MORE ON SHORT INFANT TEST

[Annette Karmiloff-Smith]

>X-Sender: ingram_wright at mail.geocities.com
>Date: Thu, 15 Apr 1999 10:05:33 +0100
>To: Annette Karmiloff-Smith <a.karmiloff-Smith at ich.ucl.ac.uk>
>From: Ingram Wright <mrpiw at leeds.ac.uk>
>Subject: Re: MORE ON SHORT INFANT TEST
>Mime-Version: 1.0
>
>Annette,
>
>Your original suggestion seemed to be about a possible short form of the
>Bayley II to be used alongside a previously obtained profile of pass/fail
>item data. The responses so far appear to be about tests which are shorter
>but may, or may not, correlate with Bayley II scores. My suggestion is to
>use the raw Bayley item data and test on a restricted set of items to
>ensure that no significant developmental change had occurred.
>
>Assuming the validity of the item scaling on the Bayley and that children
>will increase their set of passed items - Take the failed items within the
>last set administered and administer only these. If a sufficient number are
>still failed then you are still within the appropriate set of items. You
>can then modify the score according to the number of additional items passed.
>
>If all items from within the failed set are passed then you would have to
>test  on the next full set of items.
>
>The validity of doing this is based on assumptions about the course of
>development which may be problematic for atypical populations. However, if
>anyone has test-retest data on the particular population then the validity
>of using this method could be checked before testing (using that data).
>
>This may not be what you were thinking of??
>
>Ingram Wright
>
>
>-------------------------------------
>Dr. Ingram Wright
>School of Medicine
>Division of Psychiatry & Behavioural Sciences
>15 Hyde Terrace
>Leeds. LS2 9LT. UK.
>
>mrpiw at leeds.ac.uk	   http://www.geocities.com/Athens/Olympus/3788/
>