This portal was created through a partnership between the University of Iowa and Arizona State University, with the support of the National Science Foundation’s Broadband Community Capacity program (NSF Award #1338471).
BUILDING A RESEARCH COMMUNITY ON BROADBAND AND INTERNET USE
This portal makes publicly available innovative data on internet and broadband use, particularly at the subnational level, where many policy initiatives are taking place, and where there has been a shortage of reliable and comparable data on broadband, mobile and Internet use by individuals and organizations. Through this portal, we are making available to researchers and policymakers several new types of data, along with graphics and maps to visualize trends and comparisons.
Internet Use, 1997-2016
The state, county, metro and city data posted here offer unique data over time, from 1997 to 2016. Through the use of multilevel models, estimates were created for these subnational geographies based on large-sample surveys of the U.S. Bureau of the Census – the Current Population Survey Internet Supplement (1997-2012) and the American Community Survey (2013-2016). This state, county, metro and city data are also available at the Broadband Use Dataverse hosted by Harvard Dataverse.
In addition to the above data, this project also supported three small grants for demonstration projects on subnational broadband use (discussed below). This included multilevel estimates of activities online for counties and metropolitan areas; content analysis of Spanish language access for city and county government websites; and scraping of social media websites for police departments in five communities.
How is this data unique?
- Prior to 2013, estimates of Internet use from the Census were available only at the national and state level . The estimates here fill this gap, showing substantial variation by place over time.
- Beginning in 2013, the American Community Survey (ACS) provides data on Internet use and broadband for metropolitan areas and later for places with populations of more than 60,000. Our estimates drawn from the ACS go a step further, disaggregating the results by race, ethnicity, age, educational attainment, employment status, and language.
What trends are visible, with implications for research and/or policy?
In addition to providing an important resource for scholarly research, comparisons over time and across communities can help us to learn more about broadband use and needs for public policy.
Overall, there is marked variation across geographies and communities.
- This includes urban and rural differences (notable at the county level), but also dramatic variation in urban areas. In 2014, only 55% of Brownsville, TX residents had some type of Internet service at home (including mobile), in comparison with 95% of residents in Gilbert, AZ. This suggests that social inequality matters for the percentage of the population online, not just broadband infrastructure.
- Observations of Internet, broadband and mobile use at the subnational level reveal large demographic disparities in some communities; gaps that are obscured by national averages.
- There is a general flattening out of the adoption curve for broadband, or high-speed Internet at the same time that mobile increases. Yet patterns vary across communities and mobile use does not always erase disparities in Internet use.
The data made available here can be used to further explore reasons for these trends and the policy solutions to address these needs.
How can this data be used for research?
Among the many ways this data can be used, researchers can explore –
- Explanations for variation in Internet use across places and over time
- Effects of interventions for changes in Internet use (for example - Does broadband adoption at home change after the introduction of broadband infrastructure in a rural county?)
- Policy impacts of Internet, broadband, and/or mobile use (for example - Does increased Internet use in a metro area lead to changes in employment?)
We hope that researchers will share how they have used the data. Let us know about your ideas, and send us links or electronic copies of publications using this data, so that we can share it on this site.
How can this data be used for policy?
Policymakers can see how their jurisdiction compares with others, and how this has changed over time. Estimates can be read like percentages, and there is no knowledge of multilevel models or advanced statistical knowledge required to use the excel spreadsheets.
Graphs on this site show trends over time.
Links to Tableau allow users to access interactive maps and other graphics, to visualize trends and comparisons.
We provide examples on this site of how the data can be visualized in bar graphs and maps. This will help users to easily explore what can be accessed on Tableau.
 See, however, estimates for 2007 and 2009 for the 50 largest cities and metropolitan areas in Mossberger, K., C.J. Tolbert and W.F. Franko, 2013, Digital Cities: The Internet and the Geography of Opportunity (Oxford University Press).