As some readers may know I recently chaired an AAPOR Task Force that looked at online panels. We just released our report a few weeks back. One of the issues we struggled with a bit was settling on a reasonably accurate estimate of the percent of the US population that is online and therefore reachable by panel recruitments.
My friend Bill Blyth loves to tell people that the only way you can get an accurate estimate of Internet penetration is to do an in-person survey with a good probability sample. The Current Population Survey fits the bill (it has a very high response rate to boot) and the last time CPS asked the question was 2009 when they reported that 69 percent of US households had Internet access. Put another way, they report that 73 percent of adults live in households with Internet access. But do they use it? To answer that question we looked at PEW data from a telephone survey that included cell phones. They found 74 percent of adults having access (at home or someplace else like work or a library). But when we looked at how often those adults actually go online it turns out that only about 63 percent go online at least once a week.
Subsequent to release of the report the folks at Mediamark shared some of their data, also from in-person interviewing of an area prob sample. They found 70 percent of households with access, 73 percent of adults with access someplace, and 69 percent saying said they had used the Internet in the last week.
No matter which numbers you use there’s no escaping the conclusion that a significant percentage of US adults cannot join panels, do online surveys, be your friend on Facebook or Tweet. They can’t be in our samples and they are not part of the online conversation we are urged to listen to.
This likely is not news to anyone. Still, it strikes me as odd that we seem to know very little about how this substantial offline population differs from the online population and whether those differences matter in the research we do. Sure, we can all recite the demographic and socioeconomic biases in Internet use but the truth is there are people in every demographic group who are not online or online rarely, even some reasonably well-educated white people making serious money. What’s with these people? Should we just keep ignoring them or should we try to understand why they’re not online and whether it matters?