GOODIES TO GO! (tm)
September 27, 1999 -- Newsletter #47
Full Text Search: The Key to Better Natural Language Queries for NoSQL in Node.js
GOODIES TO GO! (tm)
September 27, 1999 -- Newsletter #47
Please visit http://www.htmlgoodies.com
Greetings, Weekend Silicon Warriors,
September 23 marked double digit days until the year 2000. We're under 100 to go. Be afraid. Be very afraid...or don't.
Did you hear...
>eBay caught heck for selling stock. Dave Tumey put a 10% stake in a tech company up for auction on the site over the last weekend. The starting asking price was $599,000.00. Tumey said he got permission to sell the stock, but eBay later balked when a news organization began to call and ask why stock was being sold. Tumey had his trading fee refunded.
>Farmers are on the Net! A survey of those who work the soil found that the number of farmers who use the Web for information regarding their business jumped from 13 to 27 percent. Montana is the most Web savvy with over a third of their farmers online.
>Hey! Did you hear that the Navy thinks we're sunk? According to their research, 60% or around 400 major cities will have Y2K problems with their water and electrical systems. When the story broke, apparently heavy pressure was brought to bear and now I read that the Navy isn't so sure they got it right. Well, I just hope that their retraction is right rather than their first statements. You can't pressure a computer into working to avoid embarrassment.
Now onto today's topic...
I am in a conundrum.
I want to do something, but I don't know how.
Did you see the story in mid-August about Internet Addiction? This study claimed that 6% of all people who use the Net are addicted.
I read the results of the study and, at first, thought that this was something we should be concerned about. Then I ran into how the survey was conducted and immediately lost faith.
David Greenfield, who created the study, gathered responses from over 17,000 people who filled out an online questionnaire. The results, according to Greenfield, reported that 6% of those who answered his survey are addicted. He then claimed that if you apply that 6% figure to the general population, we're talking millions of people who are technically addicted to the Web.
Scary, right? Welllll, let's look at the survey itself. First off, the survey was online so users had to know where to find it. That requires someone who is already Web savvy which leans toward people who work with the Net regularly.
Next, the majority of those who answered the survey did so because of a report run on ABC Television News. This, again, suggests that those who are more interested in the Web were answering the survey. Do you jump to see a Web site when it comes up on the TV?
Finally, the survey was a series of yes/no questions. As I understand it, there were only 10 questions and if you answered yes to half of them, you could be addicted.
So, what does this mean? To me it means that the data is skewed toward the result. The method of drawing people to the survey was skewed to get those who might be addicted instead of a random picking of the general population. Secondly, I don't know that 10 this-or-that questions can qualify someone as "addicted."
Bottom line: I don't believe the results of the survey are representative of the general population. Do I believe Net Addiction exists? You bet, but 6% of the general population? I don't think it's nearly that high.
Now that I've said that, let me comment to Dr. Greenfield himself. Good job! You took a shot at surveying what must be the most difficult group of people to solidify... Internet users. I feel for the guy. He knows the drawbacks of the study, I'm sure, and wishes he could have avoided them. But how? That's the question I am now fighting with.
I (along with EarthWeb and Macmillan) am going to undertake a survey on the Web. My big question is, how can I give everyone a fair shot at answering the survey?
Those who use the Internet make up a very interesting cluster of people. It would be great if we, the researchers of the world, could somehow survey them correctly.
It's a little more than just posting a survey and hoping people come by. Let me tell you why. When you perform research on a group of people, like the person did above, you want that group of people to be representative of the larger population. That way you can say, with a degree of certainty, that this group feels this way, thus the population probably feels this way.
In order to get that degree of certainty, you have to perform what is known as a random sample. Everyone in the population will have an equal shot at getting picked.
That's the rub of doing research on the Internet. How on earth will you make sure that all people have an equal shot at answering your survey? You can't randomly choose e-mail addresses like you can generate random phone numbers. You can't get a list of all those on the Web, like you could a corporation's employees. Even if you could, the number of people on the Web is ever-increasing in size. How would you keep up with it?
Come to think of it, you can't set out to contact people randomly at all. They have to contact you.
Every time I see a survey done on the Web, the results are always skewed because by posting a survey to the Web, a researcher immediately sets up a process whereby respondents are scooped up as they go by rather than chosen because they might be representative. It's frustrating, to say the least.
But I'm going to try it. I will put together a survey created to gather as much clean information about a subject as possible. I'll post the survey on the Goodies site, as well as other great EarthWeb sites. I'll advertise on search engines for people to come and answer the survey. I'll take steps to keep someone from answering twice. I'll try to create a volume of data so large that even though some skew may occur, somewhere inside... the truth will out. A simple volume of data that leans one way or the other will grow out of the responses and I will lay it out and proclaim it as the general thinking of the Internet population.
I will compile the data which will help people do something better and perhaps allow for a larger share of the population to be surveyed by the next person.
Then some guy who writes a newsletter will tear it apart.
And that's that. Thanks for reading. It's a weekly outlet that I enjoy putting together.
Joe Burns, Ph.D.
And Remember: "Now is the time for all good men to come to the aid of their party." If you've taken a typing class, you know that little ditty. It came from Charles Weller who created it as a drill. Contrary to popular belief, it doesn't carry every letter on the keyboard (no Z or Q). In later typing manuals the line was changed to: "Now is the time for all good men to come to the aid of their country." The reason for the change was because in this form it fills out a perfect 70-letter line if you put the period at the end.