Jump to:

  1. Introduction
  2. Data and documentation
  3. Publications



This survey experiment studies the accuracy of farm labour data in household surveys. We tested four alternative survey designs across 854 households from 18 communities in the Mara Region of Tanzania during the main 2014 agricultural season (roughly January to June). In each villages sampled households were randomly assigned to one of the following labour data collection designs:

The four designs were:

  • Weekly face-to-face surveys for the duration of the season (about 26 visits).
  • Weekly phone surveys for the duration of the season (about 26 phone calls)
  • Two different flavours of recall modules, in which a single face-to-face survey is conducted at the end of the agricultural season.

The business-as-usual scenario is to do recall questionnaires at the end of season, which the experiment aimed to contrast to visiting households weekly or calling them weekly.

The high frequency weekly data were collected at high levels of granularity. Farm labour was collected for each individual, for each day, for each plot. Data on other labour was also collected.

Data collection was implemented by EDI.


Data and Documentation

The data and documentation are hosted by the World Bank

Go to download site



The data have been used in the following publications. Pease e-mail me if you identify any gaps.


Arthi Vellore, Kathleen Beegle, Joachim De Weerdt and Amparo Palacios-Lopez. 2018. Not your average job: measuring farm labor in Tanzania. Journal of Development Economics 130: 160–172. download.


Back to bio and publications