One reason to avoid the use of the pooled variance procedure is that:
◻
it is more complicated than the other option. it assumes the populations are exactly Normally distributed.