CAN FAIRY Recycling Program Survey

The Can Fairy program delivers weekly email messages to households participating in curbside recycling. Households are randomly assigned to receive one of two types of messages: (1) educational messages that provide practical information about correct recycling and how to reduce contamination, or (2) norm-based messages that emphasize how others in the community recycle to encourage social motivation. All households continue to receive standard recycling services. Bin weights and contamination rates are tracked throughout the study to measure behavioral changes.

2025-11-01

2026-04-01

Recycling weight (kilograms per collection)
Contamination rate (percentage)
Recycling knowledge and confidence
Recycling attitudes and moral obligation
Perceived social norms
Behavioral intentions
Engagement and trust in Can Fairy emails

Recycling behavior is measured using two objective outcomes: bin weights collected during curbside pickup to track total recycling volume, and contamination audits to calculate the proportion of non-recyclable materials in bins. These measures provide concrete evidence of behavioral change over time.

Survey-based outcomes are constructed using composite indices from the Can Fairy Recycling Program survey, which is administered only to participants who voluntarily enrolled in the program. The survey captures several psychological and behavioral constructs that explain why the program may or may not have worked. Recycling knowledge and confidence are measured through questions assessing participants’ understanding of local recycling rules and their ability to sort items correctly. Attitudes toward recycling are captured through statements reflecting moral obligation, personal responsibility, and positive evaluation of recycling behavior. Perceived social norms are measured through items that assess both descriptive norms (what others do) and injunctive norms (what others expect). Behavioral intentions are derived from questions asking about participants’ plans and likelihood to recycle in the following week.

The survey also includes items on engagement with and trust in the Can Fairy emails, focusing on attention, credibility, and whether the messages made participants notice others’ recycling behavior. Each of these constructs will be created by averaging or summing standardized item scores, and reliability will be evaluated using Cronbach’s alpha. Together, these indices complement the objective recycling and contamination data, allowing the study to assess both behavioral changes and the psychological mechanisms underlying them.

Perceived barriers to recycling
Motivation for program participation
Program satisfaction and perceived usefulness
Willingness to recommend the program to others
Changes in awareness of community recycling behavior

Secondary outcomes are constructed from additional items in the Can Fairy Recycling Program survey to provide deeper insights into participants’ experiences and behavioral context. Perceived barriers to recycling are measured through items assessing logistical or situational challenges such as bin size, distance, unclear signage, weather, or coordination with household members. Motivation for participation is captured by asking participants to select reasons for joining the program, including environmental concern, social encouragement, and personal responsibility.
Program satisfaction and perceived usefulness are measured through items evaluating how helpful participants found the recycling tips, whether the program made recycling easier, and whether they would recommend it to others. Changes in awareness of community recycling behavior are captured by items asking whether participants noticed neighbors recycling more or became more aware of community expectations.

Households first voluntarily enroll in the Can Fairy program. After enrollment, participants are randomly assigned to one of two groups. One group receives weekly educational emails containing clear information on what can be recycled, how to reduce contamination, and proper sorting practices. The other group receives weekly norm-based emails that emphasize community recycling behavior and highlight shared expectations to encourage social motivation.

All participants continue receiving regular curbside recycling services. Recycling weight and contamination rates are tracked throughout the study period to measure behavioral outcomes. A follow-up survey is conducted among enrolled participants to assess recycling knowledge, confidence, attitudes, perceived norms, and engagement with the emails. This experimental design allows comparison of educational and norm-based messaging among self-enrolled households to determine which type of message most effectively improves recycling behavior.

Not available

Randomization done in office by a computer

Household level randomization is done.

No

200 Households

200

100 HH Education messaging, 100 HH Social Norm Messaging

The study is powered at 80% with a two-sided α = 0.05. Randomization occurs at the household level with 100 households per arm and no clustering. The standardized minimum detectable effect (MDE) is 0.396 standard deviations without baseline controls and 0.280 standard deviations with baseline controls, assuming that baseline outcomes explain approximately 50% of the variance (R² = 0.50). For the recycling-weight outcome (pounds per household per week), the baseline mean is 7.41 lb. Across plausible baseline standard deviations between 2 and 5 lb, the detectable differences range from 0.79 to 1.98 lb (11–27% of the mean) without baseline controls, or 0.56 to 1.40 lb (8–19% of the mean) with baseline controls. For the contamination-rate outcome (percentage of non-recyclables), the baseline mean is 1%. Assuming a baseline standard deviation between 0.5 and 1.0 percentage points, the detectable differences range from 0.20 to 0.40 percentage points (20–40% of the mean) without baseline controls, or 0.14 to 0.28 percentage points (14–28% of the mean) with baseline controls. These values indicate that the study is well powered to detect moderate behavioral effects on recycling weight and small-to-moderate percentage-point changes in contamination.