Sampling Bias

Published: 2014-06-10
Last Updated: 2014-06-10 23:29:31 UTC
by Daniel Wesemann (Version: 1)
0 comment(s)

Today, I was researching a rather complex subject, and it brought me to dozens of web sites to catch up on the latest techie clue. And what felt like half of the web pages popped up that obnoxious
 

HI. CAN WE ASK YOU A COUPLE QUESTIONS ABOUT OUR WEB SITE?

[YES]  [NOT RIGHT NOW]


insert that seems to be all too common these days. Who on earth is clicking "yes" on these?? Or, put differently, how irrelevant must the results of such "surveys" be if the respondents probably all are bored loafers who have unlimited time on their hands, and don't mind to be distracted from their work by an (end|use|point)less survey that intrudes into the thought process, clamoring for attention?

It's what statisticians call "sampling bias". Something like going to a pub to determine if people like alcoholic beverages. Surprise surprise, many of them do :). I suspect the results of such web site "surveys" are similar: WOW!! 96% of the respondents say our web page is cool!!1)

 


1) n=18 / N=1'284'154

 

Keywords: survey web
0 comment(s)

Comments


Diary Archives