Mobile methods: Explorations,innovations,and reflections

Introduction

Welcome to the Mobile Methods Special Issue. In this introduction we consider the nature of mobile methods and discuss how they often involve field-based research, the controlled collection of complex data, and attention to critical ethical questions. We then discuss highlights from each article in this issue and show how they connect to these themes in various ways.

Mobile methods are the means by which mobile communication technologies are employed to study the social world. The implications for the collection and analysis of data are vast and impact researchers working in a variety of traditions. For those working in behavioural science traditions, the large caches of detailed behavioural log data contained on mobile devices raise new opportunities along with weighty ethical and methodological challenges. Those working in these traditions can also use mobile devices to conduct on-screen surveys, experience sampling, field experiments, and other types of data collection. For those working in interpretive traditions, the role of mobile devices in many aspects of social life requires a rethinking of research contexts. Mobile devices can also be used during in-depth interviews to spur the telling of narratives and help respondents remember meaningful events.

In the inaugural issue of Mobile Media & Communication, Campbell (2013) argues that in order to define mobile communication as a field, it is critical to first define what we mean by mobile communication technology. He defines mobile communication technology as “devices and services that supported mediated social connectivity while the user is in physical motion” (p. 9). With this in mind, it is then possible to define the field of mobile communication as research that focuses on the uses and implications of mobile communication technology. By extension, we argue, it follows that mobile methods can be thought of as falling within the field of mobile communication, insofar as they incorporate mobile communication technology into the processes of data collection. As such, mobile methods are not defined and limited to traditional methodological divisions—such as the quantitative/qualitative divide—but rather they are a means of data collection.

Our focus on mobile methods within mobile communication is slightly different than how mobilities scholars from within sociology have defined them (Büscher, Urry, & Witchger, 2011). Monika Büscher and John Urry (2009) define mobile methods within a broader (im)mobilities paradigm, which focuses on the movement of people, objects, information, and ideas. However, we have delimited our approach to mobile methods to focus on how mobile communication technologies are used to study social phenomena. Most prominently this involves various smartphone apps such as texting, cameras, and maps, but also custom-made apps for data collection (e.g., Boase, Kobayashi, Schrock, Suzuki, & Suzuki, 2015; Gergle & Hargittai, 2018; Kobayashi, Boase, Suzuki, & Suzuki, 2015) or social invention (e.g., Brough, Lapsansky, Gonzalez, Stokes, & Bar, 2018). Mobile methods within mobility studies may or may not employ mobile communication technology or media and are therefore much broader than how we have defined mobile methods within this Special Section.

Our focus on mobile media and communication technology within mobile methods raises three common themes that are important for situating the contributions of the articles in this Special Issue. We suggest that mobile methods (a) enable field-based research, (b) allow for controlled complexity, and (c) require particular ethical considerations.

Field-based research

The mobility of communication technology enables researchers to go out into the world of their participants. This kind of fieldwork is often contrasted with laboratory research or experiments which bring participants into the world of the scientist. The presumed value of field-based research comes from its ecological validity, that is, the ability to study a phenomenon in situ enhances our confidence that the phenomenon “naturally” occurs rather than merely being an artifact of the research process itself.

Field-based research in this regard can take several shapes. Mobile methods research may involve collecting data that is generated in the field by mobile devices that people carry with them. For example, sometimes locational data are collected as data in and of themselves, but sometimes such information is merely used as elicitation an prompt (see Kaufmann, 2018) to engage participants in memory- and meaning-making. Here, the conversations between participants and researchers and likely transcripts that ensue are typically the data that are produced as part of the research process. The prompts generated in the field may or may not be considered data by the researchers.

Sometimes fieldwork enabled through mobile communication technology encompasses field experiments in which there is a deliberate intervention into the social world. Here, mobile devices are the means through which researchers administer an intervention or experimental treatment to participants, often in the form of material delivered through an app, or the sending of text messages. As with any field experiment, there are trade-offs regarding the complexities and messiness of being in the field, and the lack of control over the social world compared to that of the laboratory. However, the external validity of such field experiments may be greatly enhanced by mobile methods.

Field-based research through mobile methods also includes the collection of data in situ, that is, the collection of data in particular places or moments of meaning. Experience sampling method (ESM) is probably the most prominent form of these mobile methods. Part of the goal of this method is to answer survey questions or prompts about immediately preceding events or behaviours. Here again the physical and perhaps temporal closeness of the self-report with the phenomenon itself enhances validity. The longitudinal nature of mobile-based ESM also recognizes the object of study is seldom static but can best be studied through repeated measures to understand rhythms or patterns in the phenomena of interest.

Controlled complexity

Now that mobile devices have gained the capacity to run software applications and log activities, researchers have a large degree of control over the complexity of the data they wish to collect. Those seeking complex data regarding actions and reactions over time now have the possibility of creating customized apps that can interact with respondents—for example, delivering treatment materials during field experiments or encouraging them to follow scripted activities. In addition to collecting data regarding how respondents react and interact with the apps, they can further be developed to tie these resulting behaviours to times and locations. Even with small numbers of respondents, the resulting data can quickly grow much larger than data typically collected in the social sciences. And in applying time and location stamps to each resulting behaviour, the analysis of this data can be more involved and potentially complex than in more classic quantitative designs.

Even if researchers choose to collect communication or sensor data that is automatically logged by mobile devices rather than developing entirely new apps, the data they gather may be complex. As with data collected through custom software apps, the collection of time and location often implies complex, or at least nonstandard, approaches to data analysis. Perhaps more importantly, because these logs are of activities that often occur on a daily basis, when compiled, the resulting data sets are often far larger than sets typically collected through classic research designs. This gives rise to the issue that common analytic procedures—such as p star methods—no longer provide results that allow researchers to discriminate statistically “significant” associations, at least using the community standards commonly used at the time of this article. In addition to the problem of how to analyze these data, simple requirements such as having enough storage space and computational power to analyze them can become issues.

Despite the problems of data complexity, it is important to consider that researchers often have control over the data that they collect. For example, if a researcher plans to use text logs to understand interactional tendencies, it is rarely the case that the researcher will require all available logged texting data. In other words, data need not be collected simply because it is available. To address research questions about texting frequency, researchers may only require the number of texts sent each day, and not the specific times of messages or their contents. Moreover, when considering ethical obligations to respondents and practical considerations regarding the complexity and resources of processing large data sets, it is often in the best interests of the researcher and respondents that data collection be narrowed to only those data that are necessary to address their specific research questions. In short, although data complexity may be a common issue for mobile researchers, mobile researchers have control over the extent to which this issue can become a serious problem.

Ethical considerations, new and old

When considering the trove of sensitive and personal data collected by mobile devices, it may be tempting to avoid these methods and treat their use with some apprehension. These concerns are warranted, yet, it is worth considering that the handling of sensitive and personal information is also a common concern in many traditional research approaches and settings. Qualitative researchers using traditional approaches to fieldwork in communities or in-depth interviews are often faced with situations in which respondents reveal personal and sometimes compromising information. The revealing of this type of information is not always intentional, and there are often traces and records kept by researchers, either as part of a recorded interview or as part of the researcher’s field notes. Sensitive data may also be collected in traditional approaches to quantitative research, either because the topic of the research is sensitive in nature—for example, the researching of individuals with HIV or illicit opioid use—or perhaps because respondents behave regrettably during experiments.

Although the collection of sensitive and personal information is common in traditional research, it is worth considering ways in which data collection involving mobile devices is at least somewhat unique. The fact that mobile devices are carried throughout the day and often the main means by which individuals communicate can mean that data collection is also highly personal. Mobile devices can provide fine-grained information on location and behaviours, some of which can be quite compromising to individuals. Moreover, often respondents may not be fully aware of the data that their devices are collecting, and when they are aware and decide to disable or turn off their devices, these gaps in mobile records can sometimes be used to infer private and personal activities.

Fortunately, researchers working with sensitive data using traditional research methods have developed a number of strategies to safeguard respondents and themselves, many of which apply to mobile methods. First, as we point out in the Controlled Complexity section of this introduction, just because data is available does not mean that it must be collected. More to the point, researchers should avoid collecting sensitive information unless absolutely necessary, and only when informed consent has been provided.

Second, if it is necessary to collect sensitive information, safeguards should be in place to protect the identity of the respondents and ensure that identifying information is not revealed in the publication of research. In traditional research, this can be done by referring to respondents using pseudonyms, not publishing stories that would allow community members to infer the identity of a participant, and keeping interview recordings under lock and key. Similarly, mobile researchers collecting sensitive log data could mask or assign arbitrary values to all names and phone numbers, not publish specific times and dates of text messages, and encrypt data sets.

Third, as with traditional methods, researchers using mobile methods should have research proposals undergo ethical review before they begin collecting data. While the process of ethical review does not guarantee that everyone will agree with the ethical choices researchers have made, it helps to ensure that ethical norms within research communities are followed. The review process can also provide new perspectives on ethical decisions that researchers may not have considered. All of this decreases the chances that researchers will design studies that will harm participants or lead to other ethically questionable outcomes.

Fourth, with both traditional and new methods, informed consent is critical. This is especially important for scholars who wish to collect mobile device data that respondents may not fully understand or even know exist on their own devices. Researchers must explain in detail exactly what mobile data they wish to collect, how these data may be compromising, and how they intend to protect respondents. Only once respondents fully understand the nature of the data and what specifically is being collected can they provide informed consent.

In sum, while there may be no easy and universally applicable ways of dealing with ethical conundrums that arise when using mobile methods, there is much that can be learned and applied from traditional research designs. Nevertheless, the highly specific and sensitive nature of mobile data implies that researchers and reviewers should be more vigilant and cautious. Although this is obviously true for those using quantitative approaches in which logged behavioural data may be of great value, it also applies to those working in interpretive traditions who plan to integrate mobile devices into in-depth interviews and field work. Here too, researchers may be exposed to personal information on mobile devices, and discussions about mobile behaviour are likely to be personal and sometimes sensitive in nature.

Article summaries

In this section, we first provide summary highlights from each article and discuss how these highlights connect with the themes of field-based research, controlled complexity, and ethical considerations.

In the article, “Capturing Community in Mobility: Mobile Methods for Community Informatics,” Jiawei Chen and colleagues (2018) explore how mobile devices can be integrated into field experiments to better understand community informatics. They draw on four studies to show how mobile devices can be used to deliver “scripts” that structure community activities, and further collect data in the form of usage logs, user-generated content, and ESM surveys. Through these four studies, they provide several concrete ways in which mobile devices are central to a multimethod approach, allowing researchers to triangulate data in order to capture complex community dynamics. Complex data are collected with this triangulation in mind and this, in turn, results in more valid and detailed measures than standard and often burdensome self-report approaches. This is an excellent example of how the collection of complex mobile data is controlled and specific in nature. At the same time, the field-based approach used in this article increases the ecological validity of measurements.

In “Advantages and Challenges in Using Mobile Apps for Field Experiments: A Systematic Review and a Case Study,” Jingwen Zhang and colleagues (2018) perform a systematic review of 101 articles involving field experiments that use mobile apps. In their analysis they found that most of these experiments only used the apps to administer treatment materials, rather than collect data directly from respondents. This suggests that even researchers who use mobile apps in field experiments have yet to leverage the power of mobile devices. They argue that, in addition to administering treatment materials, mobile apps can serve as excellent experimental platforms. This is because they can be used to collect a variety of self-report and logged behavioural data and improve the replicability of experiments. Zhang and colleagues balanced their discussion by also considering the limitations of mobile apps in field experiments, particularly in regard to issues of generalizability (not all potential respondents may have appropriate mobile devices), retention (a common issue in field experiments), and design quality (apps can be buggy). Despite these limitations, this article makes a strong case for the use of mobile apps in field-based research and shows how the complex, multifaceted data that mobile devices can provide can be used to enhance research in this area.

“Gathering Data Through Text-Messaging to Study Question-Asking in Everyday Life” by Darren Gergle and Eszter Hargittai (2018) explores how text messaging can be used to administer ESM surveys. Although they use this method as a means of understanding information seeking, their approach to data collection should be generally useful to projects that require an in-depth understanding of complex behaviours that occur in natural settings. Using a single mobile phone and a custom-developed app, the authors were able to administer 257 text-message-based prompts to 25 participants over a period of 2 days. This resulted in 1,166 text message responses from participants, which provided a far more comprehensive understanding of complex information-seeking behaviour than could have been achieved through a standard questionnaire or interview. This approach illustrates advantages and considerations in mobile research. First, the fact that researchers were able to collect these comprehensive data with only a single mobile phone and app shows how mobile methods can provide cost-effective ways of collecting complex and large data. Second, unlikely many methods that rely on respondents having smartphones capable of collecting, storing, and processing data, basing data collection on text messaging alone significantly lowered the requirements for respondents to participate. Respondents only needed to own a mobile phone capable of sending and receiving text messages, and have the basic skills necessary to carry out this common activity. This increases the number and variety of participants that can be recruited in a field-based approach. Third, by developing their own app for data collection, the researchers pay attention to ethical concerns by ensuring a high degree of privacy to their respondents. Overall, their article is an important example of the power of field-based research, controlled complexity during data collection, and how ethical considerations factor into research design.

Jonathan Zhu and colleagues (2018) explore the issue of how researchers should quantify mobile usage in their article “How to Measure Sessions of Mobile Phone Use? Quantification, Evaluation, and Applications.” Using an open source data set in which volunteer participants installed an app to capture certain types of mobile activities, Zhu and colleagues identified 4,017 active users and developed an algorithm to identify and construct usage sessions. Logged wake up time, unlocked time, and locked/shutdown time were all used to construct use sessions that contained a simple start and end time for each user. Overall, the results allowed Zhu and colleagues to calculate the total amount of time that participants spent on their phones (a mean of 3 hours and a median of about 1.5 hours) and the number of sessions they had per day (a mean of about 24 and a median of 18). In order to detect daily rhythms of mobile use, Zhu and colleagues then grouped user sessions by an algorithm such that similar sessions clustered together. They also grouped users by sessions, such that users with similar kinds of sessions clustered together. Finally, Zhu and colleagues discuss various applications of these results. Using the same data, they investigate the hypothesis that mobile phones have increased fragmentation of daily rhythms, not finding much support, at least within a 3-month period. They also examine the regularity of mobile phone usage to find that almost half of daily sessions happen in fixed time slots. And they show the similarities and differences of mobile usage rhythms across the 13 clusters of users identified by the clustering algorithm. This paper is a strong example of how researchers can exercise control in leveraging complexity of mobile data, and it shows how secondary analysis of data collected from field-based research can help to inform critical issues in mobile communication.

In the article “The Smartphone as a Snapshot of Its Use: Mobile Media Elicitation in Qualitative Interviews,” Katja Kaufmann (2018) also uses smartphones to study mobile phone use. In particular, Kaufmann shows how the app adoption choice and the arrangement of apps on smartphones can be examined to understand personalization of the device. This in turns becomes evidence of how people incorporate mobile technologies into their everyday lives. The phone itself becomes an elicitation technique to encourage participants to think through how they use their phones in different scenarios. For example, Kaufmann demonstrates that some participants did not consider mobile shopping to be anything but purchasing through a mobile device; however, with the elicitation technique, they were able to talk through the variety of ways they use their phones as part of their shopping practices, including their “Shopping” app folder. But, Kaufmann argues, mobile media elicitation enables not just the study of mobile phone use but other phenomena as well. In particular, Kaufmann suggests geospatial log data can be helpful memory aids for interview participants. Additionally, the ever-present camera phone also enables the use and creation of visual prompts into the interview. In summary, Kaufmann argues that mobile media elicitation enhances ecological validity as a memory-aiding device, but also provides more complex data by enriching participant reflection and deeper meaning-making than self-report or log data alone.

The kinds of personalization of mobile media that Kaufmann studies in her elicitation techniques are similar to Melissa Brough and colleagues’ notion of appropriation in their VozMob project. Both papers argue that nontechnologists conform mobile tools to their needs and that these processes of personalization and appropriation are important means of understanding the ways mobile devices are taken up by individuals and communities. In “Mobile Voices: Design as a Method to Explore the Possibilities and Limitations of Community Participation,” Brough and colleagues (2018) reflect on the methodological choices, opportunities, and limitations of participatory design for an immigrant community in Los Angeles. The Mobile Voices project developed a mobile-based intervention which enabled immigrant workers and organizers tell their own life stories through a public online platform in an effort to empower them and counter the stereotypes circulating in the mass media about these marginalized communities. Drawing on participatory design, popular education, and participatory action research frameworks, the project involved members of the immigrant communities throughout the inception, design, and development stages of the intervention. Brough and colleagues identify power dynamics inherent in collaborative design, where friction and differences in perspectives become important means of revealing those power relations. For example, while the developers reported that the immigrant workers on the project influenced the technology design process and outcomes, the workers themselves did not. The workers did recognize their participation in the graphic design of the project website and in the reporting of bugs, however, their perceived lack of technological literacy and lack of time became obstacles for greater participation in the design process. In summary, the VozMob project reveals how participatory design can be applied to mobile communication technologies to support mobile storytelling for social justice. It thoughtfully engages with ethical considerations of power and agency within field-based approaches to participatory design. It also shows how mobile devices can be woven into research designs to uncover complex and rich qualitative data.

Our last article in the Mobile Methods Special Section is “Capturing Commemoration: Using Mobile Recordings Within Memory Research” by Carolyn Birdsall and Danielle Drozdzewski (2018). The researchers write about a variety of mobile methods employed in their ethnographic field research of Dutch Remembrance Day activities. The article expands the mobile methodological toolkit to include a variety of mobile communication technologies such as sound recorders, GoPro video cameras, digital cameras, and mobile phones to enable aural, visual, and textual data collection. In particular, the researchers employed a soundwalk method which enabled them to record and analyze the sounds of the 2016 Silent March through the streets of Amsterdam and coupled these data with still and moving visual images as well as thick description. The soundwalk analysis of the commemorative silence march revealed how the sounds of footsteps, the statistically predominant sound in the soundscape, became audible evidence to participants and researchers of collective participation in a commemorative act of silence. Additionally, the researchers reflect on the mobility of actually walking alongside marchers in the event and the importance of utilizing multiple mobile methods to investigate memory practices. This is a strong and original example of how mobile methods are well situated within a field-based approach.

Together, the seven articles in the Mobile Methods Special Issue represent innovative and thoughtful methodological approaches in mobile methods. As a group, they are also examples of how to successfully approach issues involving field-based research, controlled complexity, and ethical considerations.