Graphing Data in Excel: Error Bar Confusion

[Hi Im Paul]

I'm currently writing a paper over an experiment I did for class and I have a very stupid question over I need clarification on before I continue.

I am using excel to graph out data, and I have determined that the error in my Length is .001 m. However, I am graphing the inverse of length to show that my theoretical model is proportional to 1/L. Will the error bars I use be .001, or will the error bars be 1/.001? Surely it can't be 1/.001!

Also, I determined through error propegation that the error in my theoretical frequency is, let's say 17. However, I am graphing my theoretical f^2. Will my error bars by 17, or 17^2?

Could use some reassurance!

 

[kuruman]

Consider your variable to be y = 1/x. The uncertainty is δx=0.001 m. What is the uncertainty δy? Then propagate errors in f(y) knowing δy.