Health Perception and Commuting Choice 

A survey experiment measuring behavioral trade-offs between physical activity benefits and pollution exposure risks

Research Questions

1. Whether people choose to switch commuting modes to protect themselves under air pollution scenario?

2. Whether people are rational in making trade-offs between the health benefit and cost?

3. What are the determinants of the choices?

(a) Job locations of survey participants; (b) 2019 daily average PM2.5 pollution (micrograms/m3) in Zhengzhou

Survey designs

The survey was conducted in July, 2019. Our major target group is non-vehicle commuters whose job location is around Zhengzhou CBD area, with a total 2285 valid participants.

Questionnaires are designed and collected through Qualtrics. Survey takes 15-20 minutes, including an opening video introducing the iPad interface of electronic questionnaire, informed consent, four rounds of commuting choices questions, two information intervention, and some socio-demographics habits and preferences characterization questions. All respondents are asked to make four rounds of commuting choices. Each round of choices is bundled with three questions: their primary mode of commuting, whether they are willing to switch to active commuting (i.e., biking or walking) given a reasonable amount of subsidy, and what is the minimum amount of subsidy they are willing to accept for the switch. 

Results

1. Air pollution and commuting choice

After we present with people their personal pollution exposure information, we see a large reduction in respondents choosing public transit and a large increase in motor vehicle (i.e., car/ taxi). Assuming information is the only thing updated their believes, the result indicates a knowledge gap between people’s perceived pollution exposure and the reality, specifically, people seem to  underestimate the exposure in public transit.

2. Health trade-offs in active commuting

​

Though biking within 30 minutes has exercise benefit outweighs pollution cost, people who live close to the job location also intentionally switch from active commuting to other transportation modes to hedge against their subjective perception of exposure risk.

3. Implications for active commuting policies

People with counterfactual biking time 28-60 minutes value the health cost of pollution exposure the most. The differences for extensive and intensive margin can be partially explained by the selection bias, since people who adopts active commuting for biking time longer than 30 minutes usually have greater preferences for physical activities or having limited alternative transportation choice. 

Conclusion

• People have the intention of switching commuting modes as a channel of air pollution adaptation.

• People seem to overreact to air pollution.

• Both financial subsidy and green nudge policies to encourage active commuting are likely to be in vein under air pollution.
  

Team

Yichun Fan
MIT Sustainable Urbanization Lab

Juan Palacios
MIT Sustainable Urbanization Lab

Siqi Zheng
MIT Sustainable Urbanization Lab

Air Quality in different transport modes in Zhengzhou

Reporting the levels of multiple traffic related pollutants: PM, SO2, CO, NO2 in different transport microenvironments in downtown Zhengzhou.

Research Design

1. Route selection (see image on the left)

2. Time of Day:

   • Rush hours: 7am-9am, 5pm-7pm, all days of week.

   • From Sunday July 5th to July 26th:

Results overview

• NO2 , CO and  SO2 are primary pollutants and are highest in the bus and taxi, reflecting the closest proximity to the exhaust of other vehicles. PM and O3 are secondary pollutants. 

   

• The highest correlations between pollutants on different transport modes were found between PM2.5 and PM10. The bus and bike PM were more highly correlated than that of taxi and bus, or taxi and bike. 

   

• O3 and NO2 were inversely correlated, especially for bike drivers. 

• SO2 on taxi is strongly inversely correlated with CO measured on bikes, while it is weakly correlated with the CO measured in the bus. The bike and bus CO are weakly inversely related.

The most direct comparisons between travel mode were made by taking the average of individual (pairwise) run ratios between modes, because this limited the effect of temporal variations in meteorology

When comparing the average of ratios from the pairwise analysis to the ratio of the overall mode averages, we found that the rank ordering obtained is the same. 

​

In general the pairwise analysis provided greater contrasts, with an extreme increase in the bus/bike NO2 ratio, the SO2 ratios of all three: taxi/bike, bus/bike and taxi/bus

Pros and Cons:

• ​Pros:

1. We measured a range of pollutants.

2. We attempted to do a pair-wise analysis.

• Cons:

1. Our study was only conducted in one season: summer, and therefore the results cannot be generalized to other seasons.

2. We used low-cost sensors, and there could be issues with calibration

​
Future steps:

Calibrate Purple Air data + integrate subway analysis

Team

Priyanka deSouza
Department of Urban Studies and Planning, MIT

Ruining Lu
Fairsense (Beijing) Environment Technology Co., Ltd

Pat Kinney
Boston University School of Public Health

Siqi Zheng
MIT Sustainable Urbanization Lab & MIT Center for Real Estate