Accuracy: Is there a specific formula to compute Measurement Device Accuracy?

[qpractitioner]

Is there a specific formula to compute Accuracy? Is the formula below correct? if yes, what is the criteria to say the instrument is accurate or not?

Accuracy = measured value/ expected value x 100

Any help will be appreciated.

Thanks.

 

[Rameshwar25]

Re: Accuracy: Is there a formula for computation?

My perception is that accuracy is capacity of an instrument to measure the characteristic closest to actual value. In BIAS study, we check how far the averages of observations are from ref. value. In a way, while checking the bias, we are checking its accuracy too- the accuracy in the form of difference between average of readings (observed value) and ref. value (actual value).
Your formula given resembles with that given in MSA manual revision2 where the decision was taken on the basis of percentage of bias.

In the third and fourth editions, the decision of bias is taken on hypothesis basis which is the better a way to take decision.

i think it will do.

 

[qpractitioner]

Re: Accuracy: Is there a formula for computation?

Thanks for reply.

what will be the criteria for me to say the instrument is accurate? If the result based on above equation is below 99%, is it considered accurate? Sorry, I am very new to this MSA thing.

thanks again.

 

[Rameshwar25]

If you refer applicable standard for measuring instrument which are referred to judge calibration data, you will find a table for acceptance criteria of difference between master sample reading and reading taken from the instrument for different ranges. You can convert these acceptance criteria into percentage.

Following is example of one such Indian standard:
IS 3651 Part 3 : Vernier Callipers with Least Count 0.1 mm (measuring range above 1000 mm up to and including 4000 mm)