My first university job taught me about media transparency and the intricacies of political polling
There are two things you need to know before reading this story.
First: I’m American. I came to Canada in September of 2016 to study at the University of Ottawa for reasons completely unrelated to the 2016 presidential election.
Second: call centre jobs are famously awful.
I began my first year of university with approximately $1,200 in my savings account. I was proud of that number: I’d saved my tips from my waitressing job back home in Baltimore. Naively, I thought that would be enough spending money to last a full academic year.
I was broke by October.
I needed a job, and as an international student, I needed a job that would be flexible enough to accommodate the work restrictions on my visa.
The fateful posting appeared in one of the U of O Facebook groups for first-years:
“Now hiring interviewers for a local research firm. Must speak English fluently and have excellent reading skills. Pay starts at minimum wage, and goes up to $12 per hour after a probationary period.”
Hell yeah. I spoke English, and I was a theatre student who could read a script with no problem. I didn’t really understand what the job was for, per se, but I had the necessary qualifications (which I later learned were little more than having a pulse).
After a breezy application and interview process, I got the job. Better yet, the office was within walking distance of my U of O dorm room. In fact, I could see the drab brown tower from my bed.
Orientation made things a little clearer. Our small-but-mighty group of new hires would be conducting surveys over the phone. We were promised over and over that we wouldn’t be selling anything (in retrospect, this was a serious bonus, as we were screamed at enough over the phones as it was).
There were a few key things we learned in orientation, including how to turn down the volume on our headsets if (when) a respondent yelled at us.
Importantly, we didn’t learn what exactly these surveys were about right away. I suspected something political, as we were warned that some questions might be a bit “touchy,” but I figured they would be about Canadian issues – which, to a homesick American convinced she would be moving back home within the year, didn’t carry much emotional weight yet.
They gave us employee numbers and a lengthy employee agreement to sign (including a gag order and a non-compete clause that expired after one year – I left in early 2017). Apparently, we’d get our scripts during our first shift on the call centre floor.
Imagine my surprise when I logged into my computer for the first time and read that we were calling on behalf of an enormous American news network. This was the network known to have given preferential treatment to a certain blonde candidate with a killer pant suit collection, and was a contentious key player in the looming November polls.
We were administering those polls from Canada. To this day, I don’t know why.
The computer randomly generated American phone numbers that we then had to dial on a separate phone, which was connected to a (germ-laden) headset. I recognized some of the area codes: there were a few from my safely-democratic Maryland, but most of them were in key battleground states.
These surveys were serious business.
These surveys, as we all found out after Trump’s victory, were also terribly, terribly inaccurate.
Half the battle of this job was getting a real person to answer the phone: that alone was a success in itself. Then, even harder, getting that real person to listen to your spiel:
“Hi, my name’s Aly, and I’m calling on behalf of (Network). Are you in a location where you can speak freely and without distractions?”
That was the real opening line. The only clue to what the call was actually for was in the name drop of (Network), but even that was subject to tampering, as some versions of the survey omitted (Network)’s name and instead referenced the organization that was subcontracted to write and implement the polls.
Most Americans (including myself) hadn’t heard of (Network)’s little pollster sibling, and most others had serious reservations about (Network) administering these polls at all.
Here’s the problem: say you got a real person to answer. They listened to your speech and said that, yes, they were somewhere they could talk and they had a few minutes to do so. Cool. You, entry-level Aly, were great at your job and were going to make so much money doing it.
Because, entry-level Aly, you had to deal with the fact that most Trump supporters didn’t want to talk to (Network) at all. They didn’t trust (Network) to record their answers accurately, and they thought these polls were fake news. Sometimes they took it out on you personally, which hurt.
The kicker: they’d tell you they were voting for Trump, but then immediately hang up because they hated (Network) so much.
Your polling software wouldn’t let you register that as a response, because the respondent didn’t phrase it using the right language. You couldn’t say that the person associated with that phone number was voting for Trump, so instead, you had to log them as a “general callback” (which meant some other poor soul at the call centre was going to have to try them again later, and likely get screamed at for it).
Sometimes, though, you got someone patient on the phone, someone who either didn’t care about (Network)’s reputation, or even kind of liked it. They were voting blue and they were happy to tell you about it. They answered your questions the way you needed them to be answered, and even stayed on the phone long enough to answer the statistical questions about their demographics at the end of the call. Those rare wins were logged as “complete” (meaning no one would bother that respondent again), and you, entry-level Aly, were flagged as “productive” at your new job.
Those surveys were intrinsically flawed, and the United States of America paid a steep price for (Network)’s eventual miscalculations; overconfidence in Hillary Clinton was one of many factors that led to low Democrat turnout. That, coupled with under-reporting and misrepresentation of Trump support, meant the 2016 polls really were fake news: the screamers on the other end of the phone had been right.
I dreaded my shifts as Election Day got closer. I told my friends back home what I was experiencing, and shared my real fears that Trump was going to win. My then-boyfriend, still in Baltimore, said I was overreacting.
I worked on the evening of Nov.8, 2016, and left mid-shift to watch the news (my supervisors at that point knew I was American and didn’t really care if I skipped a few hours of work). I signed out shakily, ran down Laurier back to the 90 University residence, and watched Trump win state after state in abject horror. The commentators spoke gravely about miscalculated polls. For better or worse, I’d helped to facilitate those polls. The questions themselves hadn’t been the problem, but someone far, far above me in the (Network) food chain had misjudged just how hated (Network) was across the country. I felt guilty – as if I’d been somehow complicit.
I stayed at that job until April of 2017. The surveys got much more boring after the election; the call centre reset to what I was told was “normal”, and we started interviewing respondents almost exclusively about consumer electronics (to this day, I harbour a secret resentment for single-cup coffee makers because their surveys were so wordy and repetitive). For the last few months I was there, there were no more polls on post-Trump politics.
This story has haunted me for four years.
I don’t know where we go from here. I don’t know how we improve mental health in some of Ottawa’s busiest (and saddest) working environments. I don’t know how we reform election polling to accurately reflect the voice of the people, rather than the voice of the network (which is an issue not limited solely to (Network)). I don’t know how to make people care about the students, low-income workers, and recent migrants facilitating the call centre industry.
I’ve not yet been called by (network) for a phone survey leading up to the Nov. 3 election this year.
If I am, I don’t know if I’ll have the words to express my sympathy to the person on the other end of the phone – maybe a fellow American U of O student, carrying out the polling work of my home country just a few blocks down the street in Ottawa.