Faulty Polling and Distorted Coverage

I am a huge critic of polling in general, especially because they are so ubiquitous these days and so often poorly done, misleading, or pre-designed for a specific result.

This poll that I ran across today seems a prime example of the kind. It could merely be that the person writing the article is inept and trying to claim that the polling data supports something it never intended to support. But at any rate, the presentation is typically slanted.

The article/poll discusses a study done of Japanese and American attitudes towards our expecations of a World War III, the use of nuclear weapons, and general attitudes about the two countries.

As an example of its poor analysis:

According the poll, we are told that "60 years after [the U.S. used atomic weapons on Japan]...[m]ost people in both countries believe the first use of a nuclear weapon is never justified."

Now, note the present tense usage of "is never justified." Not was, is. The sentence then is not referring a polling question not about the literal first use of atomic weapons in Japan, but instead the concept of "First Use/First Strike" atomic attacks generally. This is an important distinction.

Clearly, the implication of the article's word choice is to call the U.S. bombing of Hiroshima and Nagasaki "first strike" atomic attacks and state that a majority of people disagree with their occurance.

Definitionally, there is a disconnect here. The general understanding of "first use/first strike" attacks is a strike done in the absence of previous "hot" conflict (Reading cold war literature or watching Hollywood productions from..oh, how about the last 50 years...would tell you as much).

Quite obviously, from this definition, U.S. usage of atomic weapons --coming after we had been embroiled in a world war for the better part of 4 years and suffered nearly a half million deaths--wouldn't fall under the umbrella of a first strike (I suppose technically it might be seen as a Last Strike). So why attempt to argue that the "majority-disagree" data is relevant?

Laziness on the part of the writer is a possibility. Also possible and more likely is that the author wanting the poll to be more clearly in line with the message they want it to portray, whether for ideology or to attract interest/ratings.

Interestingly, the poll does specifically ask about opinions on American usage of atomic weapons in WWII. The results are:

"Two-thirds of Americans say the use of atomic bombs was unavoidable. Only 20 percent of Japanese felt that way and three-fourths said it was not necessary. Just one-half of Americans approve of the use of the atomic bombs on Japan."

Such a result is interesting on its own (Such as how it is that 1/3 of those Americans polled can find the action unavoidable, yet still think it morally contemptible. I haven't taken an Ethics class in a couple years, but there seems a disconnect there to me).

But it's not nearly as catchy as saying that majorities in both Japan and America ostensibly opposed them since they oppose first strike attacks. Afterall, would it surprise anyone to learn that the Japanese don't approve of our bombing them?

And that gets me to the point of this post. Ultimately, who cares? The poll in question isn't really that significant or important. But it illustrates how media coverage of polling data can be shaped to create a story instead of reporting one.

The difference between saying "One in Three Americans finds President doing poor job" and "Two-Thirds of Americans satisfied with President" is more than merely rhetorical since rhetoric shapes perception.

As any reader of the former HorseRace blog, a site for in-depth statistical analysis of polling data during the last Presidential election, can tell you: Polls and the media coverage that surrounds them is deeply and unmistakably flawed.

It is unfortunate then that they are given such prominence in how we analyze our world. In times where the latest polling data on public support for our war efforts will shape how we proceed Diplomatically/Militarily, it is enormously important that our polls (and the media presentation of them) accurately reflect public sentiment. Not doing so will someday turn another Tet-Offensive victory into a defeat.

Of course, as long as the powers that be (namely politicians, the Media, etc.) are aware that perception of reality if more powerful than reality itself, there isn't much hope...