Lyshack said:
Ahem, missing records is a lot bigger problem for statistical analysis than random missing data within the database. It's a lot better to have a vibrant, growing database that everyone is happy conrtibuting to, than have a dying database that fewer and fewer people are interested in. Having 100,000 partial records is better than having 10,000 "perfect" records.
Yes, I agree! Missing records are bad, especially when they are not random, and keeping the registry thriving is a critical goal!!
(Quick statistical point: MNAR - missing not at random - within a dataset is also terrible, both your point about missing records (selection bias) and missing data points (MNAR) can lead to flaws in the analysis such that the findings can actually be *wrong*. We're better off not collecting data points than collecting them badly, so I would argue for 100,000 smaller records that are mostly complete than 100,000 larger records with lots of partial data, and I agree that we want the 100,000 not the 10,000.)
If we want to ask a specific question of the data, then maybe the solution is not to collect such data through the registry, but instead do it as part of a survey outside of the registry process. That way, people can participate or not as they choose. But that's starting to get researchy....and not very trowel-friendly....would have to be a winter exercise for sure!
I do think there may be consensus (on this thread at least) that an open-ended optional breeder's notes section might be a nice addition to the registry. I don't think that would discourage too many people, and has potential to be a helpful feature.
I would also argue to create hybridizer profiles within the data system (a one-time, free activity, and then have hyperlinks from hybridizer names on the registration display to this profile) where some useful data like zone of hybridization could be collected, since I think the question of what hybrid survives where seems to be a perennial (pun intended) theme on this forum. Hybridizers could also be incentivized to fill out this profile if there was a place for website address (another recent forum topic:
The thread "Forum Member websites featuring their daylilies" in
Daylilies forum). Since I hear all serious hybridizers eventually move to Florida, we'd have to be able to log updates to the zone...
Guess I'll never be a serious hybridizer: it's too warm there for me!
