Just read the abstracts listed, and they are predominantly unconfirming of NLP hypotheses. Those few that have significant results, are often not in the direction predicted, or they are only one result out of several which means they could be a result of comprimised confidence levels (if you test lots of things, you increase the chance that you will get one positive result by chance rather than due to a real effect). Not an impressive research database for an NLP website, which might have been expected to know about any positive research results.
Anyway I'll shut up about it now.
|