Habit has been modeled in different ways in information systems (IS) research. It is theorized to directly impact system use (SU), moderate the impact of behavioral intention (BI) on SU, indirectly impact SU through BI and other variables, mediate the effects of other variables on BI and SU, and moderate the effects of other variables on BI. Prior studies empirically examined models of habit in various settings such as different types of respondents and geographic regions. Unsurprisingly, empirical findings on the relationships involving habit have been inconsistent and mixed. This study proposes that the variations in empirical results may be due to the various models of habit and the study characteristics. An exploratory meta-analysis and review of habit and its relationships is conducted by synthesizing findings across 130 samples reported in 114 published studies. Implications for research and practice are discussed.