Case Study

Commuter Commitment

Problem

According to the U.S. Census Bureau (2019), the city of Austin, Texas has seen the nation's 5th largest population increase over the past decade. Needless to say, traffic has become one of the city’s most pressing issues. To get more Austin commuters to leave their cars behind and take public transit, a carpool/vanpool, or walk or bike to work instead, the Center for Advanced Hindsight’s government team partnered with the Austin Transportation Department.

Background

The team launched a pre-commitment commuter study together with City of Austin employees through the Smart Commute Rewards Administrative Leave (ADL) program in March 2020.  The study was launched prior to the commute interruption from COVID-19 and the team was able to examine the impact of the commute pledge over an abbreviated timeframe.  This intervention is based on the behavioral principle of pre-commitment—the idea that specifying and committing to a future action can help achieve positive behavioral change.

Method

City of Austin employees who registered for the spring ADL period were randomized into two groups, the treatment (n=304) and control (n=335) groups. Among those, a total of 401 individuals (treatment, n=195) tracked alternative commute trips over the two-week study period. Both groups were asked to visualize a reward, plan their commute, and identify barriers. Only the treatment group, however, was asked to make a commitment, or a “commuter pledge.”

Figure 1. Example of a commute pledge, presented to the treatment group of 195 individuals.

Results

Note, this study was launched in the weeks prior to COVID related shutdown. As such what was intended to be a multi month long study lasted only a few weeks before it was stopped. As such these results are interim and pending a re-launch of the study.

Figure 2a. Commuter Commitment results (N=639).

For the sample of 639 individuals, the treatment groups saw a statistically significant 6.5% increase in employees tracking at least one trip using alternative transportation. The difference in the number of alternative transportation trips between the treatment and control groups was not statistically significant.

With the short study timeframe, it appears that the presence of the commitment device led to an estimated extra 86 alternative trips tracked in the treatment group (20 people tracking, on average, 4.29 trips per person). If this were to be scaled to the fully expected population of 1,000 employees, it would lead to roughly 340 more alternative trips in a week. Future work will be needed to determine if this finding is reliable (will be replicable) or will sustain (last for any more than a week), but it is a promising start.