Such information makes best sense in context. For example, if we learn that 40% of criminals were abused in childhood, we will interpret it certain way. But if

Such information makes best sense in context. For example, if we learn that 40% of criminals were abused in childhood, we will interpret it certain way. But if we afterwards learn that 40% of non-criminals were also abused in childhood, now the previous interpretation does not make sense anymore. So we should start from the point "how many kids you know who were abused in childhood" and then ask "and how many of them later became criminals".