Background: Failure rates after arthroscopic shoulder stabilization are highly variable in the current orthopaedic literature. Predictive factors for risk of failure have been studied to improve patient selection, refine surgical techniques, and define the role of bony procedures. However, significant heterogeneity in the analysis and controlling of risk factors makes evidence-based management decisions challenging. Purpose: The goals of this systematic review were (1) to critically assess the consistency of reported risk factors for recurrent instability after arthroscopic Bankart repair, (2) to identify the existing studies with the most comprehensive inclusion of confounding factors in their analyses, and (3) to give recommendations for which factors should be reported consistently in future clinical studies. Study Design: Systematic review; Level of evidence, 4. Methods: A systematic review of the literature was performed in accordance with the PRISMA guidelines. An initial search yielded 1754 titles, from which 56 full-text articles were screened for inclusion. A total of 29 full-text articles met the following inclusion criteria: (1) clinical studies regarding recurrent anterior shoulder instability; (2) surgical procedures performed including arthroscopic anterior labral repair; (3) reported clinical outcome data including failure rate; and (4) assessment of risk factors for surgical failure. Further subanalyses were performed for 15 studies that included a multivariate analysis, 17 studies that included glenoid bone loss, and 8 studies that analyzed the Instability Severity Index Score. Results: After full-text review, 12 of the most commonly studied risk factors were identified and included in this review. The risk factors that were most consistently significant in multivariate analyses were off-track lesions (100%), glenoid bone loss (78%), Instability Severity Index Score (75%), level of sports participation (67%), number of anchors (67%), and younger age (63%). In studies of bone loss, statistical significance was more likely to be found using advanced imaging, with critical bone loss thresholds of 10% to 15%. Several studies found predictive thresholds of 2 to 4 for Instability Severity Index Score by receiver operating characteristic or multivariate analysis. Conclusion: Studies reporting risk factors for failure of arthroscopic Bankart repair often fail to control for known confounding variables. The factors with the most common statistical significance among 15 multivariate analyses are off-track lesions, glenoid bone loss, Instability Severity Index Score, level of sports participation, number of anchors, and younger age. Studies found significance more commonly with advanced imaging measurements or arthroscopic assessment of glenoid bone loss and with lower thresholds for the Instability Severity Index Score (2-4). Future studies should attempt to control for all relevant factors, use advanced imaging for glenoid bone loss measurements, and consider a lower predictive threshold for the Instability Severity Index Score.