Since many engineering projects in rock never mobilize strengths near the uniaxial compressive strength (UCS) of the rock, elastic modulus becomes a critical parameter to describe the rock's behavior under loading. There are a number of methods available for calculating the elastic modulus from laboratory test data, and each method gives a slightly different value. The objective of this study is to evaluate the most repeatable method for each of a number of rock types, and then to develop guidelines to aid the practitioner in selecting the best method as a function of rock behavior. UCS tests were performed on 78 samples of nine rock types, including two basalts, two granites, two limestones, a quartzite, a sandstone, and a gypsum. Elastic moduli were calculated using six different methods reported in the literature or modified for this study. The modified secant and modified secant-at-50-percent-strength moduli (modified by shifting the origin to best intercept the extension of the main straight-line portion of the stress-strain curve) were the most repeatable methods for rocks with elastic and plastic-elastic behavior. Elastic-plastic materials, which have a broad concave-downward stress-strain curve, are best evaluated using the tangent modulus on the upper of two distinct straight-line segments. For materials which show creep or extended plastic deformation with no sharp failure, the secant-at-50-percent-strength modulus and modified secant-at-50-percent-strength modulus are the most repeatable.