A powerful approach to analyze message testing data and avoid the pitfalls of traditional analytic techniques.

Helping campaigns test and craft messages is one of the principle responsibilities of a pollster.  Unfortunately, many pollsters use message testing techniques that rely entirely on self-reporting by respondents of how they would respond to messages.  This approach causes two important problems that can cause campaigns to adopt misguided messaging strategies:

1)      Social-desirability bias can cause respondents to over-report the effectiveness of messages that they think they should respond to and under-report the effectiveness of messages they believe they should not respond to.

2)     Because this self-reported approach does not include an importance measure it can lead to overestimating the effectiveness of messages that respondents agree with strongly but which do not really play a role in determining their vote.

Because of these concerns with the traditional approach to message testing, WRS has developed its Message Mapping ™ analytical tool. 

 

Message Mapping â„¢ uses the same message testing sequence that traditional polling uses with two key differences:

1)     We construct multiple sub-samples each of which hear different combinations of messages.  This allows us to mirror the way that researchers test medicines, by observing the difference in response (ballot movement) between those who heard a particular message and those who do not.

2)     By asking a structured recall question after the messaging we can evaluate which messages are the most memorable.

 

Analysis of the data set from the survey allows us to test each message on three attributes:

 

Effectiveness

• Effectiveness is measured by the simplest of metrics and answers the most important question: does this message alter voter behavior?
• Effectiveness of messages is tested exactly the way that other researchers test the effectiveness of medicines: by measuring the difference in response (in this case ballot movement) between those who heard a message and those who did not.

Stickiness

• Stickiness is a cutting edge variable designed to test how successful a message is at punching through a crowded media environment, grabbing a voters attention, and remaining with that voter.
•  In today’s complex media environment, voters, especially approaching an election cycle, are bombarded with information.  Making sure that a campaign is spending money on messages that voters will notice during a crowded commercial break or when sorting through many pieces of mail is the difference between a message that works in the controlled environment of a survey and one that really helps win a campaign.

Believability

• Believability is a critical, but sadly under-used, element of any message.  It doesn’t matter how much voters will agree with your message, if they don’t believe it, it won’t move them in the ballot box.

By analyzing all three of these dimensions together we can understand which messages will really help change behavior and help you win a campaign.  Below is an example of a Message Map™.

 

In this message map we see two messages which score very well on effectiveness—health care and an energy message. 

 

These are clear “winners� among the messages.  Immediate tax relief scores almost as well as those two on the effectiveness scale but is not as memorable and also may not be as believable (voters don’t think it will happen). 

 

Medical care for veterans is a classic “trap� message that would show up well in a traditional message analysis—it is memorable and voters will say they like it, but they just don’t care enough about it to respond to it on the ballot.

 

At WRS we remain committed every day to developing and improving tools that give our candidates and clients every advantage to win their campaigns by leveraging the latest developments in behavioral science and statistical analysis.  Message Mapping ™ is one of many ways we do so.

Many politicians and pundits bemoan the use of attack ads as ineffective. But, the fact is that campaign professionals use attacks because they work.  There is not a more clear example of that than in the US Presidential race.

McCain launched an attack ad last week criticizing Obama for just being another “celebrity.”  The targeted TV ads were matched with legions of surrogates with identical talking points.  Polling released this morning reveal that McCain’s full-court press has made a difference, allowing McCain to pull from eight-points down to flirting with near even with Senator Obama.

McCain’s upward movement and Obama’s negative fluctuation is a clear indication that the attacks are working.  The McCain campaign now appears to be hitting its target.

Hispanics have traditionally leaned more toward Democratic candidates than Republicans; however, between 1999 and 2006, Hispanic voters were progressively becoming more Republican with each election.  This trend has come to an end, as Hispanic voters have made a dramatic shift to become significantly more Democratic than any time in the past decade in just the past two years.

A Pew Center survey of Latino voters reveals that in 2006, Hispanics were leaning less Democratic since 1999 as 49% identify themselves as being a Democrat, and 28% Republican.  But in only two years, 65% of Hispanics now say they lean Democratic with only 26% indicating they are Republican. 

This sixteen-point shift is significant for any bloc of voters to move over a short amount of time and is a clear indication that Democratic candidates, principles, and issues are attracting more Hispanics than ever before.  The Pew survey attributes the movement to overall dissatisfaction with the county’s direction and increasing trust in the Democratic party to bring solvency to a number of Latino community issues.  However, the number of Hispanics that believe the country is headed on the wrong track (70%) is not drastically different from the overall number of voters in the US that believe the same (74%) [Source: NBC/Wall St Journal Survey - July 2008].  This means that Hispanics, like the majority of voters all over the country, are frustrated with the status quo - which includes a Democratic controlled Congress.

The fact that Hispanics are moving away from the Republican Party is another indication that the GOP brand is damaged and continuing to falter at a macro identification level; however, this doesn’t necessary mean that 65% of Hispanics will vote for a Democrat over a Republican in November.  The truth is that Hispanics, just like most swing voters, develop indirect personal connections with candidates.  These personal motivations contribute more to how these voters will cast their ballot than how they identify themselves (i.e., Democrat, Republican, Independent, etc).

Most swing voters (including Hispanics) won’t turn their full focus on the Presidential race until mid-to-late September at the earliest.  By then, both parties will have their conventions and ground operations will be in “full-swing.”  Television, radio, mail and print ads will begin to be more plentiful than interns on Capitol Hill.  These tactics will influence how these voters view and ultimately choose their next President.

Despite the economy and energy prices’ dramatic rise as a top issue among the electorate, voters indicate that the war in Iraq, illegal immigration, and continued protection from terrorism are also key concerns and will play a significant role in how they choose the next President of the United States.

The recently released AP/Yahoo News Poll demonstrates that voters are largely unsure which candidate can best fix the ailing economy (32% Obama - 28% McCain); however, they clearly believe that McCain is best equipped to deal with:

The situation in Iraq (39% McCain - 33% Obama)
Terrorism (44% McCain - 20% Obama)
Immigration (30% McCain - 24% Obama)

Clearly, voters trust McCain with overall security of the country and believe that he can reign-in the situation in Iraq.  The key for the McCain campaign now lies in proving that his experience can bring economic relief.  This is a far less strenuous task than Obama’s, who first has to prove that he is even trustworthy.  But, voters may have a hard time trusting in a candidate that is untested. 

Campaign media can do a lot, but building an alternate reality where Obama has actual leadership experience is not something I’ve ever seen political ads (even the best ads) do.

The AP/Yahoo News poll released today points to further problems voters have with Obama.  This isn’t the first time that we’ve seen voters, even Democratic voters, say they just don’t trust Obama.

A June FoxNews/Opinion Dynamics survey shows that voters trust McCain more than Obama (43%-37%).  However, the big news in that survey is that only 69% of Democrats say they trust Obama, while fully 81% of Republicans say they trust McCain.  This signals a major trust issue, even among Obama’s base voters.  But, the bad news doesn’t stop there as Independents indicate that they trust McCain over Obama by a 43%-25% margin.

Now, move forward to the AP/Yahoo News poll released today.  This survey again shows that voters are struggling with Obama, noting that he lacks experience and is perceived as being dishonest.  Contrarily, McCain’s age comes as the top attribute voters mention, but this naturally contrasts with Obama’s lack of experience.  With the exception of voters pointing to McCain as being “Bush”-like, the electorate are complementary of McCain, again demonstrating continued trust in judgement.

Here is what voters say:

John McCain:
1. Old, 19 percent
2. Military service, 9 percent
3. Record, qualifications, 8 percent
4. Bush, 7 percent
5. Strength, 7 percent
6. Insider, politician, 7 percent
7. Iraq, terrorism, 6 percent
8. Honest, 5 percent
9. Republican, 5 percent
10. (tie) Moral/good and dishonest, 4 percent

Barack Obama:
1. Outsider, change, 20 percent
2. Lack of experience, 13 percent
3. Dishonest, 9 percent
4. Inspiring, 8 percent
5. Liberal, 6 percent
6, 7 (tie). Obama’s race, young, 6 percent
8. Not likable, 5 percent
9. Intelligent, 4 percent
10. Muslim, 3 percent