The previous two posts of this digital inclusion series explored broadband access and digital distress. While broadband access remains a rural issue, digital distress is one that impacts more of the region’s urban areas. Be it broadband access or digital distress, both components represent only part of the digital inclusion story.
Socioeconomic characteristics, known to impact technology adoption, have yet to be factored into our assessment. In this post, we examine the digital divide index. This index is a robust digital inclusion metric, including both infrastructure/adoption indicators as well as socioeconomic characteristics. This index assigns a score— ranging from 0 to 100— to census tracts, where a higher score denotes a higher divide. Three scores are included in the index.
The infrastructure/adoption (INFA) score includes percent population without access to 25/3; percent of homes without internet access (not subscribing); percent homes without a computing device; and average advertised download and upload speeds.
The socioeconomic (SE) score includes percent population ages 65 and over; percent population 25 and over with less than high school; individual poverty rate; and percent noninstitutionalized population with any type of disability. These are all characteristics known to impact technology adoption.
Lastly, both INFA and SE scores are meshed, given equal weight, resulting in an overall digital divide index (DDI) score. For more information on the DDI methodology, please visit purdue.pcrd.edu/ddi.
As a general rule of thumb, if the INFA score is much higher than the SE score, efforts need to be made to improve device ownership as well as broadband infrastructure and adoption (mostly through educational campaigns and subsidized subscriptions) in the community. On the other hand, if the SE score is much higher than the INFA score, efforts need to be made to demonstrate the technology’s relevance and teach digital skills and literacy.
Figure 1 maps all 13,015 Census tracts in the upper Midwest divided into four equal groups based on their DDI score. A darker color denotes a higher digital divide. Notice how the geographic pattern is similar to the digital distress metric, yet not identical. Groups of tracts with a high digital divide are in southern Illinois, northern Michigan, and northeastern Minnesota to name a few areas.
Figure 1. Digital Divide Index Scores by Quartiles
Due to space limitations, the tract-level DDI scores were aggregated to the county-level.
The three counties with the highest SE score were (remember, a higher score means a larger divide):
|Lake County, Michigan||100|
|Alexander County, Illinois||98.45|
|Roscommon County, Michigan||91.99|
On the other hand, the three counties with the highest INFA score were:
|Alexander County, Illinois||100|
|Pulaski County, Illinois||81.87|
|Holmes County, Ohio||75.42|
Finally, the three counties with the highest DDI were:
|Alexander County, Illinois||100|
|Lake County, Michigan||85.06|
|Pulaski County, Illinois||78.84|
Download an excel spreadsheet with the DDI, INFA, and SE scores for all counties in the region analyzed.
Returning to the analysis of census tract-level data, we discover that close to one-fifth (19 percent) of the upper Midwest population (or 9.9 million people) lived in tracts with a high digital divide. On the other hand, around 30 percent (or 15.7 million) resided in tracts with the lowest digital divide.
Figure 2 explores the percent of residents in digital divide quartiles by type of neighborhoods. The higher you move up on the individual bars, the higher the divide. The more you go to the right, the more rural. What the data show is that rural constitutes a larger share of the digital divide. In particular, 41.4 percent of the population living in completely rural areas are also in areas with the highest digital divide. The share for completely urban areas was less than half at 19 percent.
A similar pattern emerges with regard to the digital distress metric. Figures 3-6 show that areas with a high digital divide had a higher share of minorities, are less educated, have a higher unemployment rate, working age residents (ages 16 to 64) participate less in the workforce, and are poorer. Worth noting is that 2.3 million children lived in areas with the highest digital divide.
Figure 7 examines all digital inclusion-related indicators in our digital inclusion three-part series by digital divide index quartiles. However, we did not include socioeconomic attributes since these were already examined. Note all indicators in the highest divide quartile (to the right) are higher compared to the region overall, as well as all other quartiles. The largest difference between lowest/highest quartile is in regards to homes not subscribing to the internet (gray bar), where almost one-third (31.7 percent) of homes with the highest divide did not subscribe compared to 8.1 percent in the lowest quartile.
Remember that digital inclusion is a complex concept that cannot be accurately measured through the use of a singular metric. While broadband access is critical, it is important that other digital inclusion components be understood and addressed as well. As described throughout this series, digital inclusion is affected by access to broadband, internet subscription type (or lack thereof), devices (or lack thereof), and socioeconomic factors.
It is important to point out that two key indicators, cost and internet use, are not included in our analysis. While cost is a key factor affecting digital inclusion, it is not available nationally. In many instances, adequate broadband is available, but too expensive in terms of monthly subscription costs or the expense associated with the type of data plans needed to make full use of various broadband applications.
Use is important, as well, and research have delineated three distinct digital divide levels. The first level focuses on access; many times determined by a dichotomous measure of yes/no broadband availability. The second level examines how internet use and utilization differs across socioeconomic groups. The final level examines how different internet uses affect social, cultural, and economic outcomes. In other words, does internet use improve the quality of life of individuals, households and/or communities and if so, in what ways? Unfortunately, much like the situation with cost, a nationally available dataset on internet use is not available.
Despite these limitations, the digital inclusion indicators and metrics that we have analyzed do offer some insights on the steps that leaders and practitioners should consider in their community, economic, and workforce development efforts. They include: