ICT-mediated agricultural knowledge transfer in Uganda: What works?

We will use video to convey information to farmers with the aim of changing adoption behaviour. We will focus on providing information on seed selection, soil nutrient management (including promoting organic fertilizer application), weeding, timely planting and plant spacing. From our conversations with experts, we learn that many farmers may already be aware of the existence and use of these technologies or practices. Therefore, simply providing information about the existence of modern technologies and recommended practices and on how to uses them may not be sufficient to change behavior. Often, access and afford ability was mentioned as a problem. In our videos, we will thus also try to alter the belief that seeds and fertilizers are “too expensive” by pointing out the costs and benefits of the different technologies and practices we promote. In addition, we will encourage farmers to start small, using fertilizer and seeds on a small area of their field to experiment and see for themselves, and reinvest in subsequent years. Inter-temporal decision making, where costs today have to be compared to uncertain outcomes in a distant future, is often challenging for poorly educated farmers. Furthermore, Duflo et al. (2011) point out farmers may have difficulties committing to fertilizer use in Kenya. In our video, we attempt to make farmers aware of this, and suggest some techniques to overcome this bias. We also pay considerable attention to the way the message is delivered. For instance, the message if brought by “peer farmers”, as it is found that farmers find communicators who face agricultural conditions and constrains most comparable to themselves to be the most persuasive (BenYishay et al., 2014). The information is also presented as a success story, which is assumed to affect a range of non-cognitive farmer characteristics such as aspirations, locus of control and self-esteem (Bernard et al., 2015).

The video starts with a farmer (a male farmer, a female farmer, or a couple, see below) introducing themselves. He talks about how he used to struggle with his maize gardens and how at one point in time, he decided things needed to change. It is shown how the farmer sells a hen, and obtains a small loan from a friend. This money is than used to buy small quantities of improved seed and fertilizer in a local shop. It is then shown that, before planting the improved seed, the farmer prepares the garden. He is shown collecting manure wherever he can find it, and applies it to a small corner of the field of 20m by 20m. Next, it is shown in detail how the maize seeds are spaced 75cm x 30cm with 1 plant per hill and how the DAP should be applied. The viewer is reminded to plant in time. The next scene depicts the field after about 10 to 12 days when the maize has emerged from the ground. At this stage, it is recommended that the farmer engages into gap filling to replace seeds that did not germinate with new seeds to preserve optimal plant density. The next shot shows the field at 18 to 20 days after planing, when first weeding is done. Particular attention is paid to identification of striga in an early stage. It is also advised to weed again two to three weeks later. The next scene zooms in on urea fertilizer application. Here, the field is shown at about 4 weeks after planting when the maize is knee high. It is shown how Urea topsoil dressing should be applied. Finally, it is recommended to do one more round of weeding around the tasseling stage of the maize.

We then spend some time comparing the costs to the benefits the different improved inputs and recommended practices. For fertilizer and improved seed, the costs for one acre is calculated and compared to the value of what is harvested. The profit is then compared to the value of what would have been harvested on that acre without improved seed and fertilizer. This would be less than half of what the profit would be when improved inputs are used. For recommended practices, we report what the expected yield increase would be if, for instance, recommended spacing was used or weeding is done in time.

Next, we try to promote a long run perspective, where the farmer is encouraged to start small (one tenth of an acre) and grow bigger over time. We also pay some attention to the commitment problem. We advise the farmer to, at time of harvest when the farmer sells most of his maize, immediately go to the farm supply store and purchase seeds and fertilizer, and store this in a safe place. If inputs are not available, farmers are encouraged to keep the money needed to buy the inputs in a separate, labeled container. It is sometimes argued that such mental accounting can be an effective commitment mechanism (Dupas et al., 2013). In a last scene, the farmer recapitulates and once more directly addresses the viewer an encourages him or her to try this as well.

A total of three such videos will be produced. They will be identical in terms of the information that will be in the video, but they will differ according to the gender of the messenger, corresponding to one of the factors in the factorial experimental setup. In one version of the video, the actor will be a man. In a second version of the video, the actor will be female. In a third version, the video will feature a couple as the messenger. Here, both male and female will feature in the different scenes, and the person talking to the viewer will be altered between the man and the women. Finally, we will use a placebo video for the control. This will be a neutral video on a non-related topic, such as the tourist potential of the region. Farmers will be shown any of these four videos according to random treatment allocation. Videos will be shown twice to each farmer in the sample, once before planing (July) and once immediately after planting (August).

2017-07-06

2017-09-08

knowledge related to what is in the videos, behavioural change of things that are featured in the video, such as spacing, soil nutrient management, maize yields, consumption expenditure and poverty. All of these will be measured at the individual level (male/female), women's empowerment in agriculture index.

This research aims to answer the questions related to information asymmetries within the household, homophilly effects in learning and the projection of farming as a household business. investigate the relative importance of (i) the gender composition of the messenger, and (ii) the gender composition of the audience for effective agricultural extension information messages to encourage sustainable crop intensification in smallholder household farms and for improving gender equity in household farming. In particular, we will compare outcomes at the household level between:

• households that are shown a video with agricultural extension information and households that are shown a random video (placebo);

• households that are shown a video with agricultural extension information where the animator is an individual “peer” farmer (ie. a man or a women) and households that were shown a video providing the same information by a couple of peer farmers (ie. man and woman who are shown to participate as equals in the family farm and deliver the message as a couple);

• households in which one adult individual (husband or wife) is shown a video with agricultural extension information and households in which the couple (husband and wife) is shown a video with the same information;

• households where the gender composition of (the individual) messenger(s) and (individual) audience is matched and households where the gender composition of the messenger and audience differ.

To assess the effect the gender composition of messenger and receiver on equity within the household, we will compare outcomes within the household between:

• households where the video was shown to a man versus households where the video was shown to the husband

• households that were shown the video that promotes a household approach featuring a couple versus households that were shown a video where a single farmer provides the messages

• households that are both shown the video that also projects a household approach to all the other households

For this research question, we will be particularly interested in how these interventions reverberate through intra-household decision-making and the allocation of time and resources to agriculture between the different individuals within the farm household. The outcomes of interest will therefore be disaggregated by gender and age, and interpreted in light of the interplay between efficiency at the household farm level, equity within the farming household and women’s empowerment.

The tools and technologies through which the information is transferred are also likely to influence effectiveness. There are many different ways of delivering agricultural extension messages through ICT. Broadly, one can differentiate between two different approaches. In one approach, extension resembles the traditional teacher-pupil model, where the farmer is assumed to absorb knowledge from experts. Showing videos to farmers or pushing information over a mobile phone would fall into this category. A second approach relies more on a consultative model, where the farmer is assumed to know his or her information needs are and requests this information from a service provider. Call centers and Interactive Voice Response (IVR) technology are examples of this second approach. In a third research question, we will compare the effectiveness in terms of changing knowledge, practices and outcome such as yields and poverty of video messages to more or less demand based options such as IVR. In particular, we will compare videos to videos augmented with IVR and videos augmented with IVR where the recipient is also provided with reminders of the existence of the IVR service.

The questions will be answered through field experiments, where farmers are randomly assigned to a group that receives particular interventions (de Janvry et al., 2017). An identification strategy that is based on randomization allows us to quantify the causal linkage between an intervention and the outcomes. In particular, we establish the causal link between extension videos and the knowledge gained, between extension videos and yield changes, and between extension videos and poverty (measured by income).

The research design itself will take the form of a mixed level factorial design. To answer the first research question, we define two different factors, each with three levels. The first factor relates to the messenger and has three levels (male, female and male+female). Similarly, the second factor relates to the recipient of the message and also has these three levels (male, female and male+female). The second research question, where video is now augmented with a demand-driven technology such as IVR, corresponds to adding an extra factor with two levels (no IVR, IVR). In practice, we therefore add the IVR treatment to half of the participants who get to see a video. Finally, we will add a pure control group to the design, such that we do not only investigate relative effectiveness of the different attributes, but also relative to someone who did not get to see a video at all.

The interventions will be targeted at the individual (household) level as opposed to group screenings. Videos will be shown in the house or in the field if necessary. This option guarantees consistency in the treatment and requires the least amount of observations. The videos will be developed in house by the research team and produced by a professional media production company such as NOTV. For IVR, we will partner with Human Network International. Human Network International’s 3-2-1develops content for farmers on a variety of crops and agricultural practices, and make the service available to farmers for free or at low cost.

computer at the village headquarter.

farm level

No

this is going to be an experiment at the individual farmer level

3600 farmers

Control=257, MM=385, MF=385, MB=369, FM=385, FF=385, FB=369, BM=342, BF=342, BB=369. In each treatment cell except control, half will receive the the treatment augmented with IVR.

power calculations are based on simulations using data on maize yields from the UNHS 2005/06. As effect size is based on maize yield (kg/acre) Control = 0, MM= +10 percent, MF= +5%, MB= +15%, FM=+1.5%,FF=+7.5%, FB=+12.5%, BM=+15%, BF=+12.5%, BB=+20%

pre-analysis plan

Agricultural advisory services generally rely on interpersonal knowledge transfers by agricultural extension agents who visit farmers to provide information. This approach is not always effective and has proved hard to scale sustainably, particularly in highly dispersed smallholder farming systems. Information and communication technologies (ICTs) have been advanced as a promising way to overcome many of the problems associated with conventional agricultural extension. We evaluate the effectiveness of an ICT‐mediated approach to deliver agricultural information in a field experiment conducted among small‐scale maize farmers in eastern Uganda. Three complementary technologies designed to address both informational and behavioral constraints to technical change are considered. First, we investigate the effectiveness of audiovisual messages (video) as a means of delivering information on input use and improved maize management practices to farmers. Second, we quantify the additional impact of complementing video with an interactive voice response (IVR) service. Third, we estimate the incremental effect of time‐sensitive short message services (SMS) messages designed to remind farmers about applying key practices at specific points during the season. We find that households that were shown a short video on how to become better maize farmers were performing significantly better on a knowledge test, more likely to apply recommended practices, and more likely to use fertilizer than households that did not view the video. These same households also reported maize yields about 10.5% higher than those that did not view the video. We find little evidence of an incremental effect of the IVR service or SMS reminders.

Van Campenhout, B., Spielman, D.J. and Lecoutere, E. (2020), Information and Communication Technologies to Provide Agricultural Advice to Smallholder Farmers: Experimental Evidence from Uganda. Amer. J. Agr. Econ.. doi:10.1002/ajae.12089

https://onlinelibrary.wiley.com/doi/full/10.1002/ajae.12089

To gain a better understanding of intrahousehold bargaining processes, surveys increasingly collect data from co-heads individually, especially on decision-making, asset ownership and labour contributions. However, answers provided by co-heads to the same set of questions often differ substantially. Recent research suggests that while some of this disagreement is due to random measurement error and cognitive bias, part also reflects non-overlapping information sets. We document differences in answers between male and female co-heads in monogamous smallholder maize-farming households in Uganda. We first confirm that not all disagreement can be explained by measurement error or bias. Using a field experiment, we then test if disagreement is due to information asymmetry between male and female co-heads. We also test an alternative explanation where discord is attributed to co-heads’ tendency to respond in line with prevailing gender norms and social customs. While the interventions did seem to reduce discord in survey response about decision-making, we do not find that information asymmetry nor reporting in line with gender norms and customs are the primary drivers of disagreement.

Bjorn Van Campenhout, Els Lecoutere & David J. Spielman (2022) Hiding or Pleasing: Spousal Disagreement Among Ugandan Maize Farmers, The Journal of Development Studies, DOI: 10.1080/00220388.2022.2113065

https://www.tandfonline.com/doi/full/10.1080/00220388.2022.2113065

Agricultural advisory services are generally biased towards men, with information targeted mainly to male members within the household, and in formats that often reinforce male dominance in agricultural decision-making. Such biases affect women’s ability to make informed decisions and limit their intra-household bargaining power. Because women’s empowerment in agriculture has many well-established benefits, designing inclusive agricultural extension and advisory services is important. In this study, we challenge the assumption that information is fully shared between co-heads of a household. We also test if portraying women as equally able farmers challenges gender norms and stereotypes in agriculture. We do this through a field experiment in eastern Uganda in which videos that provide information on recommended maize-farming practices are shown to monogamous maize-farming households. In the experiment, we manipulate who within the household is exposed to the information contained in the video. Furthermore, we vary the gender of the person delivering the information in the video. We find that targeting the female co-head alone with information increases her knowledge about recommended practices, her role in agricultural decision-making, her subsequent adoption of recommended practices and inputs, and yields on fields she manages, while the male co-head’s knowledge about the practices and his unilateral decision-making is reduced. When both co-heads are targeted, joint adoption of recommended practices and inputs increases, while the male co-head’s unilateral decision-making is reduced. We find some support that featuring female role models in the videos challenges men’s beliefs and stereotypes about women’s roles in agriculture, and encourages adoption of recommended practices by women. We conclude that if the aim is to empower women, most gains can be made by re-designing advisory services to target information exclusively to the female co-head within the household. Challenging gender stereotypes may create room for increasing women’s involvement in agriculture.

Lecoutere, E., Spielman, D.J. and Van Campenhout, B., 2023. Empowering women through targeting information or role models: Evidence from an experiment in agricultural extension in Uganda. World Development, 167, p.106240.

https://www.sciencedirect.com/science/article/pii/S0305750X2300058X