Information, Service Delivery, and Citizen Engagement: Evidence from a Randomized Trial in San Diego

Door Hanger (DH): In selected weekly bundles of street segments, segments are randomized to (i) Full-DH (all addresses on segment receive a door hanger), (ii) Partial-DH (exactly half of addresses on segment receive a door hanger), or (iii) Control-DH (no addresses receive door hangers). Door hangers provide instructions and a QR/URL for using San Diego’s Get-It-Done (GID) system.

Door-to-Door Survey (D2DS): Starting in week 2, a subset of bundles are selected for an in-person survey of adult residents to measure service beliefs, GID awareness/usage, and pothole-related perceptions. Four of six weekly D2DS bundles are drawn from the prior week’s DH bundles and two are drawn from bundles without DH but with recent pothole reports.

Pothole Fix (PF): Not randomized. We construct PF(x) indicators for whether pothole reports are fixed within x∈{1,…,7} days. For identification, we exploit standardized pothole crew allocation mechanisms by the Transportation Department.

2026-01-10

2026-08-08

Administrative (segment-level, from GID):
- weekly counts/indicators of pothole service requests
- weekly counts/indicators of non-pothole service requests
- weekly counts/indicators of any service requests
Also cumulative outcomes at 1 week, 1 month, and 3 months after treatment week.

Survey (D2DS):
- General views of City services
- Overall satisfaction with City services; perceived efficiency and fairness in the use of tax dollars.
- Awareness of the GID app/website; prior use of GID for any services; reasons for non-use.
- Whether respondents have noticed potholes on their street or nearby; whether they have noticed recent pothole repairs.
- Perceived costs and policy preferences related to potholes
- Perceived disruption from potholes; support for hypothetical tax increases earmarked for faster pothole repairs.
- Expected time to repair potholes with and without citizen reporting.
- Perceptions of how many neighbors notice potholes, know about GID, and would report potholes.
- How respondents typically travel around the city and time spent on city roads.

This study is a field experiment in the City of San Diego evaluating how information provision affects citizen engagement with municipal services. Street segments are randomly assigned to receive different levels of exposure to an informational door hanger about the City’s service request system, including a control condition with no door hanger.

Outcomes are measured using administrative data on service requests aggregated at the street-segment level and, for a subset of locations, survey responses measuring beliefs and perceptions about City services. Pothole repairs themselves are not randomized and are analyzed using observed variation in repair timing.

A pilot was conducted prior to the study and is excluded from the analysis.

Not available

Randomization done in office by a computer.

Door-hanger treatment is randomized at the street-segment level within selected geographic bundles of segments. Door-to-door survey (D2DS) locations are selected at the bundle level.

No

N/A

Administrative outcomes: ~13,000 eligible street segments (segment-week panel), of which ~1,500–2,000 segments are expected to be sampled into door-hanger field operations over the study period. Survey outcomes: ~1,400 responding households (door-to-door survey).

Door-hanger randomization (street segments): 50% Control-DH, 25% Partial-DH, 25% Full-DH (shares of sampled segments).