How to Use Polls to Judge a Political Ad’s Value

American businessman John Wanamaker once said, “I know half the money I spend on advertising is wasted but I can never find out which half.”

And it’s not just businesses which have this problem; political parties also have a hard time figuring out if their money is well spent when it comes to ad campaigns.

It’s especially difficult to ascertain if “negative” or “attack ads” are working.

That’s because negative ads have a definite stigma attached to them, meaning few people will ever admit to a survey researcher that they like or are influenced by one.

I’m bringing all this up because Abacus Data just released a report which allegedly provides an analysis on the effectiveness of recent NDP, Liberal and Conservative ads.

The method Abacus used for this analysis was simple: respondents were asked if they agreed with the points the ads made, if the ads built support for the sponsor and if the ads were fair or unfair.

In conclusion, Abacus reported that while all the positive ads resonated well with Canadians, the Conservative Party’s negative spot, the one taking aim at Liberal leader Justin Trudeau’s lack of experience, didn’t fare so well.

In fact, it was labeled “unfair” by 59% per cent of respondents who felt it “crossed the line.”

This reminds me of a similar EKOS poll from a few years ago which asked respondents to judge those initial waves of “Trudeau in over his head” Conservative attack ads.

According to that poll, 71 per cent of respondents said viewing the ad had not changed their impression of Trudeau and of those who said the ads had changed their impression, more than half said they actually felt more positive about the Liberal leader.

What’s more, respondents described the ads with words like “disgusting” and “vicious.”

At the time, EKOS President Frank Graves declared, “The ads have backfired on the Conservatives, at least in the short term.”

Yet, today’s most recent polls show that Trudeau has begun to steadily slip in public support, which perhaps means the anti-Trudeau ads are indeed working, at least in the long term.

I say “perhaps” because without proper polling data I can’t say whether the attack ads are working or not.

Nor can EKOS or Abacus, at least with the kind of polls they’re conducting.

The problem is political ads, especially the negative variety, don’t work when they are clinically examined and analyzed —they are not designed to survive under a microscope.

Rather, they are meant to be seen while a voter is half paying attention to the TV when the message can more easily seep into the viewer’s subconscious.

This is why to find out if the Conservative attack ad is truly effective you just can’t come out and ask poll respondents “Did this negative ad change your view of Trudeau?” Or “Do you think this ad is fair?”

I say that because there’s often a big difference between what a person is saying to a pollster and what they’re actually thinking.

To see what I mean, let’s imagine a dialogue between a pollster and a “Mr. Smith” on the effectiveness of TV advertising:

Pollster: Do TV ads have any influence on your buying behavior?

Mr. Smith: No, sir. TV ads do not influence me in anyway.

Pollster: OK, what brand of coffee do you buy?

Mr. Smith: I always buy Maxwell House coffee.

Pollster: Why?

Mr. Smith: Because everyone knows it’s good to the last drop.

Clearly, even though Mr. Smith says he isn’t influenced by TV ads, he is. That’s the difference between what a person says and what he really thinks.

So how would a pollster try and find the truth about a political ad?

Well, like the pollster in the above example, he or she would ask a series of questions to drill into and gauge actual attitudes on the issues raised in the ad.

For instance, in regard to Conservative attacks on Trudeau, the pollster might ask something like “Do you think Trudeau is in over his head” or “Do you think Trudeau lacks experience?” or “Do you think Trudeau is qualified or unqualified to be prime minister?”

Then the answers to those questions would be cross tabulated to match them up with respondents who also said they had seen the Tory TV ads. (So often, the real truth in a poll can be found in those hidden cross tabs.)

Anyway, if the respondents who had seen the Conservative attack also say they’re concerned Trudeau lacks the experience to be leader, it would tell the Conservatives (or the Liberals) that the attack ad might be gaining traction.

True, it takes a little more work to conduct such a poll and the results probably won’t generate eye-catching headlines, but the point of a poll should be to meaningfully and accurately reflect reality.

And when it comes to judging a TV ad, accuracy and reality matter a lot for a political party.

Only with the best available data can it tell if its attack ad spending is the half that isn’t being wasted.


Leave a Reply

Your email address will not be published.