I would say that buying in for 25$ is a bad strategy if you are a better post flop player than the average guy at the table.
I am not saying that I'm necessarily a worse post flop player than the people at NL100 (I don't really know since I've never sat at a NL100 table) but I'm advocating this simply because I don't have the BR to buy in for the full $100.

Also, Jeffry, did you find that your ROI increased when playing at a NL100 table versus a NL25 table w/ the same buy in ($20 or $25)? How about variance? Do you experience greater variance using the buy in short strategy?

Let me post some hypothetical situations. Say you have AQ in the CO, so you raise to $4. 2 people call. Flop is A49, 2 suited. They both check to you. How much do you bet here, since any sort of bet will almost commit you to this pot? Do you just push? What if the flop is all blanks? Do you use a continuation bet, keeping in mind that if you are called, you are basically commited to the pot.
Next, say a guy in MP raises to $4, you hold 44 on the button. Do you just call the raise, even though your stack is much less than 10x the raise? Thanks.