Monday, March 20, 2006

"Dewey Defeats Truman"

"Dewey Defeats Truman." Lewis Kramer has this short article on The Story Behind "Dewey Defeats Truman."

In the current media saturated world, polling is everywhere. Upsets do occur but most of the time we have a good idea who is going to win an election. With constant polling (and gerrymandered districts), there is little suspense.

The Presidential elections of 2000 and 2004 of course did not follow this. The polls showed no clear majority. Further, the USA has 51 distinct Presidential electoral districts all of which are nearly impossible to draw electoral lines around to play politics which benefit one party only.

The article noted the 1948 election and Chicag Daily Tribune, "There were many factors involved in producing this error edition. Returns were coming in slow and they were running out of time before the printing deadline. The staff, based on early returns, 'felt' Dewey would win. In addition, many of the regular Chicago Daily Tribune staff were out on strike so inexperienced people were setting the type. They did the front page, and portions of a few others, on a typewriter. Rather than erasing typos or incorrect numbers, they simply "x"ed them out with the 'x' key on the typewriter. In the far right hand column, there are even 5 lines of type upside down! All issues went out this way. "

I have to guess that in the future, despite any advances in polling, Presidential elections are going to be harder to predict. Calling people up is harder and harder to do. Many voters use cell phones or internet phones and it is difficult to find a number to call. How do you randomly get a phone number of someone using an Internet phone service provider who has no land line? The polling firms claim they are finding techiques to update their methods.

However, I am skeptical. And is this a bad thing? The media should learn constraint and stop making potentially inaccurate predictions. And does it really hurt anyone to have to go to bed without knowing the election results until morning?

The Inaugural Address of Harry S. Truman failed to mention this event. Truman was amused by the whole affair anyway. He won and that was what counted. Regardless, the whole affair provided a good lesson on how not to place an overabundance of trust in the media.

1 comment:

Jennie W said...

I have several comments on this issue:

1. This is an excellent example of why we should be wary of statistics to predict events. I use this very example in my math class to illustrate this point.

2. The media is way to trigger-happy to call elections. I grew up in Alaska and because we are the last polls to close the media was already saying who won before we get to the polls. While it is law that the loser can't concede until the last polls close in Hawaii, that doesn't stop the media from broadcasting it. My father never gets to vote until after work and would make a conscious effort on election day to avoid all media - the TV, the radio, etc, so he could go vote and not feel that his vote was useless. He shouldn't have to do this. We are taught that every vote matters, yet the media tends to ruin this. Add that to the electoral college system...sometimes its a wonder why anyone votes. Now I also firmly believe that if everyone actually voted we would see a governmental change in attitude...but that's a different story.

3. I also think that the continuous polling makes people feel that the election has already happened when it hasn't. I wonder if it doesn't sway people in their voting habits as well. I wouldn't mind seeing a law on the number of polls that can be taken and when.

4. Also people need to learn what a poll is and what is says/doesn't say. A poll is the opinion of a select group of people. While any good poll is chosen randomly all have some biases. 75 years ago a problem was telephone polls - only wealthier, or town, people had telephones and tended to be Republican...thus slanting telephone polls (I think this many have been one of the problems with the Dewey/Truman polling...but I can't find anything to back that up...I will have to do some digging unless someone else knows). Today think about internet polls - how many people in the US are still without Internet access? And even now that telephones are fairly ubiquitous, how many of us ignore calls we can't identify on the caller ID? I know I do.

Well, this got to be rather long...