Nov 25, 1997
I loved the front cover of Wired 5.12 - but the survey described in Katz's article wasn't worth the trees it was printed on.
Several things stand out:
- First, the sample size of 1,444 Americans is too small to permit extrapolation to a population of 250 million with a reasonable degree of error. Responsible reportage indicates the degree of error in a survey (for example, Gallup Surveys, in which 1024 Canadians of a (voting) population of about 15 million are accurate to within 4%, 19 times out of 20). This article does not indicate the degree of error.
- This degree of error is magnified when conclusions are drawn about the 'connected', particularly the 'superconnected' and the 'connected'. 2% of 1,444 (the 'superconnected') amounts to 28 people. This is by no means a sufficient sample size. The probability of error for this group must be enormous.
- Several people have already drawn attention to the classification system employed. I find it ironic to find myself placed among the 'semiconnected'. I don't have a cell phone (I don't need it - I'm not on the road enough) and I don't have a pager. I don't even have a laptop. My four desktop computers, three phones, and web server provide my connectivity.
The surveyors must have known that their division into such categories would provide odd data. One person above said it's because ownership is easier to measure than usage. I don't buy it. How hard is it to ask, "How much time do you spend online in a week?" So we have to look for another reason for the categorization.
Let's ask: who owns what?
- Computer? Could be anyone, from my retired father to my baby brother.
- Cell Phone? Could be anyone again, though with telephones ubiquitous ownership is a luxury, unless you are on the road a lot.
- Laptop? For the most page, could be anyone, but owners are much more likely to be well-off, and much more likely to travel a lot.
- Beeper? Nobody owns beepers unless they are in business for themselves, or are members of professions such as medicine or law.
So? Well, the conclusion I'm led to draw from this is that the definition of 'connectedness' in this survey reflects, not connectness, but rather, one's employment status. Specifically, the definition favours those who travel a lot or own their own business. Businesspeople, and especially the self-employed, tend to be more conservative than the population at large. It's hardly a surprise, then, that the 'connected', as defined in this survey, would fall into the conservative end of the spectrum, is it?
- Even give the four categories stated above, the article plays fast and loose with the categories. For example, while there are four catgories, often the results are presented split into only two: 'connected' and 'unconnected'. In another case (p.80) they are split between 'superconnected' and 'connected'. That's really odd, and the only good reason to present the information this way is to obscure an anomaly in the data.
- And what is really obscured is that a guy with a pager (but no computer, and no internet) counts as 'connected' in some of these results, thereby biasing the results of the 'connected' set in favour of the demographics of people with pagers.
- There is bias in the questions as well (I work from the assumption that the questions printed in Wired are the questions which were asked - though I have no good reason to believe that). For example, look at the way concepts are paired:
- confidence - democracy (p.71)
- confidence - free-market system (p. 71)
- control - change (p. 72)
- productive - organizations (p. 72)
- internet - responsibility (p.80)
- needs - social security system (p. 80)
- confidence - public education (p. 82)
- I want to know how many people reported the following:
- That they own computers
- That they own cell phones
- That they own lap tops
- That they use email
- That they own a pager (beeper)
- Finally, let's look at the source. I don't know Frank Luntz, and only a passing knowledge of Rudolph Giuliani. But I'm quite familiar with Newt Gingrich's political learnings. The Reform Party in Canada is a similarly conservative party. Indeed, if you would like to know Reform's political position, read the survey results! These are the people Luntz works for - and I know enough about the Reform Party to know that they won't hire a pollster unless they can be sure of the poll results. For Reform, polls are not a means of measuring public opinion, they are a means of shaping it. Such is also the case in this article, I would say.
I believe that what was presented in the survey was a distorted picture of the digital citizen. True, people on the internet are more right wing than the mainstream - in my view, that reflects their relative affluence. But they're not that far right. And by gar, they are, as Katz observes in his article, not happy with the current political system, they don't kowtow to organizations and leaders, and I'm sure they're not nearly as family and church oriented as the survey suggests.
I'm really disappointed. It could have been such a useful and informative article. As it was, it was nothing more than a waste of trees.