Attribute Sampling Planshttp://www.theiia.org/intAuditor/back-to-basics/2010/attribute-sampling-plans/
Blogen-usWed, 07 Aug 2013 03:42:19 PMWed, 07 Aug 2013 03:42:19 PMhttp://www.theiia.org/intAuditor/back-to-basics/2010/attribute-sampling-plans/2013-08-07interval confidencehow to find a 95% confidence intercal with a population of 540 loans?http://www.theiia.org/intAuditor/back-to-basics/2010/attribute-sampling-plans/2012-11-28statistical theoryIt seems the tables are based on normal distribution (z statistics). In my opinion, it is more precise to use Clopper-Pearson method. Here is a page that will give you a very good and comprehensible background:
http://www.sigmazone.com/binomial_confidence_interval.htm
And here are excel formulas to calculate:
i) the upper limit for number of errors > 1
=1-BETAINV((1-B1)/2,B2-B3+1,B3)
where B1 = confidence level (e.g.95%)
where B2 = sample size
where B3 = number of errors in the sample
and
ii) the upper limit for number of errors = 0
=1-BETAINV((1-B1)/2),B2+1,1)
where B1 = confidence interval
where B2 = sample size
Jamie Carrillo: Jamie, my guess is that it stands for the "margin of error" (length of confidence interval).
Rick D: Rick, once you test your sample and know the sample error rate you need to use the correct statistical formula (not the tables) to estimate the population error rate. Than, it does not matter what your initial guess about the population was.http://www.theiia.org/intAuditor/back-to-basics/2010/attribute-sampling-plans/2012-07-31SamplingIf I overestimate the expected deviation rate (i.e, 50% vs. 30% observed during testing) when calculating a sample size does that make my test results less reliable. A co-worker contends that the observed deviation rate must fall within the 5% +/- of the expected deviation for the results to be valid. Thanks.http://www.theiia.org/intAuditor/back-to-basics/2010/attribute-sampling-plans/2012-06-10Tolerable Deviation RateHello, what does the tolerable deviation rate equate to statistically? Is it a standard deviation?http://www.theiia.org/intAuditor/back-to-basics/2010/attribute-sampling-plans/2012-03-19audit planningi want aclear planning procedure can you help mehttp://www.theiia.org/intAuditor/back-to-basics/2010/attribute-sampling-plans/2011-10-23sample size adjustment formulaFor known populations you can use this formula to adjust the table result for a finite population. This ajusment can lend to more efficiency when 5% or more of the population is being sampled, based on the tables.
Written in Excel formula:
<strong>new ss = ss/ (1+ ((ss - 1) / population)))</strong>
where ss = the computed sample size.
http://www.theiia.org/intAuditor/back-to-basics/2010/attribute-sampling-plans/2011-07-08sample sizeThe larger the sample size the better the estimate will be.. the population of 100, the sample size should be around 55 to 60 for random selection. http://www.theiia.org/intAuditor/back-to-basics/2010/attribute-sampling-plans/2011-05-09Very high confidence level regarding a low tolerable rateIn a very special attribute sampling application, I needed extremely high confidence level (99.9% or preferably 99.99%) that a particular attributed existed in at least a low percentage (say 50%) of the population. I had an extremely hard time finding any guidance as to sample sizes regarding this application before the Internet, and I'm not having much more luck now. Do you have any suggestions as to where to look?http://www.theiia.org/intAuditor/back-to-basics/2010/attribute-sampling-plans/2011-03-30SamplingHi Dennis, since your table is based on large populations, does this mean that if I have a population of 100, I should not use a sample size of 25 if I will be doing random selection? If my population size is 100, what should my sample size be if I want 95% confidence? Thanks!