Two weeks ago, when Bernie Sanders entered the presidential race and made a big splash in the Daily Kos Straw Poll, scoring 44%, Markos reminded readers that “[Bernie Sanders] used to regularly score 53-73 percent in the 2016 cycle.“
Well, in yesterday’s poll, that’s exactly what happened. In the wake of two giant rallies in Brooklyn and Chicago, Bernie Sanders scored 55% at the 24-hour mark, which is when the poll officially closes.
But today, Markos posted an update to the poll diary, saying that Sanders supporters skewed yesterday’s results by voting multiple times, and so the poll doesn’t count:
Due to user error—mine—I failed to turn on a function that makes it harder to cheat when voting. As a result, people were able to turn their browsers to incognito mode and repeatedly rapid-fire vote—something done extensively by Bernie Sanders supporters. As a result, these results are not an accurate reflection of candidate online support, and will be discarded.
It’s too bad, really, since Sanders would’ve won anyway. There was no need to cheat. We’ll be back in two weeks, when I will presumably avoid this mistake, and we’ll have a more representative vote as a result.
Presumably this anti-voter-fraud feature was turned on in the past, but not yesterday. In the second Daily Kos Straw Poll of the 2016 cycle, Bernie Sanders scored 63%, and Markos wrote this in his analysis:
Sanders obviously has the hearts and minds of the bulk of the community, with two-thirds support. Clinton has another third. And just about no one wants anyone else, or is undecided. In fact, that lack of undecideds stands out the most to me.
So we had some big movement those first two weeks, with Sanders dropping six points and Clinton gaining seven. I wouldn't read anything into that. This is not a scientific poll after all, and it's susceptible to supporters spamming the numbers. but over time we'll have a feel for where things generally stand, as well as any trends.
By “spamming”, I assume Markos simply meant an organized effort to vote in his poll — he wasn’t suggesting that Hillary’s rise in that poll was due to cheating.
(“Freeping” is the word coined in the progressive blogosphere to mock the conservative website Free Republic for such organized voting efforts. Freeping is simply a call to action to go vote in an online poll. So if a Newsweek online poll showed Romney had “beaten” Obama, folks would say “Clearly that poll’s been freeped!” Since yesterday, though, I’ve noticed a number of folks here using the word to mean cheating by voting multiple times, but that’s never been the common usage of the word; here’s the Urban Dictionary entry.)
I wrote this in a comment this morning:
I think the main problem with all polls — online as well as “scientific” random sample polls — is that folks take them way too seriously. That’s just human nature. So folks get elated when their favorite candidate “wins” the poll (wow, I love this poll!) and dejected when their favorite candidate “loses” the poll (this poll is meaningless! It’s skewed!)
...
As doctors say, when you see hoof prints, think “horses” not “zebras” — i.e. think about the most likely cause of what you’re observing. The most likely reason a candidate does super-well in this type of poll is simply that lots of folks on the Internet are voting for that candidate — not that there’s a tremendous amount of voter fraud.
A month ago, prior to Bernie entering the race, he was lagging in the straw poll. Markos commented:
Consider this [straw poll] a leading indicator. Daily Kos is the activist class, and in a world where small-dollar donors will determine so much, this has real value.
In the “real world”, people are like “I know Biden!” and say him. It’ll be different once regular Democratic primary voters have the kind of awareness and knowledge that we do today.
So, for example, we had Obama surging early in 2008 while everyone else assumed Hillary would win the nomination.
I assume the view that this poll is a leading indicator is still operative. It’ll be interesting to see how the next poll results look with the anti-voter-fraud feature turned back on.