If one simply compares yields on SRI and conventional rice plots, the difference in our sample is insignificant. However, such a comparison can be misleading due to possible selection bias. Since farmers decided themselves whether or not to adopt the new technology, there may be systematic differences between the characteristics of adopters and non-adopters. Indeed, we find that SRI farmers in Timor Leste are those that operate under more unfavourable conditions. In other words, without SRI, they would have lower yields than other farmers. Controlling for these differences, we find that SRI adoption has a positive net yield effect of 46 per cent. This substantial gain is the average effect for the adopters. For farmers who have not adopted, switching to SRI could increase yields by an estimated eleven per cent. These findings underline that yield impacts are situation-specific, which is in line with earlier studies on SRI in different countries.

The impact of SRI on total household income is an average increase of two per cent for technology adopters.