Real-time Earth and Moon phase

Sunday, October 05, 2008

Freaque opinion polls, hiding data, etc.

Reading the "wizbangblog" today, a post by DJ Drummond on opinion polls interested me. As I have been given up on paying attention to opinion polls for a long time, his post gave me a good explanation for my behavior:
I have followed opinion polls for some years now, and I have great respect for a poll which is properly done, and properly reported. Unfortunately, there are a lot of polls which get published without providing readers or viewer with the proper context, and worse there are a number of major polls which hide vital parts of their internal data. The media, it must be remembered, is not and never has been in the business of reporting facts. Instead, they want to sell a story, in order to claim lots of attention which will increase readership or ratings. And in this election, the media has overwhelmingly been biased . . .
I stopped paying attention to polls mainly because I found the poll results and reality are totally not correlated. I sensed that they might play tricks to manipulate results. As a scientist working on data analysis, I am especially offended by " hide vital parts of their internal data."

Many years ago when I just started my research career, I was awe-struck by a young scientist who was a new Ph.D. from Stanford. He did some impressive laboratory wave experiment that verified some theoretical aspect of the wind wave studies. When I met him in a conference, I felt like I was meeting a star. Some time later I met one of his Stanford classmates in another conference. This classmate told me the truth behind his experiment made me lost some of my respect for him. It turned out that this guy worked very hard, he did hundreds, may be thousands experiments on the study. Only a few runs, by chance, fit the theory. But he only took those few case to his adviser to show that he verified the theory. He got his degree and became a rising star! He became fairly successful and well-known. I don't know what had happened to him. Somehow he disappeared from the wave research world for quite some time now. He's about my age, probably he's enjoying retirement somewhere also.

Anyway for a scientist dealing with data, hiding, disgarding, fixing, or twiddling with data are all inexcusable, bordering on falsification. Please remember it next time when you see Algore's beautiful global warming data plot again. Are they for real? Of course for pollsters, they are media instruments which in 2008 are just political instrments, they should not ever be considered as scientists by any stretch of imagination. The 2008 election season has one more month to go. The pollster are still poised to make a whole lot of money yet. Can they still lift their head up high after the election day?

No comments: