As the world embarks in a major effort to achieve the Sustainable Development Goals, there is a
high demand for high quality and timely agricultural statistics in order to design and implement
effective policies, allocate investments, monitor and evaluate progress. Most of the Governments in
developing countries are thus searching for cost-effective methods of collecting high quality data on
agriculture. Literature from small surveys conducted by research institutes suggest that using
Computer Assisted Personal Interview (CAPI) technologies could be a cost effective alternative to
the traditional Paper and Pen Interview (PAPI) method. Additionally, there is preliminary evidence
that CAPI can improve data quality, shorten interview duration, and that enumerators prefer CAPI
to PAPI. As National Statistical Systems (NSSs) operate in a very different context from research
institutes and the applicability of these findings are limited. This paper analyses paradata on cost,
interview duration, data quality, and enumerator feedback on various surveys conducted by
National Statistical Systems (NSSs) in Uganda and Tanzania. The results confirm that CAPI is costeffective
across multiple surveys. Statisticians in Uganda also indicated there were an increase in
data quality after using CAPI. Additionally, data from a questionnaire administered to enumerators
in both countries showed that they indeed prefer CAPI over PAPI. The analysis of interview
duration did not indicate that CAPI interviews are shorter, but this is likely due to factors other than
the survey instrument.