Public Policy Research Paper

II. Public Administration as a Foundation for the Study of Public Policy

III. Open-Systems Administration

IV. Stages Models of Public Policy

V. Agenda Setting

VI. Policy Formulation and Adoption

VII. Decision-Making Models and Policy Formulation and Adoption

VIII. Implementation

IX. Evaluation

X. Alternative Models in the Study of Public Policy

XI. Future Directions

I. Introduction

Policy studies emerged as an important focus in political science in the 1970s. In 1969, David Easton (1969), president of the American Political Science Association, was frustrated with the trend in political science research to study narrow questions that lent themselves to the quantitative methods expected by the behavioral movement. Thus, he called for a postbehavioral revolution where political scientists would study the most important political problems of the day even when quantitative methodologies could not be employed. Easton’s call served as a catalyst for policy research that sought to explain and predict policy patterns as well as to evaluate the relative impact of various types of policy solutions.

This research paper discusses a variety of approaches to the study of public policy and shows how the public administration and public policy subfields are closely related and at times intertwined. At the time of Easton’s call for relevance, the public administration subfield had declined as a prominent subfield in the discipline. The behavioral movement had prevailed in expectations for quantitative research, and public administration had not moved toward a grand theory or wed itself to quantitative methods. However, it had gravitated toward more policy-relevant models and concepts that were important foundations for the emerging field of public policy.

One of the policy subfield’s great advantages for those interested in government and politics is its interdisciplinary and holistic focus. Research is broad ranging and borrows heavily from the work of neighboring subfields and other academic disciplines. This can be found in its use of ideas such as systems theory, plural and elite models, subsystems concepts, and decision-making research.

Over the last four decades, a number of strands of policy research have developed as the subfield has matured. This research paper uses the policy stages framework to organize most of the literature discussed. However, a few of the newer policy frameworks follow this discussion in the alternative models and future directions sections of this research paper.

II. Public Administration as a Foundation for the Study of Public Policy

Policy and administration studies have many areas of overlapping interest. Michael Nelson (1977) suggested that the popularity of policy studies was temporal and it was more a modified version of public administration than a new subfield. A close study shows that the two fields do have substantial overlap, and a review of public administration literature is important to one’s understanding of public policy.

American scholarly attention to public administration dates back to the late 1800s when Woodrow Wilson (1887) wrote his classic essay calling for the development of a career public service. An increase in the professionalism of government administration was necessary to meet the increasing quantity and complexity of government activities. He encouraged comparative study of administration and argued that since administration is distinct from politics, the United States could examine administrative practices of European monarchies without fear of undermining its democratic form of government.

Wilson’s work was written around the same time that the path-breaking German sociologist Max Weber conducted his seminal studies on bureaucracy. Weber (1946) and Wilson (1887) each posited principles of efficiency, centralized authority, hierarchical structures, educated workers, and application of expertise to administration. Weber saw the development of bureaucracy as a natural corollary to modern government and asserted that its rule-driven decisions supported the rule of law and egalitarian values of democracy. The classical model of administration was further developed by scholars who participated in the scientific management and principles schools of administration. Scholars such as Luther Gulick, Frederick Taylor, and Leonard White reinforced the view of bureaucracy as a rational, efficient, hierarchical machine. This paradigm contributed to the sense that administration and policy were conducted in separate spheres and that organizations were controlled by the administrator at the top of the organization’s hierarchy. A careful reading of the early scholars, especially Wilson and Weber, shows that they realized the line between policy and politics was not as distinct as later scholars’ attributions. Wilson (1887) discussed the need for public opinion to be a guide for administrators but also stated that administrators should have some discretionary authority. Weber (1946) cautioned that bureaucrats would use their wealth of information and knowledge to their advantage, observed that bureaucrats were likely to categorize specialized information as official secrets, and warned that an authoritative monarch would be powerless opposite an administrative expert.

III. Open-Systems Administration

By the mid-1900s, many administration scholars challenged the classical model and its primary attention to structure, formal rules, and hierarchy within a single organization. Instead, open-systems scholars discussed the influence of other systems on the political system and how changes in the environment required organizations to adapt. Philip Selznick (1949) in his study, TVA and the Grass Roots, revealed how significantly local grassroots organizations and interests can affect an agency implementing public policy. He showed how organizations have to consider threats from external organizations and interests. One strategy to lessen or neutralize the threat was co-optation. Organizations incorporated dissident parties either formally or informally into their decision-making structures. These representatives provided increased legitimacy by expanding the perspectives that made up the decision-making body. Ideally, the representative also communicated information favorably back to the external group. When necessary, agency officials changed policy requirements to reduce external hostility to their programs. Since these policy changes occurred without participation of elected officials, a more positive view of co-optation suggested that it increased the level of democratic participation at the local level. Selznick’s contributions to an understanding of the important role of external influences, implementation, and intergovernmental complexity have been significant contributions to the study of public policy.

In addition, the open-systems model encouraged thinking about organizations as organisms rather than human machines. Thus, to understand organizations, scholars need to study both formal and informal elements rather than rely on the overwhelming emphasis that the classical model places on formal structures. Chester Barnard (1938) posited that executives and scholars must seek to understand an organization’s people, customs, myths, and values as much as the organization’s structure and rules.

The open-systems model of administration continues to contribute to policy scholarship. It helps to show that policies are not self-implementing and that the administrative variable has an independent impact on the effectiveness of programs. Policy scholars are still coming to terms with the nexus of formal and informal elements of the policy process, and the institutional and constructivist scholars are currently building on the insights of Barnard and others.

IV. Stages Models of Public Policy

As the public policy subfield was developing, it relied heavily on case studies that permitted holistic examination of a single policy. These case studies suggested important generalizations about the policy process that extended the focus of policy scholars to include the examination of the political and administrative processes that preceded and followed formal adoption of policy. Very early in the subfield, a stages heuristic became the dominant model. The stages model typically identified agenda setting, policy formulation, policy adoption, policy implementation, and policy evaluation as the sequential processes of the model (Anderson, 1975; Jones, 1970). Theoretical case studies of the stages contributed to increased understanding of the policy process, especially in the areas of agenda setting, implementation, and evaluation, which had previously received less attention from political scientists than policy formulation and adoption. As useful as the stages heuristic was in organizing the policy subfield, it fell prey to intense criticism for a variety of reasons: no causal theory, insufficient research guidance, too little multistage research, insufficient hypothesis generation, imprecise prediction, and too much linearity (Sabatier, 1991). However, it should be noted that even the scholars who were most closely associated with developing the model clearly indicated that the stages often were not distinct in actual practice and their order and characteristics could be quite varied (Anderson; Jones).

More than four decades later, the stages heuristic continues to anchor a substantial amount of policy research. The processes involved in getting the attention of the government, building coalitions of support, navigating the formal processes of policy adoption, crafting implementation, and modifying the policy over time continue to be essential elements of policy study even for those who are employing other, newer approaches.

V. Agenda Setting

Studies of agenda setting have tried to discern why some issues are given serious attention by government and others are not. Even among those that do receive serious attention, the question arises as to why some issues move quickly to reach agenda status and others take much longer. Thus, policy scholars sought to delineate the circumstances that make it more or less likely for a problem to be recognized and attended to by public officials.

Cobb and Elder (1975) observed the variation in the ability of groups to gain access to public officials and argued that this access influenced whether an issue was elevated to the formal agenda of a governmental entity. By the time public policy became a serious subfield, the importance of differential access to public officials had already been explored by the work of Schattschneider (1960). His work drew attention to the uneven participation in governmental decisions. His findings, that business interests and upper classes dominated public policy participation, flew in the face of the traditional pluralists’ claims of policy openness. It also elevated examination of how governmental leaders moved some issues onto the agenda and blocked others. Two years later, another pivotal study added to this point by developing the concept of “nondecisions” and arguing that the blocking of certain issues from advancing onto the agenda is an important type of policy power that needs to be studied even if it is difficult to observe (Bachrach & Baratz, 1962). Bachrach and Baratz assert that disadvantaged groups are less likely to demand change when existing policies benefit widely accepted and powerful interests.

How a problem is defined affects whether and how public officials address it. Schattschneider (1960) was a very early voice in discussing this dimension of the policy process in his delineation of public versus private problems. To gain legitimacy and thereby earn a spot on the agenda, issues needed to be defined as public problems. When issues were judged to be inappropriate for governmental attention, there was little chance that the issue would move beyond a private issue (Eyestone, 1978). Problem-definition research also placed perception and belief systems within the study of policy. Symbolic interactionists in sociology made the foundational contribution to this approach, arguing that human beings act on the basis of meanings they attach to things rather than on factual, objective definitions. This emphasis has found new energy in work by Stone (1988) and more recent scholars using the constructivist approach (Schneider & Sidney, 2009).

The multiple streams model articulated by Kingdon (1984) is sometimes given status as an independent approach to policy studies. It examines three separate streams of activity: problems, policies, and politics. These three dimensions of the policy process progress relatively independently of one another but occasionally couple, usually as a result of a policy entrepreneur who senses the opportunity to connect a problem with a policy proposal. When the political timing is right, the entrepreneur will push the coupled problem and solution through the policy window. Speed is important since the policy window of opportunity usually does not stay open long. Thus, agenda setting is intimately tied to an available solution and a political opportunity. This model is an excellent example of how public administration has contributed to the policy subfield since it is an adaptation of the so-called garbage can model advanced by three public administration scholars in their analysis of organizational processes and choice (Cohen, March, & Olsen, 1972).

Kingdon (1984) heightened the importance of politically capable individuals who he termed policy entrepreneurs. Policy entrepreneurs are opportunists who take advantage of crises or unanticipated events to push their policy proposals. Or they might take advantage of existing coalitions and work to soften up key policy communities. Another strategy is to modify the definition of the problem in order to take advantage of potential support to successfully push the proposal through the window. Ultimately, the entrepreneur is most successful when he or she is able to take advantage of political events or the national mood to promote a good idea that is technically feasible, consistent with core values, fiscally tolerable, and politically acceptable. However, there are many instances where issues fail to achieve or retain agenda status. Reasons for this include financial cost, lack of acceptance by the public or policy elites, opposition of powerful interests, and dominance of other issues (Kingdon).

VI. Policy Formulation and Adoption

Policy formulation focuses on the drafting and consideration of proposals by an attentive policy community while policy adoption refers to the passage of the proposal through the formal institutions that have the authority to adopt the policy. This part of the larger policy cycle received the most scholarly attention prior to the development of the policy subfield. It is closely associated with the institutions of government and is typically considered the key decision in the policy process.

In the multiple streams model, it is evident that the first three stages of the policy cycle—agenda setting, formulation, and adoption—take place more or less concurrently. Policy formulation itself focuses on the cognitive analysis and politics of designing a statute or other type of policy decision. The issue of rationality was a major focus of administration scholars in the middle of the 20th century. However, the debate between those arguing for rationality and against it continues. Today, rational choice theorists argue that a rational choice model is the best heuristic for studying, understanding, and predicting the outcomes of policy participants (Ostrom, 2007). The competing view is that scholars need to analyze language and political calculation to better understand the formulation and adoption processes (Allison, 1971; Halperin, 1974; Kingdon, 1984; Schneider & Sidney, 2009; Stone, 1988).

Many models incorporate both rational and political elements. Perhaps this is because the process itself incorporates both rational and political elements. Using a bounded rationality approach, policy analysts develop policy options, predict impacts, and oversee evaluation studies. But on the other side of the equation, political brokers approach the political landscape strategically, using data and analysis as well as other tactics to secure majorities at each stage in the adoption process.

One of the areas of greatest interest to policy scholars is the advocacy and policy design role of executive-legislative-interest group networks. Early scholars of political science advanced a number of models related to interest groups. One of the most noteworthy was David Truman (1951), who concluded that significant interests would organize to influence policies of concern to them. According to Truman and other pluralists, government’s role was often one of facilitating and legitimating group compromise. In addition, pluralists suggested that executive agencies often participated in the bargaining, protecting their interests similar to any other group.

Additional models of executive-legislative-interest group relations preceded the development of the policy subfield and more have been created since. All of these models include groups inside and outside of government and portray coalition formation organizing around a specific policy interest. Numerous terms have been assigned to these models, including iron triangles, subgovernments, policy subsystems, issue networks, policy monopolies, and advocacy coalitions. Each of these versions of policy communities assumes the importance of relationships among actors interested in the policy issue. These models are not exclusively used in the American political context. They have been broadly applied to research in multiple political systems by scholars in many disciplines. In the context of the United States, however, these policy communities thrive in part as a result of the decentralized and fragmented structure of its political institutions. Given the inability of the president and Congress as a whole to be informed and active participants on all issues, those issues that are not receiving widespread public attention and news coverage tend to be left to communities of people who have a deep interest in the specialized policy area.

In the 1960s, these communities were thought to be very stable relationships. In several case studies, the relationships were analyzed to be so stable that they were referred to as iron triangles (Cater, 1959). This conveys a pejorative view of the activity largely because the relationships among the governmental entities in the triangle and the interest groups were characterized as closed. New interests who wanted to communicate a competing set of demands and proposals were not acknowledged. Freeman (1955), in his studies of Indian affairs, showed how these policy subsystems brought stability to a policy area by bridging the executive and legislative branches with key interest groups. Each of these entities needed good relationships with the other two to enhance the likelihood of advancing its institutional and policy goals. This mutual dependence meant that once the subsystem emerged, its accommodations became the basis for determining which issues were placed on the agenda and which ones were not. These patterns of accommodation could be broken up by changes in committee leadership, widespread media attention to an issue in the policy area, or presidential interest. Once the visible interest in the policy topic waned, the subsystem political dynamic was likely to reemerge.

Hugh Heclo (1978) viewed the relationship between policy and administration as a vitally important one, especially as the role of government grew. He suggested that in the search for iron triangles, scholars sought to discover an exceptionally powerful and autonomous executive-legislative-interest group cluster of actors who dominated policy making by policy area. In the process, he believed that policy scholars ignored the more open and more commonly existing webs of people he referred to as issue networks. Issue networks are composed of those who are knowledgeable about the issue in terms of substantive knowledge as well as the history of its policy twists and turns. Policy knowledge is more heavily emphasized in this model, and it is the primary means through which additional participants can join this fluid web of relationships. In addition to being more fluid and episodic than subgovernments, issue networks, as shared-knowledge groups, have more points of view and conflicts than the iron triangle and subsystem frameworks. As aspects of the policy debate change, so do participants in the network. Thus, iron triangles and subsystems may still exist in some policy areas at some points in time, but the more typical pattern is one of a looser, increasingly complex kaleidoscope of policy (Heclo, 1978; Meier, 1985; Sabatier, 1991). For administrators and legislators, it provides a less stable and less predictable arrangement but one that permits greater maneuverability as well since these skilled politicians have the potential to split, expand, and recombine the many segments of the issue network. The irony of the issue network is that, when compared with iron triangles, it expands the number of participants involved, thereby making policymakers contend with more conflict among multiple points of view. It also accepts that involvement in the policy process is contingent on a greater understanding of the increased complexities of the policy area. Heclo (1978) suggests that this pattern of knowledge-based participation may actually increase the cynicism of the general public as the gap between activists and the public expands.

Heclo’s (1978) work was a pivotal change in the subfield’s orientation toward the nature and behavior of policy participants. Since Heclo, numerous scholars have articulated a variety of more diverse webs of policy participants, from Meier’s (1985) regulatory subsystem to Sabatier’s (1988) advocacy coalition framework. In the final analysis, though, whether one uses a looser, more episodic network or a more centralized and stable subsystem, the general public has virtually no role in any of the models. Given the level of specialized learning necessary to truly engage in the conversation among participants in the policy process, the general public’s role is minimal unless the issue somehow ignites widespread interest.

VII. Decision-Making Models and Policy Formulation and Adoption

In the early literature on decision making, rationality was elevated to a normative standard. Operations research during World War II contributed to an expectation for clear, measureable objectives, extensive research, and, ideally, evaluation of choices based on evidence. The rational-comprehensive model that emerged ideally required the clarification and prioritizing of objectives, followed by identifying a comprehensive range of options for achieving the ranked objectives, analyzing the capacity of each option to maximize ranked objectives, and choosing the alternative that best achieves the objective at the least cost. This model has continued to retain its normative appeal, but scholars have repeatedly shown that actual decisions are not made this way. Lindblom (1959) argued that the model breaks down in its first step since people find it exceedingly difficult to agree on the relative priorities of values and goals. In the public realm, policymakers must wrestle with conflicting values among the various participants involved. Even the thinking of a single participant often includes conflicting and unresolved priorities among values. Thus, these values are not usually clarified and rank ordered prior to designing a policy. Therefore, policies often embody conflicting values. Lindblom suggested that most policy is made following a process of successive limited comparisons. Analysis is truncated to a few feasible alternatives that are incrementally different from existing practice, and choice is made based on which option receives consensual support. Pluralistic preferences and bargaining processes fit the incremental model well. Furthermore, this model can be expanded to a more intentional and strategic process for achieving substantial change since it is usually easier to achieve several successive incremental changes in policy over several years rather than attempting to secure support for major change in the first instance.

One of the most important contributors to the theories of decision making is Nobel prize winner Herbert Simon (1957). He argued that there are many reasons that rational-comprehensive models are not possible. First, information is lacking and people are not likely to be able to identify all possible alternatives. Second, the ability to accurately predict the outcomes of the many possible alternative choices is unlikely. And third, humans do not have the cognitive capacity to know and remember all that is required by the comprehensive model. Given these limitations, human beings accept what Simon terms satisficing. Under this approach to decision making, it is only necessary to find a solution that meets goals at an acceptable level rather than an optimal level.

Graham Allison’s (1971) classic work analyzing the Cuban missile crisis shows how three different models of decision making each lead to very different explanations of the crisis. His rational actor model can be compared to Simon’s bounded rationality model in that the goal is to choose the alternative that advances the national interest. The rational actor model assumes that nations function as centralized unitary actors where policy choices are made to maximize the national interest. Allison uses chess analogies to emphasize the strategic elements of the choice equation. Thus, the language of the model is of an optimizing effort, but the model recognizes the knowledge limitations and additional uncertainties that constrain choice. The advantages of this model are that it is simplified and stabilizes dramatically the information one needs to make choices. Morton Halperin (1974) emphasizes that given the number of bureaucratic departments and bureaucrats involved in decision making, it is difficult to confirm the unitary actor assumption of the model. The rational actor model assumes that all policymakers agree on the interpretation of the national interest. Halperin argued that this level of unity on what constitutes the national interest is rare in American history. One period followed World War II when the emergence of the cold war and the fear of communism solidified views of the national interest. This type of consensus also emerged for a short period after September 11, 2001.

Allison’s (1971) second model examines governmental decision making as a function of the output of organizations. The organizational process model, which borrows heavily from public administration literature, argues that organizations serve as the primary actors in governmental decision making. Given their hierarchy and centralization, each organization functions in a unified manner alongside other organizations. Under this model, analysts need to consider the outputs of multiple organizations. These outputs are the result of standard operating procedures, and organizational choices are in line with Lindblom’s model of incrementalism. The best predictor for what organizations will produce in the future is to examine the status quo. Importantly, organizational decisions and behaviors are influenced by existing routines and values of the organization.

The third model Allison (1971) advanced was the bureaucratic politics or governmental politics model. This model is probably the one most closely attributed to Allison since it was the most original of the three. Under the governmental politics model, there are numerous individual participants who influence governmental choices and behaviors. Government decisions are more resultants rather than choices since decisions are a combination of actions by numerous participants in the process. Many of the participants become part of the process as a result of an organizational affiliation that they have, and they often take actions based on the values and objectives of their organization using the action channels their organizations provide them. As each of the participants takes actions advancing their personal and organizational interests, they may be involved in overt bargaining with other participants. Equally often, though, participants are inclined to take the actions permitted by their position in the system. The governmental decision is really the interaction and summation of all of these independent decisions rather than coordinated intention. Obviously, this model of decision making is not based on a single set of organizational or national values. Pluralistic values and actions dominate the process.

Just as public administration scholarship was important to the understanding of agenda setting, it also plays a central role in formulation and adoption processes. Both the subsystem and network approaches and the decision-making approaches rely heavily on public administration literature and concepts. The level of bureaucratic participation in policy making becomes even more the focus of policy making in the implementation of public policy.

VIII. Implementation

Implementation includes the administrative activities that convert a statute or other authoritative policy into a functioning program. Traditionally, implementation was characterized as a simple process of following the directives in the statute, administrative rule, executive order, or court ruling. Since early studies assumed that this took place without much delay or discrepancy from the policy’s intent, relatively little attention was paid to this process prior to the emergence of public policy as a subfield.

One study that stands out for being well ahead of the development of the policy subfield was sociologist Phillip Selznick’s (1949) study of the Tennessee Valley Authority. This case study is one of the few studies that provided guidance to the complexity of implementation and the degree to which the contours of a policy could be significantly altered based on the decisions of implementers. Since his study focused on organizations as organic systems that adapt to their external environment, his findings created an awareness of how external forces cause policies adopted in Washington to be altered in the field. Selznick showed how organizations were dependent on local support and how local opposition from powerful groups generated adaptations in policy.

Selznick’s (1949) study also challenged the traditional top-down view of implementation. Under the traditional top-down framework, bureaucrats function as instruments of the policymakers and respond to the command and control of those above them in a process similar to hierarchical lines of authority in a bureaucratic organization. However, scholars soon realized that implementation processes were much more complex and evolutionary than initially thought. Numerous studies showed that implementation was not faithful to the original plan for a variety of reasons. First and foremost, statutes and other formal policies often did not communicate with 100% clarity. There were gaps, overlaps, and contradictions. Thus, even if a bureaucrat was willing to function as a Weberian machine, it was typically the case that bureaucrats had to make significant decisions throughout the process of designing implementation. Even Woodrow Wilson (1887), whose work contributed to the policy–administration dichotomy framework, realized that all administrators operated with some discretion. Although the policy–administration dichotomy was successfully characterized by the public administration literature as unrealistic, policy scholars assumed the concept in their top-down implementation models until sufficient implementation case studies accumulated to convince most that the paradigm was inaccurate. Although there are still a variety of approaches to the study of implementation, most would agree that statutes are altered at least to some extent in the process of implementation. Policies are not self-implementing.

Another complexity in the implementation process that affects whether implemented programs align with statutes is that many national programs are carried out at the state and local levels, which dramatically increases the number of policy actors involved. This increase in the number of decision points creates huge difficulties for timely implementation and necessary communication and coordination. Pressman and Wildvasky (1973) masterfully show that as the number of participants and decision points increase, the likelihood of accurate implementation declines. Two-way interaction models were a reaction to the inadequacies of the top-down models. Bardach (1977) and Lipsky (1979) posit that statutes and other authoritative decisions made by policy-makers interact with decisions of implementers to create the actual policy. As implementers work through the issues of how to carry out the statute, they make choices that modify the policy. In addition, “street level” bureaucrats make use of discretionary authority as they engage in the day-to-day work of public policy. Accepting that implementation is an interactive process rather than a command and control hierarchy dictated by legislators focuses attention on the distance between a policy as originally designed or passed and the policy as implemented.

Majone and Wildavsky (1984) take the interactive model one step further in their work “Implementation as Evolution.” It posits that having the implementing agency modify the original policy design may actually produce beneficial results relative to the original goals of the statute. For example, if the original statute makes assumptions about cause and effect that are inaccurate, then having implementing bureaucrats modify the policy through implementation strategies would be a positive step as long as the changes they are making are in line with a commitment to the objectives and impacts the policy was supposed to produce. Both the implementation as evolution and the other two-way interaction approaches make it obvious how important it is to have agencies that are committed to the policy objectives.

Building on the work of all of these scholars, Mazmanian and Sabatier (1983) develop a synthetic conceptual framework to guide implementation studies as part of the larger policy process. They assert that it is necessary to examine the statute or other formal policy statement to assess its clarity of objectives, accuracy of causal assumptions, and effectiveness of decision rules provided to the implementing agencies. In addition, the researcher must appraise the sufficiency of financial resources, the specificity of the authority relationships, and formal access by outsiders. Variables beyond the statute and outside the implementing agency are also emphasized, including socioeconomic conditions, media attention, public and constituency group support, enthusiasm of higher level public authorities, and the commitment and skill of agency leaders. In addition to internal and external variables, there is also a recognition that implementation success is affected by the difficulty of the problem and the degree of change being sought.

Given the potential for significant adjustments to policy to occur as a result of implementation decisions, a diachronic approach of studying policies over a decade or more is recommended. Without a long-term view, the ability to understand the evolution of policy as it occurs in implementation will be incomplete.

IX. Evaluation

Most stage models of public policy end with evaluation, the systematic assessment of the policy’s impact. Importantly, policies need to be examined for both intended and unintended consequences. Evaluations are completed by the implementers themselves and by external policy analysts. Internal evaluations conducted by the implementing agency have the benefit of getting those who work with the policy on a day-to-day basis to recognize problems and propose solutions. Evaluations by outside analysts tend to bring external perspectives to the process and may possess increased legitimacy from the perspective of elected officials. One trend in the last several decades is for more ideologically affiliated think tanks to publish external policy analyses. These are usually provided with a particular point of view and thus do not necessarily offer the benefits of other external evaluations.

Distinctions should also be made between the evaluation of policy outputs and policy outcomes. Analysis of outputs draws attention to whether the administrative processes are in place through such measures as number and types of clients served. Outcomes refer to whether the policy is achieving desired results on policy goals as well as other unintended impacts.

For policy scholars, impact and outcome studies have received more interest than output studies. Scholars have also sought to understand how policy analysis leads to revisions, transformation, or termination of policies by elected officials. As could be seen in the previous discussion of implementation, revision of policy is often ongoing from the beginning of implementation decisions. However, most stage models have viewed evaluation as isolated from implementation and as more associated with the judgments of policy-impact recommendations for policy revisions. There is a tendency for this stage to look something like the feedback loop of the political systems model.

Systematic policy analysis grew as an important part of the policy process during the 1960s as government attempted to apply economic theory to policy making. Cost–benefit analysis, operations research, and various program performance measures incorporated in budgeting processes were primary tools of the effort. Thus, this part of the policy process is most closely associated with rational models of decision making. It was not unusual for program evaluation and periodic reporting to be statutorily required to assist in legislative oversight and budgeting processes. As the national government expanded the number of large domestic programs, along with detailed prescriptions and administration, collected evidence mounted that the programs were not achieving their lofty goals. Accountability continues as a common refrain today, but the capacity to systematically evaluate programs and to redesign them with effective performance measures has not been as easy or as successful as hoped. Once again, it is possible to see the tension between expectations for rational decision making as the vehicle for better policies and the reality that suggests there are significant human and organizational limitations to rationality. Even when systematic analysis is required, the uncertainty surrounding appropriate measures and the interpretation of results make evaluation as much politics as science.

Deborah Stone (1988), one of the leading critics of rational models of public policy, advanced a concept she termed policy paradox. She argued that politicians typically have dual goals: policy objectives and political objectives. Furthermore, she observed that analysis itself is political since it is rife with framing, definitions, and interpretations. Political participants frequently articulate an argument that on its face appears linear and rational but on closer examination appears constructed to achieve a political purpose. She offered the concept of political reasoning, rather than rational decision making, to understand the struggles of policy communities competing over which ideas, policy definitions, and corresponding solutions will prevail.

A closely related facet of the evaluation literature is the attention given to the use of knowledge generated through policy analysis. Generally, scholars have concluded that knowledge is not the most important dimension of policy decisions. Politics and the limitations of human and organizational capacity intervene (Simon, 1957). Even when decision makers seek information to help make policy, it is frequently questioned, interpreted in different ways, expensive, and incomplete. Although many write very pessimistically about the lack of use of policy knowledge, Carol Weiss (1977) argued that policy knowledge has been successfully used to identify problems, reconsider policy strategies, and provide an enlightenment function. Through the enlightenment process, information accumulates and causes policymakers to redefine the problem, retroactively make sense of why programs did not succeed, or readjust policy objectives to more realistic levels. Finally, policy knowledge is more likely to be used when it is timely and when participants see strategic benefit in doing so.

X. Alternative Models in the Study of Public Policy

In 1991, Paul Sabatier criticized the stages heuristic and challenged the field to develop better models. Sabatier and Jenkins-Smith (1993) put forward an alternative heuristic referred to as the advocacy coalition framework. They suggested that the stages model is too simplistic given the number of participants, institutions, and influences in the policy process. In their more complex model, they incorporate governmental and nongovernmental institutions, external conditions and events, and the multiples webs of these entities that align or compete with each other to influence policy. Furthermore, whereas the stages heuristic signals that policy follows a linear process, the advocacy coalition framework sees policies as nonlinear and rarely terminating. Multiple policies affecting the policy domain overlap and affect one another, as do policy implementations at multiple levels of government. Researchers who study a policy arena for several decades will observe policy adaptations and eventually significant change. This occurs because of learning on the part of the participant coalitions that may compete over preferences based on different resources and belief systems. However, the advocacy coalition framework posits that the most substantial changes in policy are more likely to come from external events and conditions than from policy learning. To be able to witness this, scholars must be prepared to follow a policy area for several decades since advocacy coalitions tend to be relatively stable. Some scholars (Lester & Stewart, 2000) see the advocacy coalition framework as a development within evaluation research that could be incorporated within the stages model, but Sabatier and Jenkins-Smith (1993) claim that the model provides an important and enhanced alternative model.

Another model of note is the punctuated-equilibrium framework (Baumgartner & Jones, 1993). This model is similar to the advocacy coalition model in terms of the need to look at policy over long periods of time since most policy follows incremental change and relatively stable patterns over many years before incurring major change. Major punctuations occur in the context of changed beliefs among the policy community and often a new venue for implementation. Once implemented, equilibrium is reestablished, and the policy arena goes back to a lengthy period of stability. The punctuated-equilibrium model also draws attention to the importance of institutions since they tend to help ingrain the results of the dominant coalition and to resist pressures for change.

Elinor Ostrom (2007) and many others articulate a preference for a model of institutional rational choice. This model builds on the discipline’s traditional emphasis on institutional structures and rules. Ostrom argued that institutions should be defined as rules, norms, and strategies that characterize entities with repeated processes. Rules dictate who has advantages in the pursuit of policy preferences and frequently determine who the major players are in the policy. This model also examines the hierarchical ordering of rules. Thus, constitutional rules influence options for players at other levels of policy choice all the way down to the operational, day-to-day decisions of policy actors. Rules can take the form of formal, written provisions or can be norms based on shared understandings by participants or guides that individuals develop to direct their own behaviors. These working rules help to provide stability in the midst of uncertainty. Once the study of levels and types of rules is achieved, the researcher can make probabilistic predictions using rational choice analysis. This model brings attention back to the formal versus informal distinction that was made by administration scholars many decades earlier.

XI. Future Directions

One must ask whether any of the models will rise to a position of dominance in the near future. Given that scholars agree that the policy process is complex and patterns of human behavior are varied, the search for a dominant, robust, parsimonious model is unlikely to be successful. This effort is also exacerbated by the number of contributing disciplines to the subfield. Frustration over this situation is also likely to continue. Perhaps Kenneth Meier (2009) captures it best when he complains that there are so many models, their ability to guide research is analogous to the interstate highway system’s ability to guide a vacation. A more positive point of view may come from the realization that there is a lot of overlap among the various models. For example, the advocacy coalition framework, stages framework, punctuated equilibria model, and the institutional rational choice model all describe the importance of institutional structures, political brokers, external influences, and shared meanings.

Looking at policy processes more holistically and studying policies over the long term to incorporate evolution and change are recent trends in policy research that are likely to continue in the future. The role of beliefs and the processes leading to shared meanings also appear to be increasing in importance (Schneider & Ingram, 1997). It remains to be seen whether future studies integrate policy and public administration models. The literature of the two subfields has numerous parallels. Very few policies can be implemented without the significant involvement of an administrative system. The fact that the same policy can produce such different results across different implementing subunits (e.g., states, counties, or schools) suggests that agencies and individuals make a huge difference. Leadership styles and administrative cultures may offer vital insight to explaining the variation (Hicklin & Godwin, 2009; Robichau & Lynn, 2009).

Given the large number of models and participating disciplines, the methodologies employed in the subfield will continue to range from qualitative to quantitative. The quantitative methods employed in large-N comparative studies and some evaluation studies are likely to continue, as are the more qualitative methods of case studies and language analysis approaches. This is not necessarily a negative. As Easton (1969) argued in his call for a postbehavioral revolution, the important objective is to be relevant even if it means sacrificing the quantitative methods called for in the behavioral movement. Thus, a mixture of quantitative and qualitative measures is likely to continue to characterize the policy subfield.

Bibliography:

1. Allison, G. T. (1971). Essence of decision: Explaining the Cuban missile crisis. Boston: Little, Brown.
2. Anderson, J. (1975). Public policy making. New York: Praeger. Bachrach, P. S., & Baratz, M. S. (1962). Two faces of power. American Political Science Review, 56, 947-952.
3. Bardach, E. (1977). The implementation game. Cambridge: MIT Press.
4. Barnard, C. I. (1938). Functions of the executive. Cambridge, MA: Harvard University Press.
5. Baumgartner, B. D., & Jones, F. R. (1993). Agendas and instability in American politics. Chicago: University of Chicago Press.
6. Cater, D. (1959). The fourth branch of government. Boston: Houghton Mifflin.
7. Cobb, R. W., & Elder, C. D. (1975). Issue creation and agenda building. In J. Anderson (Ed.), Cases in public policy making (pp. 10 21). New York: Praeger.
8. Cohen, M., March, J., & Olsen, J. (1972). A garbage can model of organizational choice. Administrative Science Quarterly, 17, 1-25.
9. Easton, D. (1953). The political system. New York: Knopf. Easton, D. (1969). The new revolution in political science. American Political Science Review, 63, 1051-1061.
10. Eyestone, R. (1978). From social issues to public policy. New York: Wiley.
11. Freeman, J. L. (1955). The political process: Executive bureau legislative committee relations. New York: Random House.
12. Gulick, L., & Urwick, L. (Eds). (1937). Papers on the science of administration. New York: Institute of Public Administration.
13. Halperin, M. H. (1974). Bureaucratic politics and foreign policy. Washington, DC: Brookings Institution.
14. Heclo, H. (1978). Issue networks and the executive establishment. In A. King (Ed.), The new American political system (pp. 87-124). Washington, DC: American Enterprise Institute.
15. Hicklin, A., & Godwin, E. (2009). Agents of change: The role of public managers in public policy. Policy Studies Journal, 37(1), 13-20.
16. Jones, C. O. (1970). An introduction to the study of public policy. Belmont, CA: Wadsworth.
17. Kingdon, J.W. (1984). Agendas, alternatives and public policies. Boston: Little, Brown.
18. Lester, J. P., & Stewart, J. (2000). Public policy: An evolutionary approach. Belmont, CA: Wadsworth.
19. Lindblom, C. E. (1959). The science of muddling through. Public Administration Review, 19, 78-88.
20. Lipsky, M. (1979). Street level bureaucracy. New York: Russell Sage Foundation.
21. Majone, G., & Wildavsky, A. (1984). Implementation as evolution. In J. L. Pressman & A. Wildavsky (Eds.), Implementation (3rd ed., pp. 163-180). Berkeley: University of California Press.
22. Mazmanian, D. H., & Sabatier, P. A. (1983). Implementation and public policy. Glenview, IL: Scott, Foresman.
23. Meier, K. J. (1985). Regulation: Politics, bureaucracy & economics. New York: St. Martin’s.
24. Meier, K. J. (2009). Policy theory, policy theory everywhere: Ravings of a deranged policy scholar. Policy Studies Journal, 37(1), 5-12.
25. Nelson, M. (1977, September). What’s wrong with political science? Washington Monthly, 9(7), 12-20.
26. Ostrom, E. (2007). Institutional rational choice. In P. A. Sabatier (Ed.), Theories of the policy process (pp. 21-64). Boulder: CO: Westview Press.
27. Pressman, J. L., & Wildavsky, A. (1973). Implementation. Berkeley: University of California Press.
28. Ranney, A. (1968). The study of policy content: A framework of choice. In A. Ranney (Ed.), Political science and public policy (pp. 3 21). Chicago: Markham.
29. Rivlin, A. M. (1971). Systematic thinking for social action. Washington, DC: Brookings Institution.
30. Robichau, R. W., & Lynn, L. E., Jr. (2009). The implementation of public policy: Still the missing link. Policy Studies Journal, 37(1), 21-36.
31. Sabatier, P. A. (1988). An advocacy coalition framework of policy change and the role of policy oriented learning therein. Policy Sciences, 21(2), 129-168.
32. Sabatier, P. A. (1991, June). Political science and public policy. PS: Political Science & Politics, 24(2), 143-147.
33. Sabatier, P. A., & Jenkins Smith, H. C. (1993). Policy change and learning: An advocacy coalition approach. Boulder, CO: Westview Press.
34. Schattschneider, E. E. (1960). The semi-sovereign people. New York: Holt, Rinehart & Winston.
35. Schneider, A., & Ingram, H. (1997). Social constructions in the study of public policy. In J. A. Holstein & J. F. Gubrium (Eds.), Handbook of constructionist research (pp. 189-211). New York: Guilford Press.
36. Schneider, A., & Sidney, M. (2009). What is next for policy design and social construction theory? Policy Studies Journal, 37, 103-120.
37. Selznick, P. (1949). TVA and the grass roots. Berkeley: University of California Press.
38. Simon,H. (1957). Models of man: Social and rational. New York: Wiley.
39. Stone, D.A. (1988). Policy paradox and political reason. Glenview, IL: Scott, Foresman.
40. Taylor, FW. (1911). Scientific management. New York: Harper & Row.
41. Truman, D. (1951). The governmental process. New York: Knopf.
42. Weber, M. (1946). Bureaucracy. In H. H. Gerth & C. W. Mills (Eds. & Trans.), From Max Weber: Essays in sociology. New York: Oxford University Press.
43. Weiss, C. H. (1977). Research for policy’s sake: The enlightenment function of social research. Policy Analysis, 3, 531-546.
44. White, L. D. (1926). Introduction to the study of public administration. New York: Macmillan.
45. Wilson, W. (1887). The study of administration. Political Science Quarterly, 2(2), 197-222.