Datasource: just how good are statistics?
Ok... we've heard the AHA PtB quote the statistics on the state of our 'industry', that the average age of our horses is 16, aging, demonstrating that we're in trouble, yadda yadda.
Well... I did some digging, just for fun. Yup, the Registry/PAT records registrations, and transfers (presuming they're submitted)... but guess what? There appears to be NO interest in encouraging owners to record deaths! No surprise... but the significance is that AHA has no business gleaning statistics from the database and trying to apply that to the real world. The numbers don't add up!
For example, approximately 80% of the horses foaled in 1971 are STILL LISTED AS ALIVE. Yup, it appears that we have +/-7500 34 year old purebreds out there! Yeah, right! Frankly, I suspect people have their heads in the sand (appropriate for arabian horses, eh?) regarding the impact of the industry melt-down when the tax laws changed and the bubble burst. Like the folks buying horses by the pound went to the trouble to transfer registrations on horses intended for the meat hook?
so: I've been thinking about taking on a little test case of the datasource/database situation.
The numbers are pretty ludicrous, but I took one arbitrary but manageable slice: 1971 black (and the horses/people may not be representative) and thought I might put together an inquiry to snail mail to the recorded owners. I'd be really interested to hear 1) if the letter even gets delivered to a real person, or comes back unknown 2) if the recorded owner
bothers to respond and 3) the status of the horse in question.
According to datasource, there were 88 blacks foaled in 1971. One (remember the Estabrook snafu) was registered and the papers cancelled by the Registry. So of the other 87, there were 56 mares, 6 geldings, and 25 stallions. Of those, 13 mares, 1 gelding, and 10 stallions have been reported as dead. Do the numbers! Through age 15 (or 1986), 9 of the 13
mares had been reported, the one reported gelding, and 8 of the 10 stallions. Post '86? The last mare was reported in '96, the last stallion in '94. That still leaves 44 mares, 5 geldings, and 15 stallions... yeah, right!
So... the datasource problem in miniature. Worth a little time and some postage.... I'll be interested to look at the results! Oh, and yes, I'll share them! I know this isn't a 'valid' statistical sample, but I still think it would be meaningful and probably fairly representative.
Can you use some examples how this information is used by AHA and how the errors may impact these things? Just to help us all understand better? Thanks
'boo
MB II Arabians
Colorado
American Foundation Bred
CMK/Fadjur/Azraff-Ferzon
Re: Datasource: just how good are statistics?
Ticks me off... I've been chatting with JM about this datasource problem, and she sort of understands why it concerns me so much. Look at the rationale behind the
published reasons for the sweeps overhaul! Look at the discussions regarding our breed ranking in the horse industry, or in promotional efforts, or quoting our numbers in efforts to win sponsorship, or.... Folks want to make our community look bigger, or more influential, and they're using dead horses (and dead owners, for heaven's sake) to do it.
We're using shitty numbers, and at least some of those people MUST know it! It's false and misleading! It's unethical! It's bloody stupid!
Re: Datasource: just how good are statistics?
I couldn't agree more. And on the forums how they go on about how allbreed isn't accurate...at least it's free, and yes one needs to police their entries and I've even deleted incorrect info, but it's free, and as accurate as datasource that costs 90 a year!!! Plus, being a pedigree fan on all breed pictures of the horses are available, strains, percentages, it's a way to put together as much info on any horse as one would like. That's what datasource should be pushing for. The most accurate, the most educational, the most amount of information on ANY horse, with regards to and you want to talk flagging, how about flagging horses that could be dead.
How about if a horse is adopted or bought and not transferred for whatever reason, if the date seems feasible and a person contacts AHA about a horse they had that died, then datasource should take their word for it if they have the number and have a panel of questions to ask and then if satisfied mark that horse as deceased.
Or they could have the horse flagged as bought by a new owner but records not transferred due to inability of being able to get signature of old owner, records lost, whatever, but that would go a long way to reuniting some horses with their records, rescue horses, etc. They should do all they can for the horse. Not just take your 90 bucks a year for inaccurate information. The PAT has ONE job, to register horses, and maintain the horses' information, one job, they should do it right!!!
'boo
MB II Arabians
Colorado
American Foundation Bred
CMK/Fadjur/Azraff-Ferzon
Email This To a Friend
Topic Commands
Click to receive email notification of replies 
Click to stop receiving email notification of replies