>> having 5 possible levels of the outcome variable (starved) will fragment the data (observational bee data already tend to be fragmented and noisy). Apart form the fact that beek will not be able to recognise these things. Keep it as simple as possible. I agree! And thats the problem, the average beekeeper will probably miss diagnose more times than they get it right. I won't argue the worth of a detailed or simple survey, -I'll let that rest. But I am curious what you would need to include in a survey in order to keep it simple and yet reveal something of value? Something that someone can look back 10 years from now and glean valued data? Best Wishes, Joe Waggle https://www.facebook.com/Historical.Honeybee.Articles *********************************************** The BEE-L mailing list is powered by L-Soft's renowned LISTSERV(R) list management software. For more information, go to: http://www.lsoft.com/LISTSERV-powered.html