Advice on data transformation

[KEVIN MANNING]

Dear lsit-mates,

   I could use some advice on data transformation.  My ultimate goal is to perform a factor analysis on one sample of participants who have taken a large battery of cognitive tests.
  However, for many of the variables there is a large ceiling effect (e.g., the score of one variable can range from 7-17 yet most of the participants perform perfectly on the test and score 17).  I am not familiar with transforming data and was wondering if anyone had suggestions on what procedure to use?
   Thank you for your help.
Kevin

[Alexander J. Shackman-2]

typically transforms are meant to modify the skew or kurtosis of a variable
with adequate variance (
http://psyphz.psych.wisc.edu/%7Eshackman/mediation_moderation_resources.htm#DataTransforms),
but what you describe is a lack of variance. this is akin to a situation in
digital signal processing where an amplifier "clips" because of out-of-range
inputs. there is no way to "fix" the data because there is no meaningful way
to rank the Ss (tweak the distribution of variance).

hth, alex shackman

On 6/7/07, KEVIN MANNING <[hidden email]> wrote:

>
> Dear lsit-mates,
>
>    I could use some advice on data transformation.  My ultimate goal is to
> perform a factor analysis on one sample of participants who have taken a
> large battery of cognitive tests.
>   However, for many of the variables there is a large ceiling effect (e.g.,
> the score of one variable can range from 7-17 yet most of the participants
> perform perfectly on the test and score 17).  I am not familiar with
> transforming data and was wondering if anyone had suggestions on what
> procedure to use?
>    Thank you for your help.
> Kevin
>

--
Alexander J. Shackman
Laboratory for Affective Neuroscience
Waisman Laboratory for Brain Imaging & Behavior
University of Wisconsin-Madison
1202 West Johnson Street
Madison, Wisconsin 53706

Telephone: [phone]
FAX: [phone]
EMAIL: [hidden email]
http://psyphz.psych.wisc.edu/~shackman