I'm trying to calculate the percent difference between an expected measurement and the actual measurement. I'm using the formula
Percent Difference = ((Measured value - Expected value)) / Expected value) * 100
So if I measure a value of 23, and I expected a value of 20:
Percent Difference = ((23-20)/20) * 100
Percent Difference = 15
And I get a negative value if the measured value comes in below the expected value:
Percent Difference = ((17-20)/20) * 100
Percent Difference = -15
I"m assuming this is the best way to calculate a percent difference. But what do I do if the expected value is 0 (I can't divide by zero). The above calculation gives works for larger numbers, but gives me really large values as I approach zero.
If I'm expected a value of zero, but get something close (AKA 0.5), if I approximate zero with a small number, I get a crazy value
Percent Difference = ((0.5-0.001)/0.001) * 100
Percent Difference = 49900
Or even
Percent Difference = ((0.5-0.1)/0.1) * 100
Percent Difference = 400
Is there a better way to characterize a difference with small numbers?