Perils of pressure in promoting merit goods: Evidence from a sales experiment in India
Last registered on February 16, 2017


Trial Information
General Information
Perils of pressure in promoting merit goods: Evidence from a sales experiment in India
Initial registration date
February 16, 2017
Last updated
February 16, 2017 2:54 PM EST
Primary Investigator
University of Edinburgh
Other Primary Investigator(s)
PI Affiliation
University of Wisconsin-Madison
Additional Trial Information
Start date
End date
Secondary IDs
Entities devote substantial energy to pressured sales of merit goods with the conviction that purchase stimulates usage. This method backfires if marketing pressure screens in individuals with a low propensity for usage. Through a field experiment in India promoting chlorine water treatment tablets as the merit good, we show that selling a second good alongside the merit good doubles usage by alleviating pressure and introducing a positive framing effect. We also show that the price of this second good matters, as a promotion can reintroduce pressure and undo nearly all of the usage gains induced by framing.
External Link(s)
Registration Citation
Boudot, Camille and Anita Mukherjee. 2017. "Perils of pressure in promoting merit goods: Evidence from a sales experiment in India ." AEA RCT Registry. February 16.
Experimental Details
We manipulate marketing pressure in two ways. In one condition, we market the merit good along with a second product that is well-known, frequently purchased, and cheaper. In another condition, we market the merit good along with the same second product, but this time the second product is sold at a "buy 1 get 1 free" promotion.

The merit good is a bottle of SafDrink containing 100 chlorine water purification tablets.
The second good is a packet of Good Knight fast cards containing 10 cards that repel mosquitoes when lit.
Intervention Start Date
Intervention End Date
Primary Outcomes
Primary Outcomes (end points)
Adoption and usage of chlorine tablets by experimental treatment.
Primary Outcomes (explanation)
Usage of SafDrink is measured by three distinct indicators, two measured and one self-reported. The first indicator is the household's response to whether or not they are currently using SafDrink to treat their drinking water. This is complemented by a visual description from the surveyor for the amount of SafDrink tablets present in the bottle, based on a five point scale (full, more than half, half, less than half, empty). Finally, we supplement these measures with an objective estimate for the amount of chemical chlorine residuals in the household's drinking water. We made use of AquaCheck 5-1 test strips which test for both free chlorine and total chlorine.
Secondary Outcomes
Secondary Outcomes (end points)
Secondary Outcomes (explanation)
Experimental Design
Experimental Design
Each household is randomly assigned to receive one of three offers: (i) SafDrink chlorination product alone, (ii) SafDrink and Good Knight Fast Card mosquito repellent sold at market price, or (iii) SafDrink and Good Knight Fast Card sold on a promotional offer.
Experimental Design Details
Randomization Method
Household level randomization, stratified by membership to the partner NGO.
Randomization Unit
Was the treatment clustered?
Experiment Characteristics
Sample size: planned number of clusters
980 households
Sample size: planned number of observations
980 households, divided between whether the female head of household is a member of the partner NGO.
Sample size (or number of clusters) by treatment arms
T1: Chlorine alone => 480 households
T2: Chlorine + GK at market price => 200 households
T3: Chlorine + GK at discount => 300 households
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
We use data from the Household Recode of the Demographic Health Survey (DHS) 2005-2006, which includes information on treatment practices for drinking water by the household. We limit our sample to urban households in the State of Andhra Pradesh which provides us with a total of 4102 households. Evidence from this data indicates that less than 50% of households treat their drinking water, and of those that do, usage of chlorine is approximately 0.6%. The reason for such low usage is unclear and could be influenced by factors such as price, availability and awareness. Therefore, we assume that if provided with some awareness of the product and the option of buying it at their door-step, take-up amongt the members of Safa (i.e. study population) could stand between 5-40% (10 to 80 individuals out of 200 population respectively). In fact, a population of 200 individuals, with a statistical power of 80%, will allow us to detect a minimum adoption rate of 4% which correspond to only 6 people. Similarly, a population of 200 individuals, with a statistical power of 90%, will allow us to detect a minimum adoption rate of 5% which correspond to only 10 people.
IRB Name
Institute for Financial Management and Research
IRB Approval Date
IRB Approval Number
Post Trial Information
Study Withdrawal
Is the intervention completed?
Intervention Completion Date
October 31, 2016, 12:00 AM +00:00
Is data collection complete?
Data Collection Completion Date
November 30, 2016, 12:00 AM +00:00
Final Sample Size: Number of Clusters (Unit of Randomization)
Total sample size is 980. Randomization was done at the household level. Members were randomly assigned to one of three treatments.
Was attrition correlated with treatment status?
Final Sample Size: Total Number of Observations
We completed the sales experiment with 980 households but have follow-up surveys (measures of usage) on 734 of them.
Final Sample Size (or Number of Clusters) by Treatment Arms
476 members received chlorine alone; 194 members received chlorine at market price (T2); 294 members received chlorine at discount (T3). 16/980 households received the wrong treatment (treatment allocation not the same as treatment received), hence we plan to remove them from our analyses.
Data Publication
Data Publication
Is public data available?

This section is unavailable to the public. Use the button below to request access to this information.

Request Information
Program Files
Program Files
Reports and Papers
Preliminary Reports
Relevant Papers