Defined as the ratio of market value to replacement cost of tangible assets, Tobin's q is commonly used as a performance measure in research. Although the use of the q ratio is theoretically appealing, its empirical construction is subject to considerable measurement error. In this thesis, we compare and contrast five estimators of q, ranging from the simple q to the estimator based on the Lindenberg and Ross (LR) (1981) procedure. We examine the means, medians and variances of the q estimates and investigate how robust sorting and regression results are to changes in the construction of q. We find that the empirical results are sensitive to the estimation method used in obtaining the q value. The estimator calculated using the Hall (1990) method produces the highest means and variances and has the lowest correlation with the four other estimators. The simple q and the benchmark q are highly correlated, and results from tests for the equality of their coefficient estimates reveal no significant differences. We infer that the Hall estimates are theoretically superior, and the simple q which provides good estimates of the LR q's are satisfactory given the simplicity of their calculations. Finally, we find that the firm's attributes used in Morck, Shleifer and Vishny (1988) and Chung and Jo (1996) have an influence on the q value estimates.