Understanding livestock management decisions

The treatment group will receive a bookkeeping intervention where they will be incentivized to keep records.

After the baseline survey and randomization are complete, farmers in the treatment group will
receive training on how to keep the books. They will be provided paper and pencil templates to record animal health and productivity outcomes (for both cattle and chicken) for three months between the baseline and endline survey. Specifically, the treatment farmers will be incentivized to keep detailed records on cattle and chicken and disease symptoms and death. They will also keep detailed accounts of East Coast Fever and Newcastle Disease prevention measures like application of acaricides and management of grazing lands, and on ECF and ND treatment measures like antibiotic use. Each of the two animal caretakers (cattle and chicken) in the treatment household will receive a small twice-monthly incentive of 100 KSH (approximately a half day’s wage, paid via Airtime every two weeks) to keep their respective records. Farmers in the treatment group will receive a weekly reminder via SMS to complete their records. Additionally, an enumerator will check on the books at one to three random points in time. All households will receive at least one visit, and a small subset of the treatment group will receive two or three visits. As long as their record book is up to date during the random visits, participants will continue to receive the incentive. Up to date is defined as having record book entries filled out until the past week. Enumerators will take a photo of the book pages at every random visit for verification and tracking purposes; participants will keep books in their possession. If the book is not up to date during the visit, enumerators will return the following week. If the book is up to date at that time, participants will resume receiving payments.

2022-07-01

2022-11-01

Willingness-to-pay (WTP) for Newcastle Disease vaccine (in control and treatment groups) elicited using the Becker-deGroot-Marschak mechanism, an incentivized measure (i.e., a measure with real stakes); take-up of acaricides at market prices (in control and treatment groups).

Willingness to pay: The dollar amount the chicken caretaker states is the highest amount (maximum price) they would pay for different quantities of ND vaccine. The quantities will be based on average flock size in the baseline survey and discussions with the veterinary directorate. Because we will elicit participants’ willingness to pay for multiple quantities, we will select one of these quantities at random for implementation.
Take-up of acaricides: Whether the cattle caretaker chooses to purchase acaricides at market prices from the veterinary directorate staff member at endline.

Livestock management practices, expenditures on preventions/cures/vet services, reported livestock outcomes (deaths, illness), reported livestock productivity (e.g., milk and egg production), reported income from livestock

Livestock management practices:
- Time spent on manual tick removal in the last week
- Whether cattle grazed at communal grazing lands last week

Expenditures on preventions/cures/vet services:
- Total expenditures on chicken and cattle, respectively, in the past month
- Expenditures on chicken vaccines in the past month
- Expenditures on acaricides in the past month
- Expenditures on cattle vaccines in the past month

Reported livestock outcomes:
- Reported number of chicken and cattle deaths in the past three months, respectively
- Reported number of sick chicken and cattle in the past three months, respectively

Reported livestock productivity:
- Reported average weekly egg production over the past three months
- Reported average weekly milk production over the past three months

Reported income from livestock:
- Reported total income from chicken and cattle, respectively, over the past three months

The study will be conducted with chicken and cattle-owning households in Busia county. Households are eligible if they have at least one chicken and at least one milking cow, or expect to have a calf born within the next three months. We will collect data from two individuals per household: the primary caretaker of chickens and the primary caretaker of cattle. Thus, we will sample 600 individuals in 300 households. Prior to randomization, we will administer a baseline survey (visit 1). The treatment group will receive a bookkeeping intervention. After three months, we will elicit willingness-to-pay (WTP) for ND vaccine in control and treatment groups using an incentivized measure (i.e., a measure with real stakes) and administer an end-line survey (visit 3). In addition, we will study take-up of acaricides at market prices.

Eligibility criteria

Inclusion Criteria: All households owning chickens and at least one milking cow, or expects to have a calf born within the next three months, where the primary care-takers of the chickens and cattle are sufficiently literate to keep simple written records in English and at least one household member has a mobile phone..

Exclusion Criteria: Households without chickens or eligible cattle, or households where the primary caretakers of the chickens or cattle are unable to keep written records. Households who are currently eligible for free or subsidized veterinary care through their active participation in an existing program, since this may alter willingness to pay for vaccination and acaricides. Households where no household member has a mobile phone will be excluded, since the protocol includes SMS reminders. Individuals under age 18 will be excluded from serving as respondents.


Recruitment, baseline survey and randomization (visit 1).
We will recruit eligible households into the survey by going door-to-door in target villages. Once a household is recruited into the study, we will conduct a baseline survey with the primary decision-maker for chicken and with the primary decision-maker for eligible cattle. The baseline survey will collect detailed socio-demographic information from the households.
This survey will also collect information on livestock and poultry management practices.
Furthermore, we will measure each respondent’s memory capacity at baseline using a digit-span task.

Randomization will be conducted at the household level. Treatment status will be assigned by a random number generator in SurveyCTO. Enumerators will remain blinded to the treatment status until after they administer both baseline surveys. Treatment status will be revealed once both questionnaires are complete.

Intervention (visit 1, treatment group). After the baseline survey and randomization are
complete, farmers in the treatment group will receive training on how to keep the books and
report the data. They will be provided a paper and pencil template to record cattle and chicken
health and productivity outcomes for three months between the baseline and endline survey.
Specifically, the treatment farmers will be incentivized to keep detailed records on symptoms of
ECF and deaths in cattle as well as symptoms of ND and deaths in chickens. They will also keep
detailed accounts of ECF and ND prevention and treatment measures like application of
acaricides, deworming, and antibiotic use. Treatment group farmers will receive a small financial
incentive to keep these records. Each of the two primary caretakers in the treatment household will receive 100 KSH in Airtime every two weeks if they keep records for their animals (cattle for cattle caretakers and chicken for chicken caretakers).

Random checks on treatment group (visit 2, treatment group). Participants in the treatment
group will receive a weekly reminder via SMS to complete their records. Additionally, an
enumerator will check on the books at one to three random points in time. All households will
receive at least one visit, and a small subset of the treatment group will receive two or three visits.
As long as the books are up to date during the random visits, participants will continue to receive
the incentive. Up to date is defined as having records entered within the past week. Enumerators
will take a photo of the book pages at every random visit for verification and tracking purposes;
participants will keep books in their possession. If the book is not up to date during the visit, enumerators will return the following week. If the book is up to date at that time, participants will resume receiving payments.

Willingness-to-pay. At the end of the three-month period in which treatment farmers engage in
record-keeping while control farmers do not, we will elicit from the primary chicken decision-maker
willingness-to-pay (WTP) for ND vaccination for different quantities of doses.
WTP is defined as the maximum price a farmer would be willing to pay for a product such as acaricide or ND vaccine. We will implement the Becker-DeGroot-Marschak (BDM)
mechanism for estimating the WTP for ND vaccine. In the BDM mechanism, the subject first
states her WTP for each quantity of ND vaccine. We then randomly select one of these quantities to offer a randomized subsidy to the subject. For that quantity, we randomly draw a price from within a reasonable range (not exceeding the commercial price of ND vaccine). If the randomly drawn price is lower than the stated WTP, the subject must purchase the vaccine at this random price. On the other hand, if the randomly drawn price exceeds the stated WTP, the subject is not allowed to purchase the vaccine at that price. Enumerators will be accompanied by veterinary directorate staff who will bring the vaccines with them and provide them directly to households choosing to purchase the vaccines. To study demand for acaricides, we will offer the cattle caretaker the opportunity to purchase bottles of acaricides at the market price. Enumerators will bring the acaricide bottles with them and provide them directly to households choosing to purchase the acaricides.

Randomization will be conducted at the household level. Treatment status will be assigned by a random number generator in SurveyCTO. Enumerators will remain blinded to the treatment status until after they administer both baseline surveys. Treatment status will be revealed once both questionnaires are complete.

Randomization is done at the household level.

No

0

300 chicken caretakers, 300 cattle caretakers

Households will be randomized after completion of their baseline survey by a random number generator in surveyCTO. The expected treatment share is 50% control and 50% treatment.

Each of the main analyses will be conducted on a sample of 300 individuals. This sample size allows us to detect an effect size of 0.32 standard deviations of willingness to pay, at 80% power and a 5% significance level.