Ok, I find it not just implausible but imposible that 3396 housholds responded to this poll. First, Guelph has approx 100,000 people, aka only approx 25,000 household phones. Lets assume 9 out of ten calls would be hang ups or not be home in the middle of summer when Guelph becomes a dead zone. That means that they would have to have phoned approximately 34,000 households in Guelph. I'm pretty sure that comes close to phoning every household in Guelph, something no pollster would do. I believe they phoned 3400 phone numbers. As someone who has phoned that many numbers in Guelph of Liberal members, all of whom are interested in a nomination meeting and therefore much more willing to give a response on who they support, I can 100% guarentee not more than 10% of households phoned answered that survey. So that's 340 surveys and not all of those people will be eligle to vote or likely to vote. So, if you apply any standard likely voter model for a 30% turnout for a summer by election, you get what 100 likely voters answering this poll, grosely misrepresented by saying they contacted 3400 people aka phones that weren't answered or hung up. It's still a poll but it has a serious margin of error, which doesn't even account for voter turnout, which will be the biggest issue for a summer by-election. But I will absolutely guarentee that there is no way that 3400 households answered that survey, as they would have had to phone every number in Guelph to get that response. I think the guy is probably trying to promote his company. Why wouldn't he? That's why he's in business right. And i don't see any real nefarious purpose or grand scheme in this, but I think the numbers are deliberately misreported. There's no way that 3400 people in Guelph have responded to any telephone political suvey ever, for any pollster. It's not possible or necesary.