Seeding Innovation Through Reality TV

We evaluate the national rollout of a “makeover style” reality television program in Uganda that focuses on farming as a business. The television program helps to guide small scale farmers on a range of agricultural topics in an engaging yet informative way. Examples of core topics include soil and water management, planting, and pest and disease management. Each week, the film crew visits a different farming family and encourages them to identify their key goals and challenges in crop and livestock production. The show emphasizes why there is a problem and how best to deal with it. Local experts are brought to the farm to provide customized support to the family and clearly demonstrate simple yet affordable ways in which the household can implement the recommended adjustments on their farm.

The intervention is designed to generate exogenous variation in viewership of the farming TV program among comparable groups of farmers. We encourage a (randomly selected) group of rural farmers to watch the farming show, and another (randomly selected) group to watch a competing show that is broadcast at the same time. We will also shed light on farmers’ learning and adoption processes by studying the role of information provided to urban and rural contacts of control group farmers.

2022-03-22

2023-09-03

The primary outcomes for this study will be measured under five key domains:

0. Take Up
• The extent (0/1) and frequency of viewership of the targeted farming television program

1. Improved Knowledge of Agricultural Practices and Technologies
We will use multiple choice questions (with unambiguously correct answers) to measure the knowledge of plot managers under this domain, including:
• Theoretical knowledge of agricultural practices and technologies
• Knowledge of specific practices and technologies recommended on the TV program

2. Changes in Production Practices and Agricultural Input Use.
We will measure the following choices made by plot managers under this domain:
• The adoption (0/1) of agricultural practices and technologies that are common in the study region
• The adoption (0/1) of agricultural practices and technologies specifically recommended on the TV program
• Expenses incurred (UGX) on agricultural inputs in the reference farming season
• The choice of crops grown, and farm animals reared

3. Farm Productivity, Revenues and Sales to Market.
• Production: Self-reported crop harvests in the reference agricultural season (volume)
• Revenues: Value of crops harvested in the season, inclusive of production costs
• Productivity: Crops harvested per unit of land area (volume and value)
• Sales: Value of crops sold in unprocessed form (self-reported); average weekly income from the sales of processed crops and animal products; and crop sales as percentage of total output (commercialization).

4. The Diffusion of Information and Financial Resources.
• Flows of information on financing for agricultural purposes from (and to) social networks in rural areas and towns/cities.

1. Female Participation in Agricultural Decision-Making.
This will be measured using the identity of decision makers on each plot cultivated by the household. We also measure changes in the decision rights of these plot managers, e.g., to build structures on the plot, rent it out to someone else, plant tree crops, or use it as a loan security.

2. Changes in attitudes related to women’s participation in business.
• We will collect both explicit (self-reported) and implicit measures of attitudes toward women’s participation in business tasks. The implicit measures are based on implicit association tests that will be administered to respondents in our study sample.

3. Changes in income and business aspirations.
• Subjective expectations about revenues from farming in the next agricultural season
• Farming goals (e.g., crop choice, livestock and area cultivated) in 5 years’ time
• Professional or business goals in 5 years’ time

Our project covers 10 districts in the Central Region of Uganda reached by the Luganda language broadcast of the farming TV show. We identify a population of farming households with access to TV and randomly assign them to one of the following experimental arms: (1) Watchers; (2) “non-Watchers” and, (3) Information Only. We use a symmetric encouragement design to minimize the possibility that ‘non-Watchers’ are directly exposed to episodes of the farming TV program. That is, households in group (1) are encouraged to watch the farming TV program, while households in group (2) are encouraged to watch a competing show that is broadcast on national television at the same time. The households in group (3) receive weekly farming tips that convey the same information as the farming show but through a different medium: mobile phones (with no entertainment or visual demonstration of steps for farmers to emulate).

Not available

Randomization done in office by a computer using the software Stata.

The unit of randomization is the household.

No

N/A

3400 households

Targets/Estimates only:
Watchers: 1000
Non-Watchers: 2000
Information Only: 400