[Oz-gifted] tests in the public system

[Keith McGuinness]

Ellen Hrebeniuk wrote:
> At 11:19 AM +1000 18/9/06, Fly2 wrote:
> 
>>  > I'm interested to know in exactly what ways the results are
>>>  unreliable as I presume it would be more than just a matter of a
>>>  larger margin of error.  I can make some educated guesses, but does
>>>  anyone *know*?
>> I am not sure what you are asking here exactly.
> 
> Sorry -- I meant that the reliability of testing instruments for IQ 
> is supposedly lower at 5 than at 8.  I don't know whether that's due 
> to ceiling issues, younger children's unfamiliarity with testing, 
> etc, but it would be interesting if anyone here has any definitive 
> answers.

It depends upon what is meant by "reliability".

I *think* what is being referred to here is how reliably -- ie 
how well -- does child IQ predict adult IQ.

Estimates of adult IQ are more reliable (in this sense) when a 
child is 8 than when they are 5. Putting that another way, the 
older a child is when their IQ is measured, the more reliable is 
the prediction of adult IQ.

Keith McGuinness