California’s digital divide and the specter of data uncertainty for evaluating broadband coverage

•The digital divide remains a pervasive problem in the United States.•A lack of reliable broadband data makes developing public policy solutions difficult.•Error and uncertainty in Census data result in unreliable broadband coverage estimates.•Correcting these errors can result in changes to block-level broadband analysis. The digital divide is a persistent feature in the United States. While most Americans have access to some form of broadband internet connection, the bandwidth, quality of service (QoS), and choice of providers remain highly variable throughout the country. For example, while the residents of many urban areas can choose between gigabit fiber, cable, or digital subscriber line connections from multiple providers, residents in rural areas often suffer from limited platform and provider choices. A fundamental problem with developing effective public policy for broadband in the U.S. is the lack of reliable data for estimating the geographic coverage of broadband service. This paper aims to highlight how data uncertainty can negatively impact coverage estimations in subtle and not-so-subtle ways. The state of California serves as our study area. This paper suggests a combination of uncertain broadband data combined with Census data errors co-mingle to create a variable snapshot of local broadband provision and geographic coverage. We provide several strategies for dealing with these uncertainties to improve the evaluation of broadband in California and the United States.