Radical instability and spurious branch support by likelihood when applied to matrices with non-random distributions of missing data

► Non-random missing data, without rate heterogeneity, can cause misleading results. ► Non-random missing data, w/o informative characters, can cause misleading results. ► Artifacts were found to occur frequently using 22 empirical examples. ► Partitioning based on missing data helps, but does not eliminate, artifacts. ► Artifacts are exacerbated by low quality tree searches.