ITIKI performance evaluation : bridging indigenous knowledge and science drought prediction tool in Kenya

Abstract

This evaluation assesses the effects of the ITIKI drought prediction tool on farming practices and outcomes

in Embu County, Kenya. ITIKI combines indigenous weather indicators with meteorological data to generate

localized predictions on the timing of rain onset and approximate volume of rainfall for each growing season

(i.e. ?less than normal? rainfall, ?normal? rainfall) and sends this information to farmers via text message. The

evaluator surveyed 62 registered ITIKI users, who were selected through a cluster randomized sampling

design, regarding their experience using this innovation.

User perceptions of the innovation were mostly positive. Despite drought in recent seasons, users felt

ITIKI intelligence had improved their crop yields compared to what yields would have been without the

information. The reported effects occurred through a change in farming practices in which, based on the

weather predictions, users planted early and changed to more drought-resistant crops and varieties.

However, approximately 20 percent of respondents were unfamiliar with the innovation despite being

registered users, which highlights the need to address issues with ITIKI?s recruitment and follow-up.

Quantitatively measurable effects are sparse because ITIKI has operated for a limited time in Kenya and

highly variable rainfall makes it impossible to attribute changes in yields or income across seasons to the

innovation. Nonetheless, positive user perceptions of ITIKI and reported changes in farming practices as a

result of ITIKI intelligence suggest the innovation is helping farmers cope with drought.