Hi Cathy,
I'm confused as to why you feel the need to scale your beta weights
at all. With a normalized regressor, the beta is already scaled by
the amplitude difference, that is the purpose of normalizing the
regressor.
To be more concrete, the beta weight shows the amplitude of the
response to a stimulus as a difference from the baseline. If, as
is commonly done, the data is scaled to have a mean of 100, and the
regressor is scaled to have a peak of 1, then the beta weight is
the percent change from the baseline (at the peak).
How is your scenario different? Knowing that might make it easier
to understand your question.
- rick