Adjusting range of values for R² value

Hi, I am using R² values to qualify handheld meter results against laboratory results. I have noticed that the result for a set of data that has a very small spread of readings (eg 20 results ranging from 1 - 5) gives a much worse R² value than one with a similar error, but a much greater spread of results (eg 20 results ranging from 1 - 100). Is it correct to adjust the results to get a consistent spread that relates to the specification limits you are measuring to?