0
$\begingroup$

Is it valid to z normalise data before doing Pearson's or spearman's correlation on curves? Currently I am z normalising my data and then running a spearman's correlation, from there I can generate p values.

What I'm looking is signals from a set of signals that have a statistically significant difference, and the above is the method I have used.

Improve this question
$\endgroup$
1
  • 1
    $\begingroup$ Did you try it to see how the standardization affects the correlations? $\endgroup$
    – ttnphns
    Commented Jul 13, 2019 at 11:31

1 Answer 1

Reset to default
1
$\begingroup$

Standardisation, i.e. subtraction of the mean and division by the standard deviation, is a linear transformation of your data. Neither Pearson's nor Spearman's correlation coefficient depend on the absolute values, hence neither will be affected by standardisation.

As ttnphns commented: try and see that there is no difference.

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.