Methodology: A multiple regression analysis was employed to assess the combined effect of the knowledge bases, factors such as age of the household head, land size, marketing channels and geophysical factors (such as land slope) on linseed productivity.

Findings: The findings reveal no differences in productivity between trained and untrained farmers. They further show that training and the interaction between training and experience positively influence productivity. Our findings, however, reveal farmer education to have an inverse yet insignificant effect on productivity. Furthermore, we found that factors such as the slope of the land and the choice of marketing channel also play an important role in influencing productivity.

Practical implications: Moving away from the traditional top-down approach could be the answer to avoid a mismatch between the information given by trainers and what the farmers actually need.

Theoretical implications: The study used interaction variables of knowledge bases, and attempted to demonstrate the importance of a tacit knowledge base, which is often not well documented and therefore neglected in research.

Originality/value: The research demonstrates the combined effects that interactions between several knowledge bases have on the productivity of linseed farmers.