There are some data (for my case concentration of trace metals) like < or > x. By using this type of data I am trying to get descriptive statistics, Pearson correlation, Cluster analysis, and Principal component analysis. But, it is not including the variable that contain "< x " type data for calculation and showing it as string variable.
How can I get out of this problem?
Thanks in advance.
In order to include observations below detection limit in your analysis, you have to replace each observation below detection limit by a number, and there is no exact answer as to what is best (zero, in which case you cannot transform to logarithms, or a small number such as 0.25, or a number equal to the detection limit, some authors recommend "Detection limit divided by the square root of 2"). In different disciplines there may be different customs in this regard. Check the publications in your discipline to see what they do. You might try different options and see which works best. It would be interesting to see what discussions come up from practitioners working in different disciplines.
Thank you Dr. Francois E Steffens .
I know there are several methods of dealing with Left Censored data (data below detection limit). But, I saw some research articles they are presenting the range as follows:
Thank you sir making me clear. Actually I was searching for such kind of precise answer.
Again thank you very much Sir. @ Francois E Steffens
- Badges
- Science topic