Mocking the vote
Hello, 142.xxx.xx.xxx. And hello again, and again, and again.
The Internet tells me that 142. xxx.xx.xxx is a Bell FibreOp account located somewhere in or around St. John’s — but I’m getting ahead of myself here.
This is a column about polling, and fudging the numbers. For the past few weeks, across TC Media Atlantic websites, we’ve been running a point-and-click Internet poll. Just a bit of fun, really, and, like most Internet polls, completely unscientific and methodologically meaningless.
The kind of poll that should be considered accurate zero times out of however many times you do it. The question? It looked like this, on TC websites, and on our Facebook and Twitter sites. “We at TC Media are wondering if Donald and Hillary were running in Canada, who would get your vote? The Donald of course. Hillary all the way. Can’t do it, just can’t put my mark beside one.” It ran from Oct. 5 until Oct. 18. The numbers were pretty much as you might expect in a country and region like this one, a place that runs well to the left of Texas or Arizona: Clinton was leading Trump by a two-to-one margin.
Until last weekend, and the arrival of 142.xxx.xx.xxx.
On Oct. 14 at 8:41 a.m., a computer with that IP address registered a vote for Clinton. Then, on Oct. 15, between 2:06 p.m. and 8:22 p.m., 260 votes were registered for Trump, all from that very same IP address, several only seconds apart. The next day, 30 more votes for Trump, meaning someone was taking the time to vote, refresh, and then vote again. Oh, well, perhaps it was a nice diversion from raking the leaves or cleaning out the gutters.
Perhaps it’s just a case of a little messing with a poll that doesn’t mean diddly.
But it sure does provide an interesting window into human nature. Someone wanted to hamhandedly goose The Donald’s numbers, for whatever tiny piece of traction it might have. (Funny that they thought we wouldn’t monitor their e-footprints — nothing on the Internet is truly anonymous.) It made me think of a study released earlier this week by the University of Oxford. Researchers looked at the huge volume of Twitter traffic on the Internet during — and for four days after — the Sept. 26 presidential debate. During that time, there were almost 1.8 million Tweets that were positive about Donald Trump — except that 576,000 of those came from automated Twitter “bots.” Hillary Clinton also saw bot traffic, but far less: just 136,000 bot tweets, postings made not by people, but by programs designed to manufacture tweets. Meaning someone, somewhere — deliberately or not — was pumping the numbers, making things not quite what they seem. There’s no suggestion from the study that the bots were the direct work of any campaign. But anyone who’s ever run a straw poll — and gone back through the votes — has found poll stacking. In Newfoundland and Labrador, a sitting PC government actually had members of the House of Assembly contacting supporters to get them to goose polls to support government actions. The St. John’s Telegram actually caught the politicians at it.
None of this can help raise the estimation of politics and politicians among an already-jaded electorate, whether it’s Twitter bots, faked votes, or any other deliberate strategies to try and fool voters.
It’s muddying the waters, to be sure. It’s a real shame when anyone thinks the best way to get a message across is to cheat.
And yes, 142.xxx.xx.xxx, I’m talking to you, you whose candidate of choice is always whining about “vote rigging.” Whoever you are. (And I know we could go further and find you. It’s just not worth the time.) Hope you enjoyed your weekend at the keyboard.