The year 2016 marked the launch of a totally revised set of questions on computer and internet use in the American Community Survey.

The Census Bureau generally keeps questions the same or similar to provide consistent measurement of trends over time. However, in this case, the evidence was strong that the old questions were not doing well at capturing internet access via “mobile broadband,” or cellular data plans.

When the new, revised questions were tested, there were many improvements, including a stunning increase in reported mobile broadband usage.

Figure 1 shows the old and new questions. More discussion of the changes is provided in a user note for the 2016 ACS data, and in a report on testing of the new questions.

With these new estimates in hand, it is interesting to take a look at results with a focus on smartphones and mobile broadband use.

Figure 2 shows how households with smartphones, desktop/laptop computers and tablets connect to the internet -- groups that overlap, although not completely.

Overall, most households combine mobile broadband (cellular data plans) with home-based internet connections, which include both wired connections (cable, DSL, or fiber-optic) and satellite, dialup or other types of service.

This combination represents 71 percent of smartphone users, 71 percent of desktop/laptop users, and 77 percent of tablet users.

As one might expect, people who use only one type of broadband connection (mobile or home-based) are likely to use one that matches the type of device they have.

Home-based connections were the only connection for 16 percent of desktop/laptop households, compared with 8 percent of smartphone households and 10 percent of tablet-owning households.

Using a mobile broadband connection alone was more popular with smartphone households (12 percent) than with desktop/laptop households or tablet-owning households, each at 8 percent.

Among smartphone households, 8 percent had no broadband connection at all, which was true of only 6 percent of desktop/laptop households and 5 percent of tablet-owning households.

Preference for one type of broadband or another is based on more than just the type of device in a household.

Table 1 shows how several important characteristics are associated with broadband connection types.

Age, race, education, poverty and place of residence play a role in households’ choice of mobile broadband and wired connections, factors recorded in other Census Bureau reports.

Asian households are most likely to have mobile broadband or wired connections, followed by non-Hispanic Whites, Hispanics and Blacks.

Householders under 30 are most likely to have mobile broadband in their households, while wired connections are most prevalent among households with householders aged 30 to 64.

Also more likely to have these connections are better educated households, households not in poverty, and households living in metro areas. Across types of connections, mobile broadband is more common, used by 68 percent of households versus 67 percent using wired connections.

Groups more likely to use a home-based connection than mobile broadband include households with householders 65 years or older, non-Hispanic white households, and householders with a bachelor’s degree or higher.

Older households show another interesting trend. Even though most people over 65 are not in poverty, older households are less likely to have a mobile broadband connection than households in poverty. By contrast, older households are ahead when it comes to wired connections.

Satellite service is less popular than mobile broadband and wired service, ranging between 4 and 7 percent for nearly all groups.

The one exception is the set of rural households living outside metropolitan areas: 11 percent use satellite internet service.

Dial-up or other internet service does not exceed 4 percent for any group.

Although the Census Bureau has no direct evidence on why people have one type of internet connection or another, there are several possible influences:

• For poor households, the cost of service may be important.
  
  
• For non-metro households, availability of service may be a key factor.
  
  
• For older households, the unpopularity of mobile broadband hints at differences in personal tastes or lifestyle.

The hope is that these and other relationships will become clearer as researchers and analysts look at these data in the future.