There is significant interest in quantifying force production inside cells, but since conditions in vivo are less well controlled than those in vitro, in vivo measurements are challenging. In particular, the in vivo environment may vary locally as far as its optical properties, and the organelles manipulated by the optical trap frequently vary in size and shape. Several methods have been proposed to overcome these difficulties. We evaluate the relative merits of these methods and directly compare two of them, a refractive index matching method, and a light-momentum-change method. Since in vivo forces are frequently relatively high (e.g., can exceed 15 pN for lipid droplets), a high-power laser is employed. We discover that this high-powered trap induces local temperature changes, and we develop an approach to compensate for uncertainties in the magnitude of applied force due to such temperature variations.