Skip to navigation – Site map

HomeIssues4:2Multiple Book Review Essay

Multiple Book Review Essay

Aaron Smith-Walter
p. 241-252
Bibliographical reference

Luke, Allan, Educational Policy, Narrative and Discourse, New York and London: Routledge, 2019, 303 p., ISBN 9781138502932. Hardback: £96.00, $42.00; eBook: £31.99, $42.36

Merry, Melissa K., Warped Narratives: Distortion in the Framing of Gun Policy, Ann Arbor, MI: University of Michigan Press, 2020, 225 p., ISBN 9780472131662. Hardback: $74.95; eBook: $59.95

Miller, Hugh T., Narrative Politics in Public Policy: Legalizing Cannabis, Cham, Switzerland: Palgrave Macmillan, 2020, 248 p., ISBN 9783030453190. Hardback: €58.00; eBook: €46.95

Full text


1The three books under review in this essay all analyze narrative to examine three areas of public policy. Specifically, educational policy, gun policy and cannabis policy, which are Luke’s, Merry’s, and Miller’s respective areas of interest. There are a number of shared sympathies among each author’s approach to narrative. First, they all acknowledge the importance of the study of narrative in public policy development and implementation (though the authors differ somewhat in the methods used). Second, they highlight the need for policy scholarship to effectively engage with the context in which the specific narratives circulate. Third, the authors acknowledge the importance of understanding the origins and interests behind the generation of particular narratives.

2Of course, there are also important differences in the approach to narrative taken be each author. Given the brevity of book reviews, I will focus on the most prominent difference, namely, the generalizability of findings from the study of narratives in one context to another context. I believe the crux of the disagreement between the authors whose works are reviewed in the following text is a manifestation of the larger schism between public policy scholars who lean toward more “interpretivist” approaches and those with more “positivist” inclinations (see, Jones & Radaelli, 2015). This schism, I believe, is one based on divergent conceptions of causality in social science research. In the course of this review, I hope to address both the content of the texts and point to an underlying agreement about the contributions that the study of narratives can make toward our knowledge of public policy.

3This essay will be structured in three parts: first, I will review the project and overall scope of each book. Second, I will identify similarities in the approaches each author brings to using narrative in their work. Third, I will focus on the differences in the scholars’ understanding of narrative, and how these differences might be reduced by incorporating concepts from a critical realist perspective. This is done to suggest a broader and more flexible understanding of causality in social science research than has hitherto characterized the debate surrounding the appropriate manner for conducting narrative study in social science research.

4The main thesis of this review essay is that each of these books uses narrative to undertake an exploration of a salient public policy area. While each author’s project is a valuable contribution to scholarship surrounding education, gun, and cannabis policy, respectively, the use of narrative in each of these works bespeaks an underlying unity in policy scholarship which hangs its hat, so to speak, on the analysis of narrative. As Durnová and Weible (2020) have recently pointed out, there appears to be a gulf between “interpretive policy scholars” and “mainstream policy scholars” with regard to what an analysis of narrative actually reflects, and the type of claims that can be legitimately derived from narrative studies. This review essay will explore these differences. However, it will above all, seek to identify the shared assumptions across each work which might suggest to the reader, that the distance between the two traditions is, as Durnová and Weible note, narrower than is often thought.

Educational Policy, Narrative and Discourse


5Luke’s book is the longest text of the three. As a collection of writings stretching between 1996 and 2017, it showcases a complex evolution of scholarly work which was written while serving in various positions as an academic, public servant, university administrator, and world traveler. In several chapters of the book, Luke relies explicitly on critical discourse analysis (CDA) to effectively deconstruct and decode various educational policy texts. He does an admirable job not just applying CDA, but doing so in a fashion where those new to the approach can effectively follow along with the texts. In so doing, he shows how discourse constructs various subjectivities, perpetuates gender-roles, and communicates to the children appropriate relationships to authority figures within those institutions in which they are socialized (family, school, and church).

6As Luke notes, the power of critical discourse analysis is its, “capacity to show the power relations of apparent mundane texts as work…” (p. 39), and his early chapters apply this approach with great effect, especially its use in Chapter 3, where it is applied to a speech made by Australian Prime Minister, John Howard, to the Australian Conciliation Commission in 1997. Ostensibly, this speech was supposed to feature an apology to the Aboriginal and Torres Strait Islanders for the actions (it is not too strong here to call them crimes, I think) of the Australian state as they related to “the issues of ‘stolen children’ and land rights” (p. 66) and how the actual text worked to shift blame, minimize damages, and remove any sense of responsibility from the Australian state to create a historical account “of social action without responsible actors” (p. 81). This is also the chapter in which Luke takes the strongest position in which critical discourse analysis is “openly counter-scientific” (p. 71) and states that it is a “situated political practice: a machine for generating interpretations and for constructing readings, none of which are neutral or unsituated” (Luke, 2019, p. 72).

7This chapter marks the high-tide of Luke’s “anti-science” proclamations, and later chapters take a more nuanced approach. I want to suggest that this is not because Luke ever abandons what, in my opinion, is the most important part of critical discourse analysis… the fact that it is critical of existing public policy and the outcomes it generates. Indeed, that positivist (or “mainstream”) policies studies are not inherently and explicitly critical of the status quo is a criticism that many policy scholars have failed to sufficiently address. I believe that Luke charts a promising path forward for policy studies, using both “interpretivist” and “mainstream” approaches in the middle chapters of his book, which bridge the early application of CDA and the later more auto-biographical section composed of Chapters 12 and 13.

8While the middle of Luke’s book contains interesting reading for any narrative scholar. I will concentrate here on Chapters 5, 6, and 7. The reason for this focus is that in these chapters, Luke seems to transform his understanding of the role that narrative approaches to policy studies can play in policy-making. Indeed, Luke’s journey is one which showcases a shift in his own perspective from scholar to administrator (among many other shifts) but it is a passage which he writes in Chapter 5, After the Marketplace, that I think encapsulates the need for narrative policy studies to unequivocally adopt the explicitly critical orientation championed by the “interpretivist policy scholars” while embracing more “mainstream” social science tools for analysis.

9Luke writes, “We need evidence-based social policies derived, inter alia, from a critical, hermeneutic social science that draws from a range of disciplinary discourses and fields” (2019, p. 129). He also is blunt in his assessment that policy scholarship must, “move past the binary of quantitative/qualitative, science/narrative that had been placed on the table by neoliberal policy” (Luke, 2019, p. 117). To do so, he turns to critical realism as a means to accomplish this movement and notes that it is essential for education policy research to be able to “engage qualitative and quantitative, hermeneutic and positivist paradigms in a constructive dialogue” (p. 129). Luke’s writing is echoed in the recent work of Durnová and Weible (2020) that argues for the need to bridge the gaps between “interpretivism” and “mainstream” policy studies, as he extols the value of multi-level statistical modeling to inform education policy, incorporating “ethnographic, case study and discourse analytic work to test hypotheses, to build models, and to instantiate the trends and clusters that emerge from such an analysis” (p. 129).

Connection to Broader Literature

10The call for an integration of “interpretivist” and “mainstream” approaches to the study of policy continue in Luke’s book. Here, I want to draw the reader’s attention to two key arguments that Luke makes for the adoption of a critical realist approach to narrative studies of policy. First, he notes that the dominance of neoliberal governance regimes has completely impoverished education policy discourse in the U.S., Australia, Canada, and many other nations. This has had the result of evacuating any notion of a “strong overarching normative vision about the moral and civic purposes of education” (Luke, 2019, p. 149) from the policy discourse in these nations. This means, that absent a critical dimension, education policy research is being undertaken merely in service to capital accumulation. This perpetuates historically unjust systems of militarism, imperialism, racism, and capitalism because mainstream policy studies are unwilling (not unable) to adopt a critical orientation to the purpose of their labors.

11Second, Luke is clear that the critical orientation of “interpretivist” policy scholarship is vital for advancing narrative policy studies’ ability to impact actual policy development and adoption. He is equally clear that the insistence by (what he terms “poststructuralists,” but what I am calling, following Andrew Sayer) strong social constructivists (2000, p. 62) on the dominance of discursive, to the exclusion of the material, is problematic:

Some discourses kill people, take away their livelihood, other humiliate, others marginalize and shame. Some modes and play of différance make a difference in peoples’ lives, others simply don’t matter much. In this way, the ubiquitous poststructuralist observation that we can account fully for the world through discourse, or rather for the worlds’ transient and continually deferred (in discourse) meaning, is at worst glib and at best partial.

12So, Luke’s project, it seems, arrives at a point where critical discourse analysis effectively uncovers and decodes hegemonic narratives, and narrative studies (of the “interpretive” school) can use it to discover and elevate narratives which are created and shared among the marginalized and oppressed. However, Luke wants scholars to push toward more than just narrative identification and is clear that in order to have the greatest impact and efficacy, education policy would have both “ ‘evidence’ and ‘accountability’ as the basis for social policy… “ (Luke, 2019, pp. 185-86).

Narrative Politics in Public Policy: Legalizing Cannabis


13Miller’s book uses its first three chapters to develop and layout the Narrative Politics model and its latter four to apply the model to the study of cannabis legalization in some of the states in the United States. His Narrative Politics model holds that narratives are “the most compact unit of meaning” (p. 53) which are workable in public policy studies. These narratives are products of humans and are designed to order an ambiguous reality. Narratives are also theorized to contain a connotative ideograph and a story line which collects symbolic material into a policy message (Miller, 2020, p. 1).

14In the opening chapter, Miller notes that the key concern for the Narrative Politics model is the fact that humanity’s “core concern is the problem of meaning-making” (p. 3). Meaning-making, for Miller is important in public policy due to “the ambiguity we language-using humans must negotiate in fending off one dang thing after another” (p. 3). Concretely, Miller notes that dubiety (underlying uncertainty) is a reason we should turn toward narrative inquiry since it is well-equipped to explore the contests over facts and goals which are a major feature of pluralist democracy.

15Miller also makes an important statement in this chapter with regard to his view that since all words are capable of being used to lie, dissemble, distract, and mislead, the Narrative Politics model offers “an interpretive, inductive approach to navigating the political swampland of public policy discourse…The result is not a metanarrative but another narrative, one that redescribes and recontextualizes through juxtaposition and reinterpretation” (Miller, 2020, p. 12).

16Following this foundation, Miller elaborates a key component of the narratives to be interpreted using the Narrative Politics model, the concept of the ideograph. Ideographs are “connotative-rich signs capable of forming stable images, and they can help frame a policy problem, solution, or goal in a strategically crafted way. They can gather in concepts and ideas as well as imagery, emotions and values” (Miller, 2020, p. 17). Examples of ideographs, such as welfare queen, acid rain, and death tax are provided, which serve as a good illustration of the concept to readers new to McGee’s (1980) concept.

17The ideograph is especially vital to the Narrative Politics model, because its “function in a policy narrative [is] to bring connotations, associations, and imagery into the picture to reflect political commitments in some fulsome way… [and] to normalize some view of the world or justify policy action or policy belief” (Miller, 2020, p. 19). Ideographs are arranged into story lines, which cohere as a recognizable narrative. It is these narratives which are turned loose to compete in the policy discourse and which are interpreted by the scholar employing the Narrative Politics model.

18After discussing the component pieces and theory underlying his approach, Miller applies the Narrative Politics model to cannabis policy in the United States. While some readers well-schooled in interpretive policy analysis may be comfortable with this approach, it makes the text somewhat less accessible to those interested in learning how to conduct a study using the Narrative Politics model. For example, Miller suggests that the analyst applying the Narrative Politics model is using a “curatorial approach” (p. 53) and in various places in the text, he likens the approach to “Foucault’s geneaological method” (p. 28) or one akin to the journalistic methods of Seth Abramson (p. 53). This provides some amount of clarity, but it is also one which is not explicit in its method, making the intent of the approach, not the pathway the scholar should take getting there, the focus of his explanation. He writes, “The Narrative Politics model resists the inclination to gather in the prominent, factually accurate, overlapping, or distinguishing features of the narratives in order to integrate them into a consensus summary of sorts. Instead, the point is to clarify the narratives and acknowledge their diversity, their pluralism, their idiosyncrasies, and their adaptations – a Foucauldian insight” (Miller 2020, p. 54).

19Miller goes on to note that the curatorial approach laid out in the Narrative Politics model is one where interpretations “sometimes overlap, and other curators might imagine other typologies. Moreover, it is not always obvious what contemporaneous events influence the changing discourse” (2020, p. 70). While this is an approach with high fidelity to the stated purpose of the Narrative Politics model, the lack of parameters surrounding the design of the inquiry (or the museum, if we keep with Miller’s curator analogy) generates a plethora of questions when thinking about how one might apply the Narrative Politics model to other policy issues.

20Despite this ambiguity, readers of Miller’s work will be treated to an excellent and interesting exploration of numerous narratives about and around U.S. cannabis policy. Miller names these narratives the, Nativism Narrative, Abstinence Narrative, Libertarian Narrative, Harm Reduction Narrative, Compassionate Use Narrative, Fairness Narrative, and Social Justice Narrative. He traces their rise and fall, (though it is unclear how these determinations are reached by the curator) and, of special note for students and scholars of public administration and implementation, goes further than many narrative policy scholars by explicitly identifying policy narratives which arise in response to a policy change.

21As cannabis policy developed, Miller notes that four narratives emerged from the defeat (at least in certain states) of the Abstinence Narrative; these he calls post-implementation. The first of these, is the Psychosis Narrative, which warns that the lack of scientific studies which existed prior to decriminalization of marijuana and the increase in the potency of cannabis strains available to ingest mean that there is an underappreciated danger of users developing a marijuana-induced psychosis. Another narrative gaining in force during the implementation stage Miller called the Big Marijuana narrative, which seeks to detail the dangers of a legalized cannabis, especially as it contributes to the economic goals of industry to induce users of marijuana to do so more frequently.

22Third, is the Futility narrative which seeks to build on the success of cannabis policy reform. It argues that the U.S. drug war is a failure and decriminalization (or legalization) should follow for all (or many) other scheduled drugs in the United States. The final post-implementation narrative is the Conspiracy narrative, which posits that reform has failed because powerful interests want to keep cannabis illegal, whether for reasons of protecting their profits and market-share, to weaken political enemies, or for the maintenance of bureaucratic turf.

23Miller also details the impact of an agency’s genesis narrative on the ongoing narrative discourse, which would be of great interest to public administration scholars, and opens up interesting research questions when it comes to the notion of narrative evolution, such as whether agencies ever propagate narratives which rationalize their own dissolution, and if so, when might this be the case?

Connection to Broader Literature

24Of the three texts reviewed for this essay, Miller’s most explicitly advances the case for an “interpretivist” approach to the incorporation of narrative inquiry in policy studies. As such, it serves as a trenchant exemplar of the post-positive approach. It provides the reader with a clear example of narrative scholarship that is skeptical of modern social science’s incorporation of the logic of the natural sciences. Miller’s work brings to the fore several issues in the ongoing conversation in contemporary policy studies (e.g., Dodge, 2015; Jones & Radaelli, 2015; Miller, 2015; Durnová & Weible, 2020) related to the ontological and epistemological commitments which narrative studies (may or may not) require. I will comment upon this in the closing section of this essay. However, to effectively set the stage for this commentary, I now turn my attention to Merry’s (2020) book.

Warped Narratives: Distortion in the Framing of Gun Policy


25Merry’s is the first book-length project which applies the Narrative Policy Framework (NPF) as its primary framework for investigation. Merry complements this primary theoretical foundation by explicitly including Schneider and Ingram’s (1993) work on the social construction of target populations (SCTP), and recent scholarship on critical race theory (CRT) to explore the role of race in the framing of gun policy.

26Merry explicitly acknowledges the realist assumptions which undergird the NPF. She notes, “…there is a reality independent of human perceptions, but people can assign different meanings to that reality…” (2020, p. 53). She also notes that narratives are key to human’s ability to develop and communicate meaning and that the meanings which are communicated, are not randomly patterned. Instead, she notes that narratives, “…reflect people’s ideologies, beliefs, and norms. The NPF also holds that narratives are composed of both narrative content, which varies from one policy area to another, and narrative form which is “defined by a set of structural elements, which can be identified across policy contexts” (Merry, 2020, p. 53). The structural components include: 1) characters, 2) a setting, and 3) a policy solution (Shanahan et al. 2017, p. 176). In addition, the NPF includes the concept of narrative strategies which relate to the way that different actors in the policy process arrange, favor, and deploy the various structural elements in their policy narratives.

27Importantly, Merry’s work relies on the NPF, SCTP, and CRT to develop “…expectations regarding how interest groups craft narratives when pursuing particular narrative strategies… [and]… then translate those general expectations to specific hypotheses regarding gun policy groups’ use of characters and settings” (Merry, 2020, p. 52). Leveraging content analysis in an approach to coding narratives augmented by the Linguistic Inquiry and Word Count (LIWC) software, Merry undertakes the analysis of the policy narratives featured in the communications of 15 interest groups in the U.S. gun policy arena. Nine of these groups support gun control (for example, Everytown for Gun Safety, Sandy Hook Promise), while six other groups favor gun rights (for instance, the National Rifle Association, the National Association for Gun Rights,). The scope of the data collected is sweeping, with almost sixty-seven thousand communications analyzed across press releases, blogs, emails, and social media between 2000 and 2017. This commitment to comprehensiveness and the use of existing theory to be explicit about expected findings, coupled with a method of testing said hypotheses, sets Merry’s project apart from the other texts which compose this review.

28Merry’s research approaches gun policy identifying whether interest groups are presenting narratives on gun violence in ways which differ from the reality of gun violence in the United States and, if so, how these narratives distort that reality. A key analytical assumption is that we can effectively represent the reality of gun violence in the United States. Surprisingly, reading across these texts, it does not seem to me that any of these authors object to the notion that reality can be effectively reflected (at least in certain cases) by a presentation of statistics via the analysis of quantitative models. Indeed, even Miller, whose commitment to interpretivist methods might lead us to suppose that he would be particularly unsympathetic to this approach indicates he trusts “the findings about climate change, greenhouse gas emissions, and many other questions that have entered the political realm” (Miller, 2020, p. 2). Luke, as well, cites numerous statistical figures regarding the number of Australian students attending private higher educational institutions (p. 157) as well as the economic impact of the “internationalization” of higher education on Australia’s economy (p. 221).

29These quantitative descriptions of a reality in which the policy area of interest is embedded (cannabis, education, guns) are present in each of the works reviewed here but are most important for Merry’s project.

30In essence, Merry leverages the NPF to compare the types of villains and victims found in the policy narratives propagated by gun rights and gun control interest groups with the statistics generated by scholarship and government agencies to demonstrate how the stories told by each side of the gun debate fail to reflect the reality of the situation. Merry’s findings are both illuminating and troubling, regardless of one’s own position on the issue of gun control. This is the case because her findings indicate that both sides systematically misrepresent key facts related to the identity of those who suffer most frequently from gun violence, warping public perceptions of the costs associated with the current state of U.S. gun policy.

31For instance, Merry finds that both gun rights and gun control groups endeavored to use their policy narratives to expand the scope of conflict in the policy arena, but that gun control interest groups overemphasized children as victims of gun violence, in the sense that their stories featured children as victims with a much greater frequency than their actual percentage of deaths by firearms in the United States. Merry used SCTP to explain this finding, as children represent “ideal” victims, by which she means those who are both innocent and vulnerable (Merry, 2020, p. 76). She also found that gun control groups systematically over-represent White victims, which functions to perpetuate the pattern “…of denying African Americans the status of worthy victims” (p. 76).

32Merry’s book also found that gun control groups’ narratives featured villains in their narratives much less frequently than they featured victims, and that these villains changed in their characterization in response to events of gun violence which occurred in the larger society. Merry found that gun control groups also created and advanced policy narratives which were focused on mass shootings, events which account for a very small percentage of the total number of firearm deaths each year in the U.S.

33With reference to gun rights groups, Merry found that the narratives they deployed also warped the reality of gun violence in the United States, but in very different ways. Instead of featuring people who were killed or wounded by gunshots as victims, the policy narratives of these groups cast gun owners as victims since “gun regulations or proposed regulations represented threats to the rights of gun owners” (Merry, 2020, p. 93). Villains in these narratives were often shown as criminals who threatened the safety and well-being of law-abiding people, which gave license to gun owners to defend themselves, and bolster the argument that restricting gun ownership is a threat.

34Merry’s analysis points to a more troubling possibility that the focus of each of these groups on their own version of the role gun violence plays in the country has obscured the underlying cost and actual victims of firearms in the United States. As Merry states, “both types of organizations neglected the most common cause of firearm-related fatalities, suicide” (2020, p. 119). This oversight thus obscures the fact that the majority of people killed by guns are, in fact, older, White males, who “comprise 79 percent of gun suicide victims” (p. 121). Additionally, since both gun rights and gun control groups tend to avoid any explicit mention of race in their narratives, the policy narratives also obscure the fact that where gun homicide is presented as a problem of urban violence, the victims of this type of crime, who belong to racial minority groups, are frequently presented in ways that distance them from readers of the narratives. In this way, the presentation of victims and circumstances under which their lives were lost to guns are shifted away from the most common, to the most sensational.

Connection to Broader Literature

35The last two chapters of Merry’s work use the findings from content analysis and hypothesis testing to lay out the implications on gun policy and highlight possible ways that her study can be used to refine and expand the NPF and SCTP. Merry’s effective presentation of the questions which she is studying, and the transparent presentation of her results and methodology (the omission of the dictionary lists used in the computer-assisted coding process is my only minor complaint on this account) are an exemplar, and can serve as an excellent model for scholars interested in applying the NPF to their own research projects.

36Having sketched out each of the projects, and having highlighted some of the noteworthy features of each, I will now share my thoughts with regard to how these revised understanding of causality in policy studies can help to integrate the “interpretive” and “mainstream” approaches relying on narrative-based investigations.

Conclusion: Causality, not ontological-epistemological complementarity is the core issue

37There are many things that have been said concerning this schism in policy studies, and many more that could be said. However, I will limit my comments to a single issue, both in the interests of space and in keeping with the format of a review essay, which rightly seeks to highlight the contributions made to the discipline by the authors’ work.

38First, I believe it makes sense to open this section with Miller’s stated aversion to scientism, by this, he means social science’s embrace of “…quantification, neutrality, and objectivity – an approach variously described as naturalism, positivism, or scientism…” (2020, p. 2), most especially as it manifests in scholarship that analyzes narratives, but does so using tools like content analysis to quantify narratives. Miller identifies a real danger when he worries that the understandings and discourses of the scholars employing methods of inquiry which are insufficiently sensitive to the context of the discourse under study (i.e., scientism) may miss the underlying meaning of the communication. Or worse, that scholars may substitute their own understandings for the intended meaning. Specifically, Miller states, “A content analysis that merely counts words cannot assess the meaning of those words and ignores their narrative context.” (Miller, 2020, p. 24).

39This risk, at least in Merry’s study, is mitigated by her attention to the scope and depth of discourse being carried on in the corpus of texts produced by gun control and gun policy interest groups under study. More generally, content analysis gathers a textual corpus from which the narratives under study are read. This reading is then used to generate some type of codebook, whose codes are not mere counts of words, but which capture the distinctly different meanings around the words used. Although, in many cases, words which are distinct (in the corpus) are chosen to associate a concept, object, individual, or group with a particular positive or negative connotation. For instance, the NRA often refers to gun control advocates as “gun grabbers,” and while this in theory could be a term that gun control activists use to describe themselves, they (at least to my knowledge) have not done so.

40Indeed, Merry herself notes that the narratives are distinct in their use of terms, with victims and villains in gun control and gun rights narratives being populated by different specific characters. Merry also goes to great lengths to situate the narratives in different policy settings (see her Chapter 7) which helps to contextualize the relationships and meanings of the words that are coded in her content analysis. This means that while it is possible for Miller’s complaint about content analysis to be valid, and that a word might be misunderstood, it would be surprising to me if this were the case for Merry’s findings.

41We also saw (above) that Miller’s book (2020) employs quantification and cites statistics and figures, so it strikes me that mere quantification is not the core of Miller’s skepticism. Nor is Miller’s concern necessarily tied to those narrative approaches which use formal categorization to analyze policy. Indeed, in his work he favorably cites McMorris, Zanocco, and Jones’ (2018) NPF exploration of Oregon’s Ballot Measure 97, which used both quantification and formal categorization of narrative components.

42Given this situation, I suggest that it might be beneficial for policy scholars to reconceive the disagreement as one which has emerged from a misapprehension about what must follow from the NPF’s adoption of an “objective epistemology” (Jones & Radaelli, 2015, p. 341). In short, it is not clear to me that such an epistemology requires a concomitant embrace of a model of causality based on “…gathering data on regularities, repeated occurrences” (Sayer, 2000, p. 14) coupled with the assumption that these regularities cause something to occur.

43The use of empirical approaches to the study of narrative, like content analysis, does not commit the NPF (and similar approaches) to this notion of causality. In fact, I argue, a realist approach can be understood to undergird the NPF and that this model of causality recognizes that because social science deals with open systems, overlapping structures and causal mechanisms, it becomes difficult to attribute causality to any one mechanism with certainty. As structures that have the power to affect others may not always do so, and some structures may not affect an object of interest in the presence (or absence) of an additional (perhaps, unobserved) structure, this makes the reaction to similar configurations of forces contingent (i.e., not mechanical). This is especially the case when studying humans and other living creatures who possess highly-developed forms of agency.

44As such, statistical quantification, as undertaken by Merry’s study and NPF scholarship more generally, does not require that we apprehend causality by identifying patterns using content analysis and statistical wizardry. Instead, the project of NPF scholarship, and one to which Luke’s work is also friendly, is using the power of statistical summary to identify relationships between components of narratives (settings, characters, plot, moral of the story) and one another, in addition to the manner in which these relationships are emphasized (or absent) in light of the context of the public policy issue at hand.

45This is not to say that the NPF does not hope to eventually be able to make statements about which type of narrative approach is more or less likely to be successful in crafting more persuasive, (Shanahan, Jones, McBeth & Radaelli, 2018) or more attention-grabbing, (Peterson, 2021) policy narratives. Merry’s findings, and the claims she makes derived from them, are not asserting that a series of occurrences of a particular type of character, strategy, or policy solution is likely to result in the development of a nomothetic model which can be blindly applied to policy debates devoid of context. Merry’s (2020) exploration of the policy narratives of gun rights and gun control organizations in the U.S., is an example of how high-quality NPF studies can contextualize the narratives they study. This contextualization is clearly demonstrated by her incorporation of the visual elements included in interest group communications and the presentation of her findings in tandem with a rich description of the prevailing political-social-economic conditions in the United States. As Sayer puts it, social inquiry, which relies on a critical realist approach, is undertaking an “attempt to distinguish rigorously between can and must, all and some, often and always (Sayer, 2000, p. 4).

46While some might infer that the hypothesis testing approach deployed by the NPF smacks of Popperian falsification and a commitment to the positivist theory of causality associated with it, we should be cautious. Falsification in the social science is not an instance where a hypothesis is usually completely rejected or completely accepted. Instead, parts, portions and particular instances of a theoretical proposition are often rejected as the theory itself is enriched with revisions or refinements as the theory-driven suppositions are tested against contextually-situated social situations. In addition, while Merry’s book uses content analysis and the related ability to mathematically (which is, of course, merely another human language, with its own set of assumptions, axioms, and conventions which are socially constructed) describe the similarity and divergence of given narratives, she does not draw any conclusions related to causality from them.

47Indeed, Miller (2020) himself suggests that “One sensible test for the ultimate efficacy of a narrative in a policy discourse is that it gets adopted.” (Miller, 2020, p. 26). While I do not believe that any policy scholar would disagree that this would be a sensible approach, a weakness with the recommendation is that without some mechanism by which we can isolate narratives from the fundamentally open social system in which they are constructed, communicated, and evolve, we cannot with any certainty say that the adoption of a narrative necessarily corresponds to whether a policy was adopted (or was not). Since the policy process is saturated with forces (some more easily discernable than others) which interact with one another, and since it is impossible to establish a closed-system in which we can eliminate alternative mechanisms which may have caused the adoption of the policy (Bhaskar, 1975), we are unable to rely on the metric of adoption as one which can tell us about the efficacy of a given narrative.

48This means, I think, that the NPF, Luke’s critical realist approach to the study of education policy, and Miller’s Narrative Politics model all resonate along a similar understanding of why the study of policy narratives is important. These narratives are meaning-laden, they can mobilize or dissipate support for (or against) a policy, and they are deployed and exist in a political environment. These conditions mean that they are worthy of study, and that since they possess the features that they do, we can identify that narratives differ from one another in certain ways, and that these ways are nontrivial. However, the fact that narratives can impact policy does not mean that narratives will impact policy in ways that are predictable, or that they will always impact policy decisions.

49That said, it is important to be clear that while the NPF and similar “mainstream” approaches to policy studies, when understood as rejecting a naïve positivist theory of causality (Sayer 2000), can operate in harmony and cooperation with “interpretive” approaches, to produce “evidence-based social policy derived, inter alia, from a critical, hermeneutic social science that draws from a range of disciplinary discourses and fields” (Luke, 2019, p. 129). The possibility exists, as well, that outcomes of interest could be the product of multiple causal pathways (see Fiss, 2011; Fitzgerald, 2019) indicating that equifinality and causal complexity is a real possibility in policy studies.

50However the discussion around causality plays out, there still exists the problem that “mainstream” policy studies in not explicitly critical in its practice and pedagogy. This chasm, I fear, may be wider than the miscast ontological/epistemological dispute between “interpretive” and “mainstream” strains of policy research. In the face of climate catastrophe, mass extinction of existing flora and fauna, rising levels of inequality, increasing levels of inter-state human migration, war and social instability, the re-emergence of reactionary political parties and their subsequent policy arguments, the time for “mainstream” policy scholars to abandon the illusion that one can remain “neutral” in such a political moment grows short.

Top of page


Nevin, A. (1990). The changing of teacher education special education. Teacher Education and Special Education: The Journal of the Teacher Education Division of the Council for Exceptional Children, 13(3-4), 147-148.

Bhaskar, R. (2008). A realist theory of science. Verso.

Dodge, J. (2015). Indication and inference: Reflections on the challenge of mixing paradigms in the narrative policy framework. Critical Policy Studies, 9(3), 361-367.

Durnová, A., & Weible, C. (2020). Tempest in a TEAPOT? Toward new collaborations between mainstream policy process studies and Interpretive policy studies. Policy Sciences, 53(3), 571-588.

Fiss, P. C. (2011). Building better causal theories: A fuzzy set approach to typologies in organization research. Academy of Management Journal, 54, 393-420.

Fitzgerald, J. B. (2019). Equifinality and Pathways to Environmental Concern: A Fuzzy-Set Analysis. SOCIUS: Sociological Research for a Dynamic World, 5, 1-14.

Jones, M., & Radaelli, C. (2015). The narrative policy framework: Child or monster? Critical Policy Studies, 9(3), 339-355.

McGee, M. (1980). The “ideograph”: A link between rhetoric and ideology. Quarterly Journal of Speech, 66(1), 1-16.

Miller, H. (2015). Scientism versus social constructionism in critical policy studies. Critical Policy Studies, 9(3), 356-360.

Peterson, H. (2021). Narrative policy images: Intersecting narrative & attention in presidential stories about the environment. Policy Studies Journal.

Sayer, R. (2000). Realism and social science. Sage.

Schneider, A., & Ingram, H. (1993). Social construction of Target Populations: Implications for politics and policy. American Political Science Review, 87(2), 334-347.

Shanahan, E., Jones, M., McBeth, M., & Radaelli, C. (2018). The Narrative Policy Framework. In C. M. Weible & P. Sabatier, Theories of the policy process (pp. 173-213). Routledge.

Top of page


Bibliographical reference

Aaron Smith-Walter, Multiple Book Review EssayInternational Review of Public Policy, 4:2 | 2022, 241-252.

Electronic reference

Aaron Smith-Walter, Multiple Book Review EssayInternational Review of Public Policy [Online], 4:2 | 2022, Online since 01 November 2022, connection on 26 May 2024. URL:; DOI:

Top of page

About the author

Aaron Smith-Walter

UMass Lowell, USA

Top of page



The text only may be used under licence CC BY 4.0. All other elements (illustrations, imported files) are “All rights reserved”, unless otherwise stated.

Top of page
Search OpenEdition Search

You will be redirected to OpenEdition Search