First of all, a word on hypotheses when using a t-statistics: Brooke uses the assumptions that
- we have a random sample of quantitative data
- The data is sampled from a normal population or $n>100$.
She and I just had a discussion about that. She just wants you
to have a magic number, and 100 is probably overkill. But, as
you already know, the difference between a $Z$ and a $t$ at
$n=100$ is miniscule. But there is a slight change in t-crit
numbers even if $n=36$, say -- whereas by 100, there isn't
any. So that's why Brooke chooses to focus on that number.
But the upshot is that if you're ever asked (in IMath, for
example), she's looking for 100!