Prometheus 6

me (basically quoting you): in actual fact all we can say (assume) is,  "the candidate with the MARKEDLY higher score is likely the better candidate.

you: This is at odds with basic statistics, as well as intuition.

whuuut?! i add the word "markedly" to your own statement and you disagree with the whole thing? and then throw in some mumbo-jumo about close polling data?! is this is the part of the discussion where we make like the frightened squid and cloud the issues with an inky distraction? ah, you'll not  escape so easily, friend.

i guess it's my bad for using the term "margin of error," but your example has NOTHING at all to do with my point (and it certainly fails to prove yern). you're waaay off.  almost like you done it on purpose -- i think i see what P6 be gettin at.  but i'll retract the characterization and try one mo' 'gain...

(really, shelf, you're going to have to get better at seeing the broad strokes and the general gist of things as i try to come up with cruise-missile accurate anologies)

lookit, all i meant was that 

1) determining who's best among candidates is not an exact science. it's guesswork. what you HOPE for is a deal where everybody but one pretty much sucks.  you, as even-handed hiring guy, really only earn your pay when dealing with similarly qualified candidates.

in other words, things get fuzzier, not sharper, when two (or more) candidates are close to each other. any foo can see distinctions when the scores are far apart.

2) given that any two (or more) candidates are indistinguishable (wrt suitability for role) with sufficiently close ratings, i say you can give the nod to the underepresented.

unless white male only ever planned to apply to that one gig, his chances are better'n decent that he'll latch on at some other good opportunity. the rest of us are already used to the idea that you don't win em all.

unless you begrudge your tax dollars going to anything for which you do not directly and immediately benefit, i don't see what you have to fear. please explain your fear.

I don't suggest that AA is rigorously enforced here in 2005, and I think you can find in this thread where I referred to AA's "heyday", as being in the past.

It never has been rigorously enforced. It's never had a heyday, as far as benefitting Black folks.

i guess it's my bad for using the term "margin of error," but your example has NOTHING at all to do with my point

There was nothing wrong with your use of "margin of error", but there is something wrong with your assertion that being within the margin of error implies equality.

in other words, things get fuzzier, not sharper, when two (or more) candidates are close to each other. any foo can see distinctions when the scores are far apart.

Sure.

given that any two (or more) candidates are indistinguishable (wrt suitability for role) with sufficiently close ratings, i say you can give the nod to the underepresented.

Let's be clear here memer. If you own a company and do it that way, no problem from me. If you work for a company and you do it that way, again, no problem. What we're talking about is hiring for a government job, a job where the criteria for hiring is concisely specified.

As soon as we agree that "there are good reasons for the government to prefer one race over another", we agree that the goodness of such reasons will be determined by political popularity.

the rest of us are already used to the idea that you don't win em all.

In the company I work for, a qualified black candidate for an open job like mine will be offered that job 100% of the time.   A qualified white candidate needs to out-compete a half dozen other candidates.  I'm not protesting that, it's not the government.  But the idea that white males don't get turned down when applying for jobs is bogus. It's the hardest part of searching for jobs: allowing other people to pass negative judgement over you and your efforts to impress them, and then getting up and out the next day for more of the same.

That said, I feel no desire to protect someone because they're a white male.  I'm much more motivated to just be fair to everyone. 

DW - no comment on the Pager paper? i'm shocked.

There do come some "qualities" associated with people with college degrees. They tend to be far better versed in the ways the world works. They tend to be more socialized.

Your comments about the "qualities" of an AA with a college degree are off-target. I teach at a Big Ten, and i can tell you right now that many of the students i come into contact with have NO IDEA of how the corporate world works at the clerical/secretarial level (and i'm not talking about the black ones). Why would someone who has a 4 year degree stay at an entry level secretarial job? Not to mention, what about filing, typing letters in MS Word, and faxing REQUIRES a BS or BA? Although i don't know the literature, i'd be willing to bet at least one shoe - heck, i'd bet the whole pair - that the turnover rate for AA's with BA's is higher than AA's without.

Sincerely, then tend to make better employees: they've achieved at least one important life goal which they could have opted out of

this comment was beneath you, man. you think the retention rates of Black college students reflect their decision to "opt out"? *sigh*

Here's a link to the article about race and geographical hiring preferences let's see where you go with this one...

Sincerely, they tend to make better employees: they've achieved at least one important life goal which they could have opted out of

this comment was beneath you, man. you think the retention rates of Black college students reflect their decision to "opt out"? *sigh*

I wasn't thinking about black at all when I was writing that.

But yes. The American educational system pretty well guarantees a kid will graduate from high school in all but the worst of areas.  The peer and family pressure is tremendous. High school dropouts are considered to be bad candidates for almost any kind of job.

College is the first major hurdle that a young person must face on her own.  If she wants to drop out, there's no huge infrastructure there begging her to continue.  There's the temptations of making money now rather than being a poor student.  Thus, the acquistion of that degree says something importantly positive about the graduate.  It says that on at least one occasion they followed through on a long term commitment. 

Here's a link to the article about race and geographical hiring preferences let's see where you go with this one.

g

Fascinating, CP. I found myself believing that the author had collected real data and knew how to present it.

Here's a quote:

The employers were willing to “take a chance” on people they would not normally have hired because they trusted the judgment of the LDC director with whom they had an ongoing relationship—another example of the importance of social networks.

g

The thing to contemplate about this quote (the theme of which was repeated throughout the report) is that social networks work.  They work for employees, they work for employers.  Several incidences in that report involved hiring unknown and un-vouched for people, and having dismal results.

the tears of boredom are welling up, P6.  i'm not built for these protracted circular online arguments. i will ask you (as i did Cobb) to put up a wiki expounding on all your various and sundry philosophies. makes things easier for louts like me. i can just link to a specific bit, mebbe add/reduce/remix it a bit and won't have to get into these things with shelf-types.

*rolls up sleeves* once more unto the bitch...

There was nothing wrong with your use of "margin of error", but there is something wrong with your assertion that being within the margin of error implies equality.

i most def shouldn't have used the term if it got your knickers bunched over statistics. that wasn't the point. i hope i made it more clear now.

What we're talking about is hiring for a government job, a job where the criteria for hiring is concisely specified.

There is hardly anything magical or "scientific" about a gubmint job spec. There are lots of decent-sized private and public companies that can and do create very specific matrices of duties and the supposed skills/experience to carry them out satisfactorily.  

But no matter how specific and tight you make those specs, dear dwshelf, it is rare for there to appear a candidate who fits them precisely.  Qualitative judgement calls are necessary at some point.  

It has been my experience that it is still a hit and miss thing.  Far from perfect.  Perhaps it has been your experience that gubmint hiring agents are deadly accurate and impartial in their practices, which would explain why you cling to the notion that the highest score in some arbitrary matrix -- no matter close the rest are to it -- (almost?) always delivers the best person for the job.

I say there are lots of factors that go into that important decision, and bias, a natural human tendency, can creep in. especially when it's close.  sometimes unconsciously, sometimes with most-like-me aforethought.  in a perfect, ideal world, checks and balances are unnecessary. but this ain't that world. gots to straddle, m'man -- one foot in kingdom-to-come, the other firmly planted on terror firma.  

your founding fathers had the right idea/l, inserting checks against potential (and all too human) abuse. continue with the refinements.

that's it. i'm out.