RE: /400 skills test -- MIDRANGE-L

Lance wrote:

> In a message dated 5/16/2000 7:54:24 PM EST, Buck Calabro wrote:
>
>>  some work for the new hire.  Right way -- analyse outstanding tasks,
>>  including the future for a year or so; determine skills needed to 
>>  accomplish
>>  those tasks; determine pay/benefits needed to get a candidate to hire
on;
>>  place adverts; interview/test; hire.
>
>We agree and we know what the new person will be doing and what 
>skill will be  required and the skills are in the ad.  Now, how do we 
>winnow 60 applicants - -- all of whom claim to have decades of 
>experience and vast expertise.  

"... claim to have ..."  Have you been burnt by false claims in the past?
It's an interesting position to be in, that's for sure.  Now you have
several questions to answer:  Do you need The Best person, one who is Good
Enough, the Cheapest who can get the job done or do you need someone in the
door Fast?  

>We don't care too much about certificates and degrees -- we 
>want someone who can do the work. And we want to know 
>before we hire that they can do the work.  

Doesn't one pass tests to get a degree or certification? <grin>

>At this time we cannot afford a trainee.  How does one go 
>about determining an applicant's level of expertise without 
>a test?  Thanks in advance, Lance

Hm.  I'd take the three hardest of your outstanding tasks and ask the
candidates to explain their approach to solving them.  The experienced ones
will ask a bunch of questions before starting to answer.  If you are looking
for new blood, listen for approaches that are different from the ones your
shop would take today.  If you're looking for somebody who'll "fit right in"
look for somebody who solves them the same way your shop would.

A test might weed out the outright liars of the 60 - if an RPG programmer
with decades of experience can't explain how and why to code a "page at a
time" subfile, then they probably won't fit your needs!  My contention is
that somebody else's test (generic or otherwise) won't be able to tell you
the level of somebody's experience - only their inexperience.

I truly think that the screening process depends on what you really need.
If you want The Best, then take the time to interview all 60 candidates and
have them talk programming philosophy with you - high level stuff.  Have
they done recent reading?  Do they know about mailing lists like MIDRANGE-L?
How would they tackle Project X?  That sort of thing.

If you need somebody who is Good Enough, then select all the resumes that
mention every one of your mandatory skills.  If you have 60 candidates after
that, then either your mandatory list is too generous or every candidate is
really eligible to work there.  In any event, pick 10 of the survivors and
interview them using the same technique you'd use to find The Best.  You're
looking for The Best of 10 instead of 60.  One of them is surely Good
Enough.

If Price is your filter, have HR make the same dollar offer to each of the
60, take the bottom 10 and interview them...

If speed is paramount, pick 5 at random and interview them all tomorrow.

If you intend to take ideas from the list and create a test of your own, you
may be better off than a generic test.  Just don't let a non technical
person administer the test unless the questions are so specific that there's
only one right answer (multiple guess?)  The questions should come straight
off your task/skills list, so you can be certain that you're testing for
what you need.  In any case, you'll still have the same logistical problems
issuing and analysing the results of a test of 60 candidates.

Like I said, it's an interesting situation to be in!

Buck Calabro
Aptis; Albany, NY
+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com.
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---