Varietal choice and adoption of cassava varieties under farmer elicitation and DNA fingerprinting methods : a case of Malawi

Abstract

Most studies on adoption and impact of crop technologies in Sub-Saharan Africa and in particular Malawi, have relied on farmer elicitation, crop experts or morphological descriptors for varietal identification and estimation of adoption status. These methods, as suggested by empirical evidence are error-prone and can lead to varietal misidentification and misclassification. Although the extent and effect of misclassification and misidentification has been established in empirical studies, little is known on how they affect varietal adoption and choice analyses. The current study therefore examined the effect of varietal misidentification and misclassification on cassava varietal adoption and choice. The data used in this study were obtained from Methodological Experiment on Cassava Varietal Identification and Productivity Measurement (CVIP, 2015). Using DNA fingerprinting as a benchmark for varietal identification, results showed that farmers overestimated the use or adoption of improved cassava varieties. Most of the cultivated varieties were either traditional (54.47%) or locally-selected (45.35%) and only few were improved (0.18%), suggesting a great lack of knowledge of cultivated varieties from farmers’ side. Econometric estimations on the other side, through univariate and multivariate probit models showed that determinants of adoption and choice of improved cassava varieties significantly varied when adoption was measured under either farmer elicitation or DNA fingerprinting. Moreover, results showed that farmer elicitation underestimated the extent of the effect of some variables on adoption and choice of cassava varieties. Study findings therefore support the use of DNA fingerprinting for accurate varietal identification in agricultural surveys. Results also suggest that more attention should be re-oriented to extension services for increased training of farmers on cultivated varieties for higher adoption and choice of new cassava varieties. This study therefore recommends use of DNA data in future adoption studies for identification of true adoption drivers, accurate treatment variable and correct policy recommendations.