The Art and Craft of Comparison

The following is a sample chapter from the book The Art and Craft of Comparison by RAW Rhodes (first published ).
To order your copy of this book click here
This chapter is also available as a downloadable version:


Is it possible to compare French presidential politics with, say, village level politics in rural India? The conventional wisdom in both naturalist comparative social science and humanist area studies is a resounding ‘no’. Even if this comparison could reveal interesting insights, it would be impossible to design a study that would make methodological sense.

For naturalist social scientists, the main objection would be that the contexts are too different, irrespective of the variables we might choose to investigate (that is, institutions, economic development, political culture, or religion). We might compare (say) Indian and the French political systems at the national level, or small-town politics in France with village politics in rural India. But the scales of presidential politics and village politics are just too different. In fact, they are so different that they do not even conform to a ‘most different’ design (Lijphart 1971) because there is not even one variable that might link them. In other words, it is not like comparing apples with oranges; it is like comparing an elephant with an ant. Rather than varieties of fruit, they are different species.

Humanist area studies scholars would reach the same negative conclusion though using a different logic. Comparison, from this perspective, runs the risk of sacrificing rich, nuanced and context-bound insights on the altar of parsimony, generalisability and theoretical elegance. In this tradition the concern is that these types of studies often contain inaccuracies and misunderstandings due to insufficient knowledge of the place under study. The point here is not that we cannot make this trade-off, but rather such an exercise would not produce meaningful insights.

The aim of this book is to present an alternative to these two common ways of seeing and studying the social world. French presidential politics and village politics in rural India may seem beyond meaningful comparison. But if we probe deeper, we begin to see that 2

presidents and village leaders share common dilemmas—how to mobilise supporters, how to ensure the loyalty of allies, whether to weaken or placate rivals, how to manipulate crisis and survive scandal, whether to implement change or preserve the status quo. Indeed, when we think about politics in this actor-centred way, we start to realise that the comparison we propose is not as incongruous as first appeared. We might even begin to imagine a president and a village leader sitting down together and discussing political tactics and manoeuvres, sharing what works and what does not. Rather than having nothing to offer, the findings of such a study might turn out to be interesting, even illuminating. Indeed, such research might find that some presidents have more in common with counterpart village leaders than they do with other presidents.

This example is not a thought experiment. It echoes the logic of F. G. Bailey’s (1969) classic Stratagems and Spoils. Context, for Bailey, was not a constraint, either methodologically or philosophically. Unlikely juxtaposition allowed him to render both the exotic familiar and the familiar exotic, opening space for thinking about politics as an essentially human activity in which we all take part. In doing so, he revealed both the universal and the particular; the dilemmas are similar but the strategies actors pursue and the consequences of their choices can be different (cf. Adcock 2006, p. 62).

The implications of this approach for how we undertake comparative research across the social sciences is potentially profound. As our brief illustration highlights, conventional wisdom about when, how and why to compare severely limits how we study and understand the social world. As a result, we are missing potentially rich and illuminating insights because our analysis is either too rigid, structured, and systematic, or too bespoke, detailed, and idiographic. We seek to address these limitations—and expand the horizons of comparison—by outlining a novel, clear and sound approach to comparative interpretive analysis. 3

We are indebted in this pursuit not just to a long line of innovators like Bailey, but also to a recent eruption of interest in the means and ends of comparison in social science. For example, the late Benedict Anderson’s (2016) memoir A Life beyond Boundaries challenges old presumptions about the practice and utility of small and medium-n research. Reflecting on his own career, and in particular on the slow-burning but spectacular success of Imagined Communities (1983), Anderson revels in the creative spark that enabled him to make ‘surprising’ but revealing comparisons across apparently disparate contexts. His approach bemused his naturalist colleagues in comparative politics while his contemporaries in humanist area studies regarded it with caution.

While Anderson’s reflections open new possibilities for comparative research, they are limited because he treats the work of making comparisons as a mysterious and indescribable form of alchemy. He provides no basis for different comparative strategies, and no guidance on how to conduct such research. This book seeks to demystify the process. Its chief contribution is to provide a philosophically sound and practically useful guide to a distinctively interpretive form of comparison; we are not looking for an illusory middle way. We are humanists doing comparison.

In doing so, we contribute to fast-moving debates about the universal-particular divide. In recent times, these discussions have been dominated by naturalist social scientists seeking to defend the value of qualitative research (for example: Bennet and Checkel 2015; Blatter and Haverland 2012; Coppedge 2012; Gerring 2012a; 2017; George and Bennett 2005; Goertz 2006; Goertz and Mahoney 2012). But while we are sympathetic to aspects of this cause, we challenge the underlying mantra of ‘diverse tools, shared standards’ (Brady and Collier 1994). This book attends to the question of how small facts can speak to large issues from the other 4

side of the naturalist-humanist divide. We seek not only to provide an overlapping set of tools, but also a distinct set of standards by which to conduct and judge comparative research.

This contribution is vital because interpretive social science remains widely typecast as idiographic, both among its critics and its practitioners. For naturalist social scientists, interpretivists—if they enter the conversation at all—are deemed to be in the business of ‘mere description’. They provide rich, detailed and illuminating accounts of isolated social phenomena, but the relationship to theory building and testing is tangential at best. Insofar as naturalists seek accommodation with the humanist tradition, it is by adopting a strict division of labour in which detailed descriptive inferences are reinterpreted by social scientists seeking law-like generalizability (see Gerring 2012b). Some humanist scholars who embrace, or at least are sympathetic to, the interpretive approach, invert this condescension and wear their idiographic robe with pride (for discussion, see Bates 1997). Following Geertz (1973), these scholars provide ‘thick descriptions’; that is, rich, detailed and illuminating accounts of particular social phenomena. Theoretical generalisation is seen as a fraught and fallible task (Abu-Lughod [1991] 2006; Vrasti 2008). But whether ‘mere’ or ‘thick’, whether meant as disparaging or approving, the dividing line stays the same. The common perception is that the interpretive enterprise remains rich, but context-bound, and idiographic, unable to see the wood through the trees (see Chapter 2).

We join a swelling group of interpretive scholars who are dissatisfied with this state of affairs. The group traverses diverse fields including comparative politics (for example, Wedeen 2010; Fujii 2013); organisational studies (for example, Flyvbjerg 2006); public health (for example, Greenhalgh et al. 2011); human geography (for example, Robinson 2011); international relations (for example, Pouliot 2014); and area studies (for example, Gibson-Graham 2004; Wesley-Smith and Goss 2010). We argue the dichotomy between the naturalist 5

pursuit of analytic clarity and generalisability and the humanist pursuit of rich idiographic detail is a false one (cf. Coppedge 2012; Brady and Collier 2010; King, Keohane and Verba 1994): interpretive researchers can and do make meaningful comparisons that speak to themes of general significance (see Bevir and Rhodes 2016 part 2; Adcock 2006). Our goal is to outline and promote this distinctly interpretive way of thinking about and doing comparative research.

There are at least four main reasons interpretivists might want to compare. The first is that comparison provides us with a better understanding of a particular case. So, we compare cases A, B and C because they help shed knew and illuminating light on Case D, which is the one we are interested in. Relatedly, comparison can also help us provide policy solutions—we can suggest a policy response to Case D because we know what happened in Cases A, B, and C. Much interpretive comparison that does exist is justified on these grounds. A third rationale is that we place the particular case in a broader social context (e.g. Cramer’s account of support for Wisconsin Governor Scott Walker and the global rise of populism). Nothing in this book contradicts attempts to undertake these types of comparison. But our primary rationale is that we compare because it is essential to providing ‘decentred’ explanations of the social world. To decentre is to unpack practices as the contingent beliefs and actions of individuals as the basis for explanation, as opposed to laws and rules, correlations between social categories, or deductive models. The risk of in-depth idiographic studies is that they treat context as the de facto explanation for all social and political phenomenon, rendering impossible any attempt to generate more general insights. We argue that decentred explanations can speak to general themes. But, to do so we need to rethink the nature of comparison.

We argue that the dilemmas actors experience in their everyday lives are the intellectual skeleton key that unlocks the potential of comparative interpretive research. When we ask why actors act we create an opportunity for reflection on alternative meanings and actions, and the 6

pros and cons of each. By reflecting with actors, we uncover the choices and questions they confront. By understanding how they see these choices, as a reflection of the webs of belief in which they are embedded, we are able to explain why actors do what they do. In Geertz’s (1973:15) classic formulation ‘we begin with our own interpretations of what our informants are up to, or think they are up to, and then systematise those’. When we ask whether others’ experience the same dilemmas, we necessarily explore how their experience is either similar or different, unlocking the comparative dimension of this approach. By abductively moving back and forth between the meanings and beliefs of individual actors and our sensemaking of their practices, we provide an interpretation of their interpretations. The result is a decentred approach that focuses on the social construction of a practice—theirs and ours—through the ability of individuals to create, and act on, meanings.

So, this book is a call to arms for interpretivists to embrace creatively comparative work that uses the dilemmas of situated agents as its empirical starting point to develop plausible conjectures. It justifies and explains the approach. We hope that researchers across social science disciplines might find inspiration in our attempt to reassert the place of richly interpretive analysis in broader theory-building efforts. Readers searching for a comprehensive, step-by-step guide to conducting interpretive comparative research will most likely be disappointed. There is simply not space to spell out in any detail the philosophical basis of the interpretive approach nor to provide a textbook on qualitative data collection and analysis. In fact, there is no need. Existing textbooks and handbooks do this job comprehensively (see for example, Bevir and Blakely 2018; Bevir and Rhodes 2016; Schaffer 2015; Schwartz-Shea and Yanow 2006; Wagenaar 2011). We focus on demonstrating, with concrete examples, the potential of a comparative interpretive approach. The examples come from our fieldwork and the work of scholars who have adopted a similar approach even if they would not explicitly identify as doing consciously comparative interpretive work. The argument is therefore not 7

limited to political science but applies equally to richly qualitative research in cognate disciplines in the social sciences and humanities.

This chapter consists of five substantive sections. First, we outline our basic argument. Here, we put forward the need for a consciously and explicitly comparative interpretive approach, and the creative benefits that such an approach can provide. Second, we provide a brief summary of the interpretive approach. Third, we seek to justify the rigour and sensitivity of a comparative interpretive orientation. Fourth, we foreshadow in greater depth the structure of the book and detail of its component chapters. Finally, we provide guidance for readers on how to use the book, and in particular on how to combine its insights with those stemming from canonical texts in the field.

The comparative intuition

At the heart of this book is the idea that comparing is intuitive (see Anderson 2016). People in general, and social scientists in particular, are engaged in ‘constant comparison’. Comparison is what enables us to make sense of events as they unfold across time and space. We compare to identify patterns and disjuncture in the social world. We compare new experiences with old ones to help us make sense of the exotic or unfamiliar. By doing so, we sometimes see familiar settings in a new light. We also draw on comparisons to help communicate ideas and make insights interesting for a broad audience. Human beings are comparative animals.

Seen in these terms, naturalist approaches to social science discipline and restrict the comparative intuition. Here, comparison is something to be designed into the project from the beginning. The core problem—and one to which anyone with experience of detailed fieldwork 8

can attest—is that the best laid comparative plans go wrong. Expected categories of comparison can melt away on closer inspection. Unexpected categories can emerge to take their place. As Gerring (2007, 149) concedes in highlighting the limits to even the most careful and robust process of case selection in naturalist research: ‘Not all twists and turns on the meandering tail of truth can be expected’.

Meanwhile, humanist approaches to the social sciences tend to smother the comparative intuition. The goal is to understand and highlight the rich specificities of the context under examination. Attempts to draw out comparative themes across cases can be derided as erasing or blunting these specificities. This concern is at the heart of the radical critique of efforts to draw broader theoretical lessons from bespoke settings (see e.g. Vraasti 2008). The effect, ironically given the commitment of these humanist scholars to reflexivity, is to wish away a key part of the interpretive process. Doing so prevents analysts from reflecting on, and giving voice to, how they impose their own sense making systems and categories on their data in order to render their fieldwork observations intelligible to others.

We argue that interpretivists who embrace the comparative intuition can overcome these limitations. Unlike the naturalist approach to disciplining this intuition, such a move can enable the analyst to explore comparative insights that emerge in the field after the development of authentic and in-depth case knowledge. Unlike the humanist approach to denying or negating this intuition, it can authentically present the world as the analyst sees and experiences it. Indeed, rather than flatten or misrepresent context, comparison in these terms becomes a vital means of both understanding and conveying context. It is only through comparison that new insights and experiences become meaningful, and we can communicate meaningfully to relevant audiences. 9

Embracing the comparative intuition can also help immensely in unlocking creativity. Creativity is central to all forms of social science research but little understood. Kaufman and Beghetto (2009) usefully distinguish between Big-C creativity and Pro-C creativity. Big-C creativity refers to ‘the remarkable and lasting contributions made by mavericks in some domain’ while Pro-C (or professional creativity) refers to ‘professional creators, [who] have not reached eminent status’ (and for a review of the field see Kozbelt et al. 2010). Scholars in both naturalist and humanist traditions produce work of Pro-C quality but work of Big-C quality is frustrated increasingly by the restrictive paradigms in which they work. Leading naturalists concede that, in spite of their advocacy of strict standards and rigid protocols, creativity remains an elusive but essential ingredient in good research. It enables researchers to situate their findings, to design and execute effective research programs, and, above all, to find the spark of inspiration that propels the voyage of discovery. In their naturalist bible for political science, for example, King, Keohane and Verba (1994: 14) quote Karl Popper approvingly: ‘Discovery contains an “irrational element”, or a “creative intuition”.’ Yet, in practice, any sense of creativity is all too often suppressed when writing-up findings (cf. Feyerabend 1988; Polanyi 1958; Gerring 2017). Indeed, thoughtful writers of a naturalist persuasion have acknowledged this methodological bind. For example, Collier et al. (2010: 197) write that naturalist procedures may ‘sharply narrow their substantive research questions, thus producing studies that are less important’. They identify a conflict between ‘the methodological goals of improving descriptive and causal inference’ and ‘the objective of studying humanly important outcomes’. Gerring (2017, 31-32) goes further when he states that ‘the task of discovery is a comparatively anarchistic affair. There are no rules for finding new things (Gerring 2017: 31).

If naturalist social science favours justification (i.e. verification and refutation) over discovery, humanist social science favours exploration and innovation. Leading proponents 10

typically present the opportunity to be creative as one of the chief virtues, and pleasures, of this approach to social science (see most famously Geertz 1973). The core logic is to move iteratively between an inductive account of the data and a deductive reading of the relevant literature, thus leaving space for creativity to flourish. For example, Geertz’s (1973: chapter 15) famous account of the Balinese cockfight starts by describing it as a vice raid – cockfighting is illegal. Next, he interprets it as a symbol of Balinese masculinity – the double entendre of cockfighting is deliberate. The next layer of meaning is cockfighting as blood sacrifice to keep demons at bay. It is followed by a discussion of gambling that casts cock fighting as a dramatization of status concerns in the tiered status hierarchy of Balinese society; men are allegorically humiliating one another, although their status is not changed by the cockfight’s outcome. Finally, he considers the cockfight as a typical spurt in Balinese life. Cockfighting is the story the Balinese tell themselves about themselves. Cockfighting is no longer a vice but a text which Geertz is reading to say something about Balinese culture. His account is both a creative and a masterly thick description of small facts speaking to large issues.

However, in practice, most authors remain narrowly channelled, bound by the focus on the specific context, which takes priority over all else. In the humanist tradition, rigour is defined as the richest, most bespoke forms of contextual knowledge. It produces ‘thick description’. So, interpretive researchers working in the humanist tradition feel free to forge their own creative connections to broad theoretical concerns. But, only if theory speaks to the specific context under examination. Breadth must be sacrificed for depth. Those who do manage to escape this trap, and speak to broader debates, risk scorn. Take James C Scott’s (1979, 1985 and 1990) hugely influential trilogy on agrarian politics as an exemplar; it has been humorously categorized across graduate classrooms as ‘First, peasants in Malaysia; then, peasants everywhere; finally, everyone everywhere!’ (see Rabinowitz 2014). 11

More commonly, then, it is skilful scholars working on particular issues that have a big impact on the field. Seemingly, we cannot explain or reproduce their broader impact. Again, we confront a form of magic or alchemy by which the rich insights of humanist scholarship somehow achieve broad resonance.

The key to demystifying this alchemy, we argue, is to unlock creativity in making comparisons. We use the word ‘puzzling’ to describe this creative process. It refers to solving a problem or answering a question creatively (see also Adcock 2006, p. 62). It may involve clever guesswork or a novel experiment. We are drawn to the word for two reasons. First, it contains its own contradiction. If something is a puzzle, it means it confuses us yet by puzzling over the confusion we attempt to make sense of it. The field—and especially the multiple fields entailed in comparative research—can throw up many puzzles we do not understand, so we puzzle over them to find their meanings. Fujii (2013), for example, is puzzled by why acts of extra-lethal violence (e.g. forcing victims to dance and sing before killing them, mutilation and so forth) occur given the risks involved defy rationalist explanations. Using cases from Vietnam, Rwanda and Malaysia she argues that these acts can be explained by the participants’ roles and activities , which contribute to the production of graphic effect. This type of work operates according to a ‘logic of discovery’—iterative, open-ended, evolving—as opposed to the typical social science preference for a ‘logic of justification’ in which invariant procedures are rigorously applied with the aim of producing generalizable and predictive results.

Following a ‘logic of discovery’, the second attraction is that puzzling is a process with no clear destination; we change the puzzle as we seek to resolve our confusions, often multiple times over the course of a complex comparative project. Through deep, rigorous and continuous puzzling, interpretive scholars can feel emboldened to explore and tease out comparisons that surprise and intrigue, that uncover new insights or force readers to confront 12

familiar insights in new ways. In this way, our account of interpretive comparison can provide a new set of tools and insights to seed, exploit and channel creativity in crafting effective and affecting comparison.

Interpretation and comparative intuition

Interpretive research offers a distinctive approach to channelling the comparative intuition because they consciously offer interpretations of interpretations. They concentrate on meanings, beliefs, and discourses, as opposed to laws and rules, correlations between social categories, or deductive models. An interpretive approach is not alone in paying attention to meanings. It is distinctive because of the extent to which it privileges meanings as ways to grasp actions. Its proponents privilege meanings because they hold, first, beliefs have a constitutive relationship to actions and, second, beliefs are inherently holistic.

For example, when other political scientists study voting behaviour using attitude surveys or models of rational action, they separate beliefs from actions to find a correlation or deductive link between the two. In contrast, an interpretive approach suggests such surveys and models cannot tell us why, say, raising one’s hand should amount to voting. They do not tell us why there would be uproar if someone forced someone else to raise their hand against their will. We can explain such behaviour only if we appeal to the intersubjective beliefs that underpin the practice. We need to know voting is associated with free choice and, therefore, with a particular concept of the self. Practices could not exist if people did not have the associated beliefs. Beliefs or meanings would not make sense without the practices to which they refer.

The aim of interpretive research is to decentre: to unpack practices as the contingent beliefs and actions of individuals as we just did with our short example of voting. Decentred analysis produces detailed studies of people’s beliefs and practices. It focuses our attention 13

on everyday dilemmas. It challenges the idea that inexorable or impersonal forces drive politics, focusing instead on the relevant meanings, the beliefs and preferences of the people involved. Like a kaleidoscope, decentring produces changing patterns. We use this metaphor to allude to the way that all agents are situated within multiple, intersecting social fields (age, gender, ethnicity, profession etc.). The classic social science approach is to make each field the subject of empirical analysis. Starting with agents rather than fields allows us to see how dilemmas recur and intersect. The effect is akin to looking through a kaleidoscope, with the same practices and processes given new meaning depending the agent narrating it.

Not anything goes

In seeking to unlock the creativity of interpretive comparison, we want to be clear that we do not suggest that ‘anything goes’. Underlying this scepticism is a suspicion that any attempt to bridge the particular-general divide in an interpretive approach necessarily entails a lack of rigour. We do not so much have in mind here the scepticism that might be expected from naturalist social scientists. While we will argue that such scholars might derive important insights from close engagement with our argument, we do not expect them to come all the way with us. To invert the common mantra here, we have some shared tools, but different standards. Instead, our chief target is the humanist impulse to look down on any attempt to go beyond idiography as ‘airplane ethnography’ Bevir and Blakely 2012: 94); a pale, even dangerous, imitation of the real thing. We combine breadth and depth and focus our discussion on defending the rigour of comparative interpretive research against such prejudices (and see Chapter 4).

Rigour in the humanist tradition depends on reflexivity. A long-standing and penetrating critique of the naturalist approach is that it airbrushes out persistent normative 14

biases that bump up against the naturalist pursuit of objectivity. The obvious example is the Cold War biases that permeate the literature on comparative democratisation (de Velo 2016). So, contrary to the naturalist pursuit of objectivity, the humanist response embraces subjectivity, and all its associated idiosyncratic practices and interpretations (see especially Schwartz-Shea and Yanow 2012, chapter 6). Rigour stems from a deep commitment to reflexivity throughout a research project. It is reflected in a rejection of uniform procedures and universal standards and a focus on how the subjective position of the researcher shapes their practice. Such discussion typically centres on explicit reflection on the researcher’s role in the social field of their research. It will consider such issues as their status as an insider or outsider; their normative preferences; and the disadvantages and privileges that enable or hinder access. A good example is Yanow and Schwartz-Shea’s (2006) edited volume on Interpretation and Method, which has become something of a staple on more pluralist graduate programs in political science and social policy. Each chapter begins with the authors’ personal reflections on their own experiences and commitments, and the ways in which they shaped their research journey. The handbook illustrates the reflexivity expected of interpretive research.

This practice cannot be disentangled from debates about dominant cultural norms in academic practice, especially the long tradition of Western ethnocentrism in disciplines such as Anthropology and History (Clifford and Marcus 1986). It stems in particular from the ripples of the ‘Culture War’ in Anthropology that began in the 1980s and dominated the 1990s. The point of reflexivity is to acknowledge the analyst’s social position; to foreground the total immersion that such a research project entails; to highlight the obstacles and challenges such immersion presents; and to clear the way for a richer and more authentic representation of how the subjects of interpretive research see their world. 15

We take no issue with such reflexive criticism. However, the way in which it is practiced can be a problem. The risk is a new form of academic chauvinism or gatekeeping in which only those with the requisite demographic background can perform the role of expert with authority. This stance denies the legitimacy of ‘outsider’ perspectives; for example, the criticisms levelled at Anderson for his parallels between Southeast Asia and the West (Anderson 2016: 149). Moreover, it presents only a partial view of the experiences and normative preferences that influence academic scholarship.

The researchers’ position in the social field is one thing, but what about their position in the academic field? Researchers have particular sets of beliefs and normative preferences based not just on their personal histories, but on their professional socialization. We do not to impugn the reputation of other scholars. In fact, we draw on examples of how professional norms, pressures and ambitions have shaped our own work throughout subsequent chapters. It is simply to recognise that research does not take place in a neutral professional context. The academy, as anyone who has spent any time in it can attest, is not short of tempting bandwagons and powerful cliques. Certain sorts of research attract rewards such as grants, prestige, and impact more than others. Career prospects and legacies are at stake. What we study, how we study it, and how we represent our findings can all be affected by presumptions, aspirations and doubts about how our work will be received (see especially Hay 2011).

Central to the issue that concerns us are the implications of adopting different comparative strategies. Why do we seek to compare, or refrain from comparing, in particular circumstances? Why do we choose particular ways to compare particular phenomena? What are we trying to achieve? Humanists in the idiographic tradition have typically refrained from answering these questions directly. In the hands of some of the most skilful practitioners, such choices are apparent. Take, for example, Lisa Wedeen’s (2007) study of qat chewing in Yemen. 16

Wedeen depicts the organic deliberation enabled by this cultural practice by using an implicit comparison with Habermas’s (1989) account of the emerging public sphere in Parisian salons. Through carefully crafted attribution, allusion and mimicry, she develops an account that speaks to a broad audience well beyond Yemeni experts or Middle East scholars. But not all interpretive research can be so creative and artful, and not all readers are going to be equipped to appreciate such intertextual subtleties.

We seek here to provide a more straightforward approach to reflexivity. We promote a consciously and explicitly comparative interpretive approach that can bring these design, interpretive and discursive choices out into the open. By focusing on dilemmas, we follow the humanist approach in putting people at the centre of our analysis. We commit to trying to present our research participants’ world as (we think) they see and feel it. But in identifying, developing and fleshing out comparisons, we must acknowledge ways in which their dilemmas are familiar or alien to us. Such puzzling will reveal how our own views and experiences skew our account of their beliefs and practices. So, the creative drawing of comparisons forces us to acknowledge how our personal experiences and biases inflect our practice. Second, by seeking to compare such dilemmas across contexts we must be explicit about how and why we are linking our research to major theoretical themes. In doing so, we reveal something about what we are trying to achieve politically and professionally. We acknowledge our position in the academic field by putting our own normative preferences and intentions up front.

Understanding and practicing reflexivity in this way opens interpretive comparative research up to scrutiny from more angles than either naturalist or humanist work typically allows. By consciously laying out the theoretical claims we seek to make through such comparison, we welcome the scrutiny of the generalists working on that theme. Our desire to make ‘plausible conjectures’ of general interest necessarily entails both a tight theoretical 17

argument, and much critical reflection on the broader resonance of the dilemmas at the heart of our comparative findings. Just as important, because our focus on dilemmas means we still strive to represent the research participants’ world, we still welcome the scrutiny of the specialist. Central to our account of rigour is the forensic interrogation of competing narratives by both specialists and practitioners (Bevir and Rhodes 2003: 37-40). Our comparative approach multiplies our readership and potential critics. Our accounts of the dilemmas that actors face need to ring true with scholarly and local or practitioner audiences across all the different settings under analysis.

We do not advance the bold claim that comparative interpretive research is somehow more rigorous than other forms of social science. It is just to stress that it is no less rigorous. Naturally, such an approach entails trade-offs associated with the fossilised ideas about rigour in either naturalist or humanist traditions. For the former, our approach sacrifices stable variables for more fluid and subjective dilemmas. But the theoretical insights we derive still need to resonate more broadly, and still need to stand up to scrutiny when placed against the experiences and patterns established elsewhere in the field. For the latter, our approach sacrifices some depth of contextual knowledge for breadth of insight. But the claims we make about local contexts still need to have sufficient depth and richness to resonate with experts in these fields. They still need to be accurate, nuanced, and contextually sensitive (see also Adcock 2006, p. 62).

We do not abandon rigour. We just understand rigour in an older and more encompassing sense of the term: as shared canons of accuracy and precision, rigorous argument, clear presentation, respect for evidence, and openness to criticism (see Collini 2012: 62). Our account underpins a vision of comparison not as a hard science, nor as an exercise in dangerous or whimsical speculation, but as a challenging craft, indeed an art form (see also 18

Schmitter 2009). This book will give interpretive social scientists working in the humanist tradition key tools and techniques to set about this creative work. We outline a set of principles, standards and approaches for producing richly detailed comparative research that speaks directly to issues of general interest.

Structure of the book

The heart of the book revolves around the following interrelated questions:

1. What are the principles and standards that guide comparative interpretive research?

2. What are dilemmas and why do they matter?

3. How do we design a comparative interpretive project?

4. How do we undertake comparative interpretive fieldwork?

5. How do we interpret and analyse comparative data?

6. How do we communicate the findings of comparative interpretive research?

These questions form the core chapters. We devote Chapters 3 to 7 to outlining and demonstrating the practice of interpretive comparative research. For every chapter, we do not offer rigid procedures or laws to follow. Instead, we offer various ‘rules of thumb’ that will help. Following the Oxford English Dictionary, we define ‘rule of thumb’ as a ‘method or procedure derived from practice or experience, rather than theory or scientific knowledge; a roughly practical method’. Also, we provide examples from our own comparative work: on 19

Westminster governments; policy advice to prime ministers and ministers in Australia, Britain, Canada and New Zealand; the everyday life of Chiefs of Staff to Australian Prime Ministers; policies on obesity in Australia and Britain; political leadership in small Pacific states, the court politics of aid agencies; and poverty across different communities on the affluent South Coast of England. We outline every chapter below.

To be clear, although there is some theory in the first half of the book, it is not our intention to rake over old ground on the methods and the philosophy of social science. Much has already been written on this topic and we do not feel it is necessary to rehearse the philosophical underpinnings of the interpretive approach we favour yet again (see Bevir and Rhodes 2003; 2006; 2010; Rhodes 2017b). Instead, we focus on the specific interpretive theory that underpins our approach

Chapter Two outlines briefly the difference between naturalism and humanism before providing a summary of our key concepts of decentring, situated agency, and plausible conjectures. In effect, it sets the theoretical scene for the rest of the book and provides the underpinnings for comparative analysis based on the concept of dilemmas.

An important subsidiary aim of this chapter is to challenge the naturalist mantra of ‘different tools, shared standards’ and provide an alternative account of what constitutes valuable and rigorous interpretive research. We set out a new set of criteria by which interpretive comparative work should be assessed and towards which interpretive comparative researchers ought to strive. We focus on accuracy, openness and aesthetics. We define accuracy as meaning the use of established standards of evidence and reason; so, we will prefer one narrative to another if it is more accurate, comprehensive and consistent. We define openness as taking criticism seriously and preferring positive speculative theories that open new avenues 20

of comparison and make new conjectures supported by evidence. We show that not anything goes in comparative interpretive research.

Chapter Three explains the concept of dilemmas and why we use it in preference to other ideas such process and practice. Dilemmas underpin the logic of interpretive comparison. As currently used, however, it refers mainly to ‘Big-D’ dilemmas that focus on ideational clashes between traditions such as the clash between neoliberalism and state ownership. We add the notion of ‘small-d’ dilemmas that focus on the everyday, the routine and the mundane choices, ‘court’ politics and realpolitik.

We suggest that dilemmas perform a dual function that gives them this particular analytic importance. First, the dilemmas actors experience in their everyday actions and practice are the authentic manifestation of the social tensions that typically motivate social scientific inquiry, especially interpretive research. Second, though always embedded in local contexts, dilemmas recur occur across different settings. Our reading of Bailey’s (1969, p. xi) comparison of De Gaulle and French presidential politics with village leadership in rural India shows that a concept like (in his case) ‘rules of the game’ or (in our case) ‘dilemmas’ can render social phenomena both interesting and comparable.

We suggest that empirical, comparative, interpretive social science research revolves around the process of identifying the dilemmas that actors experience, the ways they respond to them, and puzzling about whether they vary according to the traditions in which they are situated. We suggest seven rules thumb for identifying dilemmas.

The next four chapters of the book drill down into the practices of comparative interpretive research. We use our own experiences of conducting comparative interpretive research to reflect on the unique challenges and opportunities associated with this approach, and to provide clear and pragmatic guidance for those hoping to adopt it. For each chapter, we 21

identify six to eight rules of thumb to guide the research. The four component chapters follow the widespread and stylised account of the different stages of research: design, fieldwork, analysis and writing-up. However, we emphasise throughout that this is not a linear process. Interpretive researchers always move back and forth, puzzling about and refining their project as they go, a process typically called ‘abduction’ (Law 2004; Schwartz-Shea and Yanow 2012; Wagenaar 2011).

Chapter Four looks at how to design a comparative interpretive project and tackle the perennial case selection question. The problem here is one of justifying unorthodox comparison. A lot has been written on comparative case selection from a naturalist perspective but this language is often an uncomfortable fit for interpretive projects. We argue that case selection is not something that is designed into a project from inception. For interpretive research, it changes as we go. We therefore suggest different strategies of case selection for different phases of a comparative interpretive project. We identify rules of thumb to guide design choices as the project or programme evolves.

Chapter Five outlines the place of fieldwork in comparative interpretive research. Detailed qualitative fieldwork is central to most interpretive research, but practical guidance on how to navigate the field remains rooted to the idiographic tradition. The presumption is one of sustained immersion in a discrete setting. Interpretive comparison, however, necessarily requires partial immersion across a multiple sites. Crucially, the researcher must be alert to the surprises and moments of epiphany that can challenge initial assumptions and open new possibilities. We seek here to develop and illustrate key ‘rules of thumb’ that will enable researchers to manage the challenges and maximise the opportunities.

In Chapter 6, we explain, and illustrate with examples drawn from our own work, how interpretive researchers analyze comparative data. We argue that a comparative project 22

compounds the uncertainty, confusion and paralysis that can set in when confronted with a ‘mountain’ of qualitative data. We argue it is not possible to ‘somehow capture’ this full complexity. We outline and defend the need for a consciously impressionistic orientation to data analysis. Rather than searching for a ‘Eureka!’ moment that confirms or refutes a narrow theory (in naturalist or nomothetic mode) or makes sense of the whole picture (in idiographic mode), a comparative focus on dilemmas enables the use of a kaleidoscope of different analytical lenses and tools to explore complex specificness in context. We outline rules of thumb for helping along the way.

Chapter Seven looks at the craft of writing. Although we discuss the challenges of writing that confront all social scientists, we focus on the dilemmas of writing-up comparative interpretive research – dilemmas which we confront because we speak to a broader range of audiences. We focus on the dilemmas around structure, style, and substance. In doing so, we highlight the importance of seeing writing as integral to the research process, not something that starts once the research is done.

In the Retrospective, we turn our methods back on our own book and ask, ‘what are the dilemmas of using the approach we advocate? It is an exercise in professional reflexivity as we reflect on the personal dilemmas that we navigated in writing this book. Also, we take a last opportunity to impress on the reader the merits of this approach by summarising the key terms of both the interpretive approach and our comparative interpretive approach. It is a short cut for those who like to skim books before reading them. 23

How to Use this Book

The book serves different functions for different readers, and we think ought to be read in different ways by these audiences. Experienced scholars, well acquainted with the humanist-naturalist and particular-general divides, and familiar with the details of research design and practice, will be able to read it against that background. For naturalists, it is a sophisticated and grounded challenge to common presumptions about interpretive work. These scholars have been responsible for much of the most innovative and exciting work in seeking to promote the general value of case research (see George and Bennett 2005; Brady and Collier 2010; Goertz and Mahoney 2012; Gerring 2017). This flowering of ideas and strategies has served to push along sophisticated thinking in interpretive research methods, too (see, for example, Pouliot 2014). We hope that our attempt to tackle the issue from the other side of the divide might have a similar effect. For humanists, the book is not an indictment on rich idiographic work. We think such research remains important. It is simply an invitation to scholars in this tradition to be more open to the prospect of drawing comparison. And it is a how-to guide with detailed advice on innovative ways to tailor and augment their approach to do so (or as post facto justification of their long-standing deviance). Readers in either tradition may not be persuaded, but we hope at least they find our perspective interesting and valuable.

Those colleagues now learning the craft, still feeling their way through methodological debates and new to issues like case selection, data collection, and analytical techniques, may need to approach the book rather differently. They will likely benefit from reading chapters alongside orthodox textbooks and handbooks on qualitative and interpretive inquiry. Any lack of knowledge should not stop such readers understanding what we say. It is simply that they will need those nuts and bolts also to understand and justify the range of methods entailed by a complex social science project. The book is designed to be a thought-provoking and 24

accessible accompaniment to the established canon in qualitative and humanist social science. Our hope is to encourage newcomers (as well as old-timers) to be ambitious about the impact they can make by learning to practice the art and craft of comparison.