Bone quality is one of the reported factors influencing the success of bone anchors in arthroscopic repairs of torn rotator cuffs at the shoulder. This work was aimed at developing refined numerical methods to investigate how bone quality can influence the fixation stiffness of bone anchors. To do that bone biopsies were scanned at 26-μm resolution with a high-resolution microcomputer tomography (micro-CT) scanner and their images were processed for virtual implantation of a typical design of bone anchor. These were converted to microfinite element (μFE) and homogenized classical FE models, and analyses were performed to simulate pulling on the bone anchor with and without cement augmentation. Quantification of structural stiffness for each implanted specimen was then computed, as well as stress distributions within the bone structures, and related to the bone volume fraction of the specimens. Results show that the classical method is excellently correlated to structural predictions of the more refined μFE method, despite the qualitative differences in local stresses in the bone surrounding the implant. Predictions from additional loading cases suggest that structural fixation stiffness in various directions is related to apparent bone density of the surrounding bone. Augmentation of anchoring with bone cement stiffens the fixation and alters these relations. This work showed the usability of homogenized FE (hFE) in the evaluation of bone anchor fixation and will be used to develop new methodologies for virtual investigations leading to optimized repairs of rotator cuff and glenoid Bankart lesions.