# Sample Size Question

Six Sigma – iSixSigma Forums Old Forums Europe Sample Size Question

Viewing 4 posts - 1 through 4 (of 4 total)
• Author
Posts
• #23928

Forecaster101
Participant

Hello all, I’m wanting to validate new logic applied to our customer orders. It’s basically a pass or fail situation. We have approximately 2500 skus and I’m trying to determine what would be a good sample size out of the 2500 to test whether the logic is good (passes). The plan is to eventually check all, but in the short term it would be nice to validate a random sample of the skus to see whether the logic seems to be working. Any thoughts or suggestions on how to go about it? Thanks in advance!

0
#57293

Jeff
Participant

Hi,
Several considerations go into the sample size question.  When you have attribute data (pass/fail) you would obviously need more samples to determine if the results of your sample are representative of your population.  The next consideration is that you need to estimate a current defect rate in your existing (finite vs infinite?) population.  The larger the estimation of defects, the more you would need to sample.  If you estimate up to a 50% defect rate, you may find that sample size required is almost the entire population anyway.  The next consideration is how much “margin of error” are you willing to accept in your sampling (this is not the confidence interval).  A higher margin of error allowance means a smaller sample size.  So, there is a good sample size that can be calculated that will tell you within a margin of error and confidence interval how representative your sample is of the population, but start with estimating the current defect rate in the existing population.  If the plan is to eventually check all SKU’s anyway, and that is not a huge burden or cost, then by measuring the entire population, you will be 100% confident you have found all the errors!

0
#57295

Patty
Participant

For a quick and dirty sample size, I use a sample size calculator.  http://www.surveyguy.com/homepage.shtml

0
#57297

Participant

The sample size calculator from SurveyGuy that Patty referred to is a little hard to find from the main page — try this page:
http://www.surveyguy.com/SGcalc.htm
This calculator is also quite good:
http://www.raosoft.com/samplesize.html

Best regards,