EMPLOYING BINARY LOGISTIC REGRESSION ANALYSIS TO MODEL THE EFFECT OF DIFFERENT TREATMENTS ON QUALITY OF FRENCH BEANS (PHASEOLUS VULGARIS) PRODUCE

Abstract

French beans (Phaseolus vulgaris L.) is a significant agricultural product in developing countries, contributing to economic development, employment, and exports. Despite extensive research on French bean farming in Kenya, limited attention has been given to factors influencing produce quality. This study employed a binary logistic regression model to evaluate the effect of slope, soil deposition, and different mulching materials: Farmers’ Practice (no mulch), polymer mulch, tea leaves mulch, and grass mulch on French bean quality in Ol’ Lessos Ward, Nandi Hills Sub-County, Kenya. Quality was defined as the absence of disease incidence. Field experiments measured slope percentage, soil deposition (g), yield (kg/plant), and disease incidence across treatments. Results revealed that slope alone had no significant effect on quality, but mulching significantly reduced soil deposition, increased yield, and lowered disease incidence compared to Farmers’ Practice. Polymer mulch recorded the highest yields and lowest soil deposits, while tea leaves and grass mulches also improved performance. Logistic regression analysis yielded a Cox and Snell’s R2 was 0.14 and Nagelkerke’s R2was 0.63 and the Nagelkerke R2 which adjusts Cox and Snell’s value to a scale ranging from 0 to 1 was 0.628, suggested that the model had a strong explanatory power and a good overall fit. The findings underscore the critical role of mulching materials in enhancing French bean quality, independent of slope effects. Adoption of mulching, particularly polymer mulch, is recommended for sustainable production and improved marketability.