A Case for Description

Abstract

Descriptive research—work aimed at answering “who,” “what,” “when,” “where,” and “how” questions—is vital at every stage of social scientific inquiry. The creative and analytic process of description—through concepts, measures, or cases, whether in numeric or narrative form—is crucial for conducting research aimed at understanding politics in action. Yet, our field tends to devalue such work as “merely descriptive” (Gerring 2012), subsidiary to or less valuable than hypothesis-drive causal inference. This article posits four key areas in which description contributes to political science: in conceptualization, in policy relevance, in the management and leveraging of data, and in challenging entrenched biases and diversifying our field.

The emphasis on a “more scientific” understanding of politics, associated with generalization and parsimony, that has dominated the discipline since the behavioral revolution has privileged a causal approach to understanding the social world. Conversely, descriptive inquiry has a diminished role and status in the discipline. It often is regarded as messier, less important, or ancillary. Yet, scholars across methodological, epistemological, and subfield divides benefit from engaging in descriptive work. In response, we organized a series of collaborative workshops and roundtables focused on description as a social science enterprise. These gatherings served as a basis for discussing what “counts” as description in social science and what it contributes to the research process. By sharing insights from our collaborations, this article examines the contributions of descriptive research and presents an argument for its importance, in regard to both hypothesis-driven inquiry and as a standalone strategy and method.

Separate from the emphasis of causal research, descriptive research aims to answer “who,” “what,” “when,” “where,” and “how” questions. ¹ It is both a creative and an analytic process. Although “describing often involves practices which, like copying and translating, [which] tend…to be associated with lack of originality and novelty, description is neither neutral nor passive, nor does it simply entail transcribing or reenacting the given” (Weinstein 2022). This is because description involves making choices in the categorizing, sense making, and management of data.

Given the central role of description in political science, a brief discussion of its history (and eventual marginalization) is important. The behavioral revolution in political science was articulated by Merriam (1926) as an approach closely aligned with the natural sciences by Key as an effort to make research “more effective” (1950, 20); by Truman (1951, 37) as an “emphasis on empirical methods…[in] the development of a science of the political process”; and by Dahl (1961, 768) as “the modern scientific outlook.” It was a Kuhnian paradigm shift in the study of politics. Evolving through and with the expansion of computing technology in the mid- to late-twentieth century, the behavioral turn in political science is deeply linked with the application of empirical methods to data to yield causal arguments. This approach was consistent through the elaboration of the natural science model in works such as Designing Social Inquiry (King, Keohane, and Verba 1994) as well as in the Perestroika movement, which urged methodological pluralism (Monroe 2005) but did not question the underlying need for predominantly hypothesis-driven causal arguments.

This emphasis on causality has led to a decline in the prevalence of descriptive research in our field, which can be observed through the decrease of descriptive articles published in major field journals or in the phrasing that often accompanies such work as merely descriptive (Gerring 2012). We surveyed four major political science journals in 2022 ² and found a total of 32 articles (i.e., approximately 10.2% of the total articles) in which scholars used description as a primary tool of analysis. ³ Although description has been recognized as an “important step” for explanation, descriptive work still is considered in service of explanation rather than on its own (see, e.g., King, Keohane, and Verba 1994, 18). Although our survey found many scholars actively engaging in descriptive work, it often was in the service of other types of inquiry. ⁴

In the early 2000s, many scholars published work defending description within the discipline by pointing to the contributions of descriptive work in theory building (Shapiro 2002), measurement validity (Adcock and Collier 2001), and conceptualization (Wedeen 2002), among others. Yet, the discipline continues to focus on the role of descriptive work in service of making causal arguments (Gerring 2012) rather than as a standalone endeavor. Description, on its own and not in the service of inference, allegedly “loses most of its interest” (King, Keohane, and Verba 1994, 34).

We disagree. Workshop participants highlighted the importance of description in its own right to several important conceptual, theoretical, and policy-related conversations. Description is central to policy-informing research, which often answers questions about the dynamics of a policy problem and offers expertise on context and interventions. In research design, scholars spend considerable time outlining conceptualization, delimitation, and measurement to clarify the parameters of their study. This helps with sharing results and replicating studies. Description also can be leveraged to understand data collected in data-poor environments, in which data are scarce or difficult to collect; in data-rich environments in which “big data” interventions require a new understanding of the terrain of inquiry; or when major shifts have forced a reassessment of formerly settled concepts or measures. In this way, description challenges biases in the discipline.

This article focuses on moving the discussion forward by emphasizing the role of descriptive research in and of itself. First, we outline the importance of thorough description in conceptualization and how negligence limits contestation and collaboration. Second, we demonstrate the importance of descriptive work underpinning data repositories and how rich description helps scholars to understand new and emerging contexts. Third, we examine the benefits for policy makers and the media, who seek informed and curated description of contexts. Fourth, we conclude by providing a state of the field in terms of where description currently fits and ways it can enhance the discipline.

CONCEPTUALIZATION AND COHERENCE

Concepts and their application, incidence, and measurement are foundational in social science; however, good conceptual work requires description. The parameters of key concepts—for example, how many deaths make a civil war or how many electoral turnovers make for a democracy—are important to define so that coherence of research is ensured across time and place as well as among researchers. For example: What counts as a battlefield death in cross-national datasets? The counts are a descriptive endeavor, and they differ significantly between datasets (Restrepo, Spagat, and Vargas 2006). In their research comparing datasets in their counts of civilian fatalities, Broache et al. (2022, 3) found that several high-profile datasets have glaring omissions, basing their counts on dubious sources that consequently result in unsystematic undercounting in key cases. They argued that “When thousands of civilian deaths are left uncounted, political scientists’ knowledge of war—and the conclusions and implications they draw from them—remains at worst, dangerous, and at best, incomplete.” Yet, research using these data rarely engages with these debates beyond a footnote or literature-review citation.

The fundamental assumption in devaluing descriptive work is that the data are not the point of inquiry. Rather, inquiry comes after the data. This assumption limits the ability of researchers to either contest or build on findings. The potential to have works that nominally contest the dynamics of key concepts—such as civil wars, liberalism, and democracy—but fail to engage with one another because of incommensurable measures or definitions increases.

The fundamental assumption in devaluing descriptive work is that the data are not the point of inquiry. Rather, inquiry comes after the data. This assumption limits the ability of researchers to either contest or build on findings.

Descriptive work can define new concepts, such as the role of graffiti in articulating dissent and protest in authoritarian contexts (Lerner 2021) and state formation in the digital space (Yad, Raymond, and Muratbek 2022). Descriptive research also can clarify or recontextualize concepts that are foundational to our discipline such as norms (Jurkovich 2020), citizenship (O’Brochta 2022; Porisky 2022), totalitarianism (Kaul 2022), and the changing frames and identities activated in protests (Sosa-Villagarcia 2022). Establishing new concepts also is a fundamentally descriptive endeavor (Biswas 2022; Chan 2022). In the absence of descriptions that are linked to the grounded study of the social world, “…social facts are underdetermined—not enough is known to draw analytical, scientific conclusions. Previously unknown phenomena still existed prior to being described by researchers, and those phenomena were learned about through description, even if in non-scientific language” (Reiling 2022, 2).

Concepts in political science are a precursor to their application or measurement in analysis, and careful conceptualization is an inherently descriptive exercise. Scholars must consider “the question ‘What is?’ before asking ‘How much?’ Thus, meaning before measurement; quality before quantity” (Collier and Gerring 2009, 4). In a recent conceptualization article, Zaks (2023, 1) emphasized this tension as “scholars are exploring the causes and consequences of the phenomenon of rebel-to-party transition without agreement on (or debate about) what it actually is.” Moreover, although there are various defensible and valid ways to define “democracy,” a scholar using Huntington’s (1991) two-turnover test and another using Dahl’s (2008) observable dimensions of polyarchy are unlikely to be mutually intelligible without clear statements of their conceptual apparatus and supporting justifications for use. Understanding not only the definitions of the concepts we study but also their evolution and extension is necessary to make scholarship intelligible across the discipline.

MANAGING DATA: BIG AND SMALL

The processes of data collection, curation, and management are central to social scientific work across methodological and ontological paradigms. Whether in the strict empiricist sense of data as “the basic building block of knowledge”—without reference to subjectivity of the researcher or the phenomenon (Taylor 1971, 7) or in the interpretive sense as a contextually generated and intersubjective set of meaning-laden observations (Schatz 2009)—data are the descriptive foundation of subsequent analysis. Yet, the collection and curation of data rarely are discussed as the substance of political science research, often relegated to a methods section in qualitative projects or in articles debuting a new dataset for quantitative research.

However, across methodological approaches, data shape the types of projects that political scientists pursue as well as the types of results achieved by those projects. The availability of collected and curated data can fundamentally shape the questions that researchers ask (Cappella-Zielinski and Grauer 2022) because “if quality data is lacking, researchers turn away from asking and answering key research questions entirely, leaving marked gaps in scholarship on key issues” (Konken 2022).

Existing datasets, as often is acknowledged in the articles that outline their creation, are compiled using many and varied judgments, rendering the complexities of the social world as fixed data points (see, e.g., Boix, Miller, and Rosato 2013). Yet, these choices often are obscured in subsequent analysis. For example, the Eck and Hultman (2007) dataset on violence against civilians counts deaths in the United States in the context of incidences of international terrorism (e.g., the September 11 attacks) but not domestic terrorism or violence by the state (e.g., incidents of police brutality). Without robust descriptive work, such critical understanding is not possible.

When data do not exist for the questions or phenomena under scrutiny, then collecting them is a descriptive undertaking, which often is made invisible in published projects because it is considered ancillary to analysis. Yet, work on emergent phenomena—such as social media as a protest tool in Kazakhstan (Wood 2022), online democratic organizing among romance-novel fans (Fattore 2022), organizing around and against the participation of trans athletes (Murib 2023), and the emerging political-consultancy industry in India (Sharma 2022)—requires description to understand novel circumstances and dynamics. Other more long-standing but understudied subject areas—such as family-leave policies in Latin America militaries (Perera 2022) and the hyper-local politics of regulation (Wright 2022)—require new data collection. Data collection might be integral to understanding peripheral and deviant phenomena, including criminal activities and networks (Nussbaum 2022), the aesthetics of revolution (Wade 2022), the intentions of perpetrators of mass violence (Garrity, forthcoming), and the emergence of militia organizations (Avery 2022). In each case, the collection of data—including the standardizing of those observations for subsequent analysis—is a fundamentally descriptive exercise. It involves the expenditure of resources—money for travel or access and time in recording, standardizing, and coding data—but is rarely the focus of interest for publications. The invisibility of this work means that scholars who chose to pursue creating such data repositories are subject to significant additional burdens, rendered invisible by the devaluation of their descriptive work.

When data do not exist for the questions or phenomena under scrutiny, then collecting them is a descriptive undertaking, which often is made invisible in published projects because it is considered ancillary to analysis.

With the emergence of “big data” in political science (see, e.g., Brady 2019), the need to define variables or phenomena of interest is central to being able to leverage massive caches of data to understand the social and political world. Creating and maintaining datasets of densely populated phenomena—such as US campaign emails (Cepuran 2022), the proceedings of an international law commission (Holthoefer 2022), presidential agenda-setting (Eissler 2022), and international climate governance (Zebek 2022)—are a challenge not because of the lack of data but rather the abundance. Curating data that already exist also is a descriptive exercise, insofar as it involves understanding a new terrain of data where the researcher must define (or redefine within new data landscapes) new phenomena or concepts of interest.

When our discipline privileges the causal or theoretical application of the data rather than its creation, there are profound losses in terms of what is studied, valued, and ultimately understood. Whether researchers are operating within the parameters of a given dataset, creating their own from data-poor environments, or managing and leveraging large new forms or points of data, description provides the basis for a deeper understanding of the social and political world. This call to value description, however, is not intended as a license for endless digression into the minutia of data collection, a “remaking of the wheel” with every project. Rather, it is intended as an invitation to revalue the foundational work of data collection and curation as foundational—rather than ancillary—to social science.

POLICY RELEVANCE AND PUBLIC-FACING RESEARCH

Although the subject of policy or public relevance occasionally is debated as a goal of political science (see, e.g., Rogowski 2013), to the extent that there is an ambition to inform the policy-making process or the public at large, description is a central piece of the process. Informed and curated description of events and circumstances often is a front-line request by policy makers and mass media outlets, insofar as the expertise brought by political scientists can contextualize key points of tension and problems (Seay 2022; Shively 2022). Policy makers and the public want to know the who-what-when-where-how of complicated policy problems, which is currently evident especially in areas such as climate policy (Albistegui Adler 2022). Thorough qualitative description can help policy makers by cautioning them from excessive attention to outliers or in changing the framing of events with attention to cultural or institutional differences across contexts (Peh 2022).

Through generating summaries of data samples, statistical description can illuminate trends over time or illustrate the center, scope, and outliers of data samples. Additionally, experts can help to craft descriptive narratives to understand the precursors and key moments that have led to a particular point of public interest or a policy challenge (Lakin 2022) or to adjudicate among different explanations for outcomes (Harbin 2022). This work is important in part because “descriptive statistics, case studies, and similar material may be more valuable and accessible for policy makers and citizens, who typically learn and reason via discrete operating principles (or theory) as well as descriptive evidence” (Shively 2022).

THE STATE OF THE FIELD

As a field that takes seriously the social and political world, political science benefits from how description helps us to understand how we reflect or re-create the biases that exist in the world. Questions of whose voices are heard or assume prominence in the discipline are founded not only on questions of who is speaking but also who is listened to and heard. Description acknowledges the complexity of the world (Barkin 2022) and makes space for the experience of people who have been subject to or participants in the phenomena of interest to social scientists (Hooser 2022). Focusing on received concepts or assuming that default models from prior generations are universal often results in conceptual silences or in the application of deficit models to minoritized groups (e.g., Black women legislators) when they are compared with their counterparts (Brown 2014). As Avant (forthcoming) argues, “Prizing theorization over description…deters attention from what is not already represented in the field. It makes it harder to notice new insights and discourages even trying.” As Michener, SoRelle, and Thurston (2022, 164) argued, description allows for the “incorporati[on] of the voices and experiences of hard-to-reach populations…[and] collecting (and valuing) descriptive research…tells us things we do not know about people’s lives; shifting our lenses so that scholarship focused on marginalized groups is not depicted as narrow or atheoretical….”

Without descriptive work to carve out new ways of thinking through concepts and phenomena from new points of view, we risk re-creating the silences of prior generations of social scientists. As Thomas (2022) stated, “[T]he research designs and the social scientific practices we employ to describe what we observe as social scientists in our field is not separate from the ontological and epistemological ways in which the social world is framed by hegemonic sites of power, and thus, our methodological enterprises are grounded in materialist orientation and historical contexts that give rise to them.” In valuing descriptive work, especially when paying attention to many and diverse voices, we can begin to see the contours of such inherent biases and work to counter their influence (Rublee 2022).

Without descriptive work to carve out new ways of thinking through concepts and phenomena from new points of view, we risk re-creating the silences of prior generations of social scientists.

In making space for new voices in political science, descriptive work also has the potential to create new visions of our discipline as an inherently creative endeavor because “focusing on what questions, even if they cannot be entirely separated from why, primes scholars to understand more about social and political interactions. This generates greater openness to seeing the patterns that disrupt conventional wisdom….Noticing new patterns can also foster creative ways to think about them. Attention to description on its own terms helps scholars guard against trapping themselves in models that are inadequate for solving problems in the world” (Avant 2022, 4). Conversely, by sidelining description, the discipline is in danger of creating an “environment [that] encourages scholars to develop an overabundance of causal theories to justify their work as ‘innovative.’ It also creates incentives either to develop highly stylized evidence or to return regularly to the same, well-established pools of evidence rather than seeking out wider horizons” (Shively 2022). Descriptive work can broaden our scope of what scholars think of as political—or even as data—whether it is the definition, placement, and availability of greenspaces (Cantwell-Chavez 2021); the role of art in politics (Weinstein 2022); the definition and evolution of church–state relations in Africa (Longman 2022); or the contributions of emotion in foreign policy (Kowert 2022). Descriptive work also facilitates interdisciplinarity and broadens possibilities for collaboration (Hyder 2022).

CONCLUSION

Understanding the social and political world is fundamental to the enterprise of social scientific research, and the series of roundtables and conferences that we organized has called for a focus on the importance of descriptive work: in allowing our field to develop concepts, manage and leverage data, speak to policy makers and the public, and challenge inherited biases. Answering questions of “who,” “what,” “when,” “where,” and “how”—in concert with but also independent from causal inquiry—is vital to the continued relevance and coherence of our discipline in both scholarly and public-facing work. For several decades, description has been sidelined as ancillary to analytic approaches to the study of politics. Yet, as this article contends, description is essential at all stages of the research process—conceptualization, measurement, and data management—as well as in different types of research—purely academic, public facing, or policy relevant—in challenging embedded biases and making room for new visions of the discipline. It is vital to understand the world that we seek to explain throughout the enterprise.