Applying Behavioral Sciences in the Real World
Mmmmmm...the Power of Doughnuts
Mindy Hernandez, Founder, One Decision
Sometimes the most interesting questions are most are the basic. What’s the best way to get people to respond to us? We all have questions we’d like to ask or feedback we’d like clients, customers and even friends to give us. But getting people to take time out of their commuting, basketball watching, behavioral eceonomics blog-reading lives and sit down to pay attention to our questions can be surprisingly difficult. I know that I almost never respond to organizations' or businesses’ pleas to “help us improve!” or “take this quick survey." I want to help, I just never seem to have the 5 minutes.
Val Klein at Consumer Credit Counseling Service of the Delaware Valley, or CCCSDV (check out Val’s recent guest blog here), was struggling with this issue. CCCSDV offers debt management plans (DMP) (see Federal Trade Commision's Website for more information) where people struggling with debt can enroll in the program and CCCSDV will help them manage and pay-off their debt. Many people are offered the program, a few enroll, but many drop out. CCCSDV wanted to find out more about why some people never enroll, why others drop out and what they could do about it. They sent out hundreds of surveys but the response rate was low - about 10%.
There is very solid research on the power of “pre-incentives” on increasing responsiveness for things like surveys (see this very helpful meta analysis from the British Medical Journal).
What’s a “pre-incentive”? Well, a traditional incentive might work like this: You get a letter in the mail saying "complete this survey and we’ll send you a gift certificate for free doughnuts!" A pre-incentive is different. You open a letter and the gift is ALREADY inside. You don’t have to do anything - there’s a token there for the taking.
This is an obvious deviation from the traditional self-interested economic model that might say: if you want people to do X behavior you provide X subsidy, and a rational person will calculate whether his or her time is worth X subsidy and then act accordingly. By this logic you’d expect a pre-incentive to be a failure because, using strictly rational accounting, this is a great profit-maximizing opportunity: I can pocket the gift and do zero work. But this is actually the exact opposite of what the previous studies tell us and, happily, what Val and I found.
The behavioral idea is that our motivational engine is not always greed. Actually, we are wired to be extremely sensitive to ideas of fairness and reciprocity. When someone does something nice for us we feel a strong urge to reciprocate. And it can feel very uncomfortable to take a “gift” and give nothing in return.
So CCCSDV sent out almost 1,000 surveys to people who had enrolled, dropped out or rejected the DMP program in three conditions:
1. A survey asking people to respond - no incentive is given. (Click here to see a sample of this letter.)
2. A survey asking people to respond and promising a gift (a $5 Dunkin’ Donuts gift card) upon completion.
3. A survey with a “pre-incentive” included (a $5 Dunkin’ Donuts gift card), thanking people in advance for helping. (Click here to see a sample of this letter).
Among those enrolled or who declined DMP, the pre-incentive letter increased the response rate by 30 percentage points, basically more than tripling the response rate over the standard (non-incentive) letter. In addition, the pre-incentive letter increased response rates for those who enrolled in the DMP program as well as those who declined DMP - this second group had proven especially difficult to reach.
Survey Response Rates (Percentage of people who returned completed surveys) N=968
For future study: If the power in changing behavior in this case was about appealing to people's ideas of reciprocity and fairness, how much difference did the actual amount of the gift certificate make? This is not just an academic question; in the real world there are of course real costs and budget constraints to consider. So how might response rates change if the pre-incentive was a free bus pass or a Hershey's kiss? (Previous research tells us that pre-incentives that represent actual cash are more effective than small gifts like chocolate, but we don't know the actual impact for this population in this context.) What if the gift certificate was only for $2 or even $1? In both cases, my hunch is that response rates would dip, but not by much; but there's only one way to find out...