Problem

The environmental benefits of reducing drive-alone commutes are hard to feel immediately. Are there other ways that we can incentivize people to reduce their car commutes and adopt forms of public transportation? Do certain frames or prompts encourage people to rethink their commuting habits?

Research

To investigate how different frames and incentives may affect peoples’ commuting plans, 876 subjects were recruited via Amazon Mechanical Turk and randomized to see various hypothetical scenarios, then asked how the conditions might change the number of days they commute by driving alone to work. 236 participants reported they would not alter their commute habits given any incentive. Results were analyzed with and without this group of “non-movers.”

Carpool Framing

Participants were presented with one of five hypothetical emails from their employer and asked about their likelihood of carpooling afterwards. The control was an announcement of a carpool initiative. The other four groups read hypothetical emails that framed carpooling focusing on  social benefits,  quiet time, economic savings, or economic “not loss.” The Economic Savings frame resulted in a significant improvement from control, including and excluding “non-movers.”

Figure 1: Likeliness to increase carpooling behavior for various message frames.

Hypothetical Scenarios

Transit Incentive

Subjects were asked about financial incentives to take public transit. We asked participants to imagine they were given $2 per transit commute OR to imagine that they were given a pass that let them use public transportation for free. They were more willing to reduce their drive-alone trips given a free pass than the $2 per transit commute. This finding is consistent with existing behavioral science literature that shows that people are drawn to things that are free. The result was statistically significant in both the full sample and with the exclusion of “non-movers.”

Figure 2: Drive-alone reduction associated with being paid $2 per sustainable commute and with receiving a free transit pass.

Loss/Gain Monthly Commuting Payments

Participants were prompted to imagine that they were given a monthly payment of $300 if they commuted in a way besides driving alone 3+ days per week. They were broken into four groups:

1. $300 framed as a gain
2. $300 framed as a gain and asked to imagine how they’d spend the money
3. $300 framed as money they’d lose if they commuted by driving alone
4. $300 framed as a loss and asked to imagine how they’d spend the money

Examining the full sample, participants were more willing to reduce drive-alone days if the payment was framed as a loss, especially with the imagination exercise. Differences were statistically significant in both the full and exclusion samples.

Figure 3: Drive-alone reduction associated with loss/gain framing and visualization.

Lottery Versus Certain Payment for Commuting

We know from behavioral science research that people tend to overweight small probabilities, which is why lotteries are so popular. In this hypothetical, the control group was asked how much they’d reduce drive-alone days if they were given a weekly payment of $20 for using public transit. The treatment groups were offered entry into a lottery for a weekly drawing of $100 every time they used public transit, a weekly drawing of $400, or a monthly drawing of $400. None of the groups had a significant impact on how people would commute.

Leveraging Competition as Motivation

A hypothetical leaderboard explored the effects of competitiveness. Respondents were offered a free app to track transit commutes (control condition). One group was told results would appear on an online leaderboard for their community, another was told the leaderboard was for their workplace specifically. Participants were more amenable to reducing their drive-alone commutes when the leaderboard aspect was included, especially for the community leaderboard. The difference was statistically significant for the full sample, but not the exclusion sample. 

The Value of Different Forms of Incentives

In this hypothetical, respondents were told to imagine they were given a payment of $100 each month for driving alone 2 or fewer days per week. The payment was presented as either a gift card, cash, or an addition to their paycheck. There was not a statistically significant different effect between these different payment forms.

Schedule Change Options

Participants were asked to either imagine they could work flexible hours to accommodate alternative modes of transportation or that they would be permitted to leave work an hour early once per week if they used alternative transportation. This hypothetical was motivated by the behavioral concept of reducing friction to encourage positive behaviors. Flexible hours did better in the full sample and exclusion sample, but neither had a difference that was statistically significant.

Emergency Ride Home

Another friction-reducing hypothetical was about what to do if you need a car in an emergency. Respondents were offered free carshare membership, two free Lyft/Uber rides per month, or access to a vehicle owned by their employer for use in case of emergency. The work vehicle did best, but differences weren’t statistically significant.

Figure 4: Drive-alone trip reduction associated with each of the tested hypothetical strategies.  Full sample (top) and moveable population/excluded sample (bottom).

Why It Matters

The findings from this study are an important starting point for developing future pilots addressing drive-alone commutes. It’s crucial for policymakers to know not just what works, but also what doesn’t. From this survey study, we see that highlighting economic savings may be the best way to encourage carpooling, framing financial incentive programs in terms of loss and prompting participants to imagine what they could do with the money is a potentially effective design, and free transit passes appeal more than small per-commute dollar payments. Community leaderboards may also be a promising intervention. We also learned that lotteries aren’t as effective as one might expect, and that different forms of incentives of the same value do not change the incentive’s effect drastically. These results are worth keeping in mind for local governments, employers, and other key stakeholders interested in reducing drive-alone commutes.