Experimental Design
We conduct a natural field experiment in cooperation with a nation-wide energy efficiency assistance program. The main feature of that program is the ‘Refrigerator Replacement,’ which subsidizes the modernization of refrigeration appliances of low-income households with a 100 euro voucher. Upon eligibility, the household receives an information letter that informs about procedural requirements to claim the 100 euros, and average annual savings from replacement. Following the receipt of the information letter, households have to request the voucher, replace their old refrigerator against a new, energy-efficient refrigerator, and can then claim the payment of the 100 euros.
Our experiment randomizes the presentation of the information letter across three dimensions:
(1) Gain vs. Loss Frame: In the gain frame, both an icon (a purse full of money) and the framing of written information highlight the annual savings from replacing the refrigerator. In the loss frame, the icon shows a purse losing money, and annual savings are framed as lost if the household does not invest.
(2) Personalized information: Instead of presenting average annual savings based on engineering estimates, the information letter presents households the annual savings of households with the same composition as theirs. E.g., a household with two adults and one child is informed about the average annual savings from replacement of other households with two adults and one child, and sees a corresponding family icon.
(3) Reminder: In combination with the information letter, the household receives a haptic reminder (a refrigerator hanger) of the program, and/or is reminded 4-8 weeks after information letter receipt about the refrigerator replacement via mail/telephone.
In total there will be 10 treatment arms (that we will potentially pool): (1) Control/standard information letter, (2) gain frame, (3) loss frame, (4) personalized gain frame, (5) personalized loss frame, (6) gain frame and mail/telephone reminder, (7) gain frame and haptic reminder, (8) loss frame and haptic reminder, (9) personalized loss frame and haptic reminder, (10) gain frame and both mail/telephone and haptic reminder.
We study how this variation of information letter presentation and additional reminders affect (i) voucher request and (ii) refrigerator replacement probability.