Another political season is on the horizon and a relatively large field of candidates will be vying for the hearts, and sometimes the minds, of eligible voters. That means we will hear a number of slogans, sound bites and other messaging efforts to sway our allegiance. Politicians rely on certain techniques to craft these messages. Basically, they try to use to their advantage words that activate existing mental structures and metaphors.
One technique they use is called framing. The concept of framing was made popular in politics by Frank Luntz and George Lakoff, but based upon psychological (and, recently, economic) research by Kahneman and Traversky, amongst others. Briefly, framing is a strategy to use language in a message that leverages a cognitive bias most of us have; a way to manipulate thought.
For example, would you be more likely to fund a program that saves 90 percent of the people who use it, or a program that is ineffective for 10 percent of the people who use it? Although they are logically (and statistically) the same, people are much less likely to fund a program that emphasizes its failures than its successes. In other words, there is an art to how one presents, or frames, an effective message. Frames guide thought, and thought guides behavior.
The difference in response rates based upon how the issue was framed can be quite practically significant when it comes to votes, donors and the many other currencies relative to politics.
The problem is that there is no perfect algorithm for generating effective political frames. Some seem blatantly manipulative and disingenuous (for example, naming a timber-cutting bill the “Save our Forests” act), and others can backfire. Frames have to be tested, either through speeches and media “trial balloons,” or through strategies such as focus-group testing. These methods have their liabilities.
Speeches and trial balloons are public and, in an era where everything seems to be recorded and taken out of context, this exposes the candidate to being penalized as they refine their argument (flip-flopping, being unclear about their vision, etc.). Focus-group testing is time-consuming and subject to errors based on the small groups of people in the room, as well as tester bias.
Online marketers increasingly use a much more efficient method to test the efficacy of their frames. Split testing (aka A/B testing) is a powerful strategy to test different versions of an idea and compare which is most effective to guide desired behavior. For example, some companies want to know which background color is most effective to encourage further engagements. They can randomly give different versions of the web page (version A or version B) to different visitors to the website and measure which ones lead to more clicks, longer times spent on the page, etc.
The biggest advantages of split testing are related to scale. With the right amount of traffic, they can test several possibilities using thousands of visitors, and collect a vast amount of data in the background. This data has the potential to determine which versions work, or don’t, for particular segments of the target audience, and to inform future iterations.
There is an art to how one presents, or frames, an effective message.
A/B testing, fairly new to politics, has not gone unnoticed. For example, in a very interesting article published after the successful Obama campaign, Kyle Rush documented how A/B testing was used in the campaign to optimize web pages, select content and develop a productive flow. The results? Increases in the number of people following through to sign-up (161 percent) and an increase in converting website visitors to donors (49 percent). This powerful strategy, traditionally used to test formats and strategies, can also be leveraged to test frames.
We decided to run a test of two different frames related to recent political debates about political correctness. Using dynamic split testing tools, we were able to demonstrate this simple effect with more than 3,500 visitors in just 24 hours. Responders were asked to agree or disagree to one of the following statements:
- “Because people can sometimes be offended in certain situations, you should try to be as politically correct as possible.”
- “Because people can sometimes be offended in certain situations, you should try to be as kind as possible.”
Responders were significantly less likely to agree with being sensitive to others’ feelings when it was framed as political correctness than when it was framed as a form of kindness (87.46 percent versus 61.69 percent). In other words, simply framing the same concern as political correctness reduced the number of people in favor of the position by more than 450 people (1,548 versus 1,090). This method has the further advantage of allowing us to drill down to specific user characteristics (for example age, gender, neighborhood) to refine and focus the results on specific target groups.
It is easy to see how split-test methods can be leveraged for efficiently developing effective frames. The difference in the above results represents a potentially sizable yield. Assuming politicians can either drive web traffic to their sites, or can develop strategies to drive people to affiliated websites, they can utilize split testing to refine frames to guide the campaign, as well as to refine fund-raising strategies.
The difference in response rates based upon how the issue was framed can be quite practically significant when it comes to votes, donors and the many other currencies relative to politics. This kind of speed and precision, the kind Internet marketers are used to, is pivotal, especially in the fast-paced world of political campaigns.