Agenda setting in Colorado educational policy making

Material Information

Agenda setting in Colorado educational policy making
Westervelt, Gerrit
Place of Publication:
Denver, Colo.
University of Colorado Denver
Publication Date:
Physical Description:
166 leaves : ; 28 cm

Thesis/Dissertation Information

Doctorate ( Doctor of Philosophy)
Degree Grantor:
University of Colorado Denver
Degree Divisions:
School of Education and Human Development, CU Denver
Degree Disciplines:
Educational Leadership and Innovation
Committee Chair:
Murphy, Michael J.
Committee Members:
Muth, Rodney
Pipho, Chris
Cummings, Michael S.


Subjects / Keywords:
Education -- Political aspects -- Case studies -- Colorado ( lcsh )
Education and state -- Case studies -- Colorado ( lcsh )
Education and state ( fast )
Education -- Political aspects ( fast )
Colorado ( fast )
Case studies. ( fast )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )
Case studies ( fast )


Includes bibliographical references (leaves 161-166).
General Note:
School of Education and Human Development
Statement of Responsibility:
by Gerrit Westervelt.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
47848563 ( OCLC )
LD1190.E3 2001d .W47 ( lcc )

Full Text
Gerrit Westervelt
B.S., Ithaca College, 1986
M.A., University at Albany, 1991
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Doctor of Philosophy
Educational Leadership and Innovation

The thesis for the Doctor of Philosophy
degree by
Gerrit Westervelt
has been approved
Chris Pipho

Westervelt, Gerrit (Ph.D., Educational Leadership and Innovation)
Agenda Setting in Colorado Educational Policy Making
Thesis directed by Professor Michael J. Murphy
This case study investigates agenda setting in Colorado educational policy
making, to determine how educational policy issues become the subject of
governmental action. The study uses a conceptual framework that focuses
on the interactions among three largely independent streams within the
policy process: problems, policies, and politics. The study finds that
numerous variables related to problem definition, the generation of policy
alternatives, and political developments come into play in order for an
educational issue to emerge as a priority for state action in Colorado. The
mix of variables that contribute to agenda setting varies by issue, with the
fate of some issues heavily influenced by events in the problems stream,
and others governed by political or policy stream developments.

This work is dedicated to my parents, Dirck and Joanne, who have
always believed in and supported me, and who continue to set a
wonderful example of intelligence, curiosity, thoughtfulness and good

I wish to thank Dr. Michael J. Murphy and the members of my
dissertation committee for their guidance and support throughout this
process. Professor Rod Muth has been an invaluable resource for me
since I began my doctoral work, and Professor Mike Cummings has
given generously of his time, energy, and wise counsel. It has been a
privilege to work with them.
I especially want to thank Dr. Chris Pipho, a terrific friend, mentor, and
colleague, whose encouragement, assistance, and sense of humor
helped me complete this "long uphill pull." Thanks for everything.
I am indebted to all of the Colorado leaders who shared with me their
insights about the policy process. I appreciate their cooperation and
their candor.
Thanks also to Frank Newman, Kay McClenney, Bob Palaich,
Katherine Boswell, Charlie Lenth, Frank Blair, Deb Banks, and
numerous other friends and colleagues whose support and advice was
always timely and thoughtful.
I am deeply grateful for the love and support of my sister, Lisa, my
brother, Eric, my grandmother, Virginia Veeder Westervelt, Dan and
Teresa Lawlor, and Carmon and Rose Audino. Many thanks.
My warmest thanks go to my wife, Teresa, whose loving
encouragement and patient support made this experience possible. I
could not have done it without her.

1. INTRODUCTION.........................................1
Purpose and Significance of the Study..............2
Context for the Study..............................5
Overview of Methodology............................8
2. REVIEW OF THE LITERATURE............................11
The Policy Making Process.........................11
The Stages Model of the Policy Process......12
Policy Making as Pluralist Bargaining.......15
The Role of Elites in the Policy Process.....17
Policy Making and the Scope of Conflict.....19
Policy Making as an Ecology of Games........21
Agenda Setting....................................24
Agendas and Why They Are Important..........24
Kingdon's "Streams" Model of Agenda Setting..28
Punctuated Equilibrium......................33

Focusing Events...............................36
The Literature on Term Limits......................38
The Case For Term Limits......................41
The Case Against Term Limits..................45
3. METHODOLOGY OF THE STUDY..............................53
Rationale for the Methodology......................54
Sampling Procedures................................59
Survey Design and Administration...................60
Case Study Selection...............................64
Data Analysis......................................65
Conceptual Framework...............................68
4. RESULTS OF THE STUDY..................................73
Interviewees and the Colorado Context..............76
Case Study: Educational Accountability.............80
The Election of Governor Bill Owens...........83
Entrepreneurship and Accountability...........85
CSAP Scores as Agenda Drivers.................89

Case Study: School Safety.............................93
A "Must-Do" Issue...............................95
Feedback From Educators, El-Prepared Schools....96
Governor Owens, Policy Entrepreneur.............98
Values-Based Disagreements.....................100
Results on Term Limits...............................102
Administrative Turnover........................103
Term Limits and Agendas: Four Impacts..........106
Summary of Results...................................114
Educational Accountability and the Three Streams.....117
The Political Stream: The Creation of a Policy
The Problems Stream: Test Scores as Indicators.120
The Policy Stream: PoHcy Entrepreneurship......121
School Safety and the Three Streams..................124

The Problems Stream: Columbine as a Focusing
The Policy and Politics Streams: The Work of
Policy Entrepreneurs........................................128
Term Limits: The Elephant in the Room............130
The Political Stream: Musical Chairs and
Administrative Change.......................................131
The Problems and Policy Streams: More Ideas,
Less Wisdom.................................................135
The Importance of Problem Definition.............140
The Importance of Policy Entrepreneurship........143
The Importance of Turnover.......................146
The Utility of the Streams Framework.............149
Areas for Further Research.......................152
A Concluding Thought.............................156

A. Three Streams Framework....................157
B. Interview Protocol.........................158
C. Codes List.................................159

Greater than the tread of mighty armies
is an idea whose time has come.
Victor Hugo
Why do some educational policy issues capture the attention
of state policy makers, while others remain in the background?
What causes an issue or proposal to move to the "front burner?"
This case study investigates agenda setting in Colorado educational
policy making. It seeks to determine how educational policy issues
become the subject of governmental action, and what role different
people, events, and activities play in that process. In this qualitative
study, data were gathered in semi-structured interviews with
system actors, augmented by archival research. The study uses a
three-part conceptual framework adapted from Kingdon's (1995)
"streams" model of agenda setting, focusing on the interactions
among three largely independent streams within the policy process:
problems, policies, and politics. By focusing on state, as opposed to

federal, agenda setting, and by examining the development of
current educational policy issues, this research helps fill an
important gap in the research base on agenda setting.
Purpose and Significance of the Study
This research examines educational policy agenda setting in
the state of Colorado. A major goal of this research is to explain how
educational policy agendas are established in one state context, to
inform future researchers and political actors seeking to improve
public policy making. Ultimately, a better understanding of agenda
setting may help ensure that important public problems come to the
attention of key decision makers. Such an understanding may also
increase the likelihood that thoughtful, well-designed policy
proposals will get a hearing. This research matters, because agendas
shape outcomes. "The battle over public policy may well be decided
in the preliminary stages of issue emergence and agenda setting"
(Jones, 1984, p. 51). Understanding agenda setting in different
contexts fuels the hope that better solutions to pressing problems
may find their way into the policy process with increasing

Those in and around government pay more attention to
some subjects than to others. The agenda setting process "narrows
this set of conceivable subjects to the set that actually becomes the
focus of attention" (Kingdon, 1995, p. 3). Without an understanding
of how this narrowing process works, we cannot truly understand
policy making. Kingdon aptly states the case:
The phenomena involved are so central to our
comprehension of public policy outcomes and governmental
processes, yet they are so incompletely understood. The
patterns of public policy, after all, are determined not only by
such final decisions as votes in legislatures, or initiatives and
vetoes by presidents, but also by the fact that some subjects
and proposals emerge in the first place and others are never
seriously considered (p. 2).
Other scholars note that "appreciating the dynamics of
problem definition is essential to even the most rudimentary
understanding of public policy making" (Rochefort & Cobb, 1994, p.
vii). Or, as Schattschneider famously put it, "The definition of the
alternatives is the supreme instrument of power" (1960, p. 68).
Despite its importance, much of the current research in this area is
not about states, and not about educational policy. This research
helps to change that fact.
Much of the policy process literature focuses on Congress,
and on efforts by interest groups and other actors to influence the

outcome of legislation (DeGregorio, 1997; Dionne, 1996; Jones, 1984;
Kingdon, 1981). The same federal focus is evident in the sub-fields
of policy implementation research (Bardach, 1977; Bullock & Lamb,
1984; Nakamura & Smallwood, 1980; Pressman & Wildavsky, 1973);
and decision making studies (Arnold, 1979,1990; Fenno, 1973;
Smith, 1988). But states, not the federal government, have long
been the most significant players in educational policy making.
Every state constitution entrusts the state legislature with the
responsibility for providing some form of public education. In this
era of devolution of authority, state actors are making many critical
school reform decisions, including which educational problems
gain agenda status and which policy alternatives will be considered
to address them. While much attention is currently given to
studying the substantive outcomes of state policy making in various
areas, and some studies have focused on the state level (Firestone,
1989; Fuhrman, 1995; Rosenthal, 1990), very little work has been
done on the agenda setting phase of the policy process at the state
level, and none that focuses on educational policy making. By
examining agenda setting in this context, this study provides a
foothold for future researchers seeking to understand policy making
in different issue areas and governmental contexts.

This research is significant not only because it may contribute
to better policy outcomes and enhance our understanding of the
policy process, but because of its special attention on the role of
political developments in shaping agendas. Term limits are
changing the composition of state legislatures; changes in party
control of legislatures and governorships are having enormous
impacts on policy; and public problems, especially in education, are
increasingly complicated and controversial. This is an appropriate
time to focus attention on agenda setting and its political
components. Rapid and continuing political changes at the state
level may already be outpacing our ability to make sense of them.
Context for the Study
The Colorado policy environment is quite complex for a
relatively small state The first Republican governor elected in
almost a quarter century has staked his claim as an educational
reformer, virtually ensuring that educational policy issues will be of
considerable interest in the state capitol. The traditionally
Republican legislature is being transformed by term limits, enacted
by voters in 1992. And while the healthy state economy roars along,

the voter-approved TABOR (Taxpayer's Bill Of Rights) amendment
limits the ability of government to raise taxes or increase spending.
The state has been among the leaders in educational reform,
centered around the development and implementation of student
standards, assessments, and other standards-driven policies enacted
under governor Roy Romer (D) in the early and mid 1990s. There is
a strong tradition of local control of schools. Many observers believe
that the state is in the process of encroaching on this decentralized
system as it moves forward with a state-driven, standards-based
reform agenda, adding a layer of complexity to state educational
politics and policy making. There has been a long-term decline in
the state's share of total educational expenditures, increasing
pressure on local property-tax bases. In November 2000, voters
approved a constitutional amendment requiring the state to fund
K-12 education at the rate of inflation plus one percent annually, a
provision that begins the process of increasing the state share.
Nationally, the standards-based reform movement has
established strong roots in the last decade, as governors and
legislatures have pushed for state-mandated standards, assessments,
teacher training, and professional development. Increasingly,
incentives and sanctions tied to performance relative to these

standards are being implemented, prompting what some see as a
growing backlash against so-called "high stakes" tests (Schrag, 2000).
Some of the tensions between state and local actors in Colorado are
being mirrored nationally, although it is probably too early to know
whether the standards-based reform movement will begin to be
rolled back in the face of opposition.
To begin to make sense of these developments, we need a
better understanding of the policy-making process, including how
and why issues rise and fall on the agenda. As the stakes grow
higher in educational reform debates, actors need to know how the
system works in order to affect its outputs. The current reform
movement has not only evolved from its antecedents (minimum
competency testing, teacher in-service programs, etc.), but been
shaped by politicians, researchers, and other actors who have
worked hard to frame the debate in certain ways. "Agenda setting
has important policy consequences" (Baumgartner & Jones, 1993, p.
4). This study adds to our understanding of agenda setting in
educational policy, and thus enhances our ability to affect the
policies that shape our lives, our schools, and our nation.

Overview of Methodology
This research is a case study of one state's educational policy
agenda setting process. A case is "a phenomenon of some sort
occurring in a bounded context" (Miles & Huberman, 1994, p. 25).
Case studies are often used "to evaluate an event, an institution, a
process, or a program" (Krathwohl, 1993, p. 347). As discussed in
Chapter Three, this approach fits well with the subject under study.
Data for this research were collected using semi-structured
interviews with key actors in Colorado's state educational policy
community (the "policy system," Baumgartner & Jones, 1993; see
also Cobb & Elder, 1983), augmented by archival research. Data from
the interviews were reduced and coded using QSR-NUDIST
qualitative research software, and a series of matrix data displays
was created to aid data reduction and analysis. Analytic text was
then created from the displays and the coded interview data, and
revised and expanded to produce the final report. Case study
selection, snowball sampling procedures, and other aspects of the
research design are described in more detail in Chapter Three.

As a case study of agenda setting within a single state policy
system, this research has limited generalizability in any statistical
sense. However, case study findings are bolstered when a researcher
has carefully delineated the theoretical and political context of the
subject under study (Lancy, 1990, p. 149), as this study has attempted
to do. The conceptual framework used in this study is well-tested,
and the policy environment in Colorado is clearly articulated. This
political context also has much in common with that in other states,
as policy actors across the country engage in similar educational
reform debates. The issue of generalizability is related also to
expectations: "Generalizability is ultimately related to what the
reader is trying to learn from the study" (Kennedy, 1979, p. 672,
quoted in Lancy, 1990, p. 165). Readers must judge the applicability
of the findings in this study to future cases, based on their own
observations and knowledge. Hopefully, this study's findings will
have "face validity," mirroring elements of others' observations
about the policy process. The data in this study have a number of
strengths as well, including local groundedness, richness (Miles &
Huberman, 1994, p. 10), and the benefit of collection over a period of
many months, making them powerful for descriptive analysis of
an iterative process such as agenda setting.

This study examines state educational policy agenda setting,
adding to our knowledge base about how issues and ideas achieve
agenda status, or fail to do so. Special attention is paid to the role of
the political stream in Colorado and the people in it, many of
whom are now constrained by term limits. Underpinning this
research is the assumption that agenda setting mattersthat what
gets considered, or not, affects the end results of the policy process. It
is hoped that by painting a detailed picture of the agenda setting
process in Colorado educational policy making, this research will
help point the way toward more thoughtful educational policy
making at the state level.

The Policy Making Process
Scholars and students of public policy have articulated
different definitions of "policy process." Jones (1984, p. 24) describes
process generically as "a series of actions or operations definitely
conducing to an end," and policy processes as those focused on
"public problems and how they are acted on in government" (p. 25).
He also notes three types of processes relevant to the study of
governmental activity: institutional processes, associated with the
actions of legislatures, the executive and judicial branches, political
parties, and other political institutions (Rosenthal, 1990; 1998);
group processes, in which the roles of internal and external
alliances and interest groups are examined (Truman, 1951); and elite
theories of government (Mills, 1959), which see governmental
processes as controlled to a large extent by small groups of economic
and political actors who do not necessarily represent diverse
constituencies. All of them attempt to explain how political and

governmental systems operate, and seek to provide an organizing
framework for governmental activity.
The Stages Model of the P'olicyProcess
The so-called "stagaes" model Jones uses to describe the policy
process can be seen as a "traditional" model; it is similar in structure
to the types of activities ok steps that experience and common sense
tell us are typical of governmental action: a problem is identified
and becomes part of the offficial agenda, a solution is proposed,
governmental action is taken to address the problem, those actions
lead to new programs or services; these are evaluated over time and
adjustments made as the cycle repeats itself. Jones (1984, pp. 27-29)
articulates 11 policy-process elements and associated questions,
beginning with perception^/definition (What is the problem?),
aggregation (How many people think it is important?), organization
(How well organized are tliese people?), representation (How is
access to decision makers maintained?), agenda setting (How is
official agenda status achieved?). These five elements are
characterized as "getting problems to government." Three
subsequent elements constitute "action in government," including
formulation (What is the proposed solution? Who developed it

and how?), legitimation (Who supports it and how is majority
support maintained?), and budgeting (How much money is
provided?). The final phases include implementation (Who
administers it and how do they maintain support?, also referred to
as "government returning to the problem"), evaluation (Who
judges its success, and by what methods?), and
adjustment/ termination (What adjustments have been made, and
how did they come about?, or "programs returning to government"
for review).
The stages model expands on LasswelTs (1971, p. 28) seven
decision-process phases (intelligence, promotion, prescription,
invocation, application, termination, appraisal) in articulating a
model of the policy process that is cyclical (problems to government,
action in government, government to problems, programs
returning to government), but not necessarily linear (Jones, 1984, p.
28). The steps may overlap, occur out of order, or be so indistinct as
to be inseparable. For example, a failed proposal from one
legislative session may emerge as a factor in setting the agenda for
the next one. Jones (p. 29) also notes that actors are not limited to
one sphere or element of the process: legislators are not the only

ones who legitimate proposals, nor are bureaucrats the only actors
who implement policy.
More recently, scholars have moved beyond the traditional
policy process model, favoring descriptions and frameworks that
lend themselves to theory-building. Birkland (1997) argues that the
"stages" model of the policy process has fallen out of favor in the
last 20 years, primarily because it is seen by scholars as too
compartmentalized, too limiting, and not reflective of the
complexities of the policy process in many contexts. He concedes
that labeling elements of the process using the nomenclature of the
stages model has utility, as it helps to organize our thinking, even
though "it is just too linear and deterministic to serve as policy
theory" (p. 5). The activities in the policy process often overlap, and
may often lead to dead ends (greater agenda status failing to lead to
policy change, for example). With respect to agenda setting, the
traditional model does not explain any of the specific "how" or
"who" questions associated with the rise and fall of issues, and thus
must be seen as a starting point for deeper investigations into the
nature of governmental action rather than as a complete picture of
how things work (Sabatier, 1991).

Policy Making as Pluralist Bargaining
One answer to the "who and how" questions associated with
policy making is offered by pluralists, who see political resources as
diffused throughout many parts of the political system. This
dispersed authority creates a system in which "decisions are made
by endless bargaining" (Dahl, 1956, p.150). In the pluralist
conception of policy making, groups, not individuals, are the unit
of analysis. Pluralism sees an open system, in which "all the active
and legitimate groups in the population can make themselves
heard at some crucial stage in the process of decision" (p. 137). The
fact that groups can make themselves heard "does not mean that
every group has equal control over the outcome" (p. 145). It does
mean, however, that policy decisions are really a product of
continuous bargaining among diverse groups and organizations.
This system of "dispersed inequalities," in which resources are
widely distributed among many actors in the process, provides the
political system with some measure of balance, and ensures that no
one is able to exercise excessive influence in any one area of public
concern (Dahl, 1961, p. 51).

The pluralist view of policy making sees government largely
as a referee for interest-group conflict, rather than as an architect of
policy. Truman (1951) argues that groups are inextricably linked to
the shape of the political environment in which they operate. "Both
the forms and functions of government, he argues, "are a
reflection of the activities and claims of such groups" (p. 505). The
complex nature of policy making also prevents any one group from
dominating a particular issue area. A number of checks and
balances exist: groups often have overlapping memberships;
"potential groups," interests which are widely held in society, affect
the process despite the lack of an organized group to carry the
banner; and groups often lack internal cohesion, limiting their
ability to exercise power (pp. 299; 506). Policy making is essentially a
two-way street, in which the claims of organized groups are
balanced against one another, and against the limits of
governmental structure, overlapping memberships, and potential
groups in society.
Arguments for and against this group-centric view of
governance and policy making came to dominate academic political
theory for three decades. Lowi (1969) argues that "interest group
liberalism" lacks legitimacy, in part because of its reliance on the

interactions of multiple interest groups and other actors to define
due process of law. "Pluralism became the model and the
jurisprudence of the good society," resulting in a system "that can
neither plan nor achieve justice" (p. xvi). Lowi's proposed
alternative, juridical democracy, would emphasize the rule of law,
replacing interest-group bargaining with what he calls "the superior
rationality and administrative efficiency" of democratic formalisms
(p. 311). In this view, the illegitimacies created by the operations of
pluralist bargaining would be rectified by the institution of formal,
specific procedures for the operations of government.
The Role of Elites in the Policy Process
Questions about justice and fairness also influenced elite
theorists' critiques of pluralism. This view sees the policy process as
driven by the actions and interests of a select few individuals. Mills
(1959) argues that a "power elite" exercises disproportionate
influence over policy decisions, at least at the national level. The
power elite consists of those individuals "in command of the major
hierarchies and organizations of society," including corporate
leaders, military commanders, and members of the "political
directorate" (pp. 4; 18). Although its exact membership changes and

evolves over time, the power elite is always representative of those
leaders whose power is institutionalized, who share "decisions
having at least national consequence" (p. 19).
Below the power elite are middle layers of influence,
composed of a variety of political actors who share among
themselves power over certain types of decisions (Mills, 1959, p. 245;
see also Hunter, 1953). Congress and most national interest groups
are in this middle level. Decisions made at this level which are
contrary to the interests of the real elites can be counteracted by the
members of the power elite, whose interests often coincide. In short,
power is wielded in proportion to the size and importance of the
bureaucratic institution that each member of the power elite
commands. Policy making is therefore dominated by the leaders of
the largest organizations: corporations, governmental institutions,
and the military (p. 278).
In focusing on the resource inequalities present in the
American political system, elite theory directs attention to
differences in political influence associated with those inequalities.
It assumes that institutional size translates into actual power over
policy decisions. However, policy decision making can be so
complex as to prevent even an "800-pound gorilla" from winning

in all, or even most cases. For example, tobacco company
executives may have disproportionate access to policy decision
makers, but may be constrained by other factors such as the weight
of public opinion. And the nonlinear, untidy nature of policy
making provides multiple opportunities for other actors to derail
the best-laid plans of the elites. The elite model demonstrates that
not all voices are equal, but fails to account for the complex,
multivariate nature of the policy process. This complexity often
mitigates the impact of unequal resources, diluting the influence of
elite actors.
Policy Making and the Scope of Conflict
An interesting variation on the power elite model is
Schattschneider's (1960) "scope of conflict" framework. He argues
that "the most important strategy of politics is concerned with the
scope of conflict" (p. 3), and sees a system in which the outcomes of
policy debates are often predetermined by the ability of actors and
organizations to control "who can get into the fight and who is to be
excluded" (p. 20). In this view, agendas reflect the interests of those
most able to limit access to the policy arena: monied interests,
typically business. The scope-of-conflict critique ridicules the

pluralist notion that policy agendas are largely the product of an
open process of give-and-take among competing interest groups.
"The flaw in the pluralist heaven is that the heavenly chorus sings
with a strong upper class accent. Probably 90 percent of the people
cannot get into the pressure system" (p. 35).
The policy system may reflect a pro-business bias, but is not
totally dominated by elites. Rather, the key to understanding policy
making is to see who is able to control the scope of a given debate
whether the conflict is "privatized" (limited to a few actors), or
"socialized" (with widespread participation by a variety of actors and
groups (p. 12). In other words, those who control the scope of the
conflict will probably win. "If a fight starts watch the crowd, because
the crowd plays the decisive role" (p. 3). The role of government,
"the greatest single instrument for the socialization of conflict in
the American community," is to ensure that the scope is widened
sufficiently to ensure that the interests of the public are served (p.
13). The scope-of-conflict model emphasizes who is at the table, the
number and strength of their allies, and which side is best able to
mobilize others to support them.
The scope-of-conflict view of policy making also incorporates
the notion of "nondecision making." As actors work to mobilize

others in support of their cause, "some issues are organized into
politics while others are organized out" (Schattschneider, 1960, p.
71). In doing so, actors may exercise power by confining the scope of
decision making to relatively "safe" issues," which do not threaten
their position within the system (Bachrach & Baratz, 1962, p. 947).
Such "nondecision making" occurs primarily by actors reinforcing
social and political values and practices that protect their interests.
By shaping policy debates to their advantage, and limiting the
agenda to "safe" issues, actors are limiting the scope of conflict. For
example, educational interest groups that are able to keep voucher
proposals off the public agenda are shaping policy outcomes as
surely as if they had defeated a voucher bill in committee.
Nondecision making is just as impactful as traditional decision
making; it is merely less visible.
Policy Making as an Ecology of Games
Ultimately, the ongoing debate between pluralists and elitists
about the complexion of policy making fails to resolve the question
of how the different parts of a policy system may influence one
another. The "either/or" nature of the pluralism-elitism debate
tends to obscure more subtle questions about how policy actors

interact, and how these interactions shape policy agendas and
outcomes. One model that attempts to address these relationship
issues is Firestone's (1989) notion of policy making as an "ecology of
games." A given policy environment contains numerous "games"
(education, business, etc.) with different rules and norms of
behavior. They are linked ecologically, in that relationships may be
cooperative, be competitive, or have other characteristics; there is
no single, coordinating presence; and each receives input from the
others (p. 19). The games metaphor helps an observer to see both
the interactions within individual arenas or games (e.g., the
nuances of the legislative game), and the demands, resources, and
rules that flow among games (e.g., interactions between the
legislature and local school districts).
The legislative game illustrates Firestone's thesis. First, there
are players from inside and outside government, who collaborate or
compete with one another, depending on their needs. Political
demands among players constitute the inputs into the system;
enactment of policies (or not), and changes in position power, are
typical rewards for winners or sanctions for losers. The outputs of
the legislative game, such as resources or regulations, become
inputs for other games, such as local or state government

administration. Because contexts vary among the different games,
policies take on different meanings in different sectors, leading to
variations in implementation (p. 19). Legislative output in
educational policy, for example, may be treated quite differently
from transportation output.
The games model illustrates how an issue's agenda status
may be influenced not only by interactions within a legislature, for
example, but also by the disposition of other issues from other
games (e.g., interest-group pressure games, regulatory games, etc.).
The ecological metaphor promotes a systemic view of policy
making. It is often tempting for analysts to focus their attention on
the actions of one actor or group of actors, such as the govemor-as-
agenda-setter, or interest groups-as-obstacles-to-reform. While such
observations may be helpful in certain contexts, they run the risk of
missing the interrelationship forest for the single actor tree. The
ecology metaphor keeps the dynamic nature of the policy system at
center stage, allowing "the loose linkages between separate games
in the form of simultaneous interdependencies and
discontinuities" to become clear (p. 21).
In depicting this complexity, Firestone's model may
discourage the search for the "silver bullet" in policy making. Once

one understands the mutual dependencies and motivations of
actors involved in different, simultaneous games, one may be less
inclined to believe that a single, bold change will transform the
entire system. The metaphor helps explain the inherent messiness
of agenda setting: time frames differ among games, required inputs
and outputs may not materialize, and "because people operate in
different contexts, it is extremely difficult to agree on what
constitutes good policy" (p. 21). By illustrating the complexity and
decentralized irrationality of the policy process, the ecology-of-
games model supports efforts to craft more thoughtful, nuanced
policy solutions.
Agenda Setting
Agendas and Why They Are Important
No matter which lens one uses to observe policy making-
stages, pluralist, elitist, or ecology of gamesagenda setting is a
critical piece of the puzzle. Agenda setting usually refers to the early
stages of the policy process, when "some subjects become prominent
on the policy agenda and others do not" (Kingdon, 1995, p. 3).
Understanding how and why some issues become the focus of
governmental action, and why others are neglected, is important

because, agendas shape outcomes. We must understand agenda
setting in order to understand policy making (Rochefort & Cobb,
1994). Defining the policy agenda is arguably the phase of the policy
process when power is exercised with the greatest effect, but it has
not been studied as much as the more public phases, such as policy
adoption or implementation.
Perhaps the basic premise underlying much agenda-setting
research is the notion that one of the most important ways policy
leaders shape outcomes is by influencing the choice of problems to
which government will direct its attention. Knowing who plays
what role in selecting public problems and solutions is vital to a
rich understanding of policy making. The question of how public
policy agendas are set has been addressed primarily by political
scientists seeking to understand the role of agenda setting within
the overall policy process. Many scholars see agenda setting as
practically inseparable from other phases or elements of policy
making, such as consideration of policy alternatives or legitimation
of policy proposals (Jones, 1984). Although agenda setting may occur
during or as a result of many different stages of the policy process, in
general it has come to refer to the early stages of the process, when

some issues become the subject of governmental attention and
action, and others remain unaddressed.
In most contexts, there are different levels of agenda status to
be considered rather than a single agenda. Jones quotes Cobb and
Elder's definition of the systemic agenda: "All issues that are
commonly perceived by members of the political community as
meriting public attention and as involving matters within the
legitimate jurisdiction of existing governmental authority" (Cobb &
Elder, 1972, p. 85, quoted in Jones, 1984, p. 58). This definition raises
the obvious questions of who counts as a member of a given
"political community," and what constitutes the "legitimate
jurisdiction" of government. Such questions come to the fore when
issues that have been outside the legitimate jurisdiction of school
officials emerge as part of the educational policy agenda. The
systemic agenda is also called the governmental agenda, "the list of
subjects or problems to which governmental officials, and people
outside of government closely associated with those officials, are
paying some serious attention at any given time" (Kingdon, 1995,
p. 3).
The institutional agenda, by contrast, is much narrower: "that
set of items explicitly up for the active and serious consideration of

authoritative decision makers" (Jones, 1984, p. 58). These are the
issues that are the objects of the policy process; the list usually varies
by actor, so that a governor, chief state school officer, interest-group
lobbyist, and legislative committee chair may have different
perceptions of which problems deserve attention, as well as then-
scope and urgency. Kingdon refers to the narrower agenda as the
decision agenda, "the list of subjects within the governmental
agenda that are up for an active decision" (1995, p. 4). Birkland
argues that "the goal of most contending parties in the policy
process is to move policies from the systemic agenda to the
institutional agenda, or to prevent issues that they find inimical to
their interests from reaching the institutional agenda" (1997, p. 8).
The decision agenda is the "front burner" of American politics and
government, and the object of the game is to control or influence
what gets there.
Much more than decision agenda status is at stake. Kingdon
draws an analytical distinction between agenda items (issues) and
alternatives (policy options) designed to address issues. Getting an
issue on the front burner is only part of the battle; shaping the list of
potential policy options to be considered is equally crucial. Kingdon
notes that "perhaps agenda setting and alternative specification are

governed by quite different processes" (1995, p. 4). For example, chief
executives such as governors might play a greater role in setting the
agenda than they would in generating alternatives; bureaucrats or
interest groups might play a more prominent role in the process of
specifying alternatives and building or inhibiting support for them
within a given policy community. The linkage between identified
problems and viable solutions is central to agenda setting. "The
chances for a problem to rise on the decision agenda are
dramatically increased if a solution is attached" (p. 143). Thus do
actors spend time trying to influence the choice of policy solutions,
as well as the rise and fall of issues or problems on the agenda.
Kingdoms "Streams" Model of Agenda Setting
This study is concerned primarily with the decision agenda,
in order to focus attention on how educational issues gain "active"
agenda status in Colorado. Kingdon attempts to answer the question
of how issues come to be part of the decision agenda by expanding
on Cohen, March, and Olsen's (1972) "garbage can model" of
organizational choice. Cohen et al. argue that many public
organizations meet their definition of an "organized anarchy":
institutions or groups characterized by problematic preferences,

unclear technology, and fluid participation (p. 1). Such
organizations contain four independent but related "streams" that
interact to produce decisions and actions (or not): problems,
solutions, participants, and choice opportunities (p. 3). Problems are
seen as "the concern of people inside and outside the organization,"
solutions as "an answer actively looking for a question,"
participants as fluid, meaning that they move in and out of various
choice opportunities, and a choice opportunity as the decision point
or "garbage can into which various kinds of problems and solutions
are dumped by participants as they are generated" (p. 2). The
interplay that exists among these variables creates organizational
decisions, whether or not those decisions actually solve problems
facing the organization.
For example, the presence or absence of certain participants
(e.g., a senate committee chairman or an anti-tax crusader) from a
given choice opportunity (a scheduled vote on a budget bill), the
introduction or withholding of a pet solution (raising cigarette
taxes) to a defined problem (is the issue teenage smoking, or
governmental interference in the marketplace?), all influence
whether that choice opportunity will result in a decision, and how
that decision will turn out. Perhaps cigarette taxes will be seen by

the mix of decision makers present as the right answer to the
problem; perhaps the problem will get "uncoupled" from this
particular choice opportunity, postponing action on teen smoking.
"Problems are worked upon in the context of some choice, but
choices are made only when the shifting combinations of problems,
solutions, and decision makers happen to make action possible"
Kingdon's model recasts the elements in the "garbage can"
into three primary "streams" that operate simultaneously within
the policy system: problems, policies, and politics. The problems
stream consists of those problems and issues that "come to capture
the attention of people in and around government" (1995, p. 87).
Problem recognition often occurs because of a crisis or "focusing
event" that seizes public and official attention. Systematic indicators
such as governmental statistics or budgetary trends can also cause
attention to be shifted to one problem or another. In addition to
events and indicators, the problem stream also consists of feedback,
such as input from citizens, program-evaluation data, and
bureaucratic experience. Participants' perceptions of problems, and
the images associated with them, often affect problem definition,
whether or not actual conditions change (p. 110). For example,

school truancy may not capture official attention if it is viewed as a
local administrative issue; if truancy is linked to gang violence and
crime rates, its place on the agenda may change, regardless of
whether truancy rates are going up or down.
The second stream, the policy stream, includes ideas and
proposals generated by various members of a "policy community of
specialists" consisting of bureaucrats, researchers, interest-group
advocates, staffers and other players (p. 87). These policy proposals
float around government, waiting for the right moment to attach
themselves to a problem on the decision agenda. In education, for
example, proponents of private school vouchers may push their
plans in response to reports of poor test scores, or use them as a way
to attack unions and stagnant educational bureaucracies, whichever
is most likely to move their solution forward. The viability of a
policy idea can be affected by such attributes as technical feasibility,
congruence with participants' values or beliefs, affordability, and the
presence or absence of policy entrepreneurs, advocates who invest
their resources (time, money, reputation, etc.) in order to move a
particular proposal or idea.
The third stream, the political stream, "is composed of things
like swings of national mood, vagaries of public opinion, election

results...and interest group pressure campaigns" (p. 87). Other
scholars have argued that in addition to such overtly political
factors, subtler forces such as political culture and formal structure
also play a role in defining agendas by limiting the range of
"acceptable" policy alternatives and constraining actors seeking to
affect problem definition (Bosso, 1994, pp. 184,193). For example, a
change to a conservative administration can quickly elevate official
interest in tax cuts, and diminish actors' interest in expensive social
programs, regardless of whether economic data have shifted in one
direction or another. Changes in the political stream can have
profound effects on agendas, as some ideas gain currency and others
lose ground in the face of positive or negative political
Kingdon's three streams, like those of Cohen et al., are largely
independent. It is their interactions that are vital:
The key to understanding agenda and policy change is their
coupling. The separate streams come together at critical
times. A problem is recognized, a solution is available, the
political climate makes the time right for change, and the
constraints do not prohibit action. Advocates develop their
proposals and then wait for problems to come along to which
they can attach their solutions, or for a development in the
political stream like a change of administration that makes
their proposals more likely to be adopted (p. 88).

Just as in the traditional policy process model, the
interactions among the streams in the agenda-setting phase are not
necessarily linear; proposals can precede problems, and political
changes can precede the formulation of proposals. When streams
are "coupled," such as the linking of a pressing problem to a feasible
solution, a "policy window" is opened that can result in official
decision making (p. 88). External factors such as budget constraints
or the skills of "policy entrepreneurs" or other system participants
influence the chances that an effective coupling can take place or
that a "window" will open at an opportune time. For example, a
legislators willingness to compromise on a charter-schools bill may
create a reciprocal inclination on the part of a committee chairman
to "open a window" on an unrelated topic. Further, these
"opportunities for action on given initiatives" stay open for a
limited time (p. 166). Conditions in one or more of the streams
change, the chance for action is lost, and policy actors must wait for
the next window to open.
Punctuated Equilibrium
Some scholars have charged that Kingdon's streams model
pays insufficient attention to the contextual details that shape

agendas. The notion of "punctuated equilibrium" helps explain the
contexts in which issues rise or fall. Baumgartner and Jones (1993)
argue that agenda setting is characterized by such punctuated
equilibrium, in which change occurs in dramatic or sudden fashion,
rather than in a steady evolutionary pattern. In their view, the
political system "displays considerable stability with regard to the
manner in which it processes issues, but the stability is punctuated
with periods of volatile change" (p. 4). Issues "hit" suddenly, often
due to changes in an issues image or definition; new institutional
structures or "venues" are created around the new policy direction,
and supporting forces mobilize to maintain this "new equilibrium."
Over time, new ideas and other forces may destabilize the new
structure, and lead to another round of dramatic change. Periods of
stability are reinforced by policy monopolies, which are institutional
structures or venues that control policy making around an issue
and limit access to policy decision making in that area (p. 6). This
control helps create and support an "issue image" or perception of a
policy that supports the monopoly. "In sum, the American political
system is a mosaic of continually reshaping systems of limited
participation" (p. 6). Actors who are able to influence or shape the

j public image of an issue may succeed in altering its agenda status as
j well.
| For example, in recent years critics have argued that local
school boards have successfully positioned themselves as a policy
monopoly, undergirded by the powerful concept of local democratic
| control. These critics, often conservative in orientation, have
attempted to shift the image or definition of school governance
toward school failure and away from the protective cloak of local
! democracy, a shift that, if successful, could result in the kind of
disequilibrium that Baumgartner and Jones describe. "The old
j policy making institutions find themselves replaced or in
| competition with new bodies that favor different policy proposals.
j So agenda setting has important policy consequences" (p. 4). An
I important lesson implied by this model is that image matters; how
I policymakers and the public view an issue, including the values
and biases that they attach to it, can have important effects on policy
| agendas. It follows that the forces that help shape such perceptions,
| including media coverage, organized political activity, and access to
information and research, are often key elements in agenda setting.

Focusing Events
As noted above, Kingdon's "problem stream" includes
phenomena known as "focusing events," such as disasters or other
crises that call attention to a problem and push people toward action
(1995, p. 94). Birkland (1997) builds on this notion, defining a
"potential focusing event" as "an event that is sudden, relatively
rare, can be reasonably defined as harmful or revealing the
possibility of potentially greater future harms...and that is known to
policymakers and the public virtually simultaneously" (p. 22). The
extent to which an event is focal varies. He argues that such events
are distinguished most of all by their suddenness, as they become
fixed to a particular time and context. This enables actors to use the
event to mobilize support for policy change, providing an
advantage to change-oriented individuals and groups. Focusing
events such as oil spills are powerful symbols: "This power is
greatly enhanced through the reduction of these events to simple,
graphic, and familiar symbolic packages" (p. 11). He also notes that
"mass publics respond more readily to these symbols because they
are easier to interpret than are more complex stories or analyses of

public problems (p. 11). It is not just the public at large that
responds to these symbols; organized interests can use focusing
events to further their causes. Tragedies such as school shootings
are used by interest groups to move their pet solutions (e.g., gun
control, or posting the Ten Commandments in schools), or to keep
opposing groups' plans from getting onto the decision agenda.
Agenda setting, the process by which some issues become the
subject of government action, usually occurs in the early stages of
the policy process. The streams model envisions a process in which
problems, solutions, and political developments operate somewhat
independently in the policy process, coming together at critical
junctures to create policy windows that are opportunities for action.
These windows of opportunity are often transient. As the streams
continue to evolve, a given problem gets pushed aside by a crisis, a
proposal is discredited, an election changes the players, and the
window closes. The members of the policy community get busy
trying to influence the next round of agenda change. Developments
within streams, and interactions among them, are important

because agenda-setting decisions can have a profound impact on
policy outcomes.
The Literature on Term Limits
The conceptual framework for this study includes a focus on
Kingdon's "politics" stream, which includes organized political
activity, public mood, party control of institutions, and other
political developments that can affect policy agendas. The politics
stream in Colorado is profoundly shaped by the fact that in 1992
state voters enacted term limits for state-level elected officials. This
study focuses some attention on the political stream, and on term
limits in particular, to see what impact, if any, term limits may be
having on agenda setting in education. It is useful, therefore, to
review briefly the published literature on term limits, to get a sense
of the political context in which agendas are set in Colorado.
Currently, 18 states limit the terms of state lawmakers, mostly
through state constitutional amendments passed by voters
(National Conference of State Legislatures, 1998, p. 1). Most of these
states restrict lawmakers to six or eight years in office, although in
most states individuals are not prohibited from running for
another office once they have reached the limits of their current

term. Seven states (Arkansas, California, Michigan, Missouri,
Nevada, Oklahoma, and Oregon) impose a lifetime ban on service
in offices covered by the term-limit provisions, regardless of
whether the covered terms are served consecutively. All of the
term-limit provisions now in effect were adopted during the 1990s,
with all but Louisiana's on the books on or before 1994 (p. 4).
State Supreme Courts in three states, Massachusetts,
Nebraska, and Washington, have invalidated term-limit
provisions. A federal appeals court judge ruled in 1997 that
California's term limits are unconstitutional because voters were
not explicitly told that politicians would be banned from ever
seeking offices once they had reached the prescribed limits. Several
states have passed term limits for members of Congress, but the U.S.
Supreme Court has invalidated such restrictions (Heimann &
Kloha, 1998, p. 1).
In addition to state provisions, approximately 2,000 cities and
towns, including New York, Denver, Los Angeles, and San
Francisco, have adopted term limits covering more than 15,000 local
officials (U.S. Term Limits, 1998, p. 2). Organized efforts to get term
limits on the books have been behind many of the recently enacted
statutes. Most of these laws "were enacted by citizen initiatives

sponsored by state term limits organizations," according to U.S.
Term Limits, the leading national advoca te for the provisions (p. 2).
There is evidence that "term limits are enmbraced by large sections of
the electorate and are supported consistency across a wide variety of
demographic groups" (Heimann & Kloha,, 1998, p. 1). In 1992,83
percent of respondents to the American Mational Election Survey
favored "mandatory rotation in office" forr elected officials
(Southwell, 1997, p. 82). In six states (Wyoming, Nevada, Missouri,
Florida, Colorado, and Arizona), term-lirmit proposals have
garnered more than 70 percent of the vote- (p. 82). This strong
popular support appears to have held steaidy in the years since term
limits first appeared on the scene: a 1998 Colorado amendment
enabling candidates to publicly pledge to limit their own terms
passed handily.
Term-limit proposals have garnered broad support for a
number of reasons. These include public frustration with 90 percent
reelection rates, partisan gridlock, gerrymandering, perceived
interest-group dominance of certain policy" areas, and a campaign-
finance system widely seen as unfair at bes-t and corrupt at worst.
Researchers have identified a number of reasons for such reforms
gaining momentum in the public domain. Crotty (1980) cites two

reasons why political reforms emerge. First, reforms emerge as a
reaction to a nagging and persistent problem in the public domain
(i.e., at some point the public finally decides to do something about
a problem, and demands change). Second, reforms emerge as a
response to a crisis or unusual event that galvanizes public opinion
and causes a demand for change (p. 24). Term limits may be a
mechanism for ordinary citizens to "strike back" at a political
establishment that they feel has been unresponsive to their needs
for too many years.
The Case for Term Limits
The literature on term limits describes several main
arguments in support of these provisions. Proponents argue that
ensuring an infusion of new people and new ideas into public
institutions will result in more representative institutions that are
more responsive to the needs of the people who elected them. In
their view, entrenched incumbents are more interested in serving
their own career needs and the needs of special-interest groups than
they are in representing the interests of their constituents. And,
they argue, forcing turnover clears the way for more diverse
legislatures. "Term limits are the best means to open up the

political process to newcomers, women and minorities in
particular" (U.S. Term Limits, 1998, p. 2). Since California adopted
term limits in 1990, the percentage of female legislative members
has risen 25 percent, and the number of minority members has
more than doubled (p. 2). Proponents of the successful 1998
Colorado self-limiting pledge amendment argued that such a pledge
will promote behavior as "citizen legislators" who are less beholden
to special interests (Polhill, 1998, p. HI).
The "citizen legislature" argument has a long and
distinguished history. Support for the idea of rotation in office dates
back to Aristotle, who argued for the rule of "all over each and each
in turn over all" (Petracca, 1992, p. 20), and to Cicero, who noted "So
the man who obeys should have the hope that he will one day
command, and he who commands should reflect that in a short
time he will have to obey" (p. 21). America's Founding Fathers
continued this line of thinking, with Pennsylvania leading the way
by limiting the terms of all elected officials in its state Constitution
of 1776 (p. 27). Fund (1992a) notes that the first Continental Congress
limited terms to three years, but that resistance from incumbent
delegates prevented mandatory rotation from being included in the
U.S. Constitution.

A second major argument in favor of term limits is that
bringing in new blood will increase the likelihood that legislatures
will be willing to tackle difficult or controversial issues that
election-minded incumbents are more likely to ignore. Proponents
argue that legislators who know they will be going out of office will
have an incentive to try to make a difference in the time allotted to
them, and will therefore push harder for action on tough public
problems. "The term limit babies have been more willing to deal
with issues that their predecessors ignored: The 1997 [California]
legislative session was the most productive in decades" (Walters,
1998, p. 2). As Colorado's term-limits leader put it, "Deadlines
concentrate the mind wonderfully. Instead of playing games, term-
limited representatives will be eager to get to work, to accomplish
something substantial before they must turn over the baton to the
next citizen legislator" (Polhill, 1998, p. HI).
Proponents also argue that term limits encourage legislatures
to value ability over seniority in determining leadership roles. If
time horizons are short for everyone, the argument goes, those
with the ability to serve in certain leadership positions will quickly
rise to the top. And, notes Rosenthal, "The limits will not change
the members time perspectives significantly because the

perspectives already are short" (cited in Benjamin and Malbin, 1992,
p. 205). Short tenure would also inhibit "abuse of power" by senior
lawmakers by "rotating power so it could not remain long enough
with any one person for him or her to abuse it" (Fund, 1992a, p.
231). Advocates also claim that such rapid ascension will encourage
capable people to run for public office, by creating "a climate in
which talented men and women from businesses and professions
would want to nm...since they would know they would reach a
position of significant influence in a few short years" (p. 232).
Proponents also argue that term limits will result in more
competitive elections, chiefly by limiting incumbents' ability to
build up huge fundraising advantages over potential challengers,
and by stimulating voter turnout through creation of more open-
seat contests.
Proponents also point out that legislative term limits are
only bringing to the legislative branch limitations that have been in
place for the executive for many years. Beyle notes that "limitations
on gubernatorial terms based on fear of excessive executive power
have always been fundamental to the constitutional design of state
governments" (cited in Benjamin & Malbin, 1992, p. 159).
Supporters could make the argument that two-term limitations on

governors have not yet sapped gubernatorial power or shifted
influence to lobbyists or bureaucrats. However, this argument
neglects the well-known phenomenon of "lame duck" leaders, and
the perceived constraints on their leverage and influence as they
near departure from office.
The Case against Term Limits
Critics of term limits argue that forcing experienced
legislators from office merely shifts power and influence away from
elected representatives, and toward unelected staff and interest
groups. Inexperienced lawmakers "have little policy expertise
themselves which can be used as a balance against potential
misrepresentation" by other policy actors (Heimann & Kloha, 1998,
p. 1). Critics fear that lobbyists and bureaucrats will "take advantage
of the naivete of new legislators brought in under term limits and
manipulate their informational advantage for greater gain" (p. 1).
"They will become much more powerful," former Colorado Senate
Majority leader Jeff Wells says of lobbyists in the State Capitol (Dire,
1998, p. 4). A recently completed study of California term limits
found that the California legislature is paying a price for the loss of
experience, in terms of lawmakers' overall effectiveness, ability to

work together, and ability to deal with lobbyists (Hunt, 1998, p. A23).
In Maine, "observers close to the legislative scene think that newly
elected legislators will have to turn to lobbyists to gain a sense of
historical background" (Pipho, 1998, p. 101). U.S. Term Limits
argues, by contrast, that "Lobbyists are losing power in term-limited
legislatures," citing an Arkansas lobbyist who believes that new
legislators are wary of lobbyists and strive to keep them at arm's
length (1998, p. 2).
Other critics believe that term limits will give more power to
the executive branch. Rosenthal argues that "term limits will shift
power from the legislative branch to the executive, not necessarily
to staff and lobbyists" (personal communication, 1998). He also
argues that term limits will inhibit the ability of members to acquire
"political knowledge and skills" needed to move bills through the
process (Rosenthal, 1992, p. 206). hi a study of term limits and their
impact on policy decision making using computer simulations,
Heimann and Kloha found that "the signal from the legislature
becomes a bit more garbled and the agency has a harder time
adjusting in a timely fashion" (1998, p. 15). They conclude, in part,
that "critics argue that term limits will make agencies less
responsive to the demands of the legislature and thus shift power

away from the legislative body. In our simulations we saw this to be
the case" (p. 15).
In addition to shifting power toward interest groups or
toward the executive branch, term limits are also fundamentally
undemocratic, acording to critics. "Term limits narrow the range of
choices available to voters. If you like that 10-, 20-, or even 30-year
incumbent, you should be free to vote for her or him" (Loevy, 1998,
p. HI). Many term-limited officials merely run for other offices,
effectively skirting the intent of term limit laws. State Sen. Richard
H. Finan of Ohio, 1998 president of NCSL, warns that "term limits
are bad if all they do is chase some people over to the other
legislative house or rim off the most experienced hands and leave
lobbyists and bureaucrats to fill the vacuum" (Ayres, 1998, p. A10).
Loevy (1998, p. HI) argues that term limits "unfairly penalize those
with experience and proven skills in elective office." The notion
that term limits undermine the unfettered public choice that is a
cornerstone of democratic systems may be the most emotionally
(and perhaps politically) potent argument against term limits.
Critics also argue that term limits are simply not needed for
most public offices, as the two major parties are generally
competitive in most areas of the country. They also make the case

that term limits attack the symptoms and not the underlying causes
of problems in our democracy, such as state legislatures drawing
uncompetitive districts and incumbents who are able to out-raise
and out-spend challengers because of a corrupt and unfair
campaign-finance system (Loevy, 1998, p. HI.) Kesler (1992, p. 245)
says that "today's entrenched Congress is a product of the great
changes in American politics that have occurred since the late 19th
century, particularly the weakening of political parties and the great
increase in the size and scope of the federal government."
Proponents counter with the argument that term limits are at least
a step in the right direction toward fixing a system in which the
overwhelming majority of incumbent office-holders get reelected.
Perhaps the most pointed set of opinions on the early impact
of term limits comes from Schrag (1997), who argues:
But rather than bringing on the "Citizen Legislature"
promised by some of its advocates, term limits has generated
even more partisanship and incivility among its members, a
growing inability to compromise, a legislative leadership
with greatly reduced powers, and a sharp decline in
legislators' comprehension of, and interest in, the
complexities of the issues that they are supposed to deal with,
(p. 5)
However, Schrag does not offer much evidence that all of
these negative developments have been caused by term limits. A

growing inability to compromise, for example, may also be related
to partisan shiftsthe rise of the GOP in the 1990s, and the rising
influence of certain Republican and Democratic constituencies such
as the Christian right or organized labor. But California's early
experience with term limits certainly raises questions about the
overall health and credibility of representative democracy in the
United States. Does the public believe that its elected representatives
are unresponsive to its needs? Schrag argues that California citizens
have embraced rampant "neopopulism" (p. 6), to the detriment of
public services and to the state's quality of life. Future research may
indicate whether our collective philosophy of governance has
undergone a fundamental shift toward direct popular control.
At this early stage, empirical research on term limits is quite
limited. However, early evidence suggests that their impact on
policy making is mixed. Most observers concede that, for good or ill,
term limits have shaken up the status quo in many states, bringing
new people into legislatures, and perhaps invigorating otherwise
lackluster contests for elected offices as term-limited legislators
scramble to find other posts. One California commentator noted

that "term limits did change the Capitol's culture. The clout of
legislative leaders wanedespecially after the Assembly's wily and
long-serving speaker, Willie Brown, left to become mayor of San
Francisco" (1998, p. 2). It is likely that without term limits. Brown
would still be speaker, and life in Sacramento would not have
changed as much as it has under term limits.
Term limits are likely to affect state governments in a variety
of ways, with some positive changes and some negative. Everson
argues that "unanticipated consequences of term limits are almost
certain" (cited in Benjamin & Malbin, 1992, p. 190). He also predicts
that expectations will not be met on either side of the debate:
There is a corollary to the law of unanticipated consequences:
unanticipated nonconsequences. Political reform usually fails to
achieve either the heaven on earth described by its proponents
or the hell feared by its opponents. One reason for this is that
the conditions reformers wish to alter are usually the product of
multiple causes and reform usually focuses on a single factor.
(p. 190)
Crotty agrees, noting that "reformers oversimplify complex
problems and become paternalistic and self-righteous in their
pursuits. They present the problem and the solution in one fell
swoop" (1980, p. 29). He sees term limits as a way for the public to
"feel like they are more involved without having to do more than
vote once" (p. 42). Kiley argues that "term limits gives the

American people a perception of a more direct and participatory
democracy, but does not guarantee it" (1995, p. 9), and further notes
that "meeting the need of the American people to feel change has
occurred does not equate with actually securing fundamental
change" (p. 9). It is possible that term limits will turn out to be a
largely symbolic reform, with relatively little impact on the
operations or products of government.
The likely impact of term limits on the policy-making process
at the state level is unclear at this point, with opinion divided
between those who foresee greater influence for lobbyists, staff, and
the executive branch, and those who predict more representative,
responsive, effective legislatures. Neither camp appears to have
much data to support its claims. Similarly, it is too early to tell
whether policy outcomes in term-limited legislatures have suffered.
It stands to reason that complex legislation such as tax and spending
bills might be handled more effectively by experienced lawmakers,
but there is little evidence that overall "policy quality," however
that is measured, has declined in states with term limits. The actual
impact of term limits on state policy making is likely to vary
according to the specific constraints imposed (e.g., six years versus

12), the complexity of the issues, and the unique political context
and culture of each state.

j The purpose of this case study is to investigate how state
I educational policy agendas get set in Colorado. Data were collected
| using semi-structured interviews with a sample of key actors in
| Colorado's state educational policy community, augmented by an
I electronic survey, via email, which did not achieve a sufficient rate
j of return to be included as a primary data source. Additional data
i were gathered through review of government documents, research
; reports, press articles, and other archival materials. Data from the
interviews were reduced and coded using QSR-NUDIST qualitative
j research software, and a series of matrix data displays was created to
aid data reduction and analysis. Analytic text was then created from
the displays and the coded interview data, and expanded and
j revised to produce the final report. Case-study selection, sampling
\ procedures, and other aspects of the research design are described in
I more detail below.

Rationale for the Methodology
Research methodologies should match well with the nature
and characteristics of the phenomenon to be studied. For example,
complex phenomena occurring in a dynamic environment require
different approaches from simple phenomena occurring in a static
environment. The development of knowledge about agenda setting
in the policy process is best accomplished through the use of
qualitative approaches to research. In contrast to policy analysis,
which often lends itself to quantitative measurement and
correlational research, agenda-setting studies are fundamentally
about human interactions that are complex, context-sensitive, and
rooted in organizational and institutional cultures and norms.
Research tools such as case studies, interviews, and field
observations value the personal experiences, insights, and
motivations of those involved, and provide the contextuality
(Lasswell, 1971) that is critical to understanding the policy process.
"While methodology should always be as rigorous as possible, it can
be no more rigorous than the subject matter permits. A
methodology might be extremely rigorous, but if it is not suited to

the subject matter, it will not aid our understanding" (Smith, 1980,
quoted in Lancy, 1993, p. 29).
Agenda setting is enormously complex, with multiple actors
(including legislators, lobbyists, bureaucrats, and the media);
decision variables (political, personal, factual, ideological, etc.); time
frames (committee schedules, elections) and internal (institutional)
and external (non-governmental) constraints, such as the wishes of
legislative leaders or changes in the economy. As Lancy notes, "The
qualitative paradigm is ideal for phenomena that are patently
complex and about which little is known of certainty" (Lancy, 1993,
p. 9). Although much is "known" about agenda setting, there is
little agreement among researchers about the relative influence of
many of the actors or variables just mentioned, or how these
dynamics vary in different contexts.
In addition to, or perhaps because of, its complexity, the
agenda-setting process is one in which cause-and-effect
relationships are difficult to establish and measure. This difficulty
also points toward the utility of a qualitative approach to research in
this area. Quantitative methods such as correlational studies are
geared toward precise measurement, analysis of cause-and-effect
and other relationships among variables, and prediction. The

agenda-setting process does not readily lend itself to measurement
of cause-and-effect relationships. Like other complex human
interactions, it is too uncertain, political, and "messy" to study in a
way that validly ascribes X degree of causality to Y variable or that
predicts participant behavior beyond the most basic level. "People
are not billiard balls, but have complex intentions operating in a
complex web of others' intentions and actions" (Miles & Huberman,
1994, p. 145). Causality is difficult to establish under the best of
circumstances; within the highly political and often personal
dynamics of agenda setting, is next to impossible.
What is possible through interviewing and other qualitative
approaches is a thorough descriptive analysis of agenda setting,
grounded in the observations and beliefs of the participants
themselves. Qualitative research designs have a number of
methodological strengths, according to Miles and Huberman:
Qualitative analysis, with its close-up look can identify
mechanisms, going beyond sheer association. It is
unrelentingly local, and deals well with the complex network
of events and processes in a situation.... It is well equipped
to cycle back and forth between variables and processes
showing that stories are not capricious, but include
underlying variables, and that variables are not disembodied,
but have connections over time. (p. 147)

The approach used in this study enables the researcher to
achieve what Miles and Huberman refer to as "thorough local
acquaintance" (p. 146). The boundaries of the case, and the
appropriate sampling frame (the Colorado educational-policy
system) are well defined, and the framework used to organize the
research is clear, specific, and well-regarded in the field. Such an
approach is likely to lead toward a more complete understanding of
how agenda setting occurs, who has what kind of influence in the
process, and what environmental conditions or other factors affect
the selection of educational policy issues for official consideration
and action.
Qualitative approaches have their own limitations that must
be accounted for in any study of agenda setting. Two such
limitations deed with the objectivity of the data and the
generalizability of findings. The argument that the highly political
nature of the agenda-setting process calls for a qualitative approach
to research also implies that conclusions about the process in one
location or system have limited generalizability to other locations.
A partial answer to the problem of limited generalizability can be
found in rigorous efforts to ensure data validity, through careful
sampling and data triangulation (Miles & Huberman, 1994). For

example, data about the role of the governor in shaping agendas
must be gathered from many diverse sources, at different times,
with sensitivity to political contexts. Sources are likely to have
different opinions on that subject depending on their position in
the system, and what is happening politically at any given moment.
Validity is also aided by intelligent within-case sampling decisions
(p. 29), choosing informants who may provide disconfirming
evidence, an alternative viewpoint, or a new direction for the
inquiry. Such efforts help ensure that the methods employed in this
study provide a valid and reliable measurement of agenda-setting
dynamics. The methodology of this study is appropriate for the
process being investigated, and benefits from solid sampling, data
reduction, and data analysis techniques.
Relying on the perceptions of participants in the policy
process can limit the objectivity of the data they provide.
Partisanship, desire to influence public perceptions of an issue, or
other factors may influence the information people contribute.
Ensuring a valid sample of subsystem participants helps insulate
this research from the effects of subjectivity. This study is also aided
by an effort on the part of the researcher to be realistic about the
types of conclusions to be drawn from the data. The focus

throughout is on descriptive and interpretive analyses (Miles &
Huberman, p. 278) grounded in the perspectives of multiple
participants over many months, rather than more ambitious
predictive or evaluative conclusions. The analyses presented here
remain grounded in the data and well linked to the conceptual
Sampling Procedure?
The sample for the electronic survey and the interviews was
selected from the members of what can be referred to as Colorado's
educational "policy community" (Kingdon, 1995, p. 233). These top
decision makers and individuals in close contact with top decision
makers occupy key positions within the legislature, the legislative
staff, the executive branch, and influential interest groups such as
teacher's unions and educational associations. Selected members of
the House and Senate education committees form the core of this
sample, along with key legislative staffers, lobbyists, governor's
staffers, state education department officials, and others considered
by many to be knowledgable about Colorado policy making.
Throughout the study, a snowball sampling technique
(Krathwohl, 1993, p. 139) was used to supplement the core group

and add to the validity of the sample. This type of purposive
sampling involves asking people likely to know who should be
included in a given group to identify these individuals. The process
continues until the same names keep coming up and the sample is
saturated. Individuals occupying official positions of influence in
Colorado were asked to identify other influential actors who might
play a role in, or be knowledgeable about, educational policy agenda
setting. These new* individuals were asked the same question; in
this way, new names were added until the individuals so named
were those already identified in the sample. The combination of
those in official positions and those identified as influential in the
policy community generates a sample that is highly representative
of the population under study. This sample enables the study to
reflect accurately the views of those individuals who make up the
state educational policy making community in Colorado.
Survey Design and Administration
A three-question e-mail survey was administered in
November 1999 to 19 individuals occupying key positions within
the educational policy community. Several initial drafts of the
survey were adapted from Kingdon (1995), and covered all three

"policy streams" (problems, policies, politics) in the conceptual
framework. The survey was pre-tested with staff at the Education
Commission of the States, selected because of their extensive
experience in, and knowledge of, research design and
implementation. A consensus quickly emerged: politically
sensitive questions, including those about who influences agenda
status, and the impact of term limits, were not appropriate for an
electronic survey. Everyone agreed that such questions were not
likely to provoke very thoughtful or complete answers if asked in
an electronic survey. Policy actors, especially elected officials, are
likely to be hesitant about expressing in writing their opinions
about others in the policy system, for fear of their answers coming
back to haunt them later. Politically sensitive questions need to be
asked in a one-on-one interview, so the respondent can gauge the
interviewer's trustworthiness and ability to handle such
information appropriately.
Accordingly, the survey was shortened to three questions:
one asking for the three highest-priority problems facing K-12
educational policy makers in Colorado, one asking for the policy
proposals being generated or considered to address each of those
problems, and a general question asking for any other thoughts

about the educational policy agenda-setting process in Colorado.
The more politically sensitive questions about people's relative
influence, term limits, and other issues directly related to agenda
setting were left to the interview phase of the data-gathering
process. The disappointing return rate (2 out of 19 surveys
returned) confirmed my initial doubts about the utility of surveys
for this type of study, and hastened the scheduling of the interviews
that form the core of these data.
Interviewees were selected in order to get a broadly
representative picture of the politics of agenda setting. Most
interviewees were chosen according to two main criteria: a sample
of those in official positions of power (e.g., legislative education
committee members, senior bureaucrats), and individuals
recommended by others through a snowball sampling technique.
The snowball sampling was especially useful within the legislature;
several legislators were recommended by more than one colleague
as "essential" sources of good data. In selecting interviewees, I
attempted to keep in mind this question: Would this person,

through experience, formal position, or other factors, possess
insights into how educational policy agendas get set in Colorado?
Eighteen interviews were conducted, in two rounds of data
collection, including eight legislators, two executive-branch officials,
three interest-group representatives, a legislative staff member, a
Capitol reporter, and three individuals categorized as "experts"
because of their long experience in and around Colorado
government and educational policy making. The interviews were
conducted throughout the first half of 2000, including four in
February, five in May, and the rest spread across the spring and early
summer. The later interviewees were selected based in part on
feedback and recommendations received from individuals early in
the year, in accordance with the snowball sampling technique noted
above. Due to the disappointing survey response rate discussed
above, the survey results have been used only as a supplemental
data source, with the interviews serving as the main data set for
the study.
The interview protocol (Appendix B) was designed to engage
subjects in substantive discussions around the issues they identified
as priorities for state action in K-12 educational policy. It was also
designed to elicit comments about policy proposals to address these

issues, respondents' perceptions about the political and institutional
dynamics of agenda setting, and their thoughts about the impact of
term limits on agenda setting. The protocol proved quite useful in
stimulating productive conversation: nearly every interviewee
freely discussed the issues in the protocol, and much more.
Case Study Selection
Topics for the imbedded issue studies (or case studies)
emerged directly from the data. The first, educational accountability,
was cited as a top issue by nearly every interviewee. The second
issue study, school safety, was cited by nearly half of all respondents.
Two main criteria guided the selections: their broad acceptance or
identification as "priority" issues, and the likelihood of their
revealing or illustrating important aspects of the dynamics of
educational policy agenda setting. These two issues also have
saliency. The accountability bill, S.B. 186, was clearly the major
focus of the 2000 legislative session, generating enormous media
attention and controversy within the educational community.
School safety, following the 1999 Columbine High School shootings,
is both a salient topic and, it turned out, a good example of the

power of so-called "focusing events" to influence policy agendas
(Birkland, 1997).
A detailed discussion of the data on term limits was included
because they are influencing the political stream in Colorado, yet
they remain relatively imexamined at this point in time. The state
offers researchers a unique opportunity to examine the perceived
impact of term limits on state government and educational policy
agenda setting. Many members of the Colorado policy system were
concerned about the impact of term limits on the quality of policy
decisions, on the political culture and tone of the legislature, and on
the balance of agenda-setting power among legislators, staff,
lobbyists, and the executive branch. As will be seen in Chapters 4
and 5, the political stream is vitally important to agenda setting, and
term limits are essential to understanding the political stream in
Data Analysis
Qualitative data analysis includes at least three strands: data
reduction, data display, and conclusion drawing (Miles &
Huberman, 1994). Data in this study were reduced and analyzed in
stages. Field notes from the interviews were written up, creating

narrative summaries of each interview that also included my
observations, questions, and thoughts about themes and conceptual
linkages among data elements. Using the QSR NUDIST qualitative
research software package, an initial list of codes was created, using
the three-streams framework as a starting point. The software
allows researchers to code and manipulate unstructured data such
as text, and to generate reports by code, respondent, or by other
criteria. Some codes were stored in "free nodes," important
categories or concepts that did not immediately fit into the
framework, such as policy issues (e.g., accountability) or role groups
(e.g., legislator, media person). Some of these initial codes were then
combined as patterns emerged and critical concepts related to
agenda setting became clear from the data. Initial codes for student
achievement, testing, and performance, for example, were
combined into the code "accountability." Other refinements of the
coding system were made as well, eliminating unused or defunct
codes, and reorganizing the hierarchical "index tree" of codes
related to the three-streams framework. The final list of codes
appears in Appendix C.
Reports were generated that presented data by code,
respondent, policy issue, and role group. Following Miles and

Huberman's admonition, "You know what you display" (1994, p.
91), coded data including quotes and researcher comments were
then entered into a series of matrix displays that enabled me to see
graphically many of the themes, patterns, and relationships
contained in the data. For example, a display of data on the issue of
accountability, organized by respondent and by the three-streams
framework, indicated that most of the accountability data were
contained within the political stream. On school safety, most of the
data were contained within the problems stream, home of the
Columbine shootings. These displays were a great help in reducing
the data and facilitating analysis. Displays were created for the
problems, policies, and politics streams, for the top-priority issues,
and for different respondents and role groups within the sample.
Analytic text was then created from the matrices and the coded field
notes, beginning with a search for patterns, themes, relationships
among concepts, and questions or issues needing additional
In an effort to ensure validity of conclusions drawn from the
displayed data, conclusion drawing proceeded in steps. First, a quick
scan of the data was performed, to see what patterns and themes
jumped out. This "squint analysis" (Miles & Huberman, p. 242) led

to the creation of analytic text focusing on streams and variables
that seemed to be most relevant to a particular issue. Initial
observations were then examined in light of the original field notes,
a process of "checking back" that was repeated throughout the
creation of the final report. As conclusions grounded in one subset
of the data were refined, they were verified by triangulating against
data from other sources (e.g., different role groups and time
periods), and by mapping these conclusions to show where they
came from, where they fit, or did not, within the conceptual
framework, and what other explanations might be relevant.
Conceptual Framework
The conceptual framework for this study began as a rather
elaborate collection of ideas seen as central to an understanding of
agenda setting. It included such elements as Kingdon's "three
streams" model of the agenda setting process itself, along with ideas
taken from LasswelTs (1971) Social Process Model (SPM), and his
Five Intellectual Tasks; Schattschneider's 1960 work on the
mobilization of bias; and the "nondecision making" research of
Bachrach and Baratz (1962). A rough diagram of the initial
framework, since revised, appears below:

Political and Policy
Trends & Conditions
Who got what,
when, how?
Mobilization of
(Policy Windows)
As the research progressed, however, it became clear that the
contributions of Lasswell, Schattschneider, and Bachrach and
Baratz, although analytically powerful, were actually muddying the
conceptual waters, clouding the picture of agenda setting at the
heart of the study. That is not to say that Lasswell's notion of
"participants," which he defines as "all who interact in a social
context chosen for policy analysis" (1971, p. 24), is not important; but
it does not need to be included separately in the conceptual
framework because participants are included as part of Kingdon's
political stream.
The same can be said for Schattschneider's idea of the
mobilization of bias as an element in the policy process, or the
agenda-limiting "nondecision making" theory put forth by
Bachrach and Baratz. These are important concepts related to

agenda setting, but not central to organizing this research. They all
take a back seat to Kingdon's three-streams model of agenda setting
as a way of organizing and explaining that process. These other
notions are referenced throughout the paper, but they have been
excluded from the conceptual framework. Miles and Huberman
note that "as qualitative researchers collect data, they revise their
frameworks" (1994, p. 20), and that was the case in this study. The
"analysis of outcomes" section of the initial framework emerged as
superfluous the further along the study went; the heart of the
matter is the agenda setting process illuminated by Kingdon's
streams model, and that is the framework that ultimately was used
to organize this research:
Policy Window => Agenda Status
This simplified conceptual framework is descriptively useful,
yet accessible; its components are readily identifiable in real-life
situations, facilitating its application to the policy system under
study. The model helps us, in Kingdon's words, to discover "why
some subjects become prominent on the policy agenda and others

do not, and why some alternatives for choice are seriously
considered while others are neglected" (1995, p. 3).
As described in Chapter 2, Kingdon posits that three largely
independent "streams" run through government: problems,
policies, and politics. People in and around government come to
recognize certain problems as important, they generate policy
proposals and offer them for consideration, and they engage in
political activities within a dynamic political environment (p. 87).
These three streams are "coupled" at critical junctures, creating
what Kingdon calls a "policy windowopen for a short time, when
the conditions to push a subject higher on the policy agenda are
right" (p. 88). These windows of opportunity often emerge when the
political stream changes (e.g., a change in party control) or when a
focusing event or crisis (e.g., the release of dismal test scores) causes
a given problem to capture the attention of key policy actors.
The streams model of agenda setting proved quite useful in
analyzing the dynamics of Colorado educational policy agenda
setting. It enabled the identification of key political, policy, and
issue-related developments that affect the rise and fall of issues on
the state educational policy agenda. It also enabled the researcher to
place these developments in their proper context, showing how

elements of the three streams interact to create windows of
opportunity for agenda change. Chapter Four discusses the results of
the study in detail.

This chapter describes the results of this study, which
investigates state-level educational policy agenda setting within the
Colorado policy system. Respondents provided data on priority
issues in education, how these issues became priorities for state
action, the roles of different actors in agenda setting, and the
political developments and other factors that influence educational
policy agendas. The conceptual framework describes how events
and activities in three streamsproblems, policies, and politicscan
converge, creating a "policy window" that enables an issue to
achieve agenda status. Special attention is paid to activities in the
political stream, including the agenda setting impact of political
turnover and legislative term limits. In order to investigate
educational policy agenda setting in Colorado and to test the utility
of the streams framework, the interviews were structured around,
but were not limited to, the following four questions:

1. What are the top K-12 education issues or problems that
Colorado policymakers will be addressing in the next two
2. How and why did each issue become a priority? What
influenced this issue's rise to the top of the agenda?
3. What policy proposals or solutions are emerging to
address these issues?
4. What effect, if any, are term limits having on educational
policy agenda setting?
The data presented in this chapter show that two educational
policy issues emerged as top priorities for state action in 2000:
educational accountability and school safety. Respondents describe
how these issues became priorities through different combinations
of factors, with the rise of accountability heavily influenced by
political developments and indicators such as publicly reported test
scores, while school safety emerged largely because of the
Columbine High School shootings in the spring of 1999. Agenda
setting around these two issues is explored in some depth in the
two case studies that follow this section.
Respondents did not provide clear data on the third question,
about the policy proposals emerging to address these issues. With

respect to agenda setting, respondents simply do not distinguish
between problems and policy proposals, other than referring to the
two major priority bills in 2000 (S.B. 186 on accountability, S.B. 133
on school safety). Accountability and safety seemed to be defined in
actors' minds as amalgamations of problems and solutions.
Accountability was simply "the accountability question" or the
"accountability issue," rather than being seen as a clearly defined set
of problems with an accompanying set of proposed solutions. To the
extent that respondents made such a distinction, they simply
pointed to S.B. 186 and S.B 133 as the proposals on the table.
The results on the agenda setting impact of term limits,
presented beginning on page 101, centered on four themes. First,
respondents believe that term limits have created new incentives
and pressures on lawmakers, speeding up already tight timelines
within the legislative schedule. Second, they see term limits as
bringing new ideas into the policy process, which some see as a
positive force and some decry as detrimental to policy making.
Third, they believe that term limits are reducing legislators'
substantive knowledge of the issues. Finally, they believe that term
limits are prompting elected officials to run for other offices,
creating a game of "musical chairs" which, at the very least, is

subverting the original intent of term limits. System actors also
believe that the change to the administration of Governor Bill
Owens has altered the substantive focus of the state educational
policy agenda; they see administrative turnover as a powerful force
for agenda change.
The five sections below detail the results of this study
including a summary of the interview sample and the political and
policy context of the state of Colorado at the time of the research;
case studies of agenda setting on two top-priority issues, educational
accountability and school safety; a description of the results related
to the agenda-setting impact of term limits, and a brief concluding
section. Chapter Five includes analysis and discussion of these
results through the lens of the three-streams framework, while
Chapter Six discusses overall conclusions about educational policy
agenda setting in Colorado and suggests areas suitable for further
Interviewees and the Colorado Context
Interview subjects were chosen in order to create a sample
that is broadly representative of the Colorado educational policy
community. The 18 respondents included legislators, legislative

staff, governor's staff, other executive-branch officials, members of
the capitol press corps, representatives of major interest groups, and
other experts chosen for their long-standing experience with
Colorado state government and educational policy making. They
included balanced numbers of Republicans and Democrats, men
and women, and legislative and non-legislative observers. The
purposive sampling employed in this study enabled the researcher
to triangulate data across time, place, and persons, bolstering the
validity of the research.
These interviews were conducted, for the most part, at the
end of 1999 and the first half of 2000. The context for the 2000
legislative session could not have been more interesting for a study
of agenda setting in education: a new, education-minded governor
(elected in 1998 after 24 years of Democratic gubernatorial control in
Colorado) entering his second legislative session, a relatively new
chief state school officer and an increasingly conservative state
board of education, a GOP-controlled legislature being transformed
by term limits, a robust state economy generating strong state
revenues, a constitutional amendment limiting taxes and spending
in effect, and the memory of the Columbine High School shootings
fresh in everyone's mind. All of this created a complex policy

environment in which education was a dominant issue, but the
precise shape of the educational policy agenda remained difficult to
During the 2000 legislative session, the GOP held majorities
of 20-15 in the Senate and 40-25 in the House. Republicans professed
to be united behind Governor Bill Owens, elected in 1998, who had
spearheaded a successful drive for a variety of tax cuts during his
first session in 1999. Owens was widely credited with bringing GOP
moderates and social conservatives closer together, after years of
intra-party bickering. "I honestly feel we've got better unity of
thought and purpose than we did a year ago," said GOP State
Chairman Bob Beauprez at the start of the 2000 session (Brown,
2000a, p. IB). For their part, Democratic leaders were forced to adapt
to their new role as party agenda setters, without Democrats Roy
Romer or Dick Lamm in the governor's chair. "For 24 years, we
were able to rely on the governor's office to set an agenda. We were
never proactive in coming up with our own agenda," said Rep. Dan
Grossman of Denver, one of the Democrats' leaders in the House.
The Democrats' "Community Cornerstone" agenda, unveiled on
January 4, 2000, emphasized spending for education, health care,
and various public safety bills (p. IB).

For his part, Governor Owens began the 2000 session by
calling for "significant" reforms in public education, including a $19
million package that included expanded academic testing in core
subjects, public report cards for individual schools, increased
professional development, creation of alternative charter schools
for disruptive students, and grants for school improvement
(Brown, 2000b, p. IB). Owens also proposed eliminating tenure for
new teachers, but exempting existing tenured teachers. Senate
President Ray Powers, a Colorado Springs Republican, said that
education, guns and growth would be hot topics in 2000, while
House Majority Leader Doug Dean pushed the creation of special
state-funded math and science academies (Martinez, 2000a, p. 1A).
Substantive issues aside, the most important contextual
element influencing the legislative environment in 2000 may have
been term limits. Holders of three of the four top leadership posts
Senate President Powers (R), House Speaker George (R), and Senate
Minority Leader Mike Feeley (D)prepared to depart after the 2000
elections. The 1990 voter-approved constitutional amendment
creating term limits had allowed senators midway through a term
in 1990 to serve the remaining two years of that term, plus an
additional two terms, making 2000 a big year for legislative

turnover (Sanko, 2000, p. 12A). Preliminary jockeying for leadership
posts had already begun as the session got under way.
Within this political and policy context in 1999-2000, data
about priority educational issues and how they got on the decision
agenda were gathered from Capitol players and observers, along
with data on the agenda-setting impact of turnover and term limits.
The timing for this research turned out to be fortuitous: 1999-2000
was a period in which policy agendas in Colorado were in flux, and
the players in the game were in the process of changing
Case Study: Educational Accountability
Accountability was the most prominent issue on the
Colorado educational policy agenda in 2000, highlighted by nearly
every respondent as a top priority. The term refers to "the
systematic collection, analysis and use of information to hold
schools, educators and others responsible for student performance"
(Education Commission of the States, 1998, p. v). It is a broad
construct, encompassing policies and practices as diverse as
measuring student and school progress against established
standards, assessing teacher qualifications, and using rewards and

sanctions to spur improvement. Accountability systems can include
input measures, process measures, and outcome measures (p. v),
and often include elements directed at all levels of the systemstate,
district, school, and student. Perhaps more than any other issue in
education, the old rule "where you stand depends on where you sit"
applies; most people favor accountability, as long as it is someone
else who is being held accountable.
How did educational accountability get on the policy agenda
in Colorado? The data reveal four major catalysts for the rise of
accountability as an agenda item. The biggest driver was the change
to a Republican administration, which many respondents believe
had the greatest impact on the rise of accountability as a "front
burner" issue in 2000. "The governor is driving this, no question,"
said one Republican legislator. The partisan shift in the governor's
mansion was accompanied by increased assertiveness and
entrepreneurship on the part of GOP leaders in the legislature, who
shared many of the new governor's policy goals and had an interest
in his success. "This has been pushed from the top down," said one
Capitol veteran. A leading Democrat noted that "he moved his
party on this issue, and now it's moving." The partisan shift also
encouraged more vigorous entrepreneurship by business leaders

and other key actors, who sensed that they now had a sympathetic
ear in the governor's office. Finally, accountability became a priority
issue because of the public release of disappointing scores on
Colorado Student Assessment Program (CSAP) tests. A prominent
lobbyist noted that "S.B. 186 is being driven by low CSAP scores."
The results increased public pressure on schools to perform, and
provided political cover for the accountability push.
Educational accountability was consistently cited as a top
issue by respondents. Some used their own shorthand, referring to
"school reform legislation," "achievement," or "performance, but
all went on to describe the accountability-related characteristics of
S.B. 186, the governor's accountability package in 2000. But while
the language may have varied, the messages were remarkably
consistent: accountability for student and school performance was
the big issue in the 2000 legislative session. All but one of the
respondents listed accountability as a top priority. The provisions in
S.B. 186 are also seen as different from any seen previously. Former
governor Roy Romer was widely credited with pursuing a
standards-based reform agenda, including 1993's comprehensive
student standards bill, H.B. 1313. "The reform issues really grew out
of what had been done beforeH.B. 1313, the standards bill," said

one lawmaker. Governor Owens is seen as extending that agenda
into new territory. "This governor has put report cards, more
testing, and other issues on the table," said a legislator, discussing
the evolution of standards-based reform in Colorado. The
accountability issue provides a fine illustration of the agenda-setting
powers of governors, and the supporting roles played by other
policy entrepreneurs, both in setting the decision agenda and
shaping the policy alternatives to be considered.
The Election of Governor Bill Owens
Respondents cited the election of Governor Bill Owens as the
most influential factor in the rise of educational accountability on
the decision agenda. "It was Bill Owens all the way," said one
lobbyist. "If you look back at his legislative record, the roots of 186
are in there. And the same goes for school safety, too." More than
any other issue, accountability is seen as a gubernatorial initiative,
and Bill Owens' biggest educational accomplishmentor disaster,
depending on where the observer sits. "It's being driven by a change
to a Republican administration," said one legislator. "With
Governor Romer, you knew he had limits on what he would sign.
Now, the possibilities are greater." In other words, the change in

administrations is seen as breathing new life into tougher policy
alternatives that Governor Owens' Democratic predecessor would
have vetoed. Owens was no doubt aware that the public, in
Colorado and across the country, supported tougher accountability
measures and higher standards for students and schools. One
executive-branch official commented that politically, the state was
ready for the "next step" in the school-reform continuum, and the
change in administration enabled such a next step to be taken.
The advent of the Owens administration was accompanied by
a resurgent optimism among GOP leaders. Despite Governor
Owens' personal identification with accountability, respondents saw
the governor's party as playing a key role as well. "There was a big
opportunity to change things," said one Republican party activist.
Another GOP leader pointed to "a pent-up Republican impatience
with education and the slow pace of change. There are both
personal and party forces at work here." Majority-party legislators
saw their chance to enact tough reforms that never would have
been possible during the Romer years. One GOP leader said, "To
some extent, theres a political-ideological convergence going on,
and Owens facilitates this." GOP leaders were eager to see Owens'

education plans succeed, and to let him stand up and take the credit,
and the blame, for S.B. 186.
On December 1,1999, Governor Owens announced at a
Capitol press conference the package that would become S.B. 186:
testing of every student in reading in math, from third through
seventh grade; public report cards detailing every school's
performance; transportation assistance for poor children who wish
to attend a different public school; and a provision that would have
made it easier to fire newly hired teachers (Soraghan, 1999, p. IB).
Predictably, conservatives hailed the transportation "voucher," the
Colorado Education Association blasted the tenure proposal, and
school boards and school executives expressed guarded optimism
about the governor's package. Veteran lawmaker Norma Anderson
(R), a former House majority leader elected to the Senate in 1998,
agreed to lead the charge for the governor's plan, along with House
Education Committee chairwoman Debbie Allen (R). Accountability
was now officially an item on the decision agenda at the statehouse.
Entrepreneurship and Accountability
As noted above, Governor Bill Owens' election fueled the
rise of accountability on the agenda. The second and third factors

cited by respondents were entrepreneurial activities by the governor
and other actors in the system. More than half of respondents
pointed to gubernatorial entrepreneurship as a factor in getting an
issue on the agenda. As one legislator put it, "The governor is really
the only one who can set the agenda, although a lot of people in this
building think they can." One respondent noted the governor's
"natural advantages": he is a single actor, not one among many in
the legislature, he gains instant media coverage when he wants it,
and he is a focal point for important constituencies with whom he
shares some interests. A legislative staffer commented, "The
governor's proposal was really the only game in town from the
beginning. The governor is like the Microsoft of the whole process:
no one wants to compete with him."
The governor's leadership helped to energize his party, and
therefore, given GOP control of both houses, the legislature. "This is
really a party issue, rather than an outcry from the public," said one
Democrat. "There will be a lot of pressure on the Republicans to
accept all of his ideas." Owens was given credit for showing
leadership on student assessments, widely seen as the next step in
the standards-assessment-accountability continuum. A legislative
staffer said that the governor had stressed educational accountability

during his campaign: "The governor drove it, he publicly talked
about it, and it drove his party into action." In Colorado, the
governor is clearly viewed as having the first and last word as far as
his party's educational agenda is concerned. One respondent noted
that "His party is in power, so they have no incentive to get ahead
of him on this issue."
Governor Owens' election also enabled other policy
entrepreneurs to jump into the accountability fray. Respondents
pointed to effective action by other players, including business
leaders. "[Business] lobbyists worked very hard to educate members
(onS.B. 186). It was too long and too complicated." Once the
governor had announced his intention to make educational reform
his number one priority in 2000, business leaders, conservative
policy experts, and Republican activists frustrated by 24 years of
Democratic gubernatorial control saw their opportunity to push for
greater school accountability. "The business community is worried
about workforce quality," said one former official. One member of
the administration talked at length about how business leaders were
pressing the governor for actions that would help ensure the quality
of the Colorado workforce. "Business leaders .. are going outside
the country for skilled workers," he said. The governor even

enlisted the help of a new coalition of business leaders, modeled on
a similar group formed by Governor George W. Bush of Texas.
A less visible, but also important example of policy
entrepreneurship on this issue concerned the actions of a group of
local school superintendents. One interviewee discussed how 38
superintendents, led by the head of one of the larger school districts,
sent a joint letter to Governor Owens, urging him to implement a
more comprehensive student-testing program. "Owens took it and
ran with it, and knew it was a good issue for him," he said. An
educational-organization representative noted that many of them
probably did not expect the governor to use the opportunity to push
through school report cards and other measures to which they
might have objected; they simply wanted more testing. The
combination of Governor Owens' leadership on accountability,
concerted efforts on the part of his conservative and business
supporters, and an unexpected step by a group of local educators
helped drive the ideas that became S.B. 186 to the top of the
decision agenda.

CSAP Scores as Agenda Drivers
In addition to the change to a Republican administration and
effective actions by policy entrepreneurs, another factor that made
accountability an agenda priority was the 1999 release of the first
round of Colorado Student Assessment Program (CSAP) test scores.
"The problem was that the CSAP scores from last year were lower
than the governor, and everyone else, wanted to see," said an
interest-group leader. The mediocre test scores, splashed all over the
front pages, fueled public interest in school reform, and gave
political cover to policy makers, like Owens, who had a particular
solution in mind. The media stories that follow release of the scores
seem to carry great weight. Said one legislative staffer, "The test
scores in the paper drove politicians and everybody else toward
action. You have inch-high headlines, it becomes a campaign issue,
something gets done."
References to test scores, usually state CSAP scores, as an
agenda catalyst emerged in more than half of the interviews. Nearly
all of the respondents who mentioned test scores did so in the
context of poor performance by students and/or schools. Test scores

that failed to meet expectations were the number-one reason given
for the issue of educational accountability getting on the broader,
governmental agenda. This set the stage for the election of Bill
Owens to push the issue up to the decision agenda. One legislative
staff member summed up the prevailing view nicely: "There is a
sense that schools are not teaching students what they need to know
to be successful." To some extent, test scores were to school
performance and accountability what the Columbine tragedy was to
school safety: a widely recognized symbol of a larger problem. CSAP
scores gave form and substance to the amorphous problem of
school failure, crystallizing existing perceptions about school
quality, as well as helping to shape new perceptions. As one
executive-branch official put it, "We want real data that reflects
whether students are learning or not, and the CSAP is the key."
A related theme that emerges from the interviews is that
views about test scores tend to mirror individuals7 political or
ideological leanings. Conservatives tend to see test scores as
scientific proof that more radical proposals are needed to fix a badly
ailing school system; liberal policymakers tend to see them as
narrow, possibly flawed indicators that fail to capture any of the
complexities of schooling. One left-of-center lawmaker, referring to