Evaluation of data capture methods used in plant health clinics in Kenya

Plant clinics, which are context and situation specific compared to the existing extension approach in the Ministry of Agriculture (MoA), can deliver advice to farmers at the point of demand to reduce crop losses to pests. They can also address the existing disadvantages associated with manual data capture methods in the extension service. The study had three objectives: comparing the average number of errors made by plant doctors using the two hand-filled clinic register forms, to determine the factors affecting the number of errors made by plant doctors in giving crop diagnostics and to determine the differences between machine capture (OCR/ICR) and manual capture methods interm of error rate and error type made. The primary data employed was from register forms filled in through purposive sampled plant doctors who met the set criteria and are based in central and eastern provinces. The data was subjected to a paired t-test, ANOVA and independent t-test through gap computation. The study showed significant differences between plant doctors in their diagnosis of the plant health problems (p=0.002) and recommendations to farmers (p=0.040). The OCR/ICR method had a higher mean score in giving crop diagnostics (11.25) and for correctly capturing data (21.67) than manual method (3.57 and 12.38). Significant variation (p<0.001, F=39.74) was observed among the two methods of data capture. The study has shown that OCR/ICR method has increased the speed and reliability of producing plant pest data in an accurate, timely and cost effective manner. The study recommends advanced training for plant doctors to improve diagnosis and recommendation, training on data capture methods to accurately capture data.