The authors introduce and describe a new variable for interstate conflict research for use in conjunction with the Militarized Interstate Disputes (MID) dataset. This variable, First Use of Violent Force (FUVF), covers the period 1980-2001 and identifies which state involved in a violent militarized interstate dispute was the first to actually use violence rather than threatening or displaying force. This article introduces the variable, which, along with all of the supporting documentation, is now publicly available, describes its creation, discusses its utility, and uses both multivariate regression and measures of association to draw attention to theoretically interesting patterns in first use of violent force that are worthy of further exploration. The authors find that non-democratic states are more likely than democracies to be the first to resort to violent force when involved in a militarized interstate dispute; also, minor powers are more likely to use violent force first, compared with major powers. Measures of association also indicate that states that use violent force first in a MID are much more likely to be revisionist states, but that first users of violent force are unlikely to achieve victory in the dispute as a result.