IMPLEMENTING EFFECTIVE SCHOOL IMPROVEMENT POLICIES:
A SYSTEM DYNAMICS POLICY ANALYSIS
by Alan K. Gaynor
and Karl Hs Clauset, Jr.
Boston University
Background
At the last System Dynamice research conference held in the
United States, we presented a paper which described a computer
simulation model of an elementary school. The purpose of the
model was to examine the structural differences between schools
which are effective and ineffective for what we have come to call
“initially low-achieving children." In that paper (Clauset &
Gaynor, 1981), in a subsequent paper (Clauset & Gaynor, 1982),
and in a book manuscript (Clauset & Gaynor, in preparation), we
have described in varying degrees of detail tests which examined
a number of school improvement policies. Policies tested included
the following:
« Changing policies affecting time allocations
Improving teacher skills
Encouraging teachers to place more emphasis on low
achievers
Raising teacher expectations for low-achievers
Improving classroom or school-wide behavior
Changing class size
Changing the demographics of the student body (e-g+,
size, percent low achievers).
307
The central conclusion of the policy analysis was that there do
exist policies which can either erase or greatly reduce the
achievement gap for low-achievers. The most effective school
improvement strategies are those which better teacher skills,
raise teacher expectations for low-achieving students, and
maximize time available for instruction.
The work in which we are currently engaged goes beyond the
earlier work in that it focuses on the problems of implementing
policies examined earlier without regard to the demonstrated
difficulties of implementation. The earlier work drew on an
extensive review of the literature on effective teaching and
schooling and on the working knowledge of educational researchers
and practicing school administrators and teachers as a basis for
understanding the dynamics of effective and ineffective
schooling. The current work, in process, draws on the literature
about educational innovation, on empirical research into the
processes of school improvement, and, again, on the expertise of
knowledgeable actors to describe the dynamics of effective and
ineffective school improvement policie
We expect that the ultimate product of the current work will
be the marriage of our earlier model with additional sectors
currently under development which describe the implementation
dynamics. We hope to understand better how policies which seem to
work theoretically in the abstract founder in the rough seas of
practical implementation, Our focus, consistent with a
preliminary study of implementation begun several years ago
(Gaynor, 1979, 1980, 1981), {1s on the ways in which the
implementation of an educational innovation sets up response
patterns which tend toward a return to the status quo ante. We
are particularly interested in the ways in which alternative
implementation strategies engender differential systemic effects.
The paper in preparation, for which this brief description is
simply a placeholder, will, itself, constitute but a progress
report.
The School Effectiveness Model
Our initial work involved the construction of a circular
theory of causality that links student variables with
organizational and instructional variables. The causal-loop
diagrams which follow (Figures 1 and 2) highlight the essential
structural differences between schools that are effective and
ineffective for initially low-achieving children. [1]
[1] For a full description of the School Effectiveness Model
and a complete listing of model equations, see Clauset, 1982.
The firet diagram (Figure 1) illustrates the basic feedback
structure for an effective school.
GRADE LEVEL STANDARDS
+
acntzvenewt teacnen expectations
FoR ACKTEVENENT
4 + 7
v-
LEARNING BATE
s Pencerve LEARUING
fre car
. +
aerttupe vortvartow Dy
— rt
ra + dy enrnasts
ENGAGED TIME
Hear ae
AMD INTEVSITY
oF INstRUcTION
+ +
time ron txeaverton
+ +
. TEACHER EFFECTIVENESS
SCHOOL PoLICTES
FOR ALLOCATING TIME
The Basic Feedback Structure in an Effective
School.
Fig.
The causal relationship shown between achievement and
instruction is consistent with the BTES research in California on
achievement, learning rate, and academic learning time (Fisher,
et al., 1978). [2] Appropriateness and Intensity of Instruction
directly affects the amount of engaged time. It also affects
student motivation as children compare the instruction they
receive with that given to others and as they are affected
directly by instruction which is wore or less stimulating to
them. Motivation to learn is, in addition, influenced by the
child’s perception of the teacher’s expectations for him and by
his awareness of his achievement relative to grade level
standards (cf. Kolesnik, 1978).
The level of appropriateness and intensity of instruction for
a given achievement group is dependent on the amount of time for
instruction in the subject, the teacher’s effectiveness, and the
amount of emphasis the teacher places on the group. Time for
instruction is a function of: (1) school policies for allocating
(2] Partly for purposes of simplification, we chose in
constructing the model to assume equal "native learning
ability" for all children in the simulated elementary
school. Clearly, this is not likely to be precisely true;
however, a fundamental part of our thesis is that it is more
true than would appear by deduction from the variance in
actual achievement scores. An important purpose of our
modeling effort was to demonstrate theoretically that
results very similar to those obtained historically in real
schools (i.e., in terms of the divergence in achievement
scores) could be accounted for without assuming differences
in native ability. What was assumed in the model is that
students differed in their entry achievement (i.e.,
“learning readiness"). We offered no interpretation with
respect to the causes of those initial differences, neither
hereditary nor environmental. We have simply hypothesized,
on the basis of what we believe to be reasonably compelling
evidence, that many children are capable of learning far
more in school than they do and that learning is importantly
a direct effect of the appropriateness and intensity of
instruction (Bloom, 1976; Fisher, et al., 1978).
309
time among subjects and for allocating time between instructional
and non-instructional activities (assemblies, lunch, recess,
etc.) and (2) time the teacher must spend dealing with classroom
behavior problems. Time spent on behavior problems depends on the
level of student behavior in the class, the teacher’s
effectiveness, and the impact of the behavior of other students
in the school. Teacher effectiveness refers to both the teacher's
instructional and classroom management skills. Effectiveness is
mediated by class size, although more highly skilled teachers are
less affected by larger and more able to take advantage of
smaller class size.
The third component of appropriateness and intensity of
instruction is the amount of emphasis a teacher gives to a
particular achievement group. In heterogeneous classes, this
emphasis is a function of a teacher’s desired emphasis and the
competing demands of other groups. It {s central to our theory of
schooling that the perceived learning gap between teacher
expectations and the level of achievement is a major determinant
of teacher emphasis. A teacher will devote more emphasis to a
particular achievement group if the teacher perceives a gap in
achievement. If there is no gap between expectations and
achievement, there will be no effort to increase the emphasis for
a particular group. In the effective school, expectations are
based solely on standards and not on past achievement.
Consequently, there is a significant gap in achievement for the
initially low achievers and teachers want to place more emphasis
on these students to raise their achievement.
In systems terms, the basic driving force in the effective
school structure is a negative feedback loop which operates to
control the level of student achievement by adjusting the
appropriateness and intensity of instruction, The goal of this
control system is to close the discrepancy between teacher
expectations (which are based on fixed standards for all
children) and student achievement.
The school which is ineffective for initially low-achievers
is driven by a different causal configuration (Figure 2). This
difference may not be immediately apparent to the reader.
Actually, the two structures are identical except that teacher
expectations are no longer based on a fixed set of achievement
standards.
For low achieving students in the ineffective school,
teachers’ expectations respond directly to student achievement,
This is the essence of the ineffective school. Students who do
poorly are expected to do poorly. Thus, fromthe teacher's
perspective, there is no sense of a learning gap, no need to
alter the nature of instruction, no ownership of an instructional
problem, and no motivation to work at improving teaching skills.
In essence, "The kid can’t learn, or doesn’t want to. It’s his
problem, not mine."
The effect of directly linking teacher expectations and
310
student achievement in the ineffective school is to collapse the
negative feedback loop which operates in the effective school to
control the level of student achievement. What now dominates the
systen is a positive feedback loop which reinforces existing
achievement patterns. It works well enough for students who are
above average in achievement when they enter school. The system
works to teinforce their motivation, their behavior, and
ultimately their further achievement. However, for students who
enter school with poor readiness skills, this "multiplier system"
works to depress their motivation to learn, to reinforce
dysfunctional behavior patterns of "acting out" or withdrawal, to
reduce teacher emphasis on them, and to further diminish their
future achievement. Thus, instruction becomes less . appropriate
and intense in response to declining achievement.
9
ACHIEVEMENT TEACHER EXPECTATIONS +4
4 : FOR ACRIEVEMENT
3
LEARNING RATE +
rare: A rane Lani
servi — pee 4)
" 8
. teaenen EMPASTS
é
‘** BEHAVIOR +
G :
srreornzaternes
.
AND INTENSITY
a
‘TIME FOR INSTRUCTION z +
+ ie
TEACHER EFFECTIVENESS +
** scwoot porrcres
FOR ALLOcATING TIME
** ~ Potats of Iotervention
To Igprove Effectiveness
Fig. 2. The Basic Feedback Structure and Points of
Intervention in an Ineffective School.
The School Improvement
Policy Implementation Model
The purpose of the expanded model is to describe the
important structural connections between the system within which
school improvement policies are implemented and the system within
311
10
which students with differing entry characteristics learn in
schools. The depiction of the expanded School Improvement Policy
Implementation Model is intended to provide a strongly specified
theoretical base for examining the probable effects of different
policies for improving schools for initially low~achieving
students.
In developing the policy model it seems important.to consider
the responses of three system constituencies and their
counter-effects on the policy implementation sector. These
include teachers, administrators, and parents. Parents include
those of average, above-average, and below-average achieving
students. These constituencies constitute the sources of human
response in the implementation system.
A basic understanding which informs our analysis is that
typical policy objectives include the effective implementation of
actions to impact on teacher skills, teacher emphasis on
low-achieving students, teacher expectations for low-achieving
students, student achievement, and student — behavior.
Structurally speaking, the purpose of these actions is, in each
instance, to close a discrepancy between the observed level of
the variable and a desired level. This is illustrated in the
following diagram:
DESIRED
LEVEL ——
DISCREPANCY —™;
na i
oe
EFFECTS OF
ee VARIABLES
Fig. 3. The Counter-Effects of Mediating Variables on
Policy Interventions to Close Undesirable
Discrepancies.
Note that the diagram also illustrates how what are referred
to collectively as "mediating variables" represent responses to
policy actions which have the (unwanted) effect of maintaining
the status guo ante.
Mediating variables in the implementation system include the
following:
Support/resistance to the policy
Resources available to implement the policy
Conflict among constituencies
Communication and clarity about goals and expectations
Discrepancy between policy expectations and existing norms
and practices
Teacher stress
Administrator stress
312
Teacher workload
Administrator time/workload
Administrator skills
Decision functions for
=~ allocating resources
-- allocating administrator time
. Energy available.
The model assumes that implementation involves the exercise
of skills and the allocation of time and resources as available.
The model also rests upon a number of additional assumptions,
including the following:
Discrepancies between existing states and the desired
states implied by any particular policy affect levels of
support/resistance
Discrepancies in support/resistance among constituencies
produce conflict among constituencies
Resistance and conflict produce stress for teachers and
administrators
Stress reduces energy available
Energy available and the level of support/resistance
combine with skills to affect adjustment times
Conflict and administrator skills affect the quality of
communication
The quality of communication affects clarity about goals
and expectations
Clarity about goals and expectations affects perception
of discrepancies and adjustment times
13
Adjustment times affect workload for teachers and
administrators
Workload affects stress and support/resistance
Resources and administrator time are finite
Resources may be increased but only at a cost in time and
energy
+ Resources have different effects as they are allocated
differently.
A more precise progress report will be presented at the time
of the conference. An up-to-date paper will be made available to
interested participants at that time.
REFERENCES
Bloom, B.S. Human characteristics and school learning. New York:
McGraw-Hill Book Company, 1976.
Clauset. K.H., Jr. Effective schooling: A system dynamics policy
study. Unpublished doctoral dissertation. Boston University,
1982.
Clauset, K.H., Jr. & Gaynor, A.K. Closing the learning gap
Effective schooling for initially low achievers. In preparation.
Clauset, KeH., Jr., and Gaynor, A.K.. The dynamics of effective
and ineffective schooling: A system dynamics policy study. Paper
presented at the 1981 System Dynamics Research Conference,
Rensselaerville, N.Y., October, 1981.
Clauset, K.H., Jr. & Gaynor, A.K. Improving schools for low
achieving children: A system dynamics policy study. Paper
presented at the Annual Meeting of the American Educational
Research Association, New York, March 1982.
14
Fisher, C.W., et al. Teaching behaviors, academic learning time,
and student achievement: Final report of phase III-B, Beginning
teacher évaluation study. Technical report YV-1 (Summary). San
Francisco: Far West Laboratory, 1978. (ERIC Document Reproduction
Service No. ED 183 525)
Gaynor, A.K. Toward a dynamic theory of innovation in public
schools. Paper presented at the Annual Meeting of the American
Educational Research Association, San Francisco, April, 1979.
Gaynor, AK. A dynamic model of mathematics curriculum change in
an urban elementary school. Paper presented at the Annual Meeting
of the American Educational Research Association, Boston, April
1980.
Gaynor, A.K. The dynamics of stability and change in public
schools. Paper presented at the 1981 System Dynamics Research
Conference, Rensselaerville NY, October 1981.
Kolesnik, W.B. Motivation: Understanding and influencing human
behavior. Boston: Allyn and Bacon, Inc., 1978.
314
CC BY-NC-SA 4.0