According to Harris/Decima Chairman Allan Gregg “There seems to be little question that the Liberals have improved their electoral fortunes over the summer -- they have closed the gap with the Conservative lead to 3 points; moved into a virtual tie in Ontario, advanced in Quebec and significantly narrowed the Conservative lead among men. Unfortunately for them, it would appear that virtually none of this improvement is a result of voters forming a positive impression of their leader, Michael Ignatieff. In fact, while Conservative leader Stephen Harper has improved his net favourability rating (i.e. the percent of voters holding a favourable minus unfavourable impressions) slightly -- from -10 to -5 – Mr. Ignatieff still trails him at -25 and has basically not moved since May.
A more in-depth analysis presents even more disturbing news for the Liberal leader and demonstrates that he has been unable to cultivate a loyal constituency among any of the demographic and socio-economic groups we look at. In fact, his net favourable rating is worse than Harper's for every group analyzed, save Quebecers and voters with post graduate degrees and even among these two constituencies he is in negative territory. A narrow majority of Liberal voters do hold a favourable impression of Mr. Ignatieff but once again, compared to the number of Conservatives (88%) who hold a favourable impression of his opponent, it is apparent that Mr. Ignatieff does not have the ability to galvanize the Liberal support base the way Mr. Harper does his."
The vote data were gathered through teleVox, Harris/Decima’s national telephone omnibus survey over two weeks between August 12 and August 22. 2010 for 2,017 completes. A sample of the same size has a margin of error of 2.2%, 19 times out of 20. Regional margins of error are outlined below.
Atlantic – +/- 6.8%
Quebec – +/- 4.4%
Ontario – +/- 3.8%
Man/Sask – +/- 6.8%
Alberta – +/- 6.9%
BC – +/- 6.2%
The leadership data was gathered through teleVox, from August 19 to August 22, 2010 for 1,007 completes. A sample of the same size has a margin of error of 3.1%, 19 times out of 20.