Field
Trial Title
|
Before
Experimenter Effects in Experimental Program Evaluation
|
After
Evaluating a Free LED Lighting Program
|
Field
Abstract
|
Before
Are experimental program evaluation studies robust to experimenter effects? Across different recruitment waves, we study this question in the context of a free LED lighting study.
|
After
Across different recruitment waves, we study a free LED lighting study.
|
Field
Last Published
|
Before
August 02, 2019 03:30 PM
|
After
August 08, 2019 08:45 PM
|
Field
Intervention (Public)
|
Before
Recruitment intervention:
Recruitment wave 1: Households will be recruited for a free LED lighting study at their home. Recruitment will be randomly varied across two dimensions. First, whether participants receive a $10 incentive or not. Second, whether a household can opt-out of the study by checking a box on a warning given to them regarding the study.
Recruitment wave 2: Households will be recruited for a free LED lighting study on Amazon Mechanical Turk. Recruitment will be randomly varied across one dimension. Participants will either be told their participation would be a favor to the study or not.
Program intervention: A free LED lighting program that gives recipients 8 free LED light bulbs
|
After
|
Field
Primary Outcomes (End Points)
|
Before
Recruitment wave 1: Door answer, study participation, and energy usage (all hours and evening hours, 6pm to Midnight)
Recruitment wave 2: Study participation and energy usage (all hours and evening hours, 6pm to Midnight)
|
After
Energy usage
|
Field
Primary Outcomes (Explanation)
|
Before
|
After
Recruitment wave 1: Door answer, study participation, and energy usage (all hours and evening hours, 6pm to Midnight)
Recruitment wave 2: Study participation and energy usage (all hours and evening hours, 6pm to Midnight)
|
Field
Power calculation: Minimum Detectable Effect Size for Main Outcomes
|
Before
Recruitment wave 1 will be powered to observe the door answer effect observed in DellaVigna, List, and Malmendier (2012).
Recruitment wave 2 will be powered to observe the strong demand effect in de quidt, Haushofer, and Roth (2018).
Across both recruitment waves, simulations suggest this study will be powered to detect a 0.07 kWh effect of the LED lighting program during evening hours.
|
After
Recruitment wave 1 will be powered to detect the effect found in DellaVigna, List, and Malmendier (2012).
Recruitment wave 2 will be powered to observe the effect in de quidt, Haushofer, and Roth (2018).
Across both recruitment waves, simulations suggest this study will be powered to detect a 0.07 kWh effect of the LED lighting program during evening hours.
|
Field
Intervention (Hidden)
|
Before
|
After
Recruitment intervention:
Recruitment wave 1: Households will be recruited for a free LED lighting study at their home. Recruitment will be randomly varied across two dimensions. First, whether participants receive a $10 incentive or not. Second, whether a household can opt-out of the study by checking a box on a warning given to them regarding the study.
Recruitment wave 2: Households will be recruited for a free LED lighting study on Amazon Mechanical Turk. Recruitment will be randomly varied across one dimension. Participants will either be told their participation would be a favor to the study or not.
Program intervention: A free LED lighting program that gives recipients 8 free LED light bulbs
|