All extant versions carry the hallmarks of oral tradition, even the 2 broadsides, so we shouldn't expect total fidelity to reality. On the other hand they are sufficiently close despite more than 250 years between the earliest and the latest. Taking into account their closeness in text this tells me that its origins lie not too far previous to the 1759 earliest. However, many c1750 printed ballads were straight copies of c1680 ballads so we can't be too prescriptive.