For the last few weeks, things have looked pretty grim for the Conservative Party; of course, that’s assuming you believe what the mainstream media tells you.  The CBC, CTV, the National Post, and of course Pravda The Toronto Star all  seem to be actively campaigning for the NDP/Liberal coalition.

However, Ekos released a poll yesterday (well, it’s after midnight as I write this, so really, it was two days ago – September 4) which has got me wondering: are the media outlets telling us what they want us to hear, or are they really looking at the numbers and reporting the news?

News is, simply put what happened.  Unlike a lot of what we are exposed to nowadays, it’s what’s known as fact.   The trouble is, facts are boring.  They aren’t salacious.  They aren’t sexy.  They don’t attract eyeballs, and eyeballs are what sell advertising.  The more people watching, the more in-demand the advertising time on the TV station, radio station (ears, in that case), or page-space in a newspaper becomes.

It’s all about the advertising.

I’ve said before that a TV show, whether it’s news, or CSI, or Star Trek, is simply there to fill up the dead air between the commercials – it’s because TV/Radio/Newspapers are, for the most part, businesses.  While we have to buy a newspaper, the cost of the paper is still heavily subsidized by the advertising revenue.  Radio is free for the taking, and so is over-the-air broadcast TV, and, while we pay for Cable TV (or IPTV as it is now), the cost for our TV signal is still subsidized by the commercials.

So the media will tell us what they think will keep us watching and engaged.

That’s not a bad thing, mind you, but it’s good to be aware of it.

So, keeping that in mind, I have run the Ekos numbers through my prediction machine, and I’ve come up with some very interesting findings.

I ran the numbers twice.  The first time, I ran them with just the numbers supplied.  The second time, I adjusted the numbers to account for Ekos’ weird bias against the Conservatives.  They use IVR, which, as we know, can be, I believe, heavily biased against the incumbent.

In 2011, Ekos missed the results by 6%. They reported the Conservatives to have 6% less popular support than they actually received on Election Day.  So, this time around, in my second run of the numbers, I adjusted the Conservative support up by 4% (I felt generous towards Ekos), and the other parties down by 1%.

Here’s what I came up with:

First, here’s the prediction based on Ekos’ reported numbers:

[visualizer id=”1804″]

[visualizer id=”1805″]

[visualizer id=”1806″]

These results include what Ekos calls “Leaners,” meaning those are people who responded initially by saying they’re undecided, but, when asked, said they were “leaning towards” one party or another.

And here’s the map:

[scb_map pollster=”Ekos” polldate=”2015-09-04″ datafile=”PollResults.csv” basemap=”Canada” htmlpath=”Canada.Fed.Constituencies.English.Only” mainmap=”projected winners” underlyingmap=”incumbents” underlyingfolder=”Incumbents” altmap=”popular support” altfolder=”Popular.Support”]

This second set of graphs shows what happens when I adjust the Conservative support up by 4% and the other parties each down by 1%…

[visualizer id=”1807″]

[visualizer id=”1808″]

[visualizer id=”1809″]

I can only put one map on each post.  Usually, it’s not an issue, but it is for this one.  So if you want to see the map, click here.

This means the split between the NDP and the Liberals could most definitely work to Stephen Harper’s advantage.  Ekos has gone quite comprehensive in their background information, and that includes a set of second-choices for the voters.  While there is a lot of fluidity between the other parties in terms of voters’ second choices, the primary second choice of Conservative supporters is “none.”  In other words, Conservative support is rock-solid.  Conservative party supporters are the most loyal to their brand of choice, and won’t budge.

If you want to read the actual detail by Ekos, it is readily available over on their website:

7 Replies

  1. Nice and light writing, I like it. Nothing too heavy. 🙂 I too think that the CPC has a very good chance at majority simply because the NDP and LPC support is very concentrated in particular areas only. The polls do not reflect that reality.

    1. Interesting change in your tone re. “a very good chance at a majority”… you’re sounding a lot less confident than last week.

      Yes, NDP and LPC is concentrated in particular areas only such as the three of the four largest provinces, all the major cities and the four atlantic provinces.

      1. I don’t know if that’s a fair comment on my tone, Peter. Last week I was talking about what I did with regards to the plugin I wrote. Then I spent some time upgrading it to add yet another map, and make it a bit easier to navigate. I guess I was being a bit more technical than I was this time around.

        Either way, it’s me, and here online, what you see is what you get.

  2. In 2011, Ekos missed the results by 6%. They reported the Conservatives to have 6% less popular support than they actually received on Election Day.

    So naturally they will be wrong again in 2015? Instead of focusing on one polling firm, look at something comprehensive like threehundredandeight / cbc poll tracker which is showing the CPC in third.

    btw, was Nanos closest in 2011? He isn’t showing the CPC anywhere near a majority. Harper won’t get a majority (or for that matter a minority) this time – his polling numbers have no upside, those “blue” liberals aren’t moving to him this time.

    1. Yes. Ekos was off by 6% last time, and 3% the time before.

      And before you dismiss my comments out of hand, letting your obvious anti-Harper bias cloud your judgement, remember that I called it pretty damn close last time, too. Using Ekos’ poll to do it, and without relying on multiple polls like you want me to do.

      A poll is a snapshot in time. Like an election, it is its own animal. If you aggregate polls, you also aggregate margins of error, data collection methodologies, inherent biases contained within the poll, and so on. You end up with a mishmash of … well… garbage, and as they say with computer programs (which is what my Prediction Machine is), if you put Garbage In, you get Garbage Out.

  3. Interested in your update of your projections, given the current Liberal lead.

Leave a Reply

Your email address will not be published. Required fields are marked *