Online Panels and the Digital Divide
By John Block, NPD Research Sciences

Much has been made of the "Digital Divide" – the line separating those with Internet access and those without. The phrase was coined roughly seven years ago. At the time, it referred to demographic factors such as race, income, gender and education that made it more likely for some U.S. residents to be online, while others remained unconnected. In the early days of online research, the term had political overtones – that despite its original promise of equal information to all, the Web might make social stratification worse rather then better.

For online research providers like The NPD Group, the Digital Divide can have very important implications. If the research being conducted is intended to be representative of the U.S. population, demographic skews that affect the Internet population can have a direct impact on the "representativeness" of the research findings. This article considers the Digital Divide and the impact it has on today's online consumer research.

The Demographic Divide

Seven years is a long time, especially when the topic is the ever-changing Internet. So where does the Digital Divide stand today?

In its earliest days, the Internet was primarily the province of young, white males with some extra cash to spend. Since then, we have seen tremendous progress in the bridging of demographic barriers limiting access to the Internet. Broader access at public venues such as libraries and schools, online access at work, and a drastic reduction in the price of computers and Internet access at home (especially broadband access) have put the Internet within reach of nearly all who really want it. Most research firms, including The NPD Group, report Internet access to be much more widespread across demographic lines than it was even a few years ago.

So while demographic differences still exist between the online and offline populations, online panels such as those maintained by The NPD Group have been able to sufficiently compensate for the demographic differences. Because all demographic groups are now represented online, these groups can be used to represent their offline demographic counterparts. In addition, well-managed panels like NPD's can now attain response rates near or surpassing those achieved by mail access panels, which have long provided the best panel response rates in the business of consumer research. NPD's online consumer panel consistently sees response rates in the high 50% to low 60% range. This results in less weighting and greater sample efficiency.

The Attitudinal Divide

Instead of a demographic divide, recent studies are pointing towards a different type of divide: an attitudinal divide. The PEW Internet and American Life Project report released in April 2003 maintains that the divide does still exist, though not in quite the same way as was originally feared. The study, titled The Ever-Shifting Internet Population, contends that while some demographic differences still exist, they are fading and the main reason a person is not online today is most likely personal choice. Net usage today falls roughly between 58% and 61% of the U.S. population. The PEW report contends that those 40% or so of Americans who are still not online choose not to be online. An article titled The Digital Divide that Wasn't (BusinessWeek, August 19, 2003) also suggests "the divide that does exist between the Web and non-Web proficient is no longer defined simply by income, gender, race, or education."

What remains of the divide now appears to be more about people who have access to information versus those who don't. It's about people who can research an audio or video product online before going to the store, and who may, therefore, be more savvy consumers. It's about people who can go to chatrooms and get advice about home improvement and major domestic appliances from friends and strangers before going to the store to make a purchase.

If the divide is more attitudinal then demographic, is this bad for researchers? The implications of this difference would suggest rather significant disparity between measured actions and reality, yet evidence indicates the differences are subtle. For survey research, numerous "research on research" reports have confirmed that while results of online and offline methodologies have some minor differences, overall findings are the same regardless of methodology. With regards to consumer purchase trackers such as those used by The NPD Group for many of our Worlds data sources, in most cases NPD has point-of-sale data to use for comparison to consumer data. Using proprietary data calibration techniques, NPD aligns the consumer data to the POS data, then projects these relationships to non-matching consumer data to address residual biases. During the years of developing and applying this methodology, NPD has never been able to identify any systematic patterns of bias, other than those normally associated with survey research design effects.

The reasons that these attitudinal differences do not have more of an impact on survey results are not readily apparent. Good survey design and a sound understanding of the fundamental differences between research methodologies are obviously factors, as is proper handling of the data prep and delivery. In the long run, it all comes down to one question . . .

Is Online Research Viable?

The answer to that question is yes.

As the "Demographic Divide" continues to fade and techniques such as calibration allow for adjustment for attitudinal differences, the reasons to move surveys online become ever more compelling. The responsiveness, speed, cost and convenience of online research are unmatched by traditional methodologies. NPD clients can be confident that our online studies produce accurate, reliable views of who's buying what, and why – and that's what NPD's online research is all about.