Page 1 of 1

normal scores transformation

Posted: Wed May 12, 2010 1:51 am
by nfmohammed
How do I get the normal scores for the variables using eviews 7?

Re: normal scores transformation

Posted: Wed May 12, 2010 6:41 am
by startz
Would

Code: Select all

series z=(x-@mean(x))/@stdev(x)
be what you want?

Re: normal scores transformation

Posted: Wed May 12, 2010 5:07 pm
by nfmohammed
It is not really z scores, the transformation involves ranking the observations, then the ranks are being substituted by scores on the normal distribution. The scores are from the normal distribution by dividing the distribution into the number of observations plus one regions on the basis that each region has equal probability. For example, if there are six observations the normal distribution would be divided into seven equally probable parts so that the original value are replaced by normal scores (here -1.0676, -0.5659, -0.1800, 0.1800, 0.5659, 1.0676) rather than the ranks 1,2.....,6.

Re: normal scores transformation

Posted: Thu May 13, 2010 9:07 am
by EViews Glenn
series z = @qnorm(@ranks(x)/(@obs(x)+1))