Columns

Problems of Agency in Visual Risk Communication

By Daniel P. Richards | Guest Columnist

This column features ethics scenarios and issues that may affect technical communicators in the many aspects of their jobs. If you have a possible solution to a scenario, your own case, or feedback in general, please contact column editor Derek G. Ross at derek.ross@auburn.edu.

Editorial Note

Thanks to the Internet, we, as a general public, have access to powerful, often free, analytical tools. We can crunch numbers, create charts and graphs, build infographics, and map our way to anything and everything. We can track health issues with tools like Flu Near You (flunearyou.org), and we can track environmental issues like sea level rise with such tools as Climate Central’s Surging Seas Risk Finder (sealevel.climatecentral.org). All of this access to analytics is good, right?

In this month’s column, Dan Richards asks us to think about the ethics of providing expert-level analytical tools to a lay public, and to consider how tools designed for audiences with complex understandings of events and constraints might, in some ways, mislead non-expert audiences. In working through his analysis of sea level visualization tools, he asks us to consider a few important questions. For example, is there such a thing as too much agency? What are the risks of putting powerful technologies in the hands of audiences who may lack understanding of surrounding complexities?

As always, we welcome your responses, and we trust that your conversations on our columns have been engaging. Let us know your answers to the questions we’ve posed, your thoughts on our roles as technical communicators in general, or send us your own ethics cases or column ideas. Please send your responses to derek.ross@auburn.edu. Responses will be printed in an upcoming issue of Intercom as space permits.

—Derek G. Ross

IF WE BELIEVE scholars Johndan Johnson-Eilola and Stuart A. Selber, technical communication is all about solving problems. Well, I currently have a problem: The city I now call home is prone to flooding. And everyone who lives here knows it, but isn’t quite sure what to do about it. Ranked second behind New Orleans, LA, in United States cities susceptible to sea level rise (Montgomery 2014), our historical, naval-base dominated locale of Norfolk, VA—lucky as we’ve been without a major hurricane since 1933—is quite vulnerable, due in no small part to overwhelmed drainage systems, ground subsistence, and the rivers that meander so beautifully, but so piercingly, through the heart of our city.

Fortunately, the city planners and coastal managers taking charge of this problem have access to a certain type of data visualization tool that should be—and are (Kain and Covi 2013; Stephens et al. 2015; Herring 2017)—of interest to technical communicators: interactive sea level rise viewers. Emerging as a genre almost in and of themselves, these online, publicly accessible viewers afford users the opportunity to interactively overlay flood and sea level rise data on top of maps, so as to visualize the effects of water inundation of a given region or community. Planners and managers can then use these open exploration visualizations to communicate to various stakeholders in their communities about their infrastructural vulnerabilities and risks to residents.

Figure 1. Climate Central’s Risk Zone Map

On the surface, sea level rise viewers, such as those developed by NOAA and Climate Central (see Figure 1), respond to key principles of research in risk communication about the effectiveness of visualizations to augment raw data (Lipkus and Hollands 1999) as well as the importance of local “place” in the communication of risk (Scannell and Gifford 2011). They help city planners and coastal managers do better work and communicate more effectively to public stakeholders. As you might imagine, however, as publicly accessible tools, the general public might not find them that useful, or as engaging as they might be.

Responsibility of Access

When a lay user in a coastal town discovers these tools, they might soon realize that these tools were not designed for them and they are not responsive to their own needs. Take, for example, NOAA’s Office for Coastal Management’s Digital Coast, which has its own sea level rise viewer fully available to the public online. When a user enters the tool, they see a disclaimer about how it is a screening-level, planning-reference tool wherein the user accepts all risks, and has the purpose of providing “coastal managers and scientists with a preliminary look at sea level rise and coastal flooding impacts” (NOAA 2017). Other tools, such as the U.S. Geological Survey’s and Climate Central’s, provide no such disclaimer. Even so, users might not even be prone to read the technical disclaimer anyway (Siegel and Etzkorn 2013).

My own qualitative research observing how lay residents use a sea level rise viewer (Richards 2016), as well as other research done with subject matter experts (Stephens et al. 2016), indicates that people who are not coastal managers, but who stumble upon these publicly accessible websites, might not be making accurate assessments of their region’s vulnerability. The tools might be too complex for them to use.

Given the scientific look of these tools, and their engaging and cutting-edge features, residents might think their home is safe up to eight feet of inundation, when in fact it would be at risk from three feet of storm surge. They might believe they are safe during a Category 2 hurricane, but have not accounted for the accrued risk of sea level rise over the next 20 years. While the visual nature of the tool might lead to engagement, it might also lead to overconfidence; while the open exploration nature of the tool allows for personalized choice in risk data, it also allows more room for error. And given what is at stake, the last thing we need are lay residents using scientific tools to make erroneous assessments of the risks to themselves and their communities when it comes to flooding and sea level rise.

A Question of Agency

Sonia Stephens (2015) poses a useful question in regard to these tools: is there such a thing as too much agency? Stephens argues that all too often agency is constructed as inherently good, indeed, a capacity that most individuals want in a given situation. Agency means choice, agency means access, agency means power.

But are there situations in which perhaps users of a technology can have too much agency? To answer this question, Stephens differentiates between the first space of agency and the second space of agency. The first space refers to the degree of rhetorical choice afforded by a given interactive tool. Rawlins and Wilson (2015) use this first space in their analysis of interactive data display technologies to create a sort of typology of agency in interactive tools. Generally speaking, the more agency the user has, the more robust an argument a user can make in their data visualization.

The second space for Stephens, however, refers not to open data exploration or visualization but to the applied aspect of interactive tools—the aspect of these tools that allow for “making real-world decisions” (2015). She argues, in essence, that we as technical communicators should not conflate design agency with decision-making agency, not conflate use with usefulness, to borrow from the work of Barbara Mirel (2004). Having at your disposal as a user the ability to locate your city, select from demographic or property value layers, then slide an inundation meter up and down is a different type of agency than using these tools to decipher where, for example, you should buy when relocating to a new but vulnerable city.

Increasingly there is a concern in academic circles to think about agency, both in terms of nonhuman agency and in terms of designed technologies that allow people to achieve their own goals. The problem in the actual execution of this is that in providing so many inlets or opportunities for agency, there is a corresponding and correlative risk of error, or given lack of expertise by the user. And we should consider the fact that, while designed for experts, these tools are publicly accessible and sometimes offered with limited disclaimer. What sort of responsibilities do these agencies and organizations have in ensuring accurate risk assessments in these reliable-looking Web tools?

The Power of Constraints

This question is inherently an ethical one: if these complex tools exist and are openly promoted in online public spaces, accessible to all, but designed for experts in coastal management, then what sort of constraints could and should be placed on them? Is a disclaimer, which most users will not read in their rush to see visualization, enough to mitigate against this? Need NOAA and Climate Central create separate tools, setting up a bifurcation between SMEs and lay users? Should these tools sacrifice design choice in an effort to promote more accurate risk assessments?

Figure 2. NOAA’s Sea Level Rise Viewer

In speaking with the users in my own research, what most of them wanted were pre-set scenarios. They wanted cut-and-dried information about what would happen if X took place next year in their neighborhood, or what would happen if Y took place around their grandparents’ house in Florida. When faced with the sheer complexity of information, even in a visual mode attempting to simplify risk data, they wanted resonant, familiar events they could visualize.

NOAA, as of this year, has created a beta version of their sea level rise viewer, which includes a Scenarios feature on the tool (see Figure 2)—but even this feature pertains to high-level information about data projections and the effects of sea level rise on their region. Users will want more concrete scenarios, more relatable and structured scenarios to better help understand their risks and thus make more informed decisions, which is, after all, the intended purpose of these tools.

Information designers and technical communicators can play significant roles in making these tools more amenable to public use by studying how real, local residents use them, and discovering the types of scenarios or situations most on their minds when thinking about purchasing property, relocating family, or selecting schools for their children. Perhaps more concrete scenarios, such as a “Hurricane the severity of Sandy” could be included, or “What would happen if we polluted at the same rate?” Such scenarios would allow local users the ability to more accurately and realistically visualize potential situational impacts.

In The Design of Everyday Things (1998), Donald A. Norman identifies seven principles for transforming difficult tasks into simple ones. The fifth one is of particular relevance in this discussion moving forward: “Exploit the power of constraints, both natural and artificial.” I would encourage all of us working with or interested in researching these types of visual risk communication tools to heed Norman’s advice on the power of constraints in design, and think more about striking the right balance of agency between open exploration and scenarios. Ultimately, we must consider the question: Is providing influential stakeholders like coastal managers and city planners with enough slack to explore and present a range of data worth leaving lay public audiences vulnerable to committing errors in risk assessment?

References and Recommended Reading

Climate Central. 2017. Surging Seas: Risk Finder. http://ss2.climatecentral.org.

Herring, Jamie, Matthew S. VanDyke, R. Glenn Cummins, and Forrest Melton. 2017. “Communicating Local Climate Risks Online Through an Interactive Data Visualization.” Environmental Communication 11.1: 90–105.

Kain, Donna, and Michelle Covi. 2013. “Visualizing Complexity and Uncertainty About Climate Change and Sea Level Rise.” Communication Design Quarterly 1: 46–53.

Lipkus, Isaac, and Justin G. Hollands. 1999. “The Visual Communication of Risk.” Journal of the National Cancer Institute Monographs 25: 149–163.

Mirel, Barbara. 1998. “Visualizations for Data Exploration and Analysis: A Critical Review of Usability Research.” Technical Communication 45.4: 491–509.

_____. 2004. Interaction Design for Complex Problem Solving: Developing Useful and Usable Software. San Francisco, CA: Elsevier Science.

Montgomery, Lori. 2014. “In Norfolk, evidence of climate change is in the streets at high tide.” The Washington Post. 31 May. https://www.washingtonpost.com/business/economy/in-norfolk-evidence-of-climate-change-is-in-the-streets-at-high-tide/.

NOAA. 2017. Digital Coast. https://coast.noaa.gov/digitalcoast/tools/slr.

Norman, Donald A. 1988. The Design of Everyday Things. New York, NY: Basic Books.

Rawlins, Jacob D., and Gregory D. Wilson. 2014. “Agency and Interactive Data Displays: Internet Graphics as Co-created Rhetorical Spaces. Technical Communication Quarterly 23.4: 303–322.

Richards, Daniel P. 2016. “Helping Local Residents Make Informed Decisions with Interactive Risk Visualization Tools.” SIGDOC ’16: The 34th ACM International Conference on the Design of Communication Proceedings. New York, NY: ACM.

Scannell, Leila, and Robert Gifford. 2011. “Personally Relevant Climate Change: The Role of Place Attachment and Local Versus Global Message Framing in Engagement.” Environment and Behavior 45.1: 60–85.

Siegel, Alan, and Irene Etzkorn. 30 March 2013. “When Simplicity Is the Solution.” Wall Street Journal, C1.

Stephens, Sonia H. 2015. “Interactive Data Visualization for Risk Assessment: Can There Be Too Much User Agency?” SIGDOC ’15: The 33rd ACM International Conference on the Design of Communication Proceedings. New York, NY: ACM.

Stephens, Sonia H., Denise E. DeLorme, and Scott C. Hagen. 2014. “An Analysis of the Narrative-Building Features of Interactive Sea Level Rise Viewers.” Science Communication 36.6: 675–705.

_____. 2015. “Evaluating the Utility and Communicative Effectiveness of an Interactive Sea-Level Rise Viewer Through Stakeholder Engagement.” Journal of Business and Technical Communication 29.3: 1–30.

_____. 2017. “Evaluation of the Design Features of Interactive Sea Level Rise Viewers for Risk Communication.” Environmental Communication 11.2: 248–262.

DANIEL P. RICHARDS is an assistant professor of English at Old Dominion University. His scholarly interests include visual risk communication, environmental rhetoric, and writing pedagogy.