesson to be learned from the huge
situations and determine their consequences,
literature on teacher professional development it
whereas those who are less expert are more likely
is this: professional development, for the most
to miss key features, or spot them more slowly.
part, doesn’t work. But then, most of what counts
These kinds of effects have also been found for
as professional development consists of hapha-
teachers. Sabers, Cushing, and Berliner (1991)
zard, voluntary, and brief workshops that are dis-
found that experienced teachers were better than
connected from the daily work of teaching
novices at shifting their attention among multiple
(Birman et al., 2007). Such activities may suc-
views of a classroom and identifying important
ceed in making teachers more knowledgeable,
events. In our lab, we have used mobile eye
but because teaching is a complex cultural sys-
tracking methods to compare experienced and
tem, training just the teacher, in a time and place
novice teachers teaching the same subjects to
divorced from the ecology and culture in which
the same students (Miller, 2011). Experienced
they operate, is highly unlikely to improve the
teachers are more focused in their attention to
performance of the system as a whole.
the important aspects of the classroom environ-
Indeed, a small number of carefully designed
ment, such as students and curriculum materials,
experimental studies show that when professional
but some of the most interesting findings involve
development is intensive, ongoing, and job-
some of the trade-offs that beginning teachers
embedded, focused on students’ learning of the
need to make to tame the complexity of having
curriculum being taught, and aligned with the
to attend to a classroom of students at the same
school’s improvement goals and priorities, such
time they are trying to present a coherent lesson.
programs can produce significant improvements
A particular trade-off (Cortina, Miller,
in student learning (Cohen & Hill, 2000; Darling-
McKenzie, & Epstein, 2015) involves the balance
Hammond, Wei, Andree, Richardson, & Orphanos,
of attention to individual students versus to the
2009; Saunders, Goldenberg, & Gallimore, 2009;
class as a whole. Beginning teachers who are
Yoon, Duncan, Lee, Scarloss, & Shapley, 2007).
rated as being highly responsive to individual
In other words, when other critical elements of the
students do not do a good job of dividing their
system of teaching are taken into account, and
attention evenly among the students as a whole,
when teachers are given time to work on improving
as shown by eye tracking data. Experienced
implementation within the context of the system in
teachers do not show this trade-off, with some
which they are working, teaching will improve.
teachers managing to be both responsive to indi-
What does not work is training teachers apart
vidual students and attentive to the class as a
from the setting in which they are expected to
whole. Situation awareness in teaching requires
practice.
developing the kind of quick categorization of
There is a large literature on how to improve
what’s important in a classroom situation, as mea-
complex systems that include human actors, in
sured by the Sabers et al. (1991) task and the
fields ranging from automobile manufacturing to
measures developed by Kersting and colleagues
healthcare (e.g. Langley et al., 2009; Rother,
(described above). To be an expert teacher, one
2009). We also know quite a bit about what it
must be able to perceive structure in educational
takes to change cultural routines (e.g. Gallimore,
contexts, then link that structure to underlying
1996; Feldman & Pentland, 2003), as well as what
concepts and principles of teaching and learning
it takes to develop expertise across a wide variety
and to the repertoire of strategies and routines that
of domains (Ericsson, Krampe, & Tesch-Römer,
might be used to achieve the immediate goal.
1993; Hatano & Inagaki, 1986). Rarely, until With a clear understanding of the system as it
recently, however, have these strategies and meth- currently works (what Rother, 2009, calls the
ods been applied to improving teaching (Bryk, “current condition”), the next step is to establish
Gomez, Grunow, and LeMahieu, 2015). In the a “target condition,” that is, what you want to see
next part of the chapter we hope to make some of working differently (or more expertly) than it is
these connections. now. Then, obstacles are identified – what is
keeping you from reaching the target condition –
and a next step is identified: what will you work
Teachers versus Teaching as the Focus
on next to move the system incrementally closer
of Improvement
to the next target condition? In systems improve-
Feldman and Pentland (2003), in their work on ment approaches such as the one described by
organizational routines, made a useful distinc- Rother, this sets off a series of PDSA (Plan/Do/
tion. Routines, they write, consist of two related Study/Act) cycles, a process in which a change
parts: “One part embodies the abstract idea of idea is planned and implemented, then analyzed
the routine (structure), while the other part to see what is learned.
consists of the actual performances of the routine Importantly, developing the expertise of indivi-
by specific people, at specific times, in specific dual actors is not usually a goal of systems improve-
places (agency)” (p. 95). ment. The goal, in fact, is usually just the opposite:
Teaching can be thought of in a similar way. to create a system that performs reliably at a high
The structure consists of the cultural routines, and level without relying on high levels of expertise. If
the system in which the routines have evolved at all possible, the system should be designed so that
and function. Agency consists of the way the all human actors, within a normal range of skill, can
routines are implemented in different classrooms function effectively. According to Gawande (2007),
by different teachers. Improving teaching can this is the reason the field of obstetrics eventually
thus be accomplished in two distinct ways: gave up forceps in favor of the Caesarean section as
improving the system itself, including the rou- a way to deliver babies in trouble: almost any
tines, and improving the expertise of the teacher physician could successfully execute a Caesarean
as she implements the cultural routines that she section, whereas only those with high levels of
has inherited. expertise could use forceps effectively. When used
As it turns out, the processes of improving the expertly, forceps produce better results than a
performance of complex systems, changing cul- Caesarean. But when used by less skilled doctors,
tural routines, and developing expertise have a lot the results of a forceps delivery can be disastrous,
in common. In all cases, awareness is a critical and all obstetricians must begin as novices.
part of the process – seeing clearly what the The advantage of improving the system of
current system looks like. Improvement scientists teaching (as opposed to only the expertise of
(cf. Langley et al., 2009; Rother, 2009) talk about teachers) is obvious: changes in the system, even
learning to “see the system,” which is never easy, if small and incremental, can lead to long-term,
but it is even more difficult when the system large-scale improvements in student outcomes.
includes cultural routines. Building awareness Teachers come and go; most actually stay in the
of current routines (or skills in the case of exper- profession for only a few years (Simon & Johnson,
tise) is often better accomplished in collaboration 2015; Ingersoll & May, 2012). But system
with others – peers, team members, stakeholders, improvements that can be captured and accumu-
or even just a coach – because it requires that the lated over time have a far longer lifespan. Thus, it
routines be described explicitly – put into words. makes sense to shift at least part of our focus from
teachers to teaching. But how can we accumulate usually requires a set of training tasks that are
system improvements so that they can be shared outside the current comfort range of the learner,
with each new generation of teachers? but not so far that they can’t be mastered in a
Morris and Hiebert (2011) put forth an matter of hours.
interesting proposal, that instructional objects According to Ericsson (2006), deliberate prac-
(e.g. curriculum materials, lesson plans, etc.) be tice always involves conscious concentration on
seen not only as important components of an the skill, and informative feedback, both from the
instructional system, but as a mechanism for stor- performance itself and from a coach or peers.
ing professional knowledge. In their vision, educa- This requirement for concentration “sets deliber-
tors will work together to continuously improve the ate practice apart from both mindless, routine
system of teaching, capturing what they learn, as performance and playful engagement, as the lat-
much as possible, in improvements to the actual ter two types of activities would, if anything,
instructional materials themselves. The shared merely strengthen the current mediating cogni-
instructional objects thus become a sort of knowl- tive mechanisms rather than modify them to
edge base to guide teaching. allow increases in the level of performance”
Yet, despite the importance of improving the (Ericsson, 2006, p. 694). Hatano and Inagaki
system of teaching – the curriculum, resources, (1986) argue that the connection of concepts
methods, routines, and so on – there will always and understanding to the practice of routines is
be a need, at the same time, to improve the exper- what makes the routines adaptive and able to be
tise of teachers. As pointed out earlier, teaching is applied effectively in novel situations.
more like driving to work than it is like shooting a There are a number of characteristics of teach-
rocket to the moon: successful teaching will ing that work against successful implementation
always include a large element of implementa- of deliberate practice. First among these might be
tion. Because teaching is so highly contextual, the cultural assumption that teaching is not some-
and because so much judgment is required, it thing generally subject to improvement. This
will always be necessary to develop individual belief might arise from a number of sources.
teachers’ expertise to effectively implement rou- Unlike other domains of expertise, the outcomes
tines for all students in all situations. of teaching (e.g. student learning) are affected by
Experience by itself will not lead to expertise. numerous factors, the teacher being only one.
Instead, opportunities for deliberate practice will Especially in environments where achievement
be required, to which we now turn. is low, it may be difficult to convince teachers
that the work they do can have a significant and
direct impact on outcomes. If teachers don’t see
Creating the Conditions for Deliberate
their own skill as something that needs to be
Practice of Teaching
improved, it will be hard to engage them in delib-
A large body of work has documented the critical erate practice.
role of deliberate practice in the development of No matter how motivated teachers are to
adaptive expertise. Deliberate practice is not to be improve, there are other things that make it hard
confused with repetitive practice. Anyone with for them to engage in deliberate practice of teach-
lots of experience has also gained a lot of repeti- ing. Chief among these is the fact that, unlike
tive practice. But deliberate practice is something many domains of expertise, teaching is one in
else. Deliberate practice is usually a designed which most of the teacher’s time is spent in per-
experience – often by a teacher or coach – not formance, not in practice. Fadde and Klein (2010)
one that happens naturally (Ericsson, 2006). It point out that in music and sports, for example,
performance happens rarely, leaving lots of time Lewis, Perry, & Murata, 2006; Stigler & Hiebert,
for practice in between performances. But many 1999). In lesson study, teachers meet regularly in
professions, for example, management, primarily groups of four to six to work on planning, imple-
consist of performance on a daily basis. In these menting, and improving specific lessons from the
kinds of professions – and teaching is clearly one – curriculum. The goal of lesson study is not just to
Fadde and Klein (2010) propose that deliberate improve a particular lesson. Instead, the lesson
performance, not deliberate practice, is the most becomes a vehicle for working on an improve-
feasible route to expertise. ment goal that goes well beyond a single lesson.
Fadde and Klein (2010) define deliberate per- So, for example, a lesson study group might be
formance as “the effort to increase domain exper- working on deepening students’ understanding of
tise while engaged in routine work activity” a particularly challenging concept, or on improv-
(p. 6). Starting with features of deliberate practice ing their ability to elicit students’ thinking
(e.g. repetition, timely feedback, task variety, and through questioning.
progressive difficulty), they propose four types of Similar practices are at work in other Asian
exercises that professionals can engage in while countries, and the practice is also growing in
on the job. These exercises include estimation popularity in the United States. Lesson study is
(e.g. predicting how many students will use relevant here for at least two reasons. First, it
each of two different strategies for approaching appears to have been inspired by the early work
a math problem), experimentation, extrapolation, of Deming, a pioneer of improvement science
and explanation. Fadde and Klein also propose who, coincidentally, spent much of his career
that coaching is a much-needed resource for working in Japan (Gabor, 1990; Kenney, 2008).
anyone embarking on a path of deliberate As practiced in Japan, lesson study is first and
performance. Coaching would help to bring foremost a research and development process,
hidden routines to awareness, to make explana- much like the PDSA cycles of improvement
tions explicit, and to provide feedback in domains science (Langley et al., 2009). Second, lesson
where feedback is often difficult to come by. study appears to be aligned almost perfectly
Thus, it is difficult but not impossible to create with the requirements for deliberate practice. In
the conditions for deliberate practice of teaching. this sense it has the potential to greatly accelerate
Our challenge is to create settings in which tea- the development of expertise in teaching, as well
chers can practice the implementation of a teach- as the system of teaching itself.
ing routine or strategy; see the concrete results of The lesson functions similarly to the designed
their actions in evidence of students’ learning; practice tasks described by Ericsson (2006). In
reflect on the cause–effect relationships between lesson study, the practice of teaching is slowed
their teaching, the learning opportunities they down, often for weeks, as teachers discuss, first,
create for students, and the evidence of what what they want students to take from the lesson,
students learn; and get feedback from knowledge- and then what kinds of evidence they can collect
able others. We turn now to two examples of to indicate students’ thinking and learning during
where this has happened. the lesson. The group then plans, in great detail, a
lesson that they hypothesize will achieve the
goals they have set, and develops explicit hypoth-
Lesson Study as a Lab for Deliberate
eses and predictions for both the processes and
Practice
outcomes they will observe. Once they have cre-
Much has been written about the Japanese prac- ated a detailed plan – a process that often takes
tice of lesson study (Lewis, Perry, & Hurd, 2004; place over several weeks – one of the teachers in
the group will teach the lesson while the others Analysis as the Key to Developing
observe, clipboards in hand, collecting evidence Expertise in Teaching
that then will be used to revise and improve sub-
Even if lesson study can provide a setting for
sequent versions of the lesson.
deliberate practice of teaching, it does not appear
Debriefing sessions follow the implementation
to provide enough opportunities for such practice
of the lesson. In these sessions, teachers share the
to produce high levels of expertise over time. But
evidence they have collected, and analyze its sig-
what if the main things being practiced during
nificance for both understanding and improving
lesson study are not the skills of teaching gener-
the learning process that unfolds during the les-
ally, but the specific skills teachers would need to
son. They focus special attention on places where
learn from their own experience? In other words,
their predictions were wrong, and seek to explain
perhaps lesson study is providing teachers with
these deviations with cause–effect hypotheses.
skills that can transform their daily work in the
For example, they may observe some subset of
classroom into further opportunities for deliber-
students whose work indicates a misconception
ate practice – something similar to Fadde and
of the concept being taught. They then may think
Klein’s (2010) theory of how deliberate perfor-
carefully about the task that was assigned, and
mance might yield growth in expertise.
even the specific questions the teacher asked as
A primary candidate for what these skills might
the work progressed, seeking to explain how the
be are analytic skills: teachers who participate in
misconception could have emerged, and to find
lesson study are learning to observe and analyze
an incremental improvement that they can test in
practice in ways that could directly improve their
the next version of the lesson, which generally
ability to read a classroom situation, select an
will be taught by a different member of the group.
appropriate strategy, and implement the strategy
Through all of this the focus is kept on the lesson
effectively, all while monitoring evidence of stu-
that the group has designed, not on the individual
dents’ thinking and learning. Specifically, teachers
teacher who taught it, a strategy that helps to
in lesson study are practicing:
mitigate any defensiveness teachers may feel
that could slow down their own learning. • Careful observation and analysis of students’
Ericsson has noted that premature automaticity thinking and learning during a classroom lesson.
may well be one of the biggest enemies to the • Generation of cause–effect theories (what
development of expertise. Teaching, because of Lipsey, 1993, calls “small theories”) that link
its heavy reliance on cultural routines, may be a teacher’s actions to students’ thinking and
even more subject to this barrier to improvement learning during the lesson, and that may explain
than are other domains of expertise. Lesson study failures in learning.
provides a means of disrupting the normal “good • Generation of alternative teaching strategies
enough” routines of teaching. By spending weeks that may, if one’s theory is correct, lead to
or months on a single lesson, each and every part improved outcomes for students.
of the lesson – both the parts that are planned • Testing of alternative strategies, and using what
ahead of time and the parts that are adapted at the is learned to revise one’s own theories of class-
time of implementation – is subject to deep ana- room teaching and learning.
lysis, revision, and practice, all in a context in
which feedback is readily available. Lesson study If these specific skills can be applied by
is thus a context or lab in which deliberate prac- teachers to their own teaching, then daily
tice can take place. experiences in the classroom can become a
site for deliberate practice (or deliberate per- Conclusion

formance) and thus a possible mechanism for
Teaching is a complex socio-cultural system,
the continuous improvement of teaching over
both hard to see and hard to change. Improving
time. Teachers become more aware of their
teaching requires that both the routines of teach-
own teaching routines, and then develop and
ing and the expertise of the teacher be improved.
test changes in routines that might better
Although the aim of improving teaching routines
address the needs of their students. Through
is to enable an average teacher to produce the
iterative application of these skills, teachers
desired outcomes, the nature of teaching will
will become better at reading situations,
always necessitate a certain level of adaptive
deciding which strategy to use, and adapting
expertise on the part of the teacher. Teachers
the strategy to meet their specific instructional
must have knowledge (knowing that); they must
goals.
have skill (knowing how); and they must have
Evidence supporting this theory comes from
judgment – the ability to size up a situation, see its
work, reviewed above, by Kersting and col-
structure the way an expert physicist sees the
leagues in which teachers’ analyses of class-
structure of a physics problem, and then bring
room video clips are used as indicators of
the right knowledge to bear so as to achieve the
pedagogical content knowledge (Kersting,
instructional goals.
Givvin, Sotelo, & Stigler, 2010). In this sim-
The models of teaching and of the development
ple paradigm, teachers are asked to watch a
of expertise in teaching proposed here share the
series of short video clips depicting authentic
key feature that both require cycles involving ana-
classroom events, and are then asked to com-
lysis of a problem, planned activity, and assess-
ment on the interactions among the teacher,
ment of the results that then informs future
students, and mathematics. Teachers’ analytic
practice. This approach has proven successful in
skills, measured in this way, have been shown
improving performance in a number of important
to significantly predict students’ learning from
domains, and we believe that it, rather than the
pre- to post-test in a variety of mathematical
promulgation of a decontextualized set of best
domains, and the effect has been shown to be
practices, holds the most promise for improving
mediated by an observational measure of
the quality of teaching and learning.
instructional quality (Kersting et al., 2012).
Perhaps because of the complexity and cultu-
Although the relationship is correlational – i.e.
rally nested nature of teaching, the fields of tea-
nothing was manipulated in the study – other
cher education and teacher professional
research lends further support to the idea that
development have lagged in bringing principles
teachers’ analytic skills directly lead to improve-
of systems improvement and development of
ments in instructional quality and student learn-
expertise to the task of improving teaching.
ing. Roth et al. (2011), for example, showed that
Teaching presents a challenge for models of
teachers who participated in a year-long profes-
expertise – it is difficult to identify experts and
sional development program focused on the ana-
the realities of teaching make it difficult to engage
lysis of lesson videos (project STeLLa: Science
in deliberate practice. The concept of deliberate
Teachers Learning from Lesson Analysis) pro-
performance and work on improving perfor-
duced high quality instruction and higher levels
mance in other complex fields suggest that there
of student learning than did a group of compar-
are promising ideas from other fields that can be
ison teachers. And analysis of practice has
usefully applied to understanding and improving
become a key component of many teacher learn-
the development of teachers.
ing programs.
References Reflections on a decade of work and some hopes

for the future. In R. J. Sternberg & L. Okagaki
Ball, D. L., Bass, H., Hill, H., Sleep, L., Phelps, G., (eds.), Influences on children (pp. 147–80).
& Thames, M. (2006). Knowing and using Hillsdale, NJ: Erlbaum.
mathematics in teaching. Paper presented at the Bryk, A. S., Gomez, L. M., Grunow, A., & LeMahieu,
Learning Network Conference, Washington, P. G. (2015). Learning to improve: How America’s
DC, January. schools can get better at getting better.
Ball, D. L., & Forzani, F. M. (2007). 2007 Wallace Cambridge, MA: Harvard Education Press.
Foundation Distinguished Lecture – What makes Campbell, D. T. (1979). Assessing the impact of
education research “educational”? Educational planned social change. Evaluation and Program
Researcher, 36, 529–540. Planning, 2, 67–90.
Ball, D. L., Thames, M. H., & Phelps, G. (2008). Cavalluzzo, L. C. (2004). Is National Board
Content knowledge for teaching: What makes certification an effective signal of teacher
it special? Journal of Teacher Education, 59, quality? Alexandria, VA: CNA Corporation
398–407. Working Paper.
Berliner, D. C. (1986). In pursuit of the expert Clotfelter, C. T., Ladd, H. F., & Vigdor, J. L. (2007).
pedagogue. Educational Researcher, 15, 5–13. Teacher credentials and student achievement:
Berliner, D. C. (1990). What’s all the fuss about Longitudinal analysis with student fixed effects.
instructional time? In M. Ben-Peretz & R. Economics of Education Review, 26, 673–682.
Bromme (eds.), The nature of time in schools: Cohen, D. (2011). Teaching and its predicaments.
Theoretical concepts, practitioner perceptions Cambridge, MA: Harvard University Press.
(pp. 3–35). New York: Teachers College. Cohen, D., & Hill, H. (2000). Instructional policy
Birman, B. F., Le Floch, K. C., Klekotka, A., Ludwig, and classroom performance: The mathematics
M., Taylor, J., Walters, K., . . . & O’Day, J. (2007). reform in California. The Teachers College
State and local implementation of the No Child Record, 102, 294–343.
Left Behind Act. Volume I, Title I: Teacher quality Correa, C. A., Perry, M., Sims, L., Miller, K. F., &
under NCLB: Interim report. US Department of Fang, G. (2008). Connected and culturally
Education (ERIC Document Reproduction embedded beliefs: Chinese and U.S. teachers
Service No. ED497970). talk about how their students best learn
Bjork, E. L., & Bjork, R. A. (2011). Making things hard mathematics. Teaching and Teacher Education,
on yourself, but in a good way: Creating desirable 24, 140–153.
difficulties to enhance learning. In M. A. Cortina, K. S., Miller, K. F., McKenzie, R., & Epstein,
Gernsbacher, R. W. Pew, L. M. Hough, & J. R. A. (2015). Where low and high inference data
Pomerantz (eds.), Psychology and the real world: converge: Validation of CLASS assessment of
Essays illustrating fundamental contributions to mathematics instruction using mobile eye tracking
society (pp. 56–64). New York: Worth. with teachers. International Journal of Science
Black, P., & Wiliam, D. (1998). Inside the black box: and Mathematics Education, 13, 389–403.
Raising standards through classroom assessment. Csikszentmihalyi, M., Rathunde, K., & Whalen, S.
Phi Delta Kappan, 80, 139–144. (1997). Talented teenagers: The roots of success
Bond, L., Smith, T., Baker, W., & Hattie, J. A. (2000). and failure. Cambridge University Press.
The certification system of the National Board for Cuban, L. (1990). Reforming again, again, and again.
Professional Teaching Standards: A construct and Educational Researcher, 19, 3–13.
consequential validity study. Greensboro, NC: Darling-Hammond, L., Wei, R. C., Andree, A.,
Center for Educational Research and Evaluation, Richardson, N., & Orphanos, S. (2009).
The University of North Carolina at Greensboro. Professional learning in the learning profession.
Bransford, J. D., Goldman, S. R., & Vye, N. J. (1991). Washington, DC: National Staff Development
Making a difference in people’s ability to think: Council.
de Groot, A. D. (1965). Thought and choice in chess. Gabor, A. (1990). The man who discovered quality:
The Hague: Mouton. How W. Edwards Deming brought the quality
de Groot, A. D. (1966). Perception and memory versus revolution to America. New York: Penguin Books.
thought: Some old ideas and recent findings. In B. Gallimore, R. (1996). Classrooms are just another
Kleinmuntz (ed.), Problem solving. New York: cultural activity. In D. Speece & B. Keogh (eds.),
John Wiley. Research on classroom ecologies: Implications for
D’Mello, S., Lehman, B., Pekrun, R., & Graesser, children with learning disability (pp. 229–250).
A. (2014). Confusion can be beneficial Mahwah, NJ: Erlbaum.
for learning. Learning and Instruction, 29, Gallimore, R., & Tharp, R. (2004). What a coach can
153–170. teach a teacher, 1975–2004: Reflections and
Endsley, M. R. (1995). Toward a theory of situation reanalysis of John Wooden’s teaching practices.
awareness in dynamic systems. Human Factors, The Sport Psychologist, 18, 119–137.
37, 32–64. Gawande, A. (2007). Better: A surgeon’s notes on
Ericsson, K. A. (2006). The influence of experience and performance. New York: Metropolitan Books.
deliberate practice on the development of superior Goldhaber, D., & Anthony, E. (2007). Can teacher
expert performance. In K. A. Ericsson, N. quality be effectively assessed? National board
Charness, R. R. Hoffman, & P. J. Feltovich (eds.), certification as a signal of effective teaching.
The Cambridge handbook of expertise and expert Review of Economics and Statistics, 89, 134–150.
performance (pp. 685–706). Cambridge Hanushek, E. A. (2003). The failure of input-based
University Press. schooling policies. Economic Journal, 113,
Ericsson, K. A. (2008). Deliberate practice and F64–F98.
acquisition of expert performance: A general Harris, D. N., & Sass, T. R. (2009). The effects of
overview. Academic Emergency Medicine: NBPTS-certified teachers on student
Official Journal of the Society for Academic achievement. Journal of Policy Analysis and
Emergency Medicine, 15, 988–994. Management, 28, 55–80.
Ericsson, K. A., Krampe, R. T., & Tesch-Römer, C. Hatano, G., & Inagaki, K. (1986). Two courses of
(1993). The role of deliberate practice in the expertise. In H. Stevenson, H. Azuma, & K.
acquisition of expert performance. Psychological Hakuta (eds.), Child development and education
Review, 100, 363–406. in Japan (pp. 262–272). New York: W. H.
Fadde, P. J., & Klein, G. A. (2010). Deliberate Freeman.
performance: Accelerating expertise in natural Hill, H. C., & Ball, D. L. (2004). Learning mathematics
settings. Performance Improvement, 49, 5–14. for teaching: Results from California’s
Feldman, M. S., & Pentland, B. T. (2003). mathematics professional development institutes.
Reconceptualizing organizational routines as a Journal for Research in Mathematics Education,
source of flexibility and change. Administrative 35, 330–351.
Science Quarterly, 48, 94–118. Hill, H. C., Rowan, B., & Ball, D. L. (2005). Effects of
Feldon, D. F. (2007). The implications of research on teachers’ mathematical knowledge for teaching on
expertise for curriculum and pedagogy. student achievement. American Educational
Educational Psychology Review, 19, 91–110. Research Journal, 42, 371–406.
Fisher, C. W., Berliner, D. C., Fully, N. N., Marliave, Hoetker, J., & Ahlbrand Jr, W. P. (1969). The
R. S., Cahen, L. S., & Dishaw, M. M. (1980). persistence of the recitation. American
Teaching behaviors, academic learning time and Educational Research Journal, 6, 145–167.
student achievement: An overview. In C. Denham Hoffman, R. R. (1998). How can expertise be defined?
& A. Lieberman (eds.), Time to learn (pp. 7–32). Implications of research from cognitive
Washington, DC: US Department. of Education, psychology. In R. Williams, W. Faulkner, & J.
National Institute of Education, Program on Fleck (eds.), Exploring expertise (pp. 81–100).
Teaching and Learning. New York: Macmillan.
Hogan, T., & Rabinowitz, M. (2009). Teacher expertise Kersting, N. B., Sutton, T., Kalinec-Craig, C., Stoehr,
and the development of a problem representation. K. J., Heshmati, S., Lozano, G., & Stigler, J. W.
Educational Psychology, 29, 153–169. (2016). Further exploration of the classroom video
Ingersoll, R. M., & May, H. (2012). The magnitude, analysis (CVA) instrument as a measure of usable
destinations, and determinants of mathematics and knowledge for teaching mathematics: Taking a
science teacher turnover. Educational Evaluation knowledge system perspective. ZDM, 48, 97–109.
and Policy Analysis, 34, 435–464. Kirschner, P. A., Sweller, J., & Clark, R. E. (2006).
Jackson, C. K. (2012). Recruiting, retaining, and Why minimal guidance during instruction does
creating quality teachers. Nordic Economic Policy not work: An analysis of the failure of
Review, 1, 61–105. constructivist, discovery, problem-based,
Jackson, C. K., Rockoff, J. E., & Staiger, D. O. (2014). experiential, and inquiry-based teaching.
Teacher effects and teacher-related policies. Educational Psychologist, 41, 75–86.
Annual Review of Economics, 6, 801–825. Lampert, M. (2003). Teaching problems and the
Kane, T. J., McCaffrey, D. F., Miller, T., & Staiger, D. O. problems of teaching. New Haven, CT: Yale
(2013). Have we identified effective teachers? University Press.
Validating measures of effective teaching using Lampert, M., Franke, M. L., Kazemi, E., Ghousseini,
random assignment. Seattle, WA: MET Project of H., Turrou, A. C., Beasley, H., . . . & Crowe, K.
the Bill and Melinda Gates Foundation. (2013). Keeping it complex using rehearsals
Kane, T. J., & Staiger, D. O. (2012). Gathering to support novice teacher learning of ambitious
feedback for teaching: Combining high-quality teaching. Journal of Teacher Education,
observations with student surveys and 64, 226–243.
achievement gains. Seattle, WA: MET Project of Langley, G. J., Moen, R., Nolan, K. M., Nolan, T. W.,
the Bill and Melinda Gates Foundation. Norman, C. L., & Provost, L. P. (2009). The
Kellman, P. J., & Garrigan, P. (2009). Perceptual improvement guide: A practical approach to
learning and human expertise. Physics of Life enhancing organizational performance (2nd
Reviews, 6, 53–84. edn.). San Francisco: Jossey-Bass.
Kenney, C. (2008). The best practice: How the new Lewis, C., Perry, R., & Hurd, J. (2004). A deeper
quality movement is transforming medicine. New look at lesson study. Educational Leadership,
York: Public Affairs. 61, 19–22.
Kersting, N. B., Givvin, K. B., Sotelo, F. L., & Stigler, Lewis, C., Perry, R., & Murata, A. (2006). How should
J. W. (2010). Teachers’ analyses of classroom research contribute to instructional improvement?
video predict student learning of mathematics: The case of lesson study. Educational Researcher,
Further explorations of a novel measure of teacher 35, 3–14.
knowledge. Journal of Teacher Education, 61, Lewis, C., & Tsuchida, I. (1998). A lesson is like a
172–181. swiftly flowing river. American Educator, 22,
Kersting, N. B., Givvin, K. B., Thompson, B. J., 12–17.
Santagata, R., & Stigler, J. W. (2012). Measuring Lipsey, M. W. (1993). Theory as method: Small
usable knowledge: Teachers’ analyses of theories of treatments. New Directions for
mathematics classroom videos predict teaching Evaluation, 57, 5–38.
quality and student learning. American Lortie, D. C. (1975). Schoolteacher: A sociological
Educational Research Journal, 49, 568–589. study. University of Chicago Press.
Kersting, N. B., Sherin, B. L., & Stigler, J. W. (2014). Miller, K. F. (2011). Learning from the experience of
Automated scoring of teachers’ open-ended others: What education can learn from video-based
responses to video prompts bringing the research in other fields. In M. Sherin, V. Jacobs, &
classroom-video-analysis assessment to scale. R. Phillip (eds.), Mathematics teacher noticing:
Educational and Psychological Measurement, 74, Seeing through teachers’ eyes (pp. 51–65). New
950–974. York: Routledge.
Morris, A. K., & Hiebert, J. (2011). Creating shared Shulman, L. (1987). Knowledge and teaching:
instructional products: An alternative approach Foundations of the new reform. Harvard
to improving teaching. Educational Researcher, Educational Review, 57, 1–23.
40, 5–14. Simon, N. S., & Johnson, S. M. (2015). Teacher
Mueller, F., & Dyerson, R. (1999). Expert humans or turnover in high-poverty schools: What we
expert organizations? Organization Studies, 20, know and can do. Teachers College Record
225–256. 117, 1–36.
Nye, B., Konstantopoulos, S., & Hedges, L. V. (2004). Skemp, R. R. (1987). The psychology of learning
How large are teacher effects? Educational mathematics. London: Routledge.
Evaluation and Policy Analysis, 26, 237–257. Stigler, J. W., & Hiebert, J. (1999). The teaching gap:
Resnick, D., & Resnick, L. (1977). The nature of Best ideas from the world’s teachers for improving
literacy: An historical exploration. Harvard education in the classroom. New York: Simon &
Educational Review, 47, 370–385. Schuster.
Rivkin, S. G., Hanushek, E. A., & Kain, J. F. (2005). Stigler, J. W., & Hiebert, J. (2004). Improving
Teachers, schools, and academic achievement. mathematics teaching. Educational Leadership,
Econometrica, 73, 417–458. 61, 12–17.
Rockoff, J. E. (2004). The impact of individual teachers Thorndike, E. L. (1906). Principles of learning. New
on student achievement: Evidence from panel York: A. G. Seiler.
data. American Economic Review, 94, 247–252. UNESCO Institute for Statistics (2015). A Teacher for
Roth, K. J., Garnier, H. E., Chen, C., Lemmens, M., Every Child: Projecting Global Teacher Needs
Schwille, K., & Wickler, N. I. (2011). Videobased from 2015 to 2030. October, Number 27, page 1.
lesson analysis: Effective science PD for teacher Montreal: UNESCO. Institute for Statistics. http://
and student learning. Journal of Research in uis.unesco.org/sites/default/files/documents/fs27-
Science Teaching, 48, 117–148. a-teacher-for-every-child-projecting-global-tea
Rother, M. (2009). Toyota Kata: Managing people for cher-needs-from-2015-to-2030-en.pdf.
improvement, adaptiveness and superior results. US Department of Education (2016). Digest of
Boston: McGraw-Hill. Education Statistics, 2014 (NCES 2016–006).
Sabers, D. S., Cushing, K. S., & Berliner, D. C. (1991). Wang, Z., Miller, K. F., & Cortina, K. S. (2013). Using
Differences among teachers in a task the LENA in teacher training: Promoting student
characterized by simultaneity, multidimensional, involvement through automated feedback.
and immediacy. American Educational Research Unterrichtswissenschaft, 4, 290–305.
Journal, 28, 63–88. Whitehead, A. N. (1929). The aims of education and
Saunders, W. M., Goldenberg, C. N., & Gallimore, R. other essays. New York: Macmillan.
(2009). Increasing achievement by focusing Williams, B. A. (1998). Thought and action: John
grade-level teams on improving classroom Dewey at the University of Michigan. Ann Arbor:
learning: A prospective, quasi-experimental study University of Michigan Bentley Historical Library
of Title I schools. American Educational Research Bulletin #44.
Journal, 46, 1006–1033. Yoon, K. S., Duncan, T., Lee, S. W. Y., Scarloss, B., &
Schwartz, D. L., Chase, C. C., Oppezzo, M. A., & Chin, Shapley, K. L. (2007). Reviewing the evidence on
D. B. (2011). Practicing versus inventing with how teacher professional development affects
contrasting cases: The effects of telling first on student achievement. Issues and Answers. REL
learning and transfer. Journal of Educational 2007-No. 033. Austin, TX: Regional Educational
Psychology, 103, 759–775. Laboratory Southwest.

Expertise in Teaching

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Expertise in Teaching

Uploaded by

Copyright:

Available Formats

24 Expertise and Expert Performance

Introduction domains of expertise. But in education, teachers

long-term outcomes such as career advancement The Nature of Teaching

within the normal limits of variation for their Expertise in Teaching

Knowledge. Much research has focused on what

site for deliberate practice (or deliberate per- Conclusion

References Reﬂections on a decade of work and some hopes

You might also like