This thesis examines why some governments and rebel organizations but not others recruit children to be child soldiers. The theory posits that if a country fights in a civil war of long duration, armed groups are more likely to recruit children as soldiers. I find that the probability of child soldier recruitment increases when a country experiences following conditions: a longer duration of civil war, a large proportion of battle deaths, a large number of refugees, a high infant mortality rate, and the presence of alluvial diamonds. An increase in education expenditures and civil liberties would decrease the probability of child soldier recruitments.