I guess it is a simple question, but would be happy to get a statistical advise over here: I have the total concentration of metals in my samples and their standard deviation. I did a linear combination fit according to their EXAFS and got the best fit for a certain composition. I am aware of the limitations of this interpretation; still want to get some quantitative estimation of each component in my sample. I am multiplying its fraction (according to LCF) in the total concentration... that's easy.. but how should I treat the standard deviation of each fraction in this case?