Increased productivity has been cited as one of the benefits of scientific collaboration. While some studies have identified a positive relationship between collaboration and publication productivity, others have found a negative or no clear relationship between the two. A number of factors determine the relationship established, ranging from the context under which research is carried out that shapes the scientists' collaboration and productivity behaviors and practices, to the methods and measures used. Few studies have examined the factors underlying the said relationship, more so in studies of developing areas. In this article we present the empirical findings of a study seeking to establish the relationship between collaboration and productivity of academic scientists in Kenya, including factors determining the established relationship. The study uses mixed methods research design employing self reported measures of collaboration and productivity. The study establishes a significant, though weak relationship between collaboration and publication productivity for academic scientists in Kenyan Universities. We find that factors affecting level of collaboration, including disciplinary area, academic qualification, and the national and institutional context within which research is done have an effect on productivity of academic scientists. The results presented serve to enhance our understanding of factors determining scientist's productivity, and methods best suited to investigate collaboration and productivity in developing areas.