# Measurement error influence

Let's say we have 260 daily measurements of variable X. The picture in the link shows graphically those measurements. This variable has an average measurement error of ±1 units.

I want to prove that this error is not significant. So, I thought to generate random +1 and -1 values and add these values to each of the 260 measurements. So, the idea is:

X1'=X1 ± 1
X2'=X2 ± 1
.
.
X260'=X260 ± 1


Then I want to plot these Xi' values in the same way as in the picture in order to show that the error does not influence the trend of the line and that the two graphs would look the same. But this visualization approach can be a bit subjective. Do you think my approach makes sense? If not, can you think of a more numerical way? https://dl.dropbox.com/u/71799266/Variable%20X.bmp

-