Data envelopment analysis (DEA) is a relative measurement technique used to evaluate the efficiencies of a homogeneous group of samples with multiple inputs and/or outputs. DEA can be highly effective when right variables are chosen. The objective of this study is to identify the most appropriate variables for DEA to evaluate stock performance and find the efficient ones from a pool of stocks. Evaluation of stocks are carried out either by assessing their financial strength or by assessing their past price behaviour in the secondary market or both. In any case, it is imperative to use suitable variables to evaluate the performance of stocks. For this purpose, three different combinations of variables were tested on 69 non-financial stocks listed in the National Stock Exchange (NSE), which were selected based on their market capitalization. The results obtained suggest that all the three sets of variables taken for the study help in the identification of efficient stocks. The average returns of the stocks selected in all the three cases are higher than the market return. Among the three sets, stocks identified using the past price behaviour give a higher return when compared to the other two sets. The study can help academicians and investors to percolate efficient stocks from a large pool of stocks. The selected stocks can be further analysed to construct an effective portfolio.