How many teleworkers are there today? Who knows? As I noted in a previous blog it is hard to count teleworkers. One problem is that the definition of teleworker/telecommuter varies from country to country and counter to counter. Full-timers working at home for a distant employer are the easy core. It’s at the edges where the counting gets difficult — and the definitions differ.
Another, growing problem in the teleworker counting business is the simple fact that contemporary technology has made it really hard to count reliably. An excellent overview of the problem is given in an article in the New York Times by Cliff Zukin of Rutgers University. Titled What’s the Matter With Polling?, the article paints the picture of the rapid decrease in accessibility of your average pollee. The polling business is changing for the worse
The last poll that I ran was in 2000, covering the status of telework in the United States. At that start-of-the-century period we did things the old-fashioned way: a random number land-line-telephone survey that involved operators calling thousands of telephones over several days in order to get enough responses to be statistically significant. Each successful call needed the callee to respond to a number of questions about his/her experiences with telework. The number of calls required depended on three things: whether
- an actual person answered the phone at the number called;
- the person was a teleworker of some sort; and
- the person was patient enough to answer all (or most of) the questions.
This is very expensive. There are the costs of preparing the survey questions, phone bills, the salaries of the interview team, the number crunching needed to do the statistical analysis, and report preparation and editing. This sort of survey is not something to be attempted on a whim.
And that was when such surveys were relatively cheap. Now it’s a different story. For starters, in 2000 most households in the US had land-line phones so the likelihood was high that your sample would be representative of the general US population. Our response rate was around 30%. Today, according to Zukin, response rates are at or below 8%. In 2000 the phone dialing was done by automatic dialers. Today, because people have cut the cord and rely on cell phones that are nor accessible to automatic dialers, phone calls have to be dialed by hand. Therefore each call, successful or not, takes longer. All of this means that the price per successful call is at least four times higher than it was 15 years ago.
So go to the Internet, right? Not so fast. Not every American of working age has access to the Internet even though most teleworkers do; Internet users are concentrated in certain age groups so it’s hard to get a sample that’s representative of the whole population. In this case, teleworkers might be overrepresented, giving a too optimistic view of reality. Translation: you can’t be sure the results are meaningful.
The consequence of all this is that my estimates of the number of teleworkers in the US are based on data collected in 2000 plus a computer model based on history up to and including 2000. That’s the bad news. The good news is that US data on teleworking are better than that available for other countries.
The worse news is that forecasts based on polls for elections are among the shakiest of all these days, according to Zukin.
So how many US teleworkers are there today? I estimate that there are somewhere north of 43 million of them, collectively preventing about 57 megatons of CO2 from entering the atmosphere.
But who knows?