1
Introduction

Why this book?

This book introduces the skills that uphold the use of digital methods in social research. As an introduction, it serves researchers and students inside and outside academia, who are interested in digital methods but do not know where to start. As a short introduction, it provides a brief guide for a quick start. Focusing on the different practices associated with digital methods, it touches upon the epistemic and political consequences of these methods, but with the purpose of providing hands-on advice for their use.

This book falls squarely between two previous digital methods books that situated this approach historically and conceptually (Rogers, 2013) and described recipes and tools for studying web archives, search engines, and social media platforms (Rogers, 2019). This book focuses on the principles that sit behind the practice of digital methods and that are often taken for granted in previous works, while avoiding discussion of specific software tools, application programming interfaces (APIs) or data dashboards, which are ever evolving and often discontinued or revamped. As we write, a crucial data source for Facebook and Instagram is being shut down (CrowdTangle) and replaced by another (Meta Content Library), as happened again five years prior - in both cases with significant consequences for researchers (Grevy Gotfredsen & Dowling, 2024).

Digital methods are introduced in the form of a digest, without the footnotes of software settings or the cataloging of digital studies to date. Rather, we present short takes on collecting and wrangling data by querying, prompting artificial intelligence (AI), text-mining, scraping, crawling, and calling APIs. Alongside those, we discuss preferred visualization approaches as well as a rationale for placing these approaches in the middle of the analysis rather than as a culmination.

As it will be clear across the chapters of this book, far from being a single or simple research protocol, digital methods are a complex toolbox composed of a multitude of techniques for collecting, preparing, analyzing, and visualizing digital records. This variety is unified, however, by four core methodological principles: repurposing data and methods, following the medium, foregrounding biases, and diversifying social research.

Repurposing data and methods

From the outset, digital methods defined themselves in opposition to the notion of "virtual" both methodologically and topically. Methodologically, they distinguished themselves from the project of "virtual methods," the social science effort to "digitize" and migrate existing methods to the online environment, for example, turning ethnographies into virtual ethnographies or surveys into online surveys (Rogers, 2009). Digital methods were instead interested in the so-called methods of the medium, i.e., natively digital methods that have their first or only implementation within digital media. In a sense, digital methods are shorthand for digital media methods, or to "digital media, digitally analyzed" (Rose, 2022).

Likewise, in terms of research topics, digital methods saw new media not only as a site for the study of online cultures, but also as a medium to inquire into broader cultural and societal dynamics. They examine the online world not as a virtual realm in and unto itself, or as a cyberspace of imaginaries, but rather a panoply of records about social life.

This effort to focus on digitally native methods and to use them to study societal phenomena is best captured by the idea of repurposing. Repurposing means starting from data and methods developed for non-academic purposes (e.g., user engagement metrics developed for marketing) and using them as tools for social research. In practice, it means that when using web data for research, one should first undertake some "technical fieldwork" (Rieder et al., 2015) to identify the kinds of data available online and consider how they could be used for social research. Repurposing means that digital methods are a form of "second-hand research" that sits atop online services and mainstream and alt-tech platforms, which have not been designed for scientific data collection.

Repurposing Wikipedia, for example, can mean extracting the table of contents of the "same entry" in different languages and placing them side by side to study cultural specificity or competing interpretations of events, or snowballing the hyperlinks in the "see also" section of different pages to create a conceptual network demonstrating close and distant content relationships. In both instances, the repurposed or second-hand research outputs become material for cultural and social analysis that is broader than the platform under study.

This approach is in line with an early definition of new media as "putting things on top of other things" (Rogers, 2013). The repurposing of digital methods has its origins in the remix culture of new media, and especially in the art of mashing up songs shared on Napster and elsewhere during the years of the copyleft debates (Shiga, 2007). Mashup culture extended to the mixing of data streams or data overlays to produce "composite applications" (Liu & Palen, 2010; Daniel & Matera, 2014). Possibly the most well-known was the Katrina Information Map, a Google Map that was layered or annotated with pinpricks of disaster relief centers (Crowe, 2005). To make such a service or application, one examines which sources (or data) are available to mix and inventively produce new outputs of interest for research and societal relevance.

The focus on repurposing distinguishes digital methods from digital humanities, which tend to focus on digitized materials from historical or documentary archives. Whereas digital methods are oriented towards online and digitally native data, digital humanities have long been in the business of converting historical records into machine-readable material. As one scholar puts it: "Though not 'everything' has been digitized, we have reached a tipping point, an event horizon where enough text and literature have been encoded to both allow and indeed force us to ask an entirely new set of questions" (Jockers, 2013: 4). Drawing on a more established tradition of archiving and preservation, digital humanities are now engaged in developing the techniques of distance reading necessary to harness the "great unread" (Moretti, 2000) they amassed.

Following the medium

The second principle, following the medium, means that research with "web data" should focus on the "web" as much as on the "data." Websites, social media platforms and apps are neither neutral recorders nor mere conduits through which opinions and actions flow unfettered. They are complex bundles of technological effects and vernacular cultures, which privilege certain sources and practices while depressing others. Content is recommended and moderated, both manually and algorithmically. Online publishing and social media posting have styles and rhythms for the user to fit in and stand out (Carmi, 2020). In digital methods these media effects and media biases are never just a noise to be cleaned out - not only because such cleaning is impossible, but also because there is value in keeping the web, the search engine, or the social media platform as a part of social research accounts.

There are at least two ways to retain the medium in an account of a social or cultural analysis. One is to situate the data collection in the history of the platform in terms of the algorithmic and vernacular culture. Twitter/X of 2025 is a different platform to Twitter/X of 2020. The feed has changed from posts by followed accounts in reverse chronological order to a personalized list based on a "for you" algorithm first popularized by TikTok. Metrics have been renamed: retweets having become reposts. The culture of the platform has also changed with certain subcultures (academic Twitter/X, for example) disengaging from the platform and other (right-wing trolls, for example) becoming more prominent.

Another way of retaining the medium is in the story of data collection. There have been, for example, multiple styles of Twitter/X data collection, depending on where the data were retrieved: the search and streaming APIs; the academic API; scraping; and the API compliant with the European Digital Services Act. Given their research affordances, each of these methods of data collection could produce profoundly different data sets.

Following the medium means considering its research affordances: i.e., the digital objects and metrics provided by a platform or a web service and available for repurposing. For example, Facebook native digital objects are posts, pages, and metrics of interactions. One may make a collection of posts from a series of pages concerning a social group (e.g., the Somali diaspora) or a movement (e.g., the Alt Right) and amongst them determine which posts have received many (and few) interactions, thereby demonstrating what animates the group or movement "on Facebook." To include the platform in the account, means considering, for example, the effect of content moderation. Studying an Alt Right movement would, for example, yield less material today than a decade ago. As a case in point, one scholar found that a significant percentage of...