Using behavioural interventions to promote the adoption of improved sweetpotato varieties in Uganda

Study area
The study will be implemented in Amuria district, one of the leading sweetpotato producing districts in Uganda, and will cover the entire district. This district has not benefitted from significant promotion of ISVs (especially the biofortified varieties). It is located in Teso sub-region of Eastern region of Uganda and is comprised of 15 sub-counties, 91 parishes, and 539 villages. Each village has approximately 100 households. Nearly all rural households in this district grow sweetpotato.

The technology: improved varieties and vine sources
The study will use 4 varieties of sweetpotato namely, Ejumula, Tanzania (Osukut), NAROSPOT 1 (New Dimbuka) and NASPOT 13 (Joweria). Unlike Ejumula and Osukut, New Ndimbuka and Joweria are new introductions in the study district. They will also be sourced from sources inspected and approved as SPVD-free. The characteristics of the varieties are described in Figure 3.

Figure 3: Varieties of sweetpotato to be used in the study
Quality-certified seeds/vines of these improved sweetpotato varieties (ISVs) will be used in the study, sourced from trained community or group vine multipliers. However, there are currently no established vine multiplication enterprises in Amuria district. Thus, all the vines of ISVs supplied to the study villages will be sourced from registered government-inspected multipliers in nearby districts such as Kamuli, Bukedea and Serere. Figure 3 provides characteristics of the promoted ISVs. Ejumula and Tanzania are both widely grown varieties in the Amuria district, hence are not new to sweetpotato farmers in the district. However, due to repeated planting, the vines of these varieties have accumulated a heavy load of the sweetpotato virus diseases (SPVD) that greatly reduce their yielding ability. It is estimated that SPVD can cause up to 98% yield losses in sweetpotato (Gasura and Mukasa, 2010). Hence, there is a need to obtain vines for use in the study from sources that will be inspected and approved to be free of these viruses. The other advantage of obtaining vines used in the study from outside the district is that it enables the research team to maintain close control of movement of vines and hence avoid contamination of quality vines with those infected with viruses and pests. Besides the government inspection, samples of vines from the vine multipliers will also be collected by the research team and tested/inspected for SPVD and weevil infection before being distributed to the study villages.
Baseline Survey and Randomization
In the first part of the study, a baseline survey will be implemented targeting all 1,200 farmers recruited into the behavioural nudges study. Specifically, the survey will target a sample the 10 farmers who have grown sweetpotato in the past one year in each of the 120 villages. Data will be collected on purchase (yes/no) of quality seed of ISV, quantity purchased, number of times quality seed of ISV were purchased, farmers’ socio-demographics including farmer data, sweetpotato production practices including acreage and varieties grown, varietal, trait preferences, nutrition and food security indicators, risk, and vine sourcing behaviours before the randomization and intervention phases of the study. Data will also be collected on the sales transactions at the salespoint. Further, data on the characteristics of the salesperson will also be collected. The last two data sets will generate variables that can be used to control for confounders. Following the baseline survey, the villages will be randomized into the treatment and control arms.
The baseline will be followed by randomised assignment of the 120 villages that have been participating in on-going nudge study into treatment and control. The process of assignment is described in Figure 4. Randomization will be done on R (a computer program) using the village GPS data and will consider the distance between pairs of the villages. An average radius of 4Kms will be maintained between the villages in the control group and those in the treatment to reduce the likelihood of control villages being contaminated with treatment information.



Figure 4: Sweetpotato RCT design


As shown in Figure 4, the 64 nudge villages will be cross-randomly assigned into a treatment group that receives trial pack of sweetpotato seed (i.e., vines) and to the control group (no vines). The cross-randomization will take into account distances between villages to avoid a treatment village being very close to control – hence reducing the risk of contamination. This will result 31 villages receiving the treatment (trial pack) while 33 don’t, hence serve as control. Similarly, the 56 villages that did not receive nudge interventions (i.e., non-nudge villages) will be cross-randomized into treatment and control, again considering distances, to yield 28 treatment (trial pack recipients) and 28 non-treatment (no trial packs).
For the second part of the study involving consumer testing, through cross-randomization, and as shown in the last boxes of Figure 4, 15 of the 31 nudge + trial pack villages and 14 of the non nudge + trial pack villages will participate in the consumer intervention (as a treatment). On the other hand 17 nudge only and 14 no nudge + no trial pack will not participate in the consumer intervention. The last group thus serving as pure control group. The trial pack comprised of 25 pieces (each 30-centimers long) of each of improved sweetpotato varieties. Th cooking intervention entail cooking and tasting the roots of improved sweetpotato varieties.

2023-07-03

2024-12-15

Adoption and extent of adoption respectively measured by dummy (1,0) for adoption, and the quantity of vines of improved sweetpotato varieties purchased (a continuous variable) for extent of adoption.

Household food security, food consumption, and dietary diversity

The 64 nudge villages were initially randomized into 64 treatment and 56 control villages. In each village a salespoint was set up at central location and supplied with quality vines of ISVs at the beginning of each rain season. These 64 nudge villages will be cross-randomly assigned into a treatment group that receives trial pack of sweetpotato seed (i.e., vines) and to the control group (no vines). The cross-randomization will take into account distances between villages to avoid a treatment village being very close to control – hence reducing the risk of contamination. This will result 31 villages receiving the treatment (trial pack) while 33 don’t, hence serve as control. Similarly, the 56 villages that did not receive nudge interventions (i.e., non-nudge villages) will be cross-randomized into treatment and control, again considering distances, to yield 28 treatment (trial pack recipients) and 28 non-treatment (no trial packs).

For the part of the study involving consumer cooking and tasting, through cross-randomization, 15 of the 31 nudge + trial pack villages and 14 of the non nudge + trial pack villages will participate in the consumer intervention (as a treatment). On the other hand 17 nudge only and 14 no nudge + no trial pack will not participate in the consumer intervention. The last group thus serving as pure control group. The trial pack comprised of 25 pieces (each 30-centimers long) of each of improved sweetpotato varieties. Th cooking intervention entail cooking and tasting the roots of improved sweetpotato varieties

Study area
The study will be implemented in Amuria district, one of the leading sweetpotato producing districts in Uganda, and will cover the entire district. This district has not benefitted from significant promotion of ISVs (especially the biofortified varieties). It is located in Teso sub-region of Eastern region of Uganda and is comprised of 15 sub-counties, 91 parishes, and 539 villages. Each village has approximately 100 households. Nearly all rural households in this district grow sweetpotato.

The technology: improved varieties and vine sources
The study will use 4 varieties of sweetpotato namely, Ejumula, Tanzania (Osukut), NAROSPOT 1 (New Dimbuka) and NASPOT 13 (Joweria). Unlike Ejumula and Osukut, New Ndimbuka and Joweria are new introductions in the study district. They will also be sourced from sources inspected and approved as SPVD-free. The characteristics of the varieties are described in Figure 3.

<<Figure 3 here>>

Quality-certified seeds/vines of these improved sweetpotato varieties (ISVs) will be used in the study, sourced from trained community or group vine multipliers. However, there are currently no established vine multiplication enterprises in Amuria district. Thus, all the vines of ISVs supplied to the study villages will be sourced from registered government-inspected multipliers in nearby districts such as Kamuli, Bukedea and Serere. Figure 3 provides characteristics of the promoted ISVs. Ejumula and Tanzania are both widely grown varieties in the Amuria district, hence are not new to sweetpotato farmers in the district. However, due to repeated planting, the vines of these varieties have accumulated a heavy load of the sweetpotato virus diseases (SPVD) that greatly reduce their yielding ability. It is estimated that SPVD can cause up to 98% yield losses in sweetpotato (Gasura and Mukasa, 2010). Hence, there is a need to obtain vines for use in the study from sources that will be inspected and approved to be free of these viruses. The other advantage of obtaining vines used in the study from outside the district is that it enables the research team to maintain close control of movement of vines and hence avoid contamination of quality vines with those infected with viruses and pests. Besides the government inspection, samples of vines from the vine multipliers will also be collected by the research team and tested/inspected for SPVD and weevil infection before being distributed to the study villages.
Baseline Survey and Randomization
In the first part of the study, a baseline survey will be implemented targeting all 1,200 farmers recruited into the behavioural nudges study. Specifically, the survey will target a sample the 10 farmers who have grown sweetpotato in the past one year in each of the 120 villages. Data will be collected on purchase (yes/no) of quality seed of ISV, quantity purchased, number of times quality seed of ISV were purchased, farmers’ socio-demographics including farmer data, sweetpotato production practices including acreage and varieties grown, varietal, trait preferences, nutrition and food security indicators, risk, and vine sourcing behaviours before the randomization and intervention phases of the study. Data will also be collected on the sales transactions at the salespoint. Further, data on the characteristics of the salesperson will also be collected. The last two data sets will generate variables that can be used to control for confounders. Following the baseline survey, the villages will be randomized into the treatment and control arms.
The baseline will be followed by randomised assignment of the 120 villages that have been participating in on-going nudge study into treatment and control. The process of assignment is described in Figure 4. Randomization will be done on R (a computer program) using the village GPS data and will consider the distance between pairs of the villages. An average radius of 4Kms will be maintained between the villages in the control group and those in the treatment to reduce the likelihood of control villages being contaminated with treatment information.



<< Figure 4 here>>

As shown in Figure 4, the 64 nudge villages will be cross-randomly assigned into a treatment group that receives trial pack of sweetpotato seed (i.e., vines) and to the control group (no vines). The cross-randomization will take into account distances between villages to avoid a treatment village being very close to control – hence reducing the risk of contamination. This will result 31 villages receiving the treatment (trial pack) while 33 don’t, hence serve as control. Similarly, the 56 villages that did not receive nudge interventions (i.e., non-nudge villages) will be cross-randomized into treatment and control, again considering distances, to yield 28 treatment (trial pack recipients) and 28 non-treatment (no trial packs).
For the second part of the study involving consumer cooking and tasting, through cross-randomization, and as shown in the last boxes of Figure 4, 15 of the 31 nudge + trial pack villages and 14 of the non nudge + trial pack villages will participate in the consumer intervention (as a treatment). On the other hand 17 nudge only and 14 no nudge + no trial pack will not participate in the consumer intervention. The last group thus serving as pure control group. The trial pack comprised of 25 pieces (each 30-centimers long) of each of improved sweetpotato varieties. Th cooking intervention entail cooking and tasting the roots of improved sweetpotato varieties.

Deployment of the interventions and follow-up surveys
Following the randomization and in the first study intervention, nudges will be deployed for 3 seasons, with salespoints constructed in each of the 120 study villages prior to the onset of the first season rains. Each village will have one salespoint that will be hosted by one salesperson recruited within the village from among progressive farmers in that village. The salesperson will be recruited with the help of local administration and frontline extension workers. The criteria that will be used for recruitment and the process of onboarding of salespersons are in Annex 1 & 2. Salespoints will be constructed in locations within the village that are easy to access and easy to see.

Following the construction of salespoints, the salespersons will be supplied with quality vines of the four sweetpotato varieties (namely, Tanzania, Ejumula, NASPOT 13 and NAROSPOT 1 – see details below) at the onset of the rains. The vines will be delivered to the salespoints in two rounds over a period of 1 month in season 1. In the first round, each salesperson will receive two bags of each variety, hence a total of 8 bags. Two weeks later, they will receive a second round of delivery comprising one bag per variety, hence a total of 4 bags per village. Thus, during the first season, each salesperson will receive 12 bags of quality vines of ISVs. The same process will be followed in subsequent seasons. The bags will be displayed for sale on stand at the salespoints. For the treatment group, in addition to receiving the bags of vines, the salespoints will have a poster mounted in a prominent easy-to-see place on the sales stands. These interventions will be repeated during the second and third sweetpotato growing seasons.
For the treatment group, text message reminders will be sent to all the farmers who purchased vines in the first season. The messages will be in the local vernacular language and will be sent to the phone owned by the purchaser, their family member or their nearest neighbor. Each season every vine purchaser will receive 7 messages per week for 2 weeks, with one message going out each day during the first week. The same messages will then be repeated (one per day) during the second week. Hence, a previous season purchaser will get the same message twice over the duration of two weeks. The messages will be customized to communicate the benefits of the ISVs over the local variety to the recipient. Midline and endline surveys will be conducted to assess the effect of the nudges on i) the decision to purchase, ii) vine purchases and iii) repurchase of vines. The 1200 farmers interviewed during the baseline will be revisited and resurveyed during the midline and endline.

R simR package in computer

administrative village

Yes

120

960 households

59 treatment villages receive trial packs, 61 villages control

Power calculations for the initial nudge RCT were conducted using the R simR package. The outcome variable of interest in all calculations is the adoption of improved sweetpotato varieties (a dichotomous variable). The other outcome of interest in this study was the extent of adoption measured by the quantity of vines of ISVs purchased, which is a continuous variable. In most behavioural studies, dichotomous variables generally require larger samples to detect effects than continuous variables, hence we focused only on the binary adoption variable. We used a base rate of adoption of 9%, which implies a standard deviation of approximately 0.29. Estimates were based upon a mixed logit analysis, with households nested in villages. We assumed a correlation of adoption within a village of 0.10. An economically significant result requires an increase in adoption of approximately 20 to 25% points. We also assumed a control and one treatment condition, targeting rejection of the null that the treatment is equivalent to the control. We found that with n=1000 (100 villages, 10 households per village) we would detect an increase in adoption of 20% points with 97% probability and 15% points with 80% probability. In order to cater for potential drop-offs and missouts due to unavailability, we targeted 120 villages (clusters) with about 60 villages implementing the full program, and the balance serving as a control. We surveyed 10 randomly selected households per village to obtain more detailed information on adoption behaviour, including the use of second or third-generation seeds, demographic information, and exposure to the messaging/interventions. This allowed us to estimate the broader impact of the program, and also to examine the populations most affected. For the trialpack intervention, power analysis allows us to detect, reasonably well, the effect of trialpack with a random sample of 8 households per village. With this sample of 8 households per village in the 120 villages, and maintaining nudge treatment and control and the other parameters above, we will be able to detect a 50% effect with probability 0.73. Going with 8 households per village/cluste results in a total sample size of 960 households, for the second part of this study. As before, the outcome variable of interest is the adoption of ISVs.