Do election ads really work? Testing 2015’s pre-writ ads

INNOVATIVE will be testing political ads throughout the 2015 election campaign. In our first release we have tested the TV ads that aired in the months leading up to the writ drop.


Political ads serve two key purposes; they can motivate the sponsor party’s supporters to vote and they can persuade target voters to vote for the sponsor party. Ads accomplish these goals through two mechanisms; they can “prime” or raise awareness of something people already know or feel that gives the sponsor party an advantage over its competitors or the ads can “persuade” by providing new information or framing existing information in a new light to change how people feel about the sponsor party and/or its opponents.


Do political ads really work? Answering that question involves addressing two key challenges:

  • You need to examine reactions according to voters’ initial views. The views voters hold before they see an ad influence how voters see that ad. Generally the supporters of a party are predisposed to like that party’s ads and not to like the ads of parties they oppose. Unaligned voters may be open to several parties. To assess the effectiveness of ads, you need to control for voters’ initial attitudes.
  • Whether people say they like an ad or not really doesn’t matter. What does matter is if the ad moves the numbers. However, if people have already seen the ads, the effect of the ad will already be factored in to initial attitudes. So we need to find people who have not already seen the ad and see how they respond.

We will use the following approach to test political ads throughout the campaign:

  • First, we will identify the respondents initial views including Party Identification, their current vote and how respondents feel about the leaders and time for a change. These are our pre questions.
  • We will show them an ad.
  • We will ask if they have seen the ad before and what they feel about that ad. We call these diagnostic questions.
  • We will re-ask their vote preference and how respondents feel about the leaders and time for a change. These are our post questions.
  • We will show several others ads and ask if they have seen the ad before and what they feel about that ad to increase the sample size for the diagnostic questions.

The key test is whether people who are seeing the ad for the first time change their responses on the pre/post test. Does the ad “move the numbers”. We then use the diagnostic questions to understand why the numbers are moving.


It is important to understand that an ad can make a positive difference to a campaign even if people say they don’t like it. For instance, people may say they don’t like an ad that provides negative information about another party, but if they feel the information is credible and informative, it can still move the numbers.

Click here to download the Full Results and Methodology
Click here to see the results from our second wave of testing


NDP – Bring change to Ottawa

Liberal – True Story

Liberal – Enjoying the Playoffs

Conservative – The Interview

Conservative – Proven Leadership

Engage Canada – Neglect

Working Canadians- Middle class