IBM Click Here!
| Navigational map -- for text only please go to the bottom of the page ||Forums|
Readers' Choice Awards (siteadm) Fri, 13 Jun
What Poll? (forum) Sat, 14 Jun
Another poll was held?? (up_yours) Mon, 16 Jun
Big Name Computer Company (rparker) Mon, 16 Jun

Bias.

Posted by: acagle
Date posted: Mon, 16 Jun 1997

I'm not sure I buy the no-bias argument. We got into this the last time, but I think it's a bit more complicated than having a representative sample of the readership. I say this because this introduces it's own bias into the results. Eg.: just restricting the discussion to the OS realm, if you take a simple random sample, you will get a sample that's skewed a priori towards Windows simply because there are more Windows users out there. Generally speaking, this will only tell you that a lot of people use Windows, whether by personal choice or because their employer mandated it. In other words, your sample is already weighted towards one OS simply because of its market share, so you are really only measuring market share rather than, shall we say, "preference". A bit like taking a random sample of car owners and asking them which they like best. You can bet GM would get a huge vote, not because most people really like them, but because most people can afford them. Something like Alfa-Romeo won't even make the cut because they aren't represented.

In either case, bias is relative to the problem you're interested in, which has never been provided anyway.


Respond | Search | Help


For questions or comments, email the Forums Editor

kristin_kueter@infoworld.com