Closing the Intention-Action Gap
Will you floss your teeth today? Will you exercise tomorrow? Will you visit a museum within the next six months? Surveys that ask questions like this about future intentions—even when there is rigor built into the methodology—tend to be an unreliable indicator of behavior. While an imperfect measure, such studies can still offer some valuable benefits.
Often represented in a survey as “I intend to…” or “How likely are you to…” followed by a Likert scale, behavioral intention is a way to measure how likely someone is to perform a given action. Unfortunately, we are generally poor predictors of our own future behavior and people do not always act on their intentions. This breakdown between what people say they want to do and what they actually do is called the “intention-action” gap. As we are all aspirational creatures by nature, surveys that ask about future intentions tend to draw out responses that are steeped in optimism and framed by an idealized version of the future.
Surveys that ask about future intentions are also susceptible to “social desirability bias,” or the tendency to answer questions in a way that will be viewed more favorably by others, which can skew results to exaggerate socially desirable behaviors and attitudes (for example, giving to a cause or being the kind of person who joins an art museum). While social desirability bias is usually associated with highly sensitive or embarrassing topics such as drug use or sexual behavior, people will also adjust their responses about the most trivial of subjects based on what they think the researcher or organization wants to hear. In some cases, respondents may not even be aware that they are lying to themselves out of guilt or an idealized self-identity.
So, if asking people about their intentions doesn’t provide a valid prediction of future behavior, why ask them at all? While asking about future intentions can lead to biased results, there is a strong business case to ask current and prospective customers about what’s important to them, if they intend to do certain things, and if they plan to purchase products and services in the future. Behavioral intention can, in fact, be used to predict behavior through (1) the influence of attitudes, (2) the “mere measurement effect,” and (3) predictive analytics.
The Influence of Attitudes
Our attitudes are a compound construct made up of three distinct components: affective (how people feel), behavioral (what people intend to do), and cognitive (what people think). The affective component involves a person’s feelings such as “I’m afraid of snakes.” The behavioral component influences how we act or behave like “If I see a snake, I will run away.” The cognitive component involves a person’s knowledge or beliefs such as “I believe snakes are dangerous.” To put attitudes into the context of membership, consider a botanic garden member’s attitude toward being a member:
In terms of affect: I feel proud when I support the garden’s mission.
In terms of behavior: I regularly attend lectures about conservation.
In terms of cognition: I believe being a member is the right thing to do.
Attitudes are powerful, and researchers have put forth compelling evidence that specific behaviors can be predicted when an attitude is aligned with intentions. For example, if a person holds an attitude (e.g. donating blood save lives) and they say they will do something related to that attitude such as “I will donate blood this week,” there is a greater likelihood that they will follow through on their stated intention. That is, the predictive accuracy of behavioral intentions are strengthened when there is a strong tie between the attitude and the behavior we are seeking to predict. In fact, a person’s beliefs about a particular behavior is the most critical determinant of whether or not they will perform that behavior. Thus, if a person believes that “membership is important to supporting the museum’s mission” they are more likely to behave in a way that is consistent with that attitude (e.g. joining).
The Mere Measurement Effect
In a nationally representative study of more than 40,000 participants, researchers investigated whether or not simply asking someone if they intended to buy a new car would increase actual purchase rates. It did, by a whopping 35 percent. This and numerous other studies show that answering survey questions can actually shift your behavior—especially if the behavior is seen as socially desirable. Dubbed the “mere measurement effect” (because the mere act of measuring a person’s intent makes it more likely they will act on it) has been observed across a broad range of behaviors, from disease prevention and consumer purchases to prosocial behaviors such as blood donation and voting. Interestingly, studies have shown that the mere measurement effect can even be amplified by asking people to describe how and when they plan to take a specific action.
While the vast majority of studies have shown that the mere measurement effect can be harnessed as an effective intervention to increase uptake of certain behaviors, it should be noted that a failure to control for targeting and response biases in surveys can lead to overestimating the impact of the mere measurement effect. In one study, researchers found that asking customers about their intent actually decreased purchase frequency, spending per order, and total customer spending—yikes. Moreover, the mere measurement effect did not occur (or generated the opposite effect) if the survey was clearly sponsored by an organization that would likely want to persuade respondents to take up a certain behavior (e.g. a survey sponsored by the American Fruit Growers Association asking if you intend to eat less meat). These findings indicate the need for further investigation into the potential benefits of the mere measurement effect—particularly with respect to membership in the arts and culture sector.
Predictive Analytics
Museums can leverage “big data” to develop predictive models and make predictions about future behavior. Through techniques such as data mining, advanced statistical analysis, and machine learning, predictive analytics relies on transactional and CRM data such as membership sales, visit frequency and recency, email behavior, donations, and renewal behavior as well as third-party data to uncover patterns and dispel assumptions. When paired with predictive analytics, behavioral intention provides a valuable input that can strengthen a predictive model. For example, predictive analytics may reveal that visitors who say they intend to join within a week after their first visit are more likely to join at a higher level than those who do not express that same intent. Such information would be invaluable in developing a membership marketing strategy and directing resources toward first-time visitors.
When properly designed (and appropriately applied), behavioral intention surveys can provide valuable insights to inform business strategy and may even encourage desired behaviors by activating the mere measurement effect. However, before stated intentions can be used to predict behavior, audiences’ attitudes toward museums and membership must be measured and understood. To be effective in influencing behavior, museums must focus investment upstream towards shaping attitudes and reinforcing beliefs about the importance of membership.
To close the intention-action gap, museums need to intentionally design systems and environments to combat barriers to action. For example, museums can help visitors and members to take action on their best intentions by eliminating the barrier of transportation, addressing concerns about guest safety, and offering a more affordable monthly payment option for membership. Further, museums should ensure that surveys use detailed questions about how and when the person intends to take an action to increase the likelihood that they will follow through on their intentions. For example, if we are interested in predicting renewal behavior, the survey should include questions that are specifically focused on the action of renewing such as “How likely are you to renew your membership online next week?”
Finally, while the old adage “the best predictor of future behavior is past behavior” still holds true, in times of massive change and uncertainty, museums will need to prepare for a future that does not follow past patterns of behavior. As it is impossible for anyone to know how profound the impact of the COVID-19 pandemic will be on visitor and member behavior once things return to “normal,” it will be important to take audience responses about stated intentions with a grain of salt.*
References
https://www.amazon.com/Attitudes-Personality-Behavior-2nd-Ajzen-ebook/dp/B0019388T2/
Morwitz, V. G., Johnson, E. J., & Schmittlein, D. (1993). Does measuring intent change behavior? Journal of Consumer Research, 20, 46-61.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4931712/
https://marketing.wharton.upenn.edu/wp-content/uploads/2016/10/anderson_hansen_tripathi.pdf
https://www.researchgate.net/publication/228269675_When_Consumers_Do_Not_Recognize_Benign_Intention_Questions_as_Persuasion_Attempts
* To take a statement “with a grain of salt,” means to maintain a degree of skepticism about its truth. One origin theory of the idiom is that in 77AD Pliny the Elder (a natural philosopher under The Roman Empire) translated an ancient cure to poison, in which he wrote that it should be taken with “a grain of salt,” suggesting that salt was important in counteracting the effects of the poison. https://www.bloomsbury-international.com/student-ezone/idiom-of-the-week/1326-take-it-with-a-grain-of-salt/
Share your ideas, comments, and questions with fellow choice architects!
Is your museum considering using predictive analytics to predict future behavior? How are you using behavioral intention data in your membership program? Have you seen the mere measurement effect in action?