The following survey was undertaken in 2005 to assess the effectiveness of professional training activities in international agricultural research organizations that were undertaken between 1999 and 2002 at ICRAF (International Centre for Research in Agroforestry), now World Agroforestry Centre, Nairobi. Trainees were randomly selected from various professional fields in the agroforestry domain. Survey questionnaires and telephone interviews were adopted. Descriptive statistics, logit regression, cross-tabulations and bi-plot analysis were used to analyze the data. The Kirkpatrick's training evaluation theory provided the methodological framework for the study. Male participants were more educated and had longer average professional experience than females. Trainees had good recollection and learning of various aspects of the training. There was high potential for skill transfer and practical implementation of training skills among all trainees, but lack of resources was a major limitation. Female workers faced more constraints during workplace implementation of skills than male workers. The study has practical implications for current and future design and evaluation of training in agricultural domains. The work also contributes to knowledge building in training evaluation within agricultural institutions in Africa, which is poorly documented or lacking in certain specific settings. The paper is original because scientific evaluation of training activities in agricultural practice in Africa is rarely or poorly documented, thus adding value to agricultural research. (Contains 4 figures and 13 tables.)