Using Stickers to Improve Customers' Water Payment Behavior in Kosovo.

There are six sticker types in a 3x2 design. Half of the stickers highlight the positive consequences of paying, whereas the other half highlights the negative consequences of not paying. The stickers will be attached to the invoices of customers. Customers are informed that they may use the stickers as reminders to pay, by putting the sticker and bill on the inside of their door. Further, the stickers intend to strengthen the link between payment behavior and self-concept of customers. Two types of self-concept treatments are tested: Persuasion (e.g., "Please be a responsible citizen") and attribution (e.g., "You are a responsible citizen").

The treatments:
Untreated baseline condition

Positive frame - Reminder
Positive frame - Reminder + Self-concept persuasion
Positive frame - Reminder + Self-concept attribution

Negative frame - Reminder
Negative frame - Reminder + Self-concept persuasion
Negative frame - Reminder + Self-concept attribution

2017-06-06

2017-08-31

Payment behavior and water consumption behavior

One particular research interest is to investigate between and within treatment differences among customer types and their past payment behavior as well as long-term effects.

Subjects will be randomly assigned to one of six treatment conditions or a untreated baseline condition by day and reader. To implement the intervention, over a period of three months, a student research team accompanies eight members of the staff of the water company ("readers") during their work. The readers read the water meters throughout the month on a rolling basis and then print the bill with a handheld mobile printer and directly deliver it to the customer. The amount due for this month and the total amount due are highlighted with a marker. The sticker ist hen attached to the bill and put on a customers door, mailbox (if available) or delivered personally (if customer is met).

All treatments are randomized by a computer program. All treatments are balanced by day and by reader.

The level of randomization is at the individual level (customer).

No

The eight readers are assigned to bill over 15000 customers each month. However some customers are inactive, their location unknown, the meter cannot be read (damaged, flooded) or commercial customers (which are not treated), etc. The number of eventually treated private customers will thus be substantially lower and is unknown until the coding team has digitalized all values.

The eight readers are assigned to bill over 15000 customers each month. However some customers are inactive, their location unknown, the meter cannot be read (damaged, flooded) or commercial customers (which are not treated), etc. The number of eventually treated private customers will thus be substantially lower and is unknown until the coding team has digitalized all values.

The eight readers are assigned to bill over 15000 customers each month. However some customers are inactive, their location unknown, the meter cannot be read (damaged, flooded) or commercial customers (which are not treated), etc. The number of eventually treated private customers will thus be substantially lower and is unknown until the coding team has digitalized all values.

Although all treatment arms are randomized and balanced by day and by reader, this does not guarantee an equal share of treated customers between the treatment arms. Treatment arm sample size may vary because the reader may for example get sick, has connection problems with his printing device or goes on vacation.