Boulderar15
03-24-2013, 08:49
http://www.denverpost.com/opinioncolumnists/ci_22843876/real-skewed-polls
In other words they didn't like what they heard so it has to be rigged!
More than 100,000 people voted in one of The Denver Post's online opinion polls last week, which is a significant number.
But the poll's result — with 70 percent of respondents favoring a ballot measure that would overturn a ban on high-capacity magazines — is largely insignificant.
Certainly the result shows passion for one side of the issue, but it also shows something else — ballot stuffing.
On a typical day we'll have a few thousand readers weigh in on opinion poll — maybe 25,000 in a good week. But in the last week, we've counted more than 200,000 "votes" in a series of polls.
Noticing the trend, a reader from Centennial wrote to ask if the polls "accurately reflect the opinions of the people of Colorado?"
He was pointing specifically to a poll asking whether Democrats in the statehouse were overreaching.
When we posted results of the overreach poll in the March 19 edition, 91 percent of 33,681 respondents said "yes." What's notable there is both the percentage who answered in the affirmative, but also the number who participated. The previous high was about 31,000 in a July 2012 poll in the wake of the Aurora theater shootings.
On March 16 we asked for opinions on the bid to overturn a bill limiting gun magazines to 15 rounds.
Watching those results online, one reader remarked:"It's pretty obvious to me that someone is using their computer skills to manipulate the poll."
He's right.
This is not the "skewed polls" argument that challenged the weighting and methodology of many scientific polls during last fall's presidential election.
This phenomenon is real.
Changing results of an online poll is simple and a common practice. If someone is passionate about an issue, they can — and often do — weigh in multiple times on a single question by voting from different devices. They can do it whether they live in Grand Junction or the Grand Caymans.
An organized effort to influence results of an online poll is called "freeping." The name is a nod to an online forum called the Free Republic, whose members are called "freepers." They're notorious for pushing results of online polls in one direction or another — notably on Second Amendment issues.
In fact, they did it to our gun-magazine poll on March 15 (and to other polls on our site). After 24 hours online, the magazine question had received 26,704 votes, with 69 percent favoring a vote to overturn the law. Anti-overturn forces then mounted a campaign that pushed that figure down to nearly 51 percent.
"Basically someone has hacked it and are adding an average of 26 votes per minute to the 'keep the ban' column," one freeper lamented.
"After Dynachrome pings the list it'll look better," another remarked. (The user known as Dynachrome maintains a "ping list" of users summoned to help sway the vote in online polls — and shortly after he or she was engaged, support for overturning the ban climbed to 61 percent.)
The poll also drew the attention of prominent conservative blogger Michelle Malkin and magazine maker Magpul, which posted a link to the poll in its Twitter feed.
As of this writing, 112,000 votes (and climbing) have been received.
That doesn't mean gun-rights supporters are alone in rallying people to vote for their cause in an online poll.
After the presidential debate at the University of Denver, almost 11,000 people voted in our poll, with 40 percent saying they thought Barack Obama won. That's not just statistically irrelevant, it's ridiculous.
So, you're probably wondering, if the polls are easily manipulated and non-scientific, why run them?
Primarily because they engage readers in our product and, at some level, force people to think about issues we think are important.
Online polls are, at best, informative. They are not — and should never be considered — statistically valid on the order of a scientific poll. Like many things you find online, they should be viewed with skepticism.
In other words they didn't like what they heard so it has to be rigged!
More than 100,000 people voted in one of The Denver Post's online opinion polls last week, which is a significant number.
But the poll's result — with 70 percent of respondents favoring a ballot measure that would overturn a ban on high-capacity magazines — is largely insignificant.
Certainly the result shows passion for one side of the issue, but it also shows something else — ballot stuffing.
On a typical day we'll have a few thousand readers weigh in on opinion poll — maybe 25,000 in a good week. But in the last week, we've counted more than 200,000 "votes" in a series of polls.
Noticing the trend, a reader from Centennial wrote to ask if the polls "accurately reflect the opinions of the people of Colorado?"
He was pointing specifically to a poll asking whether Democrats in the statehouse were overreaching.
When we posted results of the overreach poll in the March 19 edition, 91 percent of 33,681 respondents said "yes." What's notable there is both the percentage who answered in the affirmative, but also the number who participated. The previous high was about 31,000 in a July 2012 poll in the wake of the Aurora theater shootings.
On March 16 we asked for opinions on the bid to overturn a bill limiting gun magazines to 15 rounds.
Watching those results online, one reader remarked:"It's pretty obvious to me that someone is using their computer skills to manipulate the poll."
He's right.
This is not the "skewed polls" argument that challenged the weighting and methodology of many scientific polls during last fall's presidential election.
This phenomenon is real.
Changing results of an online poll is simple and a common practice. If someone is passionate about an issue, they can — and often do — weigh in multiple times on a single question by voting from different devices. They can do it whether they live in Grand Junction or the Grand Caymans.
An organized effort to influence results of an online poll is called "freeping." The name is a nod to an online forum called the Free Republic, whose members are called "freepers." They're notorious for pushing results of online polls in one direction or another — notably on Second Amendment issues.
In fact, they did it to our gun-magazine poll on March 15 (and to other polls on our site). After 24 hours online, the magazine question had received 26,704 votes, with 69 percent favoring a vote to overturn the law. Anti-overturn forces then mounted a campaign that pushed that figure down to nearly 51 percent.
"Basically someone has hacked it and are adding an average of 26 votes per minute to the 'keep the ban' column," one freeper lamented.
"After Dynachrome pings the list it'll look better," another remarked. (The user known as Dynachrome maintains a "ping list" of users summoned to help sway the vote in online polls — and shortly after he or she was engaged, support for overturning the ban climbed to 61 percent.)
The poll also drew the attention of prominent conservative blogger Michelle Malkin and magazine maker Magpul, which posted a link to the poll in its Twitter feed.
As of this writing, 112,000 votes (and climbing) have been received.
That doesn't mean gun-rights supporters are alone in rallying people to vote for their cause in an online poll.
After the presidential debate at the University of Denver, almost 11,000 people voted in our poll, with 40 percent saying they thought Barack Obama won. That's not just statistically irrelevant, it's ridiculous.
So, you're probably wondering, if the polls are easily manipulated and non-scientific, why run them?
Primarily because they engage readers in our product and, at some level, force people to think about issues we think are important.
Online polls are, at best, informative. They are not — and should never be considered — statistically valid on the order of a scientific poll. Like many things you find online, they should be viewed with skepticism.