Monday, 10 November 2014

Probit/Ekos Oct 28 2014 Poll: Leading Questions Guarantee Invalid Results

Text of Oct 31, 2014 e-mail to Probit/Ekos


On, October 28, 2014, I participated in a Probit/Ekos online poll concerning the federal political landscape.

Much to my disappointment, the survey was loaded with leading questions and partisan favouritism/nonsense. 

In this regard, I have copied and pasted on to the attached Word document, a particular set of the survey questions of the online Probit/Ekos survey in question.

The section to which I am referring is divided into 3 sub-sections, one for each party leader and his party.  These sub-sections open with the line "The following are some things people have said about _____(party leader). Please rate each from 1 to 9, where 1 means it makes you a lot less likely to vote ___ (party) , 9 means it makes you a lot more likely to vote ____ (party), and 5 means it has no impact on your vote."

In these sub-sections, the totals for positive and negative statements regarding "some things people have said," broken down by party, are as follows:

NDP: 3 positive and 9 negative statements 
CPC: 4 positive and 8 negative statements
LPC: 9 positive and 5 negative statements

The aforementioned partisan slant is extraordinary and something that no survey firm with the slightest amount of integrity would ever do.  As such, this survey will produce statistically invalid results.

Equally important, the problem with "some things people have said about" Mulcair, Trudeau and Harper, is that most of the statements are not only are leading questions but also are gratuitous unsubstantiated remarks and often outright lies, that risk influencing the less politically astute person's answer to the questions.  

By contrast, a poll with integrity, solicits one's political preferences, without leading questions, but may solicit responses from survey participants based on party policies and positions, leaving it to survey participants to decide whether the policies/positions are credible.  In the event party policies/positions are provided, a "normal" survey subsequently asks questions based on the solid information provided.  

As per the this survey, ONE SHOULD NOT GET LEADING QUESTIONS LIKE "Trudeau represents the next generation of Canadians" -- There is absolutely no evidence of generational policy positions coming from Trudeau.  "The NDP track record in provincial governments shows that they can't be trusted in power." -- A nice planting of an unsubstantiated thought ---- Imagine if the statement was instead, "NDP provincial governments have the best track record on balancing budgets."(can be substantiated) -- This would be an equivalent planting of a thought.

The Probit/Ekos double standards are staggering!

Clearly this poll has been designed to skew public opinion.

Regarding the option of "No impact" it doesn't apply for me as valid answers for the leading questions because I was outraged at Probit/Ekos for these unprofessional and inflammatory questions.

In another section of the survey, not attached, there was a statement to the effect that one knows very little about Mulcair.  Apparently, the Probit/Ekos team never stepped outside of Toronto to know enough that, for Quebec residents such as myself, this makes no sense given that Mulcair is the most popular politician in Quebec, all levels of government combined.  

While Toronto isolationists may think a one size fits all survey makes sense, I can only recommend that these Probit/Ekos people crawl out of their Toronto bunker and live in French in Quebec for a minimum of one year as an essential part of their training. 

For all the above-mentioned reasons, this survey demonstrates incompetence on the part of Probit/Ekos.

Accordingly, I have decided send this e-mail and the attachment to various media over the course of the next week. (in addition to those cc'd above).

Will Dubitsky

(Tech problem: Word attachment of questionnaire/survey can't be copied onto blog)

No comments:

Post a Comment