The primary objective of this dissertation was to identify and estimate the relative importance of factors used by the Tax Court in deciding hobby loss cases. This was accomplished in two steps. The first step was to review the Treasury Regulations, cases and literature pertaining to hobby losses to determine the relevant factors used by the Court in deciding this issue. The second step involved using probit analysis to identify which of these factors actually influenced the Court in deciding hobby loss cases and to determine the relative importance of the factors. A secondary purpose was to to explore the probit model's ability to predict decisions likely to be appealed. The probit model was based on the analysis of 219 post-1954 Tax Court cases involving determination of whether activities were or were not engaged in for profit. By application of log-likelihood techniques, it was determined that the model developed was stable over time and across lines of "business," the implication being that cases decided both before and after the passage of Section 183 and involving all types of activity should have precedential value in conflicts between taxpayers and the Internal Revenue Service. Five factors were found to be significant predictors: manner of operation, level of expertise, time expended, history of income and loss, and presence of elements of personal pleasure. It is important to note that the two factors not susceptible to tax planning--success in other activities and financial status of the taxpayer--were insignificant discriminators between business and hobby outcomes. The implication is that with careful tax planning, one can organize and operate a given activity so that it is likely to receive favorable tax treatment. The probit model proved to be unable to predict decisions likely to be appealed. The probability that a particular case would be classified by the Court as a business did not appear to be a significant predictor of whether a taxpayer would appeal an adverse decision.