Sunday, May 10, 2015

Federal Political Public Opinion on Vancouver Island May 5th - 8th

The survey was was conducted between May 5th and 8th using Google Surveys to sample public opinion.  Here are the results:

I need a better tool for making graphs, please suggest one for me to use

Party  Vancouver South of North of
       Island    Malahat  Malahat
NDP     33.3%     33.8%    32.6%
Greens  22.5%     24.8%    18.5%
Cons    20.5%     14.7%    28.0%
Liberal 20.5%     23.5%    15.4%
Other    4.0%      3.2%     5.4% 

815 weighted responses - weighting based on age and gender.
247 weighted responses were undecided
568 weighted decided voters for all of Vancouver Island
318 weighted decided voters south of the Malahat
237 weighted decided voters north of the Malahat
Google Surveys were used as a the sampling method
Because this is not a probability sample of the population of Vancouver Island a statistical margin of error is not applicable.  Getting a true probability sample is virtually impossible to achieve, getting even close no longer possible.  You can read a basic explanation of sampling methods at this link

----------------------------------------------------------------------------------
I have been getting back into the world of polling.   Back in the 1980s I ran polls for BCPIRG and then in the later 1990s to early 2000s from time to time for various First Nations.   

I have been unhappy with the quality of random data the various methods of sampling methods most pollsters are using.   I have been experimenting with various models since the 2013 BC election but found limitations to the various different ideas that made them less than ideal solutions.   About a month ago I came across Google Surveys for the first time.  I have run a number of tests of the Google Survey model since then and I have been pleasantly surprised at the statistical quality of the results.  I am still working through the math of the statistics and comparing the results from the surveys with the 2011 census demographics.   

I am comfortable enough with the results from Google Surveys from my tests that I have confidence that the results are a good statistical representation of the public

My intention is to run regular polls of federal political opinion in BC between now and October 19th.   The frequency will depend on how successful the crowdfunding campaign will be to pay for the work.

No comments: