Comparative analysis of eye estimation method and GPS technique of crop area estimation. a case of area under maize in Mubende District

Abstract

This study aimed at comparatively analyzing the two methods of crop area estimation; the eye estimation method which is a subjective technique for collecting current agriculture statistics and the more objective method of GPS measurement. This was motivated by the fact that crop area is one of the most important variables in estimating crop production, and while numerous methods have been developed on how to estimate crop area, both objective and subjective methods, the collection of data on crop area is still a problem. The subjective methods do possess the challenge of bias, while the objective methods are either expensive to employ or require technical knowledge and specialized equipment to use. The study aimed at establishing specifically the equality of the means from these two methods, the homogeneity of variance for the two approaches, an analysis of systematic measurement bias, and then fitting a regression model to predict GPS given the eye estimates by agriculture extension staff. In order to achieve the study objective, a sample of 209 households was selected from maize farming households in Mubende District using a three-stage stratified simple random sampling method. Agriculture extension staff were used to estimate area covered by maize on nineteen plots each in their areas of jurisdiction and a total of eleven extension staff participated in the study and those same plots were also measured using a GPS device. This gave us a paired sample or repeated measurement on the same plot, i.e, the same plot was first estimated by an agriculture extension staff using the eyes estimation technique, then the same plot was measured by the researcher using a GPS. The study findings revealed that agricultural extension workers can accurately estimate the crop area subjectively as close as using GPS with a minimal error of 1%. The mean and variance difference tests were done to establish whether measurement by GPS differed from eye estimation by agriculture extension workers; however, the results were not statistically significant. However, it was also revealed that the existing difference between GPS and eye estimation is more pronounced in small plots than in large plots. As plot size increases, the two measurements tend to converge due to the decreasing mean difference according to plot size. Crop areas that were 5 acres and above reported a mean difference of only 4.4% compared to areas of 0.5–1.0 acres, which had a mean difference of 22%. Therefore, I recommend the adoption of the agricultural extension workers as key crop area data collectors in the sector of agriculture with the use of cheap method of eye estimation to determine. The National statistical system can therefore develop or adopt a national standardized units and use them as statistically true values. This will save the country from expensive methods and the unavailability of the current agriculture statistics, especially on crop areas, which is a very important variable for production estimation and policy formulation.