Tuesday, January 04, 2011

Party brands, the bimonthly view

The "top of mind" word clouds from Abacus Data seem to have generated a lot of attention yesterday. The word cloud derived for the Liberal party, in particular, garnered a lot of attention and was quite prominently displayed on the Globe's Politics page. The word "liars" was featured at the very top of the word cloud picture presentation, the most noticeable word, with "corruption" standing out in it as well.

Accompanying the Liberal word cloud was a rather dramatic statement about the Liberal brand from David Coletto, the Chief Insight Officer of Abacus. That too stood out:
Based on first mentions, it is clear that the Liberal Party brand is severely damaged and the negativity associated with the party extends beyond Michael Ignatieff’s leadership. These results suggest that the Sponsorship Scandal continues to hurt the Liberal brand and the residual effect is a lasting association with corruption, arrogance, and dishonesty.
Most troubling for the Liberal Party is that very few respondents mentioned positive phrases or words about the Liberal Party. Instead, comments described the party as having weak leadership, being dishonest and corrupt, and divided.
"Severely damaged," "sponsorship scandal," (really?) "a lasting association with corruption, arrogance, and dishonesty." What's more, very few positive phrases, words. What to make of all that.

In particular, a question that came to mind was how to reconcile that kind of analysis with the poll results released by Abacus on December 15th, 2010, which also dealt with party brands. It seems to have been done with the same group of Canadians (both 1,361 respondents, taken December 3-6th). The December 15th poll employed a different methodology, however, by presenting respondents with statements to associate with each party, a bit more deliberative rather than visceral. That one was not as harsh on the Liberal party:
The Liberal Party brand, like the Conservative one, is highly polarizing. However, Canadians who would not currently vote for the party did select some of the more positive brand attributes, which rarely happened among non-Conservative supporters.
The different methodologies appear to have produced different results. That Abacus poll was not exactly a glowing result for Liberals either, but if you read it, it's a far cry from "severely damaged," etc. that Coletto states in yesterday's word cloud poll.

A difference is also seen in the Conservative results with the word cloud poll producing largely favourable analysis:
Although many Canadians regard the party as arrogant, the Conservative Party is seen as a good economic manager that keeps taxes low, is good for business and is honest. It seems that the notion of a Conservative “hidden agenda” is not top of mind with Canadians. Instead, the Conservatives have developed a “what you see is what you get” relationship with Canadians which may help to explain the polarizing nature of the Conservative Party brand."
But again, the December 15th poll produced different results:
The Conservative Party is the most polarizing political party brand in Canada. The difference between Conservative supporters and non-supporters is large. This exercise clearly demonstrates that the party has a lot of work to do to appeal to a broader base of support as a large number of Canadians feel the party is out of touch and will promise anything to get elected.
Again, the two polls with their different takes on party brands seem to have been done with the same group of Canadians. Both polls rely on 1,361 randomly selected Canadian adults with the identical count and weighting breakdowns by region (page 2 of both surveys). It would have helped to know that in the Globe report, at least by making a reference to the earlier (released) results which were less dramatic. (Even if they weren't the same group of Canadians, the differing results would still warrant comparison.)

A further point here on the "top of mind" word cloud model. It (like the other Abacus poll) was taken from December 3rd to 6th, 2010. If you look at the weighted respondent count for Quebec, it represents the second highest weighting for all regional respondents (Ontario is first). Even without the weighting, Quebec respondents are the second highest count. This might have impacted some of the "top of mind" first responses to the question of what comes to mind when you think of the Liberal party in particular.

If you look back to December 3rd to 6th in Quebec, dominating the news was the December 5th Charest appearance on Tout le monde en parle. It was widely viewed (1.5 million/week), with much media coverage and drawing attention to such words as "corruption." Additionally, during the week of this Abacus word cloud poll, Charest had announced a new Quebec "anti-corruption unit." Then "corruption" happens to feature prominently enough on the Liberal word cloud and figures fairly prominently in Coletto's analysis. Coincidence? It may have had some influence. This is the point that Nelson Wiseman was getting at in the Globe piece yesterday about news items influencing such "top of mind" surveys. It raises doubts about the Coletto point on the "lasting association with corruption" being attributed to the sponsorship scandal in particular or to the federal party to the degree Coletto does. Of course, it's not spectacular news, either, that the Quebec Liberals could possibly be tarnishing the federal brand but the word cloud poll is presented as meaningful for the federal brand, when that may in fact not be the whole story.