That's consistent with Villani 2011's ( https://www.gwern.net/docs/catnip/2011-villani.pdf https://www.gwern.net/docs/catnip/2011-villani.pdf ) finding of consi

That's consistent with Villani 2011's ( https://www.gwern.net/docs/catnip/2011-villani.pdf https://www.gwern.net/docs/catnip/2011-villani.pdf ) finding of considerable measurement error in catnip response: you have to try a cat multiple times to be sure if they are a responder. This, I suspect, may be an issue with a number of catnip studies, since researchers administering catnip may be strangers or not be on good terms with the subject cat, and definitely none of them mention trying to control or correct for this (or any awareness of this beyond Villani). My suggestion is that the response is not totally uncontrollable and so anxiety/uncertainty/stress can 'shut down' the response, kind of like the socially-dependent response to many drugs in humans, like alcohol or MDMA - boozing with your friends at a cheery party with no inhibitions is very different from meeting a stranger & maintaining your vigilance.