A study of alley-cropping data from Northern Brazil. II. A comparison of methods to estimate sample size

Data from a Brazilian Inga/maize/bean alleycrop experiment are used to illustrate patterns of variation between and within crop rows and their possible effects in the determination of a sampling strategy. Three methods for the determination of sample sizes in agroforestry experiments are compared: the combined plant technique, the formula of Snedecor and Cochran and variance component estimation. Results show that variability in the edge rows was generally greater than variability in the inner rows and suggest three sampling strata – one for each of the two edge rows and the third for the inner rows. For all three sampling methods a sample size greater than ten plants per stratum per block gave little extra precision. Increasing the number of blocks improves precision and may permit a smaller sample size per row to achieve the same precision. The comparison of different methods of sample size determination shows that variance component estimation is the most flexible and efficient approach.