1
$\begingroup$

Afternoon,

I have a number of different large data capture processes with populations from 100,000 up to 4.5million forms.

I am wanting to determine the error rate in the capture process so am going to manually verify a number of forms against what was captured. I am using an assumption of 1% error rate, a confidence level of 99% with a confidence interval of 0.5% to calculate a sample size.

For the small process (100,000) the calculator comes out at 2440 forms. For the large process (4.5 million) the sample size goes up to 2499 forms (only 59 more).

Why is that number (59) so small? Is there a simple explanation for this?

Improve this question
$\endgroup$
2
  • 1
    $\begingroup$ could you share more detail on the calculation you are using? $\endgroup$
    – ReneBt
    Commented Mar 28, 2018 at 13:28
  • $\begingroup$ nss.gov.au/nss/home.nsf/pages/Sample+size+calculator Confidence level: 99% Population Size: 100,000/4,500,000 Standard Error: 0.01 $\endgroup$
    – Robert
    Commented Mar 28, 2018 at 22:04

1 Answer 1

Reset to default
1
$\begingroup$

The population size is largely irrelevant to power analysis, unless the sample is a large portion of the population. The calculations are based on the assumption of an infinite population and there is a finite population correction for estimating parameters:

$$FPC = \sqrt{\frac{N-n}{N-1}}$$

where N is the population and n is the sample size. When N is much larger than n, the FPC is so close to 1 that it can be ignored. The difference you got (59) is the part that would often be ignored.

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.