The Scientific Value of Unscientific Polls ( talks about DU)

[Kadie]

The Scientific Value of Unscientific Polls
10.05.04

As I write, we await the start of the vice-presidential debate between Dick Cheney and John Edwards. Word has leaked out that the Bush campaign has issued orders for its troops to hit the Internet (and other media) full blast after this one. Apparently they believe they were outworked last time, allowing Kerry’s folks to control the post debate spin. Not this time! The GOP faithful have their orders -- send letters to the editors, call opinion lines, jump onto the online chat boards and, neither last nor least, hammer the unscientific Internet polls.

Not to be outdone, the Democrats have also sprung into action. Over at the Democratic Underground, for example, the administrators have posted a thread with direct links to more than 60 online polls. Charge!

This is nothing new, of course, to anyone who spends any time at all at partisan discussion boards. One of the most common topics at these places is this or that online poll that needs to be hammered by the faithful.

All of which raises an interesting point. I’ve always laughed off these "unscientific polls" as, well, unscientific. But maybe that’s not true anymore. Maybe in today’s Internet world unscientific polls have scientific value -- not in judging public opinion -- but in judging organizational strength. Who knows? Maybe the side that can best hammer the polls, is the same side that can best hit the ground running on election day to get out their vote.

more...
http://www.poppolitics.com/

[Skinner]

1. I have to kick this.

I can't believe nobody responded. :shrug:

[GiantCranes]

2. I agree

They are increasingly showing effectiveness of organisational strength, but they are still practically useless as an indicator of true public opinion. I wish the media would not use them as much as they do.

[spooky3]

4. welcome to DU, GiantCranes!

[meganmonkey]

6. So-called 'scientific polls'

Maybe this will get people questioning the validity of polls which claim to be 'scientific', as if that makes them accurate...
Wishful thinking on my part, I suppose, expecting people to question much of anything these days...

[bemildred]

3. Ok, this is correct.

They are not polls and not scientific, so as indicators of how things
will go on election day they mean little. But OTOH they are "facts",
as factual as you can get, the "poll" was put up and that many people
voted that way (or cheated that way), and that certainly means that
there are people out there on the internet with "nothing better to
do". And that does mean something. The "outrage" of the media when
their lying, manipulative "polls" don't give the desired results means
something too ...

[WilliamPitt]

5. Scientist Discovers Water Wet. Amazement Ensues.

http://www.americanpolitics.com/20010609GodFreepers.htm...

God Sees The Freepers
by William Rivers Pitt

----------->June 9, 2001<-------------- (APJP) -- There was a striking moment during the interregnum last November/December that has stayed with me ever since.

Inside Politics was running 24 hours a day on CNN, you will recall. I was watching one evening, several days into the theft, and there was Judy Woodruff interviewing conservative columnist Bob Novak. The question of the hour was whether or not Al Gore should just quit and go home.

On this night, Novak was pointing to a public poll that had been running on the CNN.com web site. You know these polls: log on to a news site and give your opinion on whatever happens to be the headline of the day. The poll Novak referred to asked the question: "Should Al Gore concede?"

The results showed that some 89% of the American population who found their way onto CNN.com voted "Yes" to this question. The count of those who voted numbered in the tens of thousands.

Novak flapped this poll all around the studio as indisputable proof that a large majority of the American people saw Gore as a thief and a usurper and a sore loser who should just go away.

I never forgot that night, and never lost the sneaking suspicion that something shady had occurred. Somehow, someone had flooded that poll with "Yes" votes to skew the result. I had no proof, and the theme song to X-Files was sounding in my head -- I was mortally sure that something was rotten in Denmark.

Now, after all these months, I have figured out what happened that night.

That CNN.com poll was 'Freeped'.

(snip)

A common Freeper tactic is to post on the FreeRepublic forums a notice that a poll exists somewhere which asks a question dear to the conservative heart, such as: "Should the Congress pass more gun control legislation?" or "Is Bill Clinton the illegitimate spawn of Satan and Baal?" The URL to this poll is provided, and the Freeper legions swarm to vote... say, "no" on the first and "yes" on the second. There are a lot of Freepers, and many of them will vote multiple times. This obviously skews the result.

This is how a poll is 'Freeped.'

[sendero]

7. I agree with this guy's point...

... they are not useful in absolute terms but they definitely are in relative terms.

They are a general indication of 1) the organizational savvy of either party and 2) the extent to which people are motivated enough (give a crap either way) to be bothered to participate in these polls.

I'm surprised not to hear much about this, but it is easy for the people running the polls to tell "where" the respondents are coming from. For example, if you click a link in a thread here, the poll you visit will log a "referrer" of "democraticunderground.com". It would not be technologically difficult to essentially toss out the votes from a particular referrer or set of referrers. I've seen no evidence of that happening, yet.

Also, someone could definitely be gleaning some useful information from these "referrer logs". You can defeat this by copying the URL (link) into another browser's location (address) bar and visiting the site this way. That visit would have no referrer.