Reproducibility package for In Tablets We Trust? Validation Of The Use Of Tablet-Based Tools For Land Area Measurement In Household Surveys

Accurate and scalable methods for land area measurement are essential for estimating key agricultural indicators, such as crop yields, as well as understanding agricultural relationships. A significant body of literature has demonstrated systematic bias in farmer estimates of land area, the most inexpensive option, and at the same time provided evidence supporting the use of handheld GPS units for objective measurement. Despite the successful adoption of handheld GPS measurement in many survey operations, new tablet-based tools offer potential for smoother, more cost-effective measurement of land area. Using data collected in a methodological experiment in Uganda, this study evaluates the accuracy and operational trade-offs of tablet-based area measurement tools, both with and without the support of GPS-boosting devices, relative to handheld Garmin GPS units. Findings demonstrate that both tablet-based approaches are highly correlated with handheld GPS measurements, with relatively small but statistically significant differences in means. On average, the boosted and unboosted tablet measures exhibit an absolute percentage error of approximately 11 and 12 percent, respectively, with underestimation more pervasive than overestimation. Importantly, findings highlight that even small differences in land area measurement can lead to substantial variations in yield and production estimates, underscoring the critical role of measurement method for agricultural productivity analysis and policy. From an implementation perspective, the tablet-based approach offers clear advantages, with a significantly reduced burden of data cleaning and processing compared to handheld GPS data and enumerators facing few technical challenges.