The ability to compute accurate free energies by molecular simulation is necessary for many applications. Methods for performing these calculations are well established, but are often misused or misinterpreted. We review ways to avoid and uncover systematic errors in free-energy calculations through simple and appropriate analysis of the simulation averages. One particularly difficult area of application of free-energy methods is in the analysis of solid-phase behaviors. Free energies cannot be computed as easily in solids as in liquids, owing to the inability to perform particle insertions into solids without introducing a defect. We describe our attempts to develop new and efficient methods to circumvent this problem, and show how the analysis of free-energy methods can help to understand the behavior of the newly proposed methods. We examine the method and analysis applied to a simple system of hard rods in one dimension, selected because all of the properties being measured by the simulation method can be evaluated analytically, including finite-size effects.