data transformation for normal distribution

[always-good]

Hello!

How to transform data through the log transformation to ensure a normal distribution for the somatic cells count variable, please?

is there any test to test the normality of the data before?

Finally, is there a procedure or other way to do the log transformation?

Thank you!

[whymath · Accepted solution]

To transform a variable to its logarithmic form, you just need to apply "log()" function, this function returns the natural (base e) logarithm. You can also use "log2()" (base 2) and "log10()" (base 10).

data _null_;
  x=10;
  y1=log(x);
  y2=log2(x);
  y3=log10(x);
  put y1=/ y2=/ y3=;
run;

y1=2.302585093
y2=3.3219280949
y3=1

To test the normality of data, you can use "proc univariate" with a "normal" option.

proc univariate data=sashelp.class normal qqplot;
  var height;
run;

The result window will show you test result like this:

You can also use qqplot statement in this procedure, to get a Q-Q plot of analysis variable.

[Ksharp]

For single variable to test the normality try NORMATEST option:

proc univariate data=sashelp.heart normaltest;
var weight;
run;

For multiple variables to test normality try @Rick_SAS 's blog:
https://blogs.sas.com/content/iml/2012/03/02/testing-data-for-multivariate-normality.html

For transforming data to conform normal distribution ,try Box-Cox transformation:
https://blogs.sas.com/content/iml/2022/08/22/box-cox-transform.html
https://blogs.sas.com/content/iml/2022/08/17/box-cox-regression.html