Im a firm believer in differing synth route = different feeling off the MDMA that is produced. And I also remember around the summer of 97/98 there was a change in pills. Less empathy, less loved up. I have occasionally had the odd pill or MDMA in the last few years that gets me in the same state but the majority of MDMA i have these days does not have the same quality.
There's a study of
MDMA methods on Erowid. Three methods were investigated; the Merck method involving safrole / methylamine produced no impurities. The other two produced significant impurities.
This I assume is in lab conditions. Illicit drug manufacturing is not exactly known for good lab conditions, cleaning up their products or having the most controlled reactions.
The "best" synthesis method for consumers, therefore, would be the one that easiest forms a pure product.
It's possible, I think, that impurities in less-than-perfect synthesis methods contribute to changes in "feeling" between now and then. Some of the impurities in the referenced study are known to be active. But if the impurities are relatively small, I wouldn't think it would be *that* different.
Personally, I assume that the biggest change in MDMA between then and now is: A) Pill strength, as I have heard 1990s MDMA was very strong, and B) the huge amount of of "fake MDMA" going around, as well as perhaps some flat out poor quality synths that go beyond small impurities.
Methylone, the most numerous fake MDMA product of the moment, for instance goes quite well with beer, whereas on MDMA and MDA (and, so not to not bash all RCs, 6-APB for sure) such a thing is unthinkable. (For the people wondering why people might drink on "MDMA" now.

)
There are some things popping up on ecstasydata that we have no idea what they do (eg, "MDMA Methylene homolog", what the fuck?). As ecstasydata is a simple library match type GC/MS I doubt they will catch every impurity in the world, and even with what they got there's some nice ones. I liked the pill that was mostly MDP2P (very bad synthesis obviously), but at least that's known. The library match in one MDMA pill for a large amount of 2,2,3,4-tetramethyl-5-(3,4-methylenedioxy)phenyloxazolidine (what the hell is that?) along with MDBZ was even more bizarre. The atomoxetine sold as MDMA made me WTF.