For the purposes of making recommendations, micronutrient soil or tissue tests are basically classification procedures. The empirical test variable usually groups samples into three areas, i.e. a group containing mostly samples with no micronutrient problems, a second group containing mostly samples with the micronutrient problems, and a third group containing samples both with and without micronutrient problems. The third group described above can be thought of as occurring at a “transition zone” where the correct status of a sample is unidentifiable by the micronutrient test. Statistical methods using an interaction Chi‐Square are presented for identifying the boundaries of the “transition zone.” It is proposed that a micronutrient test be evaluated in terms of the proportion of successful recommendations. A successful recommendation results when the proper status of a sample is identified by the test. Under the assumption of random sampling from a stable population, the above evaluation scheme follows a binomial distribution. The computation of variances and confidence intervals is shown and discussed for the proportion of successful recommendations. A statistically valid method for comparing different empirical tests, such as extracting solutions, is shown and contrasted to previously used methods.