Confusion about Census numbers

July 6, 2017 11:51 AM

I've always been confused about how the total number of people doing something is determined. For instance, this site says 88% of people in the U.S. use the Internet and slightly more than 10% don't.

But if the population is divided roughly by thirds; 1/3 under 18, 1/3 between 18 and 65, and 1/3 over 65, then how could, especially, very young children and very old seniors be included in that 274-million that are said to be users? When I see such a large number, are researchers forgetting that among the people they include may be babies or dementia sufferers who, in no way, could use a computer?

posted by CollectiveMind to Science & Nature (4 answers total)

Best answer: So, that particular one has methodology. It's not about whether they do use the internet regularly, it's about whether they are theoretically able to. Here is their blurb about the methodology, linked to at the bottom of that page, all stress theirs:

User
An individual who has access to the Internet at home. This indicator does not record use, or frequency of use, but only access. In order to have access, the hardware equipment must be in working conditions, the Internet subscription service must be active, and the individual household member must have access to it at any time (there must be no barriers preventing the individual from using the Internet). The hardware equipment may or may not be owned by the household. There are no age limits (minimum or maximum), so an Internet user can be of any age. There can be multiple devices and services within the household. The data is collected through annual household surveys administered by individual countries based on ITU guidelines.[1] The United Nations Statistics Division has recommended collection of data on households accessing the Internet also outside of home [2], but this is not a Core ICT Indicator. [3] An "Internet User" is therefore defined as an individual who can access the Internet, via computer or mobile device, within the home where the individual lives.
posted by brainmouse at 11:58 AM on July 6, 2017 [4 favorites]

Best answer: Great explanation from brainmouse as to this particular survey. Also, this isn't Census data -- it looks like it comes from yearly househould surveys (the Census is only taken every 10 years). Every survey will have different methodology -- for example, many are going to be limited to adults 18 and over, and in any case, all are limited to those capable of filling out a survey (so someone with severe dementia is by definition not going to be included because they can't fill out a survey). If you wanted to study babies or people with dementia, you'd have to use a different type of methodology than a regular survey -- perhaps interviewing/surveying parents, or getting information through nursing homes or hospitals, or something else.
posted by rainbowbrite at 12:03 PM on July 6, 2017

Best answer: When I see such a large number, are researchers forgetting that among the people they include may be babies or dementia sufferers who, in no way, could use a computer?

The census doesn't ask about your internet connectivity. But many researchers do. And there are a lot of different things counted that all get glossed as the same thing

- access to internet
- access to broadband (and definition of broadband)
- population surveyed (some count adults, some count basically everyone. Not everyone who is old has dementia, but you can assess populations from that)
- access to internet at home (which is what this survey is about)
- internet users (sometimes there are some people in the house who use the internet and some who don't)
- smartphone owners (and looking at 3G/4G/LTE/whatever coverage areas)
- service areas of ISPs

I pay a lot of attention to Pew Research Center's Internet's reports on this sort of thing because they've been doing it over time and are not trying to sell anyone anything. They're also very forthcoming about their methodology so it might help you to poke around some there.
posted by jessamyn at 12:15 PM on July 6, 2017 [1 favorite]

Also, this isn't Census data -- it looks like it comes from yearly househould surveys (the Census is only taken every 10 years)

This is sort of a pointless nit, but there are Census data that aren't the census. The American Community Survey is a Census program and product that's taken (at least) monthly, though they don't release data anywhere near that frequently. It's what used to be the long form of the decennial census that they spun off into its own thing. ACS data are Census data (data from the Census Bureau) but not census data (data from the decennial enumeration).
posted by ROU_Xenophobe at 1:29 PM on July 6, 2017 [3 favorites]