Being digitally invisible: Policy analytics and a new type of digital divide

Being digitally invisible: Policy analytics and a new type of digital divide

Citation: Longo, J., Kuras, E., Smith, H., Hondula, D. M. and Johnston, E. (2017), Technology Use, Exposure to Natural Hazards, and Being Digitally Invisible: Implications for Policy Analytics. Policy & Internet. Early View doi:10.1002/poi3.144

(this blog post also published by Dr. Justin Longo)

Policy analytics involves the combination of new data sources – e.g., from mobile smartphones, Internet of Everything (IoE) devices, and electronic payment cards – with new data analytics techniques for informing and directing public policy.

The concept of the digital divide has been around for some time now. Whether it focusses on basic ownership and access to digital tools, or the ability to use them effectively, the digital divide means that some people are not able to send information into or receive information across digital channels. If you haven’t got a computer, you can’t tweet about it.

The access part of the digital divide has diminished in recent years (mostly because of the falling cost of technology needed to get online, and efforts by corporations and governments to put mobile technology into people’s hands at low or no upfront costs), but the broader concept of who is represented online is still of concern to researchers and policymakers.

With the rise of new data sources (often referred to as “big data”) driving the possibilities for policy analytics, work undertaken at the Center for Policy Informatics at Arizona State University in early 2015 came to focus on those who do not use or own devices like smartphones, IoE devices and transaction cards.

We explored the possibility that people may be rendered digitally invisible if the signals from their daily actions are not generated or captured because they don’t carry the devices that “big data” presumes, and therefore don’t figure into policy analytics. Failing to observe the lived experience of those outside the “big data” world may result in policy analytics being biased, and policy interventions being misdirected as a result.

With my CPI colleagues Evan Kuras, Holly Smith, Dave Hondula, and Erik Johnson, we set out to determine whether the concept of the digitally invisible could be shown empirically by conducting an exploratory study with the participation of homeless individuals in Phoenix and the Phoenix Rescue Mission, in the context of extreme heat exposure.

The results of that work have been published in a special issue of the journal Policy & InternetIf you don’t have access to the online version at the publisher, the published version can be accessed here.

Do the digitally invisible exist? Perhaps surprising to some, homeless individuals in the United States have very good coverage in terms of mobile phone usage (this is partly a result of government programs, and partly because a mobile phone becomes a crucial technology when you don’t have a fixed addess). And public libraries and other access points provide computer resources and Internet access, leveling the digital playing field and lowering cost barriers.

Yet policy analytics is based not on active participation, as is the focus of the digital divide literature, but instead is based on passive data contributions (through “big data”). We think this is the key idea that distinguishes digitally invisible from the digital divide.

For those without a smartphone, without a bank account or credit card, without regular and ubiquitous Internet-connected computer access, living beneath and beyond the network of sensors, monitors and data capture points, their existence is being rendered increasingly invisible, with policy developed using a policy analytics approach biased against them, even if unintentionally. As a result, policymaking is blind to their existence and policy based on incomplete evidence will not reflect their reality.

We’re at the early stages of the policy analytics movement. But we argue that a contextual awareness and humility should guide the developing policy analytics approach, understanding that it offers only a partial picture of a reality that is influenced by the values we bring to the analysis. We recommend being vigilant in looking for those who are hidden and will do the same in our future work.

We look forward to your comments.