OEA Working Paper 51
Abstract: High-speed Internet access, or broadband, is critical to economic opportunity, job creation, education, and civic engagement. Yet, there is a digital divide between people who have access to high speed Internet/advanced telecommunications, and those who do not. Currently, most of the discussion regarding the digital divide focuses on access to fixed broadband networks. In this paper, we examine the digital divide as it relates to mobile broadband. Specifically, we explore the following two research questions. First, is there a digital divide in how certain groups access mobile broadband as measured by the mobile connection technology? Second, is there a digital divide in the quality of their mobile broadband as measured by download and upload speeds? We investigate the first question by running a multinomial logit of the type of on-air connection (WiFi vs. 3G, Non-LTE 4G, or LTE) on U.S. county demographics and characteristics, as well as technological variables; and we investigate the second question by running separate OLS regressions of log download and upload speed for each technology on U.S. county demographics and characteristics, and technological variables. Connection technology and speed data from the last six months of 2016 are obtained from the Ookla Speedtest app, and U.S. Census data are used to provide local demographic information and other county characteristics. Overall, we conclude that the mobile digital divide does exist across certain dimensions. Rural areas are somewhat more dependent on non-WiFi mobile technology and experience slower speeds on their mobile connections. We also find that counties with higher minority populations are more likely to use older mobile technologies and experience slower speeds. Counties with older populations are more likely to use mobile technologies and are more likely to have slower speeds. Counties with larger households are more likely to use WiFi and also have faster WiFi. Indicators of economic health, as well as technological and infrastructure related variables, have mixed and complicated effects. These complex results suggest that future research and on-the-ground data are necessary to further examine the nature of the mobile digital divide.