Department of Agricultural, Food, and Resource Economics Feed the Future Innovation Lab for Food Security Policy

William J. Burke, Stephen Morgan, Thelma Namonje, Milu Muyanga, and Nicole M. Mason, 2019. Beyond the "Inverse Relationship": Area Mismeasurement Affects Actual Productivity, Not Just How We Understand It, Feed the Future Innovation Lab for Food Security Policy Research Paper 159.

Abstract

Although measurement error in agricultural field area and productivity data for developing countries
is widely acknowledged, there is still a shortage of evidence on what the errors imply for researchers,
and even less evidence on what the implications may be for farmers. In this study we investigate
field size measurement errors on Zambian maize fields to examine the nature of these errors and the
implications they have for: 1) our ability to understand productivity, 2) actual productivity, and 3)
our broader understanding of total land use. Using a nationally representative dataset on Zambian
smallholder maize plots, we compare self-reported (SR) and global positioning system (GPS)
measures of land area on a farm’s largest maize plot to assess how measurement error might affect
productivity estimates and farmer input use. Consistent with other studies, we find strong evidence
that the land area of relatively smaller fields is overstated, and relatively larger fields is understated.
However, correcting for this measurement error using GPS measurements appears to strengthen the
evidence of an inverse relationship between field size and productivity. Additionally, we find strong
evidence to suggest farmers themselves believe the area figures they report to enumerators and that
their input use is more closely aligned with the reported field sizes than actual field sizes. Based on
these results and insights from semi-structured interviews with farmers and extension agents, we
argue that measurement error may affect real productivity in addition to productivity estimates.
Strengthening extension efforts to improve farmer understanding of land area measurements may be
an important and affordable way to improve productivity. Moreover, improving the accuracy of data
collection for area seems feasible and will improve researchers’ understanding of productivity.