Accurately estimating reliability of individual components in multi-component systems is challenging when only system-level failure data is observable. This paper develops maximum likelihood techniques to estimate component reliability from right-censored lifetimes and candidate sets indicative of masked failure causes in series systems. A likelihood model accounts for right-censoring and candidate sets. Extensive simulation studies demonstrate accurate and robust performance of the maximum likelihood estimator despite small samples and significant masking and censoring. The bias-corrected accelerated bootstrap provides well-calibrated confidence intervals. The methods expand the capability to quantify latent component properties from limited system reliability data. Key contributions include derivations of likelihood models and validation of estimation techniques via simulations. Together, these advance rigorous component reliability assessment from masked failure data.