Sunday, May 20, 2007

Ron Paul fans gaming the polls? No they're not

The conventional wisdom is that Ron Paul's visible success in most online polls and networking site activity is the result of coordinated multi-voting and page-refresh schemes being used to create the illusion that his support online is much greater than it actually is. (Note: It's odd that this conventional wisdom is not applied universally to other candidates receiving high marks in online polls. Why?).

While there is no way to verify the integrity of the many online polls or video view counts, without knowing whether they use cookies or IP addresses to verify participants, or use nothing at all to filter multiple votes from the same computer, there are many other metrics in the sea of data to conclude Paul's support is real.

1. Number of blog posts. Since the first debate on May 5, Paul has consistently outranked all the other GOP candidates in the number of posts being generated in the blogosphere. This is not an easily automated process, but one that requires much more than 1 minute per post to accomplish, an eternity in internet timeframes. In order to be counted in the rankings, the blog would already have to exist, and have some "authority" by technorati standards, which can't be conjured up in a matter of hours out of thin air. This makes blog postings a good metric to use to gage the attention Paul is getting in the blogosphere.

Granted these post totals do not differentiate between positive and negative commentary, but they do indicate the interest level in the subject matter at hand. Let's look at the media's assertion that Guiliani won the second debate because of his "smackdown" of Paul's root cause of anti-US terrorism remarks, even though Paul finished a strong second in Fox News' own poll, 6 points ahead of Guiliani. Oh, it generated alot of buzz across the internet and indeed in the mainstream media that stretched on for several days, but if you look at the blogs, G-man didn't have half the attention Paul had after that little dust-up.

Blog posts over the next three days following the debate, per data from technorati.com:
May 16:2700 Paul
1500 Guiliani
May 17:2000 Paul
1100 Guiliani
May 18:1450 Paul
850 Guiliani


2. Web traffic. Web traffic is the number of uniques visitors to a site in a given period. Most traffic analyzers limit users to their IP number, or a cookie installed in their browser. In the case of Alexa.com, whose data we use in this example, the user is identified by his or her computer on the basis of a toolbar the user has consciously installed, meaning that each computer is counted as one unique visitor.

These Alexa.com graphs show that the number of unique visitors to the candidates' websites are real visitors, and the vast majority are assumed to be seeking more information about the candidates. Taking into consideration the support for Guiliani gained from his remarks in the second debate, his site traffic remained relatively calm and stagnant, while traffic to Paul's site literally soared. (Note: click here to view graph if you can't see it below)








An important point to make about internet users is that those who know how to game the system generally don't install these web tracker toolbars on their computers, thus further increasing the likelihood that the numbers on Alexa's charts are genuine.

3. Contributions. The best way to gage support for a candidate is to look at the source of his campaign contributions. Unfortunately we've got another month and a half to go before the current quarter results are released. It is my hope that Paul will conduct some online fundraisers very soon, where the contributions are tracked and reported in realtime, in much the same way that Howard Dean transformed his internet presence in 2004 into a formidable powerhouse campaign. In the event Paul does launch an online fundraiser with realtime reporting, I predict a more resounding earthquake irreparably undermining the mainstream media's credibility once and for all. For more enlightened commentary on the Ron Paul effect on the media establishment, see Butler Shaffer's post at Lew Rockwell's blog: Ron Paul 1, Establishment 0.

If you haven't contributed to Paul's campaign, please do so now to prove to the skeptics that his support is very real.

3 comments:

Anonymous said...

http://digg.com/2008_us_elections/Vote_Poll_2008_Presidential_Election

IP protection is on, and I check logs as much as I can to look for duplicate IP's.

There is no refreshing or multi-voting.

Anonymous said...

Another question I have is: Are Ron Paul supporters the only ones smart enough to vote multiple times?

as an example of what I mean, consider that Mitt Romney, and Ron Pual won about 25% of the vote each, but John McCain was at 2% last time I looked.

You could say that Mitt and Ron had some one pulling the lever for them, but where's the McCain supporters?

Also, has anyone noticed that Gallup hasn't published any poll numbers since May 13th?

Chris said...

I recall reading on one of the liberal blogs last week that McCain had put out notices to rally online support in response to the debate, for people to hit the online polls and vote him up. Apparently his efforts didn't work so well.

And Romney had launched a sign up America campaign just two days before the debate in part to get more visible online support for his theatrics. I'm sure that if one or two overly-enthusiastic Paul supporters hit the button a couple extra times, you can trust the same holds true for the other candidates as well.

I don't know what to make of the traditional offline polls. They limit their contact to under 500 people via landlines generally, which makes the poll results highly unbelievable. I think they should henceforth qualify all poll participants by asking whether they've seen one of the debates or not and limit the pool to those who are well informed of the choices.

There's a growing gap between the polling outlets and the online support base. What will it take to merge the two worlds into something reliable and credible? I don't know, other than having some way to qualify the participants.