Making the most of the frame: developing the analytical potential of frame analysis

ABSTRACT Frame analysis has been developed and applied across contexts in several disciplines such as policy analysis, where the perspective has proven fruitful to carve out essential differences in the construction of meaning and to understand the responsiveness of the strategic use of ideas. However, this article argues in line with other scholars that the analytical potential of frame analysis is not fully utilized in most empirical studies. The article addresses two points of critique raised against frame analytical perspectives: the limited view of the framing process and the limited understanding of frame effects. We suggest two analytical dimensions that help to develop the analytical potential of frame analysis in policy analysis and beyond: firstly, the institutionalization process of frames which can capture the struggle of meaning within policy processes and also distinguish between the varying influences of different frames over space and time. Secondly, the extension of frame effects that through a reconceptualization of frame effects can capture how a frame has an effect on actors other than the audience and beyond its immediate purpose.


Introduction
How can we analytically capture framing processes in which certain policy frames gain influence as others lose their dominance? And how can we make sense of the fact that certain frames produce effects even though they have not been effective? These questions arise from two empirical case studies of Swedish prostitution policy (Erikson 2011) and the Nepali civil war and peace process (Björnehed 2012). While these two studies are empirically diverse, they nevertheless share a common theoretical point of departure in that both employ frame analysis in order to understand the construction of meaning and its effects. They also share a point of frustration concerning the lack of analytical tools in existing approaches that are needed in order to answer the questions raised above. This article aims to address this problem.
Frame analysis is an analytical approach within the constructivist tradition that addresses not only the construction of meaning, but also the roles of actors in such processes. It has been used to investigate ideas and discourses in such varied fields of research as framing in media and communications (De Vreese 2012;Entman 1993;Scheufele 1999); individual actors' cognitive and psychological frames (Kuhnberger 1998;Levin, Schneider, and Gaeth 1998); framing in social movements Gamson 1988; Morris and Mueller 1992;Snow and Benford 1992); frame conflicts in international relations (Drake and Donohue 1996;Mintz and Redd 2003); gendered frames (Verloo and Lombardo 2007;Verloo, Lombardo, and Meier 2009); and framing in policy processes (Schön and Rein 1994;Stone 1989;van Hulst and Yanow 2014). The strategic use of ideas has also been of interest to scholars who utilize frame analysis, including such issues as how strategic framing influences public opinion (Chong and Druckman 2007), assists in recruiting members to social movements (Snow and Benford 1992), and contributes to successful election campaigns (Rhee 1997).
However, the scope and versatility of frame analysis can also be a shortcoming, and the approach has faced criticism in this regard for being conceptually unclear (Benford 1997;Entman 1993;Vliegenthart and van Zoonen 2011). Entman called for greater conceptual precision over two decades ago (1993), while Benford argued some years later that 'frame analytic methods remain underdeveloped ' (1997, 412). More recently, Vliegenthart and van Zoonen have remarked that 'frame/framing articles contain a cacophony of new definitions, divergent operationalizations and a wide, often incompatible range of empirically established content features' (2011).
Although there have been a number of important contributions with respect to the development of tools that are useful in frame analysis (De Vreese 2012; Kangas, Niemelä, and Varjonen 2014;Matthes 2011;van Hulst and Yanow 2014), our position is that there continues to be a need for a further improvement in frame analysis with respect to two of its key aspects, namely, (1) the framing process and (2) frame effects.
Our intent in the present discussion is to contribute to this endeavor. We will draw upon insights provided by our previous empirical research, existing frame analytical approaches, institutional theory, and policy studies in the effort to outline two new analytical dimensions that may be useful for frame analysis in various fieldsthe frame institutionalization process and the expansion of frame effects.
The theoretical contributions provided by these new dimensions reside upon (1) a more detailed understanding of the framing process, in which frames gain or lose influence depending upon their level of institutionalization and (2) a more comprehensive understanding of framing effects, which are taken as analytically distinct from their effectiveness.
The discussion below is structured as follows. First, a theoretical foundation is presented concerning the need to develop the two key aspects of frame analysis noted above. Second, each of these facets of frame analysis is discussed and analytical tools are specified for both of them. Finally, these analytical tools are applied separately in two different empirical case studies in order to demonstrate their usefulness. Note that these studies are presented here not in order to provide new empirical findings, but rather to illustrate the usefulness of these new analytical tools in our previous research.
Two key tasks of frame analysispoints of critique and a way forward Frame analysis often involves two key analytical tasksuncovering the construction of meaning within a given context, or the framing process, and identifying the effects frames have on, for example, actors' behavior and political outcomes. Snow and Benford's frame analysis, Entman's framework, and Schön and Rein's approach all include these two analytical dimensions in different ways. We will discuss these two issues in order to describe the limitations of currently existing approaches and indicate how we propose to ameliorate them.
With respect to the framing process, most frame analyses utilize the analytical tools of diagnosis and prognosis in order to describe the problems and solutions associated with the various frames at stake within a given context (Schön and Rein 1994;Verloo and Lombardo 2007). While such analyses have often addressed frames on a number of different levels that range from individual issue frames to broad meta-cultural frames, they have frequently focused on the discursive content of frames rather than on the framing process itself. Although such discursive analysis is an essential element of frame analysis, criticism from within the field has pointed to a tendency to conclude analyses with a description of frames (Carragee and Roefs 2004;Reese, Gandy, and Grant 2010;Snow and Benford 1988). This limits the potential of frame analysis as a method for capturing processes of meaning construction. Benford warned against what he termed a static tendency in framing literature as early as 1997 (1997), while Benford and Snow later argued that, from a social constructivist perspective, the element of analytical interest is not the frame per se, but rather the process of framing (Snow and Benford 2000, 3). The focus should therefore be on 'human interaction, discourse, and the social construction of reality' (Benford 1997, 420), that is, on the framing process as a 'set of dynamic, negotiated, and often contested processes' (Snow and Benford 2000, 3).
Merlijn van Hulst and Dvora Yanow adopt the same point of departure in a recent article in an attempt to render analyses of framing more dynamic and political (2014). But although their discussion constitutes an important contribution to the literature, it is primarily concerned with the framing act, not the process, which is the key issue in Snow and Benford's critique.
We propose that combining frame analysis with an institutional perspective provides a means for analytically capturing the framing process over time. Even though institutional assumptions are a fundamental element of frame analysis, they have not been the most analytically prominent element in existing approaches. We argue in this regard that it is useful to take into consideration the degree to which the frames being analyzed have been institutionalized insofar as this serves to capture the process in which frames gain or lose influence and new meaning is established.
The second aspect of framing that we will address concerns frame effects, which has traditionally been understood within frame analysis as a specific and predetermined effect involving an identified audience Buzan, Waever, and de Wilde 1998;Entman 1993;Levin et al. 2002). For example, Benford and Snow discuss the mobilization of a target audience as a frame effect, while securitization theory, which is arguably a particular form of frame analysis, examines an audience's acceptance of the use of extraordinary measures by the political leadership as an effect of a security frame (Buzan, Waever, and de Wilde 1998;Weaver 1995). There has recently been a growing interest in frame effects with respect to their importance for gaining an understanding of the social impact and importance of frames, with studies investigating such issues as the duration of frame effects (De Vreese 2012) and the effects of various types of frames (Levin et al. 2002;Kangas, Niemelä, and Varjonen 2014). However, this interest has also served to illustrate that frame effect research is regarded as underdeveloped, which has prompted a general call for more empirical studies of the dynamics of frame effects (De Vreese 2012).
We agree with this critique and argue that the conventional view of frame effects as predetermined with respect to given audiences creates a premature endpoint in an analysis of frame effects, which restricts both the type of frame effects that can be captured and the means available to do so. The reason for this particular conceptualization of effects, along with its analytical shortcomings, can be traced to the analytical troika of success, resonance, and audience that comprises the basic fabric of most frame analysis frameworks. In order to remedy this problem, we propose that the effects of frames should be analytically distinct from their effectiveness.
We will first turn our attention to an examination of the frame institutionalization process before addressing the expansion of frame effects.

Frame institutionalizationcapturing framing processes
Scholars working in the theoretical fields of policy studies and new institutionalism have emphasized that ideas can be institutionalized to varying degrees (cf. Hajer 1995;Schmidt 2008;Schneider and Sidney 2009, 106), and that the related processes are marked by path dependency (Cox 2004;Hay 2006). However, they have seldom specified or operationalized the character of these processes (Erikson 2015). We take this theoretical insight as a point of departure in this article since institutional perspectives often comprise an implicit element of frame analysis insofar as it is assumed that frames can become part of the context, exert an influence, and generate effects. Schön and Rein emphasize in this regard that 'interpersonal discourse must have an institutional locus within some larger social system ' (1994, 31), and that frames are grounded in the institutions and discourses that sponsor them (1994, 31;Snow and Benford 2000, 628). Existing versions of frame analysis also draw upon institutional assumptions in that they acknowledge the existence of several levels of frames with differing influences and functions, such as is the case in Schön and Rein's framework regarding policy frames, institutional action frames, and meta-cultural frames (1994, 33;cf. Snow and Benford 1992, 38). Nevertheless, there is no analytical model that ties these various levels to each other and captures the dynamic framing processes in which a new understanding of an issue may become the dominant frame. More precisely, existing analytical frameworks have not specified how frames ascend or descend from one level to another and thereby gain or lose influence.
Consequently, we argue that taking the level of frame institutionalization into consideration provides a fruitful means for capturing the dynamic framing processes in which new meaning is created. Moreover, this perspective makes possible the development of tools that grasp not only how a given frame acquires increased influence, but also how other frames simultaneously lose theirs. In order to advance frame analysis in this direction, we explore the frame institutionalization process as an analytical dimension that provides a heuristic model capable of illustrating the collective framing processes in which the influence and status of frames change with respect to space and time. This dimension is also important for distinguishing between the anticipated effects of different frames (on different levels) in that the influence and effects of a given frame depend upon its level of institutionalization. Although institutionalization is often used in the literature in reference to institutional change at the macro level, frame institutionalization is addressed in the present discussion as the process in which a frame gradually gains influence and regulative functions. For example, a particular frame possesses virtually no regulative functions when it is merely held by an individual actor, but it begins to influence how actors behave and express themselves on the issue in question when it comes to being dominant and advocated by many individuals.
While specific contexts influence in different ways how frame institutionalization processes are operationalized, it is important to note that an institutional setting is a common denominator for all contexts. Since all institutional contexts share certain common features, the analytical model that we present here is relevant for most settings insofar as it is possible for a given frame to be institutionalized to differing degrees within a given institutional environment, both formally and/or informally (cf. Schmidt 2008).
The four-step ladder outlined below illustrates how a frame institutionalization process can be operationalized within the context of formal decision-making, although the analytical approach it represents can be readily applied in other institutional settings as well. Note that these steps provide a set of analytical tools for capturing policy processes that may advance either gradually or abruptly. It is also significant that although the model has certain features in common with the policy formalization process, its main analytical focus concerns discursive shifts in how policy issues are constructed, not how support for a certain (fixed) policy proposal ascends or descends within the policy process.

The frame institutionalization ladder
(1) Reaching the political agenda. The first critical step is for a frame to reach the political agenda and be explicitly discussed in relevant venues. (2) Support from a coalition of actors or key actors. At this step, either a coalition of actors or a few key actors express the frame, and various actors advocate either for it or its core elements. Such actors include ministers and spokespersons of political parties in the cases we present below as examples of frame institutionalization. (3) Official acknowledgment. This can occur either formally, when a particular frame is expressed in such official statements as commission reports or governmental directives, or informally, when all actors who participate in a given debate acknowledge a particular frame by relating their statements to it in one way or another. (4) Formal institutionalization. The fourth step is for a frame to become expressed in formal institutions, which often involves legislation with respect to policy.
Capturing the influence of different framings of a given issue within a particular policy process involves analyzing the discursive constructions and changes over time concerning that issue within the argumentation expressed by individual policy actors and utilized in debates, policy documents, and reports. This requires an analysis of all the competing constructions of a given issue expressed at different points in time insofar as many framings of an issue may be articulated simultaneously. In addition to traditional frame analysis, which is concerned with discursive differences between distinct frames, this model identifies discursive shifts over time as framings ascend or descend from one level of institutionalization to another. Although it is clearly possible to imagine a fifth or sixth step of institutionalization that would capture the even broader influence of a particular frame, such as complete consensus in its support or the rendering of other frames as virtually unthinkable, the four steps above suffice for an analysis of policy outcomes. One limitation of constructivist perspectives, including frame analysis, is that the associated analytical models usually do not contain factors related to actors' preferences, interests, and values that may help explain how there can be different framings of one and the same issue. These models instead focus primarily on policy discourse and meaning construction. The institutionalization ladder is useful for analyzing the dynamics of collective framing processes in terms of the temporal sequences of discursive changes, including how early events shape later outcomes. Scholars working in the field of new institutionalism have observed that the concept of path dependency is particularly fruitful when examining the roles of ideas in social and political processes insofar as ideas are always path dependentthey are influenced by previous ideas rather than being entirely new products (Cox 2004;Hay 2006).
Furthermore, tracing the movement of frames from one stage to another constitutes a tool for identifying critical junctures in a given process. For example, ascending to a higher level, which is in itself an important change in the process that alters subsequent development, may at times be the result of a formative event or strategic action that can be further analyzed through other approaches.
The logic of appropriateness provides a useful mechanism for gaining an understanding of the mechanisms at work in such processes (March and Olsen 2008) inasmuch as the social pressure of what is regarded as the 'appropriate' way in which to understand an issue that can lead actors to change their view and framing of a given issue. At certain times this is instrumental in character, while at others it reflects a substantial change in actors' core beliefs. In addition, the discussion concerning social learning provides useful insights into how actors may change their understanding of a given policy issue as they engage in negotiations and accommodate the views of others in a policy framing process (Berman 2013;Hall 1993).
In summary, this tool assists us in garnering insights into the dynamics of collective framing processes that lead to the creation of new meaning. It is also useful in analyzing the broader effects of various frames insofar as the potential effects of a given frame are closely related to its degree of influence.
We will illustrate how this approach can enrich and develop analytical frame perspectives by describing the way in which it was employed in gaining an understanding of changes in Swedish prostitution policy. Before doing so, however, we will address the second aspect of frame analysis that we have noted above, namely, the expansion of frame effects.

The expansion of frame effects
The theoretical foundation of frame analysis presumes the involvement of rational actors, which is arguably a strength of the framework. However, a rational actor perspective is associated with a predisposition to focus on the strategic use of frames, which tends to reduce potential research questions concerning frame effects to questions about the success of a frame with respect to the constructed self-interest of the framing actor Schimmelfennig 2001;Blyth 2003). As a result, the notion of effects is reduced to a single dimension, which can be summarized as 'Does a specific frame succeed in gaining support from a predefined audience?' Since what is regarded as success is clearly context dependent, frame analysis is particularly useful within such actor-focused research areas as policy studies. Nevertheless, this focus on success renders the notion of frame effects one dimensional.
Resonance with an audience constitutes the mechanism that links an attempt at framing with success, regardless of what is meant by success. Benford and Snow, who regard resonance as connected with frame effectiveness, argue that 'the concept of resonance is relevant to the issue of the effectiveness or mobilizing potency of proffered framings, thereby attending to the question of why some framings seem to be effective or "resonate" while others do not ' (2000, 619). However, a more in-depth look at the concept of resonance reveals a lack of analytical clarity such that the category of effectiveness becomes empirically fused with that of audience effect. For example, frames are associated with a specific effect in Benford and Snow's analytical framework (activism) and in the securitization framework (consent to extraordinary measures), and both of these particular actions on the part of the target audiences are linked with frame resonance. The resonation of an attempt at securitization within the audience is regarded as equivalent to the acceptance of the use of extraordinary measures, and the resonation of a social movement frame within an individual is regarded as equivalent to that individual becoming mobilized.
But this fusion of the concepts of effectiveness and effect hinders analysis because it constructs an analytical endpoint, so to speak, concerning how we should study the effects of particular frames. Viewing effects as effectiveness does not challenge, but rather enforces an analytical focus on whether or not a specific frame succeeds in producing a predefined effect in the target audience.
In contrast, we argue that it is more fruitful to distinguish between effectiveness (resonance) and effect (actions influenced by the frame). 1 Taking these two concepts as distinct makes it possible to address broader issues concerning frame effects by allowing for a theoretical expansion of the investigation that not only includes actors other than the specific audience, but also raises questions about possible effects that cannot be foreseen, such as the unintended consequences of framing. Stated briefly, frame analysis possesses greater inherent potential for analytical nuance than merely producing binary outcomes concerning success or failure.
Our proposed solution for the reduction of the scope of frame effects to binary outcomes involves the expansion of frame effects upon the basis of an alternative definition that views frame effects in terms of perceived maneuverability (Björnehed 2012). This definition makes it possible to develop analytical frameworks within frame analysis that are capable of incorporating not only effects upon the audience, but also effects upon any actors who are related to the frame.
The inclusion that this new conception allows for is important because a theoretically constructed co-dependent relationship between success and audience, on the one hand, and resonance and effect, on the other, generates a preconceived endpoint in the framing process that often masks the empirical effects a given frame produces even in the absence of resonance. As a result, a number of effects upon actions in the social world that are traceable to a specific frame would be entirely missed by this type of truncated analysis (Björnehed 2012). This theoretical reconceptualization of frame effects also reaffirms the actor-centered approach of frame analysis. In addition, it is consistent with the social constructivist foundation of frame analysis in that one of the central characteristics of ideas is that there is no ownership of them as suchonce an idea is formulated and propagated, it acquires a life of its own and can produce effects outside any specific segment of the population. Furthermore, this way of thinking about frame effects increases the fruitfulness of frame analysis within policy research insofar as it provides a framework for investigating policy effects beyond resonance and effects upon actors other than the audience.

Empirical demonstrations of two new analytical dimensions
The frame institutionalization ladder in policy-making We have argued that tracing frame institutionalization processes comprises a new analytical dimension that is useful for capturing framing processes as such. The empirical case study of Swedish prostitution policy mentioned above was characterized by a framing process in which the understanding of prostitution in Sweden gradually changed from the 1970s, when prostitution first appeared on the parliamentary agenda, to 1998, when the purchase of sexual services was criminalized (Erikson 2011;. In this section, we will illustrate how tracing frame institutionalization processes provides valuable insights into the framing process associated with client criminalization that are not available to other approaches, including the gradual discursive changes that take place when the influence of a given frame increases or decreases. For example, a social framing enjoyed strong as well as broad support in both legislation and the parliamentary debate at the beginning of the period noted, while a competing criminal framing was expressed at that time by only a handful of women in parliament and by a small number of extra-parliamentary women's organizations. At the end of the period, however, the power relation had become reversed and the criminal framing was institutionalized in a new prostitution policy. Before we provide examples of how the criminal frame ascended from one stage to another in the institutionalization process, we will briefly describe the content of the two frames in question. In the social framing that was dominant during the 1970s and 1980s, prostitution was framed as a societal problem caused by the individual's social situation, such as drug abuse or exposure to sexual abuse, as well as by such broad social structures as class inequality and patriarchy (Erikson 2011(Erikson , chapter 3, 2012. The existing legislation supported this view, and neither selling nor buying sex was criminalized. The government's stance since the 1960s regarding women in prostitution was that society should only provide supportive measures for their protection, mainly within the social services and healthcare sector, and that compliance with such measures should be voluntary (Govt. Bill 1981/ 82:187;SOU 1981:71, 55). A common view at the time, which was highlighted in the government's 1977 prostitution inquiry, was that 'the importance of social measures to decrease prostitution can hardly be overestimated' (SOU 1982:61, 120).
While the competing criminal framing of prostitution shared the overarching construction of prostitution as a problem in itself, it nevertheless described prostitution primarily as an expression of gender inequality. It viewed prostitution as an unacceptable phenomenon with such negative consequences for gender equality in society at large, particularly for the women involved, that buying sex should be regarded as a criminal act. The cause of prostitution within this framing is particularly significant insofar as the individual men who buy sex were viewed as responsible for the existence of prostitution. This was completely different from the social framing, which regarded the cause as non-personal, with society being responsible for the problem of prostitution (cf. Debate Protocol 1983/ 84:152).
The criminal framing reached the political agenda of the parliament at the beginning of the period indicated, but it was expressed only by a very small number of female MPs in their speeches and motions (the first step in the institutionalization model). The first important discursive shift was evident at the beginning of the 1980s, when a significantly larger number of MPs from different political parties in parliament framed the problem in criminal terms both in motions and in the debate in the chamber. For example, Kerstin Anér (Liberal party) argued in support of client criminalization that [I see] that kind of criminalization as a desirable, public preventive measure aimed to reduce prostitution and as a marker that society does not allow that type of contempt for women. (Debate Protocol 1983/84:152) By 1985, the women's sections of four out of the six political parties in parliament expressed support for a criminal framing (Erikson 2011, 90-108), which marked the ascent of the criminal frame to the second level of institutionalization. Although the influence of the social framing remained strong at this point, this ascent made possible further changes in favor of the criminal framing. The criminal framing gained further influence by the end of the 1980s, when two standing committees framed prostitution in their reports in terms of a gendered problem of unequal power relations and described male clients as the cause of the problem. For example, the Health and Welfare Committee stated in 1987 that This constituted a significant change from the social framing that parliamentary committees had previously expressed in their official reports.
Committees were thus more positive than they had been concerning criminalization as a solution for prostitution, but they were still not convinced on the issue. However, this discursive change meant that the client criminalization frame had formally attained official acknowledgement (step 3), at least with respect to problem framing. It is significant that the criminal frame had also gained informal recognition in the debate at this venture, with the opponents of criminalization responding to its advocates in a new way. Criminalization arguments were taken much more seriously than they had been previously, and the opponents of criminalization felt it necessary to devote time to discussing why criminalization was not a feasible or efficient solution instead of simply dismissing the claim, as had been the case (Erikson 2011, 98).
The fact of such formal and informal recognition was used by female MPs in support of criminalization as a stepping stone to the appointment in 1992 of a new prostitution inquiry that was assigned the task of investigating the feasibility of criminalization as a measure to be used to combat prostitution (SOU 1995:15). This inquiry was to be decisive for the final advance of the criminal frame to the fourth step on the institutionalization ladder, which was expressed in the 1998 policy change that resulted in client criminalization. 2 The influence of the social framing decreased as the criminal framing advanced, which also can be analyzed as an institutional process. The social framing went from the highest step on the institutionalization ladder (step 4), being expressed in law, to an increasingly marginalized position that was questioned by parliamentary committees and supported by ever fewer actors. This reversed institutionalization process concerning the social framing was noted informally before it was expressed formallywhich is often the caseinsofar as policy changes lag behind discursive changes.
Previous research indicates that gradual discursive shifts were imperative for the final policy outcome in that they conditioned not only the course of the process, but also the actors involved (Erikson 2011, 165). However, it is also necessary to address the strategic roles of actors, particularly how they managed to gain support for a new framing through strategic work, in order to fully explain the final outcome.
Frame analysis has previously not provided the tools needed to understand the dynamics and shifts of the collective framing process described above. Existing approaches stop short at describing the frames at stake on different levels, and cannot account for the dynamic processes in which relations between frames change. We argue that the institutionalization ladder provides a tool that is useful for capturing the dynamics of framing processes, including identifying important discursive shifts and revealing the path dependencies of ideas. It is also fruitful in a broader analysis of framing effects insofar as a frame's level of institutionalization affects the frame effects that can be anticipated. The model has already proven productive within the context of policy processes (Erikson 2015) and in explaining institutional change (Erikson 2017), but it can readily be operationalized in different ways and adapted to different contexts.

The expansion of frame effects in the case of the Nepali civil war
We will use two examples from an empirical study concerning the recent Nepali civil war and peace process to illustrate how reconceptualizing frame effects in terms of perceived maneuverability makes it possible to capture effects that a conventional conceptualization of frame effects cannot, such as the unintended consequences of framing as well as effects upon actors other than the audience.
The conventional approach to frame effects directs analysis towards the audience with whom the frame resonates, and it considers resonance to be a prerequisite for a frame to have an effect. In cases where the object of frame analysis is a specific issue, such as migration, abortion, or deforestation, it is likely easier to regard a framework that has a restricted focus upon success-audience-resonance as adequate.
However, the need to look beyond the audience for important frame effects becomes apparent when the object framed is not an issue, but rather an actor who is not a member of the audience. For example, a recent study of the Nepali civil war (Björnehed 2012) investigated the effects that the terrorist label had on the Maoist guerrillas, which were not accessible to a conventional frame analysis since the Maoists were not the audience of the framing attempt. As such, they could not be addressed within a conventional analytical framework because whether they accepted or rejected the frame (resonance) was supposedly of no consequence for determining the success of the framing attempt. A conventional framework could only focus on whether or not the terrorism frame resonated with the relevant audience, which it arguably did since it resulted in the parliament's adoption of the Terrorism and Ordinance Law. Nevertheless, it was not possible from a conventional perspective to examine whether and how the terrorism frame had an effect upon the Maoist guerillas.
However, the terrorism frame clearly did have a profound effect on the Maoists, who related in interviews how they felt forced to counter the frame, in both words and actions, in order not to be regarded as terrorists (Dahal 2010;Gurung 2007;Magar 2010).
Oh, yes well because you know, why would we call ourselves the terrorist? Nobody likes to be called a thief, or criminal or corrupt man or something like that when you are a political force and to have that very tag … [trails off]. 3 But this process within the Maoist party was not simply a frame dispute. In the interviews, the Maoists repeatedly justified their course of action in relation to the terrorism frame, even though they simultaneously denied that it had exercised any effect upon them (Dahal 2010;Magar 2010;Shreshta 2008). It is thus clear from the Maoists' own statements that their perceived maneuverability was determined by the terrorism frame in the political discourseit had an effect upon them insofar as they deliberately sought to act in opposition to it (Björnehed 2012).
The effect that the frame had, which could be demonstrated with respect to how the Maoists acted and justified their actions, was obviously not tied to the success or failure of the terrorism frame from a conventional perspective. The Maoists clearly refused to give any credence to the terrorism frame, and it would be absurd to maintain that it resonated among the Maoist cadres. It is also doubtful that the terrorist frame was successful among the wider electorate since support for the Maoist movement in fact grew after the attempt was made to define them as a terrorist organization (Thapa 2002, 84).
Briefly stated, redefining frame effects in terms of perceived room for maneuver, rather than with respect to success with a given audience, reveals the effects frames have other than those made visible by analytical frameworks based upon the conventional conceptualization of an effect.
Another type of effect that cannot be grasped by the conventional conceptualization concerns the unintended consequences of framing insofar as conceptualizing effects in binary terms of success or failure creates a predetermined analytical endpointresonance does or does not occur in the audience, and the intended effect does or does not materialize. An analysis of terrorist frame effects in Nepal from the conventional perspective could then conclude that the framing attempt was successful since a law was passed, and it would be possible to determine whether the frame had resonated among the wider population by conducting a large-scale survey. The analysis could conclude at this point, regardless of the results, since there is no analytical space within the conventional conceptualization of effects to ask what happened afterwards. Although a new frame analysis could obviously be performed, doing so would be interesting only with respect to a new or competing frame, such as a counter-frame expressed by the Maoists.
Nevertheless, the inability of existing frameworks to provide continuity with respect to a failed framing attempt would create an empirical blind spot in the frame chronology, where what happens next is not associated with the original frame. Moreover, focusing on the success or failure of a frame obscures any effects it might have upon actors who are not members of the audience. In contrast, regarding effects as perceived maneuverability makes it possible not only to delay the endpoint in frame analysis, but also to include other actors in the analysis since this broader definition does not limit an effect to a given audience. It is thus possible not only to investigate actors who are the object of a given framing attempt, such as the Maoist guerillas who were to be labeled terrorists, but also to analyze the actors who are responsible for the framing.
Directing the spotlight upon the framing actors, which cannot be done if we accept the conventional conception of effect, is particularly useful for capturing the unintended consequences of framing. We will illustrate this point by a brief discussion of the framing that marked the Nepali peace process.
There were two parties to the peace process in Nepal, the Seven Party Alliance (SPA) and the Maoists, both of whom were involved in framing attempts concerning themselves, their visions of what peace in Nepal involved, and the policies they sought to implement in order to realize those visions. In essence, two versions of a peace frame existed in the country in 2006, and while they shared such central elements as peace and democracy, different priorities were placed upon the latter by the SPA and the Maoists. The outspoken priority of the SPA was the re-establishment of democracy through the restoration of parliament, the causal logic being that reestablishing a legitimate government (democracy) was necessary for the resolution of the conflict (Arjun KC 2008;Jhala Nath Khanal 2010;Ram S. Mahat 2010;Sher Bahadur Deuba 2010). Stated otherwise, the SPA regarded bringing peace to Nepal as a two-step process that consisted of, first, the restoration of democracy, which had been suspended by the king on 1 February 2005, and second, the subsequent negotiation of a peace agreement with the Maoists.
The Maoists' version of a peace frame presented the opposite logic, whereby peace had to be established before democracy could be realized. The Maoists also had a more clear and detailed vision than the SPA of what peace entailed, which included the key demands they had made since 1994 concerning elections for a constituent assembly and the establishment of a republic. Their position was that the formation of a constituent assembly would pave the way for the social changes they sought to implement (Hisila Yami 2010). The Maoists thus established a clear discursive connection between peace and their own principal demands.
It was the Maoist peace frame that finally prevailed (Björnehed 2012)their peace frame came to dominate the political landscape, and they transformed their decadelong demands into policies. Representatives of civil society as well as international observers both concluded that the Maoists in fact dominated the peace process (Malla K. Sundar 2010;Markus Heiniger 2008).
A conventional frame analysis would stop at this point by virtue of the resonance that was established with respect to a clearly dominant frame. The peace frame of the Maoists had been successful and their policies were well on the way to being implementedthe king was ousted and a constituent assembly had been formed. But if one understands a frame effect as perceived maneuverability, and includes the actors who performed the framing in the analysis, a starting point is established from which to uncover the unintended consequences of the peace frame and thereby attain a deeper understanding of the behavior of the various actors. Investigating the effects produced by the peace frame from the perspective of the SPA makes these points clearly visible.
The SPA had not simply framed peace in terms of democracy followed by conflict resolution, they also made the effort to define themselves as peacemakers by orienting their statements in the media in defense of their actions concerning the negotiation process towards their commitment to peace. For example, when former Prime Minister Koirala was criticized for his decision to meet with the Maoists on foreign soil (India), he retorted that 'I will even go to hell if necessary to bring peace in the country' ("We Had Face-to-Face with Maoists" 2005). It is clear from the interviews conducted with top-level SPA politicians that the respondents considered it important to portray their party as committed to peace and present themselves as advocates of a peaceful solution, even though Koirala and the Nepali Congress (NC) had previously pushed for a military solution to the conflict and supported labeling the Maoists as terrorists. The SPA became heavily engaged in portraying themselves as peacemakers, and they placed the highest priority upon bringing peace to Nepal. But because of the success of the Maoist peace frame, the SPA found themselves tied to a notion of peace in which they did not believe, but were nevertheless forced to supportthey had to abandon their monarchical sympathies and accept a republican constitution. They theoretically had the option to fight the vision and policies put forward by the Maoists, but the SPA could not oppose the Maoist frame in practice since they feared that doing so would bring a fate worse than changing the stature of their partybeing viewed by the public as against peace (Prakash Sharma Mahat 2008).
This sentiment was voiced by Bimalendra Nidhi, General Secretary of the Nepali Congress Democratic (NCD), who stated that In 1990, during the peace movement, was not for the abolition of the monarchy, but to bring back the monarch within the constitution. At that time if anybody spoke about the abolition of the monarchy, they were branded as extremist and anti-democracy, not helping the people's cause. But now in this movement [Jana Andolan IIauthors] if anyone spoke a single word in favour of the monarchy they would be branded anti-peace. This is there. 4 The experience of the SPA was that the members of their own parties, civil society, and the Maoists all held them to their own self-definition as peacemakers. As a result, the parliamentary parties felt that they would be branded as anti-peace if they proposed political alternatives that fell outside of the Maoist-constructed peace frame, particularly with respect to policies concerning the role of the monarchy (Gyawali 2008; Mahat 2010; Prakash 2008). The SPA had thus become trapped by their own frame.
Being able to investigate how framing actors decide what courses of action are and are not open to them within a particular discursive context makes possible a broader and more clear understanding of frame effects. While previous frameworks could only examine a limited range of effects, conceiving of frame effects in terms of perceived maneuverability opens up a wider range of possibilities concerning what motivates behavior, regardless of whether framing is performed strategically or occurs unintentionally. Since this permits investigation of a more inclusive set of effects, and involves both the audience and other actors as well, researchers are able to access new aspects of frames through frame analysis, such as their unintended consequences, and address new issues.
Operationalizing the concept of frame effects as perceived maneuverability provides frame analysis with a new analytical dimension and a broader range of application.

Conclusion
In this discussion, we have drawn upon our previous empirical research in order to develop two new dimensions of frame analysis and thereby contribute to enhancing the value of frame analysis for political and social analysis. We found that existing approaches were inadequate when we applied frame analysis to the cases of Swedish prostitution policy and the Nepali peace process since they lacked the concepts and tools needed to capture the changing influences of frames over time and the effects of frames beyond those predicted.
We agree with other scholars that frame analysis has been in need of development and analytical clarification, and we have focused in this regard on the framing process and on frame effects. We have proposed two amendments to frame analysisthe frame institutionalization process and the expansion of frame effectsthat may be particularly fruitful for studies in the field of policy analysis concerned with the construction of meaning and its effects.
The frame institutionalization process is operationalized when frames ascend or descend from one level to another, which leads to changes over space and time in their influence. This analytical dimension is imperative for gaining an understanding of how different processes of meaning contend with each other in different settings. We have provided a four-step model of this process within the context of formal decision-making that can be readily employed in other settings as well by adapting the steps indicated to the particular features of a given context. It is perhaps easier within the formalized context of the parliament to specify the various stages of frame institutionalization than it is in more informal contexts, where institutionalization processes tend to occur in a more informal manner. In most contexts, however, legitimization, support by key actors, and official recognition will likely be relevant even if they are expressed in somewhat different terms.
This dimension also facilitates placing the collective processes involved in the construction of meaning, which existing approaches have not been able to analytically specify, at the center of analysis. This contribution to capturing the dynamics of framing processes constitutes an advancement in the capabilities of frame analysis, even though it is not exhaustive and cannot fully capture the dynamics of complex and multifaceted framing processes. The latter task will require the use of additional analytical tools in order to reveal, for example, frame constitution mechanisms and the roles played by various actors in the overall process (see Erikson 2015).
The importance of an expanded conceptualization of frame effects is evident in comparison with the limitations of the conventional notion of frame effects, which cannot capture effects on actors other than the audience and thereby overlooks such analytically significant aspects of framing as unintended consequences. The assumption that the participants in framing processes are rational actors has led to a focus on the success of frames, while the traditional conceptualization of frame effectiveness (resonance) has been analytically fused with frame effect. This creates a predetermined endpoint in the analysis of frame effectswhether or not a frame has resonated with the audience becomes the generic research question when looking for frame effects. We have argued that this type of restriction prevents an investigation of frame effects outside of the audience, even though doing so could help explain how ideas affect actors.
Distinguishing between effectiveness and effect through a reconceptualization of effects in more general terms as perceived maneuverability removes any previously constructed endpoint from the analysis and expands the horizon of frame analysis with respect to frame effects. The conceptualization of frame effects that we have presented in the present discussion, with examples based upon the Nepali peace process, is applicable to a very broad range of empirical contexts.
The two new analytical dimensions that we have presented in this article can be applied separately in individual studies that address the process of framing or frame effects, as we have illustrated, or they can complement each other in a more comprehensive approach to the creation of meaning. They are also fruitful for furthering our understanding of policy processes. Tracing the frame institutionalization process provides insight into how ideational changes precede policy changes and why certain policy attempts stop short and fail, while a broader understanding of frame effects is productive for understanding the strategic actions of policy actors in a more comprehensive way.

Notes
1. This article defines frame effect as perceived maneuverability for action in order to separate it from effectiveness in terms of frame resonance. Similar separation between effect and effectiveness can be found in other fields, for example, in research on policy instruments and legitimacy through differentiating between substantive and procedural legitimacy, respectively (see Scharpf 2002Scharpf , 2003. Although both procedural and substantive policy instruments are argued to have an effect on the behavior of policy actors, i.e. have an effect on action (see Howlett 2000, 414, 418) making the distinction between effect and effectiveness less clear.

Notes on contributors
Emma Björnehed is a researcher and university teacher at the Tactical Warfare Division at the Swedish Defence University. Her research interests include security on the strategic, operational and tactical level and methodological issues. She defended her PhD, an analysis of the Nepal conflict and process from a frame analysis perspective, in 2012 at the Department of Government, Uppsala University, Sweden.
Josefina Erikson is a researcher and university teacher in the Department of Government, Uppsala University Sweden, where she defended her PhD thesis in 2011 on Swedish prostitution policy. Her research interests include educational policy, gender and politics and methodological issues. In 2017 she published the book Criminalizing the client in which she engages with processes of gendered institutional change.