Vincent Canfield's Blog

How to Rig an Election, Get Published by JP Morgan, and Get Threatened With Waterboarding in Just 12 Hours


Category: Farage In My Garage

How hard could it be to get people to believe a fake political poll?

I was looking at the Wikipedia page "Opinion polling for the United Kingdom European Union membership referendum", which is pretty much where all the cool kids get their Brexit polling data:


I noticed all of the sources of these polls are the original polls themselves. What do the sources look like? Some look like this:


Wow, fucking boring. Who wants to fake tens of cells of tabular data? Not me! So I found a "poll" whose source was literally a Mailchimp campaign. I couldn't pull up a screenshot for the original campaign as they've deleted all of the pages for their previous polls. But in any case I signed up for a Mailchimp account, and made my own poll:


I wanted to see how long this could last, so I didn't include any birds with arms or pictures of weev, but I did set the sample size to "1420". So I got my poll published on the Wikipedia page and it was up until I went to bed:


The poll lasted about 10 hours before being removed. By the time I woke up, I faced multiple news articles, accusations of trying to rig 30 million fake british dollars in betting markets, and threats of waterboarding.


|2016-06-15-16-14-12| (link)

|2016-06-15-16-14-56| (link)

|2016-06-15-16-18-36| (link)

|2016-06-15-16-16-55| (link)

|2016-06-15-17-40-47| (link)

|2016-06-15-17-40-51| (link)

Simple enough. Here are some good quotes from these articles:

The poll appeared to have first been posted by the BritainElects account on Twitter, which was then included in a report by the *Guardian* and included in JP Morgan's analysis of the referendum on Wednesday.

.. raw:: html
Researchers at the U.S. bank based their calculations on polls from ComRes, TNS and BMG. .. raw:: html