According to earlier research farmer's crop orientation in developing countries mainly depends on farm size, large-scale farmers prefer cash crop while small-scale farmers prefer subsistence crops. The first aim of this study is to see if this hypothesis can be applied on six households in Babati District in rural Tanzania. The second aim is to investigate if other factors than farm size affect crop portfolio choice and the final aim is to see if those crop portfolio models can be improved. A case-study research design and qualitative interviews are used. The primary data is based on a fieldwork that took place from the 18th of February until the 7th of March 2009 in the study area.

From a theoretical perspective the underlying assumptions of the Marcel Fafchamp's model Crop portfolio choice under multivariate risks is discussed in connection to the result of the study.

Interviews were made with six households of different farm size. The result of the study indicates that both small-scale and large-scale farmers are using cash crops. The fact that all crops can be used for selling, gives also small-scale farmers in season with higher prices, an opportunity to sell a large share of their crops. It's thereby not possible to state that large-scale farmers devote a larger share of their land for cash crop than small-scale farmers do.