The step transaction doctrine is one of the judicial doctrines used by the courts to interpret tax law. The doctrine requires that a series of transactions be treated as a single transaction if the transactions share a single, integrated purpose. Many authors believe there is a great deal of uncertainty as to when the doctrine will be applied. Uncertainty and inconsistency in the application of tax law add to the complexity of the law. One of the most complex areas of tax law is Subchapter C of the Internal Revenue Code, which governs corporate formations, redemptions, liquidations, distributions, and reorganizations. The purpose of this study was to determine if the step transaction doctrine is being consistently applied by the Tax Court and what variables affect the judges' decision in these cases. Hierarchical logit analysis was used to derive a full model and two restricted models. The full model was used to determine the predictive power of the variables that were identified and to explain the extent to which the individual variables affect the judges' decisions. One restricted model was used to test temporal stability. The other was used to test consistency when different issues of tax law are involved. The data source was the sample of step transaction cases involving Subchapter C issues decided by the Tax Court and its predecessor, the Board of Tax Appeals. Eight variables were identified to evaluate the factors discussed in the literature and major court cases involving the doctrine. Four of the variables were found to be statistically significant. The full model correctly predicted the outcome of 79.5 percent of the cases. The restricted model to test temporal stability correctly predicted the outcome of 86.3 percent of the cases. The restricted model to test the consistency of the decisions relating to a specific topic correctly predicted the outcome of 85.7 percent of the cases.