[nagdu] How long is "successful"

Mon Jun 10 13:41:15 UTC 2013

Robert's solution is excellent!  And so elegantly simple!
I don't think one can assume that, because school X puts out 500 dogs and
school Y puts out 1000, that Y is better.  Y may have more experience, but
X may give students more individual attention, and thus have a higher
percentage of long-working teams.
I don't know that's true, but I'd love to see the data and compare.
I vaguely recall seeing in TSE's Annual Report how many teams were
graduated, and maybe how many were active.  Am I imagining that, or might
annual reports be a place to look for data?
Tracy

> Robert,
>
> You are my hero!
>
> everything you said is absolutely spot on.  Statistics aren't ever going
> to
> tell you the whole story, they aren't meant to, but they do a very good
> job
> of giving a snapshot of the situation.  I'd want to know the stats on
> success rates for the same reason I check the weather before I head out.
> I
> want to know what I am likely to encounter and  what is possible I might
> encounter. Maybe that means I change my plans and maybe that means I grab
> a
> jacket, but maybe it just confirms what I already thought.
>
> Julie
>
>
> -----Original Message-----
> From: Robert Hooper
> Sent: Friday, June 07, 2013 2:33 PM
> To: NAGDU Mailing List, the National Association of Guide Dog Users
> Subject: Re: [nagdu] How long is "successful"
>
> Hello All:
>
> I've had a few thoughts on this subject, and have arrived at a few
> different
> possibilities.
> The first is to ask why such a statistic exists--ultimately, one would
> answer, to measure the success of the guide dog program in general.
> Therefore, to measure general success, I would suggest looking at (1) how
> many active teams a program has and (2) how many teams a program turns out
> a
> year. This takes the question of success on a personal level and makes it
> irrelevant. For example, if guide dog school x has 500 active teams and
> guide dog school y has 1000 active teams, then there are some hypotheses
> you
> can draw with these numbers. Here are a few:
> 1. Guide dog school y is more efficient than guide dog school x with
> regards
> to resource use, class size, dog placement, etc.
> 2. Guide dog school y has a better marketing strategy, as it has seemed to
> reach more prospective clients than guide dog school x.
> 3. The amount of teams put out by guide dog school y puts them further
> ahead
> with regards to experience (and thus, one would hope, competence).
>
> One can continue to make such hypotheses--which reminds me--these are not
> factual statements one can make about these schools. As anybody using
> statistics properly can tell you, statistics are tools that can be
> misappropriated and poorly used. Thus, it is important to call the above
> statements hypotheses--for example, when reviewing the schools, make these
> hypotheses, and, to the best of your ability, test them scientifically.
>
> I have another possibility, and this one is more of a statistical one. If
> you are wanting a time-based method of success, do the following.
> Take a representative sample of data points (N) where each data point
> represents the duration of a guide dog team. This can be measured by
> taking
> the length of time from dog acquisition to team termination (due to
> death--of either party--retirement, etc.). If N is large enough, and
> assuming the sample comes from a statistically normal population, you
> should
> get a bell curve representing a normal distribution. This can help you
> figure out a lot of things. I would take the mean of team duration and
> call
> that "successful," just to have a realistic expectation of guide dog team
> duration. If you convert your sample to z-scores, then where z=-1.0 and
> z=1.0 is where you will find approximately 2/3 of the population.
> If one must have a time-based method of determining success, then I
> believe
> that this would be a more accurate means of obtaining such, rather than
> just
> assigning an arbitrary number of years. Of course, this is purely a
> measure
> of how frequently schools put out long-lasting partnerships. Whether that
> constitutes personal success is a different issue entirely. I think that
> the
> word success, as originally defined in this thread, is too broad and
> encompasses too many factors. Instead, get more specific--for example,
> call
> it "successfully long-lasting partnerships." If that's what you want to
> measure, then nobody can argue that those statistics don't measure that
> variable. The problem arises when the variable is too generalized--a word
> like "success" is sort of a scientific catch all. I enjoy this discussion,
> and feel free to inquire about my various points--I'm writing this on
> break,
> so this message was a bit hurried.
> Sincerely,
> Robert Hooper, Undergraduate Research Assistant, HDFS Early Childhood
> Development Lab
> Hooper.90 at buckeyemail.osu.edu
> The Ohio State University, Department of Psychology; Department of
> Neuroscience
> 166H Campbell Hall
> 1787 Neil Avenue
> Columbus, Ohio 43210
> Cell: (740) 856-8195
>
> -----Original Message-----
> From: nagdu [mailto:nagdu-bounces at nfbnet.org] On Behalf Of Tracy Carcione
> Sent: Friday, June 7, 2013 3:09 PM
> To: NAGDU Mailing List, the National Association of Guide Dog Users
> Subject: Re: [nagdu] How long is "successful"
>
> I'm not comparing it to hospital stays; I was just trying to give an
> example
> of helpful statistics.  Rebecca's example of divorce statistics is a
> better
> comparison.  Nor would I expect it to be the whole picture, more another
> piece of data to use.  For instance, when I was choosing a place to get
> trained in computer programming, my first question was "What percentage of
> your graduates got work in the field after completing training?"  That's
> another kind of success statistic.
> I feel that stats on bad matches or unhealthy dogs would be helpful to
> both
> school and prospective student.
> Anyway, I'm sure I've used up my 5 list messages, so I'm off.
> Tracy
>
> ----- Original Message -----
> From: "Margo and Arrow" <margo.downey at verizon.net>
> To: "'NAGDU Mailing List, the National Association of Guide Dog Users'"
> <nagdu at nfbnet.org>
> Sent: Friday, June 07, 2013 2:15 PM
> Subject: Re: [nagdu] How long is "successful"
>
>
>> Tracy, I don't know if we can statistically do this.  I'm sure there are
>> ways to do it but I do not believe statistics show the entire opicture
>> and
>> I
>> also think comparing this to hospital stays is like comparing apples and
>> oranges.
>>
>> Margoa nd Arrow
>>
>> -----Original Message-----
>> From: nagdu [mailto:nagdu-bounces at nfbnet.org] On Behalf Of Tracy
>> Carcione
>> Sent: Friday, June 07, 2013 1:19 PM
>> To: NAGDU Mailing List, the National Association of Guide Dog Users
>> Subject: Re: [nagdu] How long is "successful"
>>
>> I think we're mixing up personal success with statistical success.  I'd
>> agree that, on a personal level, getting your dog to do what you want
>> with
>> a
>> minimum of effort, and being safe together, constitutes success.  But
>> how
>> do
>> you measure that, if you want broad statistics?  The only way I can
>> think
>> of
>> is to see how long the team stays together.  There will be
>> outliers--people
>> who retired a dog young because of an attack, for instance, but it's the
>> only way I can figure to measure what I want to measure. If someone has
>> a
>> better idea, I'd love to hear it.
>> Now, maybe you don't care.  You're happy with your dog and your school,
>> and
>> that's all that counts.  I can dig that.  But I'd also like some level
>> of
>> assurance that, when I go to class, I have the best chance of getting a
>> dog
>> I can work with for a long time.  Right now, all a person can do is talk
>> to
>> other people from their school of choice and see how they've done.  I
>> just
>> wonder if there isn't some more objective measure that could be added
>> into
>> the mix, to give the prospective student another way to compare schools,
>> and
>> to show schools if their training is working as well as they would like.
>> After all, the hospital I work for analyzes data on bad outcomes, so we
>> can
>> do better.
>> Tracy
>> ----- Original Message -----
>> From: "Larry D. Keeler" <lkeeler at comcast.net>
>> To: "NAGDU Mailing List,the National Association of Guide Dog Users"
>> <nagdu at nfbnet.org>
>> Sent: Friday, June 07, 2013 12:42 PM
>> Subject: Re: [nagdu] How long is "successful"
>>
>>
>>> Well, the point is that in my neighborhood, I know where those small
>>> curbs are at. She does stop at the big ones and whenever there are
>>> obstacles in the way. But, for me, the little ones where I live are
>>> not that important so I don't get on Holly to do them. I could have
>>> but its really not that important. Some folks might not consider that
>>> successful but, i don't really mind. I do  care if obstacles are there
>>> and if she didn't stop for them! Some folks will tolerate there dogs
>>> eating things they shouldn't or sniffing other folks for example. What
>>> I consider successful is that if I want Holly to do something, I can
>>> get
>> her to do it with a minimum of fuss.
>>> If I want her to stop at those little curbs, she will! And, if we are
>>> somewhere that I don't know, she will anyway. I guess success for me
>>> is knowing what your dog will do when you do something and what your
>>> dog knows you will do! And, if your dog listens to your commands and
>>> you listen to the dog.
>>> ----- Original Message -----
>>> From: "Star Gazer" <pickrellrebecca at gmail.com>
>>> To: "'NAGDU Mailing List,the National Association of Guide Dog Users'"
>>> <nagdu at nfbnet.org>
>>> Sent: Friday, June 07, 2013 11:25 AM
>>> Subject: Re: [nagdu] How long is "successful"
>>>
>>>
>>>> Larry,
>>>> Your post about your dog not stopping at curbs as a good example of
>>>> how
>>>> difficult this is to deal with.
>>>> Reading your post, I was thinking "I couldn't deal with that
>>>> behavior".
>>>> Y
>>>> Ou feel differently. You love your dog. You and she have a history. I
>>>> don't
>>>> know your dog, and have no history with her.
>>>> I'm wondering if the statistics used on marriage and divorce rates
>>>> would
>>>> serve as a good model?
>>>> We all know people who have been married for 60 years and are
>>>> miserable.
>>>> Yet, for all kinds of reasons they stay married.
>>>> Likewise, we all know marriages that ended quickly for any number of
>>>> reasons.
>>>> And as with dogs, we all have a friend where we think "How does she
>>>> put
>>>> up
>>>> with *that*".
>>>>
>>>>
>>>> -----Original Message-----
>>>> From: nagdu [mailto:nagdu-bounces at nfbnet.org] On Behalf Of Larry D.
>>>> Keeler
>>>> Sent: Friday, June 07, 2013 11:19 AM
>>>> To: NAGDU Mailing List, the National Association of Guide Dog Users
>>>> Subject: Re: [nagdu] How long is "successful"
>>>>
>>>> Aggreed! My point is that too many variables exist to have a perfect
>>>> team.
>>>> You have to use some kind of continuum scale to measure. And, what
>>>> success
>>>> if for one is not the same as it is for another. If you use saftey as
>>>> you're
>>>> standard most folks I know at least have that covered.
>>>> ----- Original Message -----
>>>> From: "Margo and Arrow" <margo.downey at verizon.net>
>>>> To: "'NAGDU Mailing List, the National Association of Guide Dog
>>>> Users'"
>>>> <nagdu at nfbnet.org>
>>>> Sent: Friday, June 07, 2013 10:53 AM
>>>> Subject: Re: [nagdu] How long is "successful"
>>>>
>>>>
>>>>> Well, I'd say that even if a team works for one month and does well,
>>>>> they're successful.  I say this because after one gets home, things
>>>>> could happen.
>>>>> Dogs get sick, humans get sick, dogs die, humans die, circumstances
>>>>> change, etc., etc., etc.
>>>>>
>>>>> I just don't think we can put too much of a figure on it.  I figure,
>>>>> though, one can begin to tell how successful a team is after they get
>>>>> home and work a bit.  One can also tell during class if a team might
>>>>> be successful or not.
>>>>>
>>>>> Margoa nd Arrow
>>>>>
>>>>> -----Original Message-----
>>>>> From: nagdu [mailto:nagdu-bounces at nfbnet.org] On Behalf Of Tracy
>>>>> Carcione
>>>>> Sent: Friday, June 07, 2013 8:04 AM
>>>>> To: NAGDU Mailing List, the National Association of Guide Dog Users
>>>>> Subject: [nagdu] How long is "successful"
>>>>>
>>>>> Darla asked how long a team has to be out to be "successful".  I'd
>>>>> say
>>>>> at least 2 years, just to put a number on it.  Or possibly 3; I could
>>>>> argue either way.
>>>>> I'd be real curious to see numbers from schools of teams graduated,
>>>>> and partnerships that lasted 3 years or more. I think that should be
>>>>> a
>>>>> pretty good indicator as to how well the school is doing. I mean, if
>>>>> school X put out 500 teams, and 300 of them stayed together, that's
>>>>> only a 60% success rate, and not so good.  But if 400 of them worked
>>>>> 3
>>>>> years or more, that's 80% success, which is pretty good.
>>>>> Tracy
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> nagdu mailing list
>>>>> nagdu at nfbnet.org
>>>>> http://host.nfbnet.org/mailman/listinfo/nagdu_nfbnet.org
>>>>> To unsubscribe, change your list options or get your account info for
>>>>> nagdu:
>>>>> http://host.nfbnet.org/mailman/options/nagdu_nfbnet.org/margo.downey%4
>>>>> 0veriz
>>>>> on.net
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> nagdu mailing list
>>>>> nagdu at nfbnet.org
>>>>> http://host.nfbnet.org/mailman/listinfo/nagdu_nfbnet.org
>>>>> To unsubscribe, change your list options or get your account info for
>>>>> nagdu:
>>>>> http://host.nfbnet.org/mailman/options/nagdu_nfbnet.org/lkeeler%40comc
>>>>> ast.net
>>>>
>>>>
>>>> _______________________________________________
>>>> nagdu mailing list
>>>> nagdu at nfbnet.org
>>>> http://host.nfbnet.org/mailman/listinfo/nagdu_nfbnet.org
>>>> To unsubscribe, change your list options or get your account info for
>>>> nagdu:
>>>>
>> http://host.nfbnet.org/mailman/options/nagdu_nfbnet.org/pickrellrebecca%40gm
>>>> ail.com
>>>>
>>>>
>>>> _______________________________________________
>>>> nagdu mailing list
>>>> nagdu at nfbnet.org
>>>> http://host.nfbnet.org/mailman/listinfo/nagdu_nfbnet.org
>>>> To unsubscribe, change your list options or get your account info for
>>>> nagdu:
>>>>
>> http://host.nfbnet.org/mailman/options/nagdu_nfbnet.org/lkeeler%40comcast.ne
>> t
>>>
>>>
>>> _______________________________________________
>>> nagdu mailing list
>>> nagdu at nfbnet.org
>>> http://host.nfbnet.org/mailman/listinfo/nagdu_nfbnet.org
>>> To unsubscribe, change your list options or get your account info for
>>> nagdu:
>>>
>> http://host.nfbnet.org/mailman/options/nagdu_nfbnet.org/carcione%40access.ne
>> t
>>>
>>
>>
>> _______________________________________________
>> nagdu mailing list
>> nagdu at nfbnet.org
>> http://host.nfbnet.org/mailman/listinfo/nagdu_nfbnet.org
>> To unsubscribe, change your list options or get your account info for
>> nagdu:
>> http://host.nfbnet.org/mailman/options/nagdu_nfbnet.org/margo.downey%40veriz
>> on.net
>>
>>
>> _______________________________________________
>> nagdu mailing list
>> nagdu at nfbnet.org
>> http://host.nfbnet.org/mailman/listinfo/nagdu_nfbnet.org
>> To unsubscribe, change your list options or get your account info for
>> nagdu:
>> http://host.nfbnet.org/mailman/options/nagdu_nfbnet.org/carcione%40access.net
>>
>
>
> _______________________________________________
> nagdu mailing list
> nagdu at nfbnet.org
> http://host.nfbnet.org/mailman/listinfo/nagdu_nfbnet.org
> To unsubscribe, change your list options or get your account info for
> nagdu:
> http://host.nfbnet.org/mailman/options/nagdu_nfbnet.org/hooper.90%40buckeyemail.osu.edu
>
>
>
> _______________________________________________
> nagdu mailing list
> nagdu at nfbnet.org
> http://host.nfbnet.org/mailman/listinfo/nagdu_nfbnet.org
> To unsubscribe, change your list options or get your account info for
> nagdu:
> http://host.nfbnet.org/mailman/options/nagdu_nfbnet.org/julielj%40neb.rr.com
>
>
> -----
> No virus found in this message.
> Checked by AVG - www.avg.com
> Version: 2012.0.2242 / Virus Database: 3199/5891 - Release Date: 06/07/13
>
>
> _______________________________________________
> nagdu mailing list
> nagdu at nfbnet.org
> http://host.nfbnet.org/mailman/listinfo/nagdu_nfbnet.org
> To unsubscribe, change your list options or get your account info for
> nagdu:
> http://host.nfbnet.org/mailman/options/nagdu_nfbnet.org/carcione%40access.net
>