The Role of Social Signaling in Community Mass Deworming: Evidence from a Field Experiment in Kenya
Last registered on May 18, 2017

Pre-Trial

Trial Information
General Information
Title
The Role of Social Signaling in Community Mass Deworming: Evidence from a Field Experiment in Kenya
RCT ID
AEARCTR-0001643
Initial registration date
October 02, 2016
Last updated
May 18, 2017 7:43 PM EDT
Location(s)

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

Request Information
Primary Investigator
Affiliation
UC Berkeley
Other Primary Investigator(s)
PI Affiliation
Evidence Action
Additional Trial Information
Status
On going
Start date
2016-09-05
End date
2017-12-31
Secondary IDs
Abstract
Can social signaling incentivize adults to take up deworming treatment? Working with the Kenyan Government, we implement a new Community Deworming Program that offers free deworming treatment to adults and explicitly emphasizes the public good aspect of deworming. We test two types of social incentives in the form of colorful bracelets and ink that adults receive upon coming for deworming. Different to most incentives that are material or private in nature (e.g. food, cash transfers) the bracelets and ink make the decision to deworm or abstain from treatment observable and allow adults to signal to others that they contributed to protecting their community from worms. We further introduce a calendar as private incentive that is comparable in its consumption value to the bracelet but cannot easily be observed by others. Communities are randomized into being eligible for the ink, calendar or bracelet or receive no incentive when coming for deworming. In a second stage, we offer free text messages to a random subset of adults living in these communities, reminding them of the availability of free deworming treatment and providing information about deworming treatment take-up in their community. The cluster randomization identifies the combined effect of social signaling, learning and reminders. The text messaging treatment allows us to separately identify the extent to which deworming decisions are driven by (i) reminder effects, (ii) social learning or (iii) the desire to signal to others. We further measure the effect of social incentives on individuals' knowledge and beliefs about others' deworming choices.
External Link(s)
Registration Citation
Citation
Karing, Anne and Karim Naguib. 2017. "The Role of Social Signaling in Community Mass Deworming: Evidence from a Field Experiment in Kenya ." AEA RCT Registry. May 18. https://www.socialscienceregistry.org/trials/1643/history/17782
Experimental Details
Interventions
Intervention(s)
Working with the Government of Kenya we offer free deworming treatment to adults in 3 counties (Busia, Kakamega, Siaya) in Western Kenya. Community Health Volunteers (CHVs) provide the treatment at central locations and sensitize households about the upcoming deworming treatment. CHVs were identified as ideal agents to sensitize communities and give out deworming drugs, both because of their experience in providing health services to communities and because of their familiarity and trusted relationship with community members, allowing them to credibly convey new information (e.g. about the externality aspect of deworming) and address concerns (e.g. that deworming drugs might sterilize women). CHVs inform all study communities one week prior to the start of the deworming treatment about the social benefits of deworming, when and where free deworming treatment will be available and if applicable what type of incentive will be given to adults when coming for treatment. Deworming treatment is subsequently offered at central locations for a period of 10 days.

We identified 150 central locations in the 3 selected counties. Central locations and their catchment areas are randomly assigned, in equal numbers, into one of four intervention arms:
1) CHVs mark adults' fingers with ink;
2) CHVs give adults a green wrist bracelet;
3) CHVs give adults a one-page calendar for 2017;
4) Only deworming treatment is provided and no incentive.

We mapped communities that fall within a 2.5 catchment radius of the central location and randomly selected one community per catchment area. We conducted a census of all adults in those 150 communities and randomly selected 15 individuals in each control arm community to receive reminder text messages, and 15 (for control arm) or 25 (for ink/bracelet/calendar arms) individuals to receive reminder text messages that also contain information about deworming take-up in their community. Adults that signed up for the free text message service receive a text message on the day before deworming starts, on the 2nd, 4th, 6th, 8th and 10th day of deworming treatment. The text message service is free and individuals can unsubscribe at any time at no cost. The reminder messages contain the following content: "Free deworming now at [Central Location]. Reply 111 to stop texts. All texts are free." The reminder plus information about deworming take-up messages contain the following content: "Free deworming now at [Central Location]. [No/few/almost half/half/more than half/almost all/all] of your village came, that is [X] in 10 adults. Reply 111 to stop texts. All texts are free." To verify that adults are reading the text messages and encourage them to do so, we offer an airtime reward of 50 Kenyan Shillings (50 US Cents) to a random subset of subscribers conditional on texting back and confirming the receipt of the text message. The following reward text is sent to adults: "Thank you for signing up for this text message from Evidence Action. To receive your 50Ksh airtime reward, message 123 to [Phone number]. The text is free."
Intervention Start Date
2016-10-03
Intervention End Date
2016-11-04
Primary Outcomes
Primary Outcomes (end points)
Primary outcome: number of adults coming for deworming treatment. Secondary outcomes: individual knowledge and beliefs about others' deworming choices.
Primary Outcomes (explanation)
Secondary Outcomes
Secondary Outcomes (end points)
Secondary Outcomes (explanation)
Experimental Design
Experimental Design
A mapping was created of the 3 counties and locations for central locations were identified such that 157 central locations could be fit into the geographic area and central locations had a catchment radius of 2.5km. Subsequently a census was conducted of all central locations and the communities falling inside the 2.5km catchment radius. A GPS mapping was created, resulting in 150 central locations. One community per central location catchment area was randomly selected, with half of the communities being less than 1.25km from the central location and half of the communities being more than 1.25km away located from the central location. All selected communities are therefore located <2.5 km from their central location and >3.5-4km from all other central locations. Central locations were then randomized into the four different intervention arms outlined in the "Intervention" section.

A census of all adults (18 years and above) was conducted in the selected communities. Based on the census three random samples were drawn: i) baseline respondents, ii) endline respondents and iii) adults eligible for text messaging treatment. Details for the individual level randomization for the text message intervention are outlined in the "Intervention" section.
Experimental Design Details
Not available
Randomization Method
The randomization was conducted using a computer random number generator.
Randomization Unit
Central locations and their selected communities were randomized into either (i) control, (ii) ink, (iii) bracelet or (iv) calendar incentive. Individuals in selected communities were randomized into either receiving (i) reminder, (ii) reminder + deworming take-up information text messages or (iii) no text messages.
Was the treatment clustered?
Yes
Experiment Characteristics
Sample size: planned number of clusters
150 central locations and communities.
Sample size: planned number of observations
9,935 adults of which 2,250 adults are surveyed at baseline, 3,750 adults surveyed at endline and 3,935 adults selected for text messaging intervention.
Sample size (or number of clusters) by treatment arms
37 central locations for control, 37 central locations for ink, 39 central locations for bracelet, 37 central locations for calendar incentive. For 37 control clusters, per cluster 15 adults receive reminder text messages and 15 adults receive reminder + deworming take-up information text messages. For the 37 ink, 39 bracelet and 37 calendar clusters, per cluster 25 adults receive the reminder + deworming take-up information text messages.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
Given an intracluster correlation (for the random effects logit latent model used in power analysis simulations) of 0.1 we will be able to detect a 10 percentage point difference in deworming take-up between any incentive treatment arm and the control group, and a 8 percentage point change between any two incentive treatment arms. We can further detect a 7 percentage point difference in deworming take-up within treatment arms between adults who received and did not receive a text message reminder/information about deworming take-up in their community, and a 9 percentage point change in demand between any two arms for adults that did receive the text messaging treatment.
IRB
INSTITUTIONAL REVIEW BOARDS (IRBs)
IRB Name
Kenya Medical Research Institute
IRB Approval Date
2016-08-15
IRB Approval Number
KEMRI/RES/7/3/1
IRB Name
Committee for Protection of Human Subjects (CPHS) UC Berkeley
IRB Approval Date
2016-08-16
IRB Approval Number
2016-05-8735