|From||David Airey <email@example.com>|
|Subject||re: st: Clustered dataset question|
|Date||Fri, 19 Sep 2003 16:19:28 -0500|
For the threshold of interest there is no doubt a population threshold around which there is variation due to individual differences. For traits influenced by multiple genes and environmental sources of variation (like perception), this will always be a reasonable description. Try a nested model of some kind, maybe "nlogit"?This is right except for the bit about "nlogit", which I'm not really sure about. Do the 36 measures represent the same stimuli for each person such that stimuli are crossed with person? Post or email me your solution when you get one. I'd like to know it too.