[SamCPA]

I know for attribute testing tolerable misstatement, expected error rate and risk of assessing control risk too low all affects sample size. However, for variable sampling there are different factors that affect sample size. What are those factors that are unique to variable sampling?

[JamesBJames]

In terms of factors that affect sample size, it's pretty similar to attribute sampling. Sample size will increase when expected misstatement increases, population variably increases, tolerable misstatement decreases, or the acceptable level of risk decreases (as in, we have to be 99% confident instead of 95%).

There's also the concept of stratification which is unique to variables sampling, and the whole purpose of it is to reduce variability such as to lower the sample size.

(Also, gonna be a little nitpicky here: it's tolerable deviations for attributes sampling)

FAR: May 1st, 2014 - 91
AUD: May 29th, 2014 - 97!
BEC: July 16th, 2014 - 91
REG: August 29th, 2014 - 88

Licensed December 2015

Feel free to add me on LinkedIn by clicking my username!

[JamesBJames]

In terms of factors that affect sample size, it's pretty similar to attribute sampling. Sample size will increase when expected misstatement increases, population variably increases, tolerable misstatement decreases, or the acceptable level of risk decreases (as in, we have to be 99% confident instead of 95%).

There's also the concept of stratification which is unique to variables sampling, and the whole purpose of it is to reduce variability such as to lower the sample size.

(Also, gonna be a little nitpicky here: it's tolerable deviations for attributes sampling)

FAR: May 1st, 2014 - 91
AUD: May 29th, 2014 - 97!
BEC: July 16th, 2014 - 91
REG: August 29th, 2014 - 88

Licensed December 2015

Feel free to add me on LinkedIn by clicking my username!

[SamCPA]

Thanks. Would risk of incorrect acceptance or incorrect rejection affect it? If so, which one or both?

[SamCPA]

Thanks. Would risk of incorrect acceptance or incorrect rejection affect it? If so, which one or both?

[JamesBJames]

That goes back to the acceptable level of risk (or confidence level, if you prefer that term). The acceptable level of risk encompasses audit risk, the risk of incorrect acceptance, and the risk of incorrect rejection.

Changing any of these parameters should affect the sample size, as far as I'm aware. For example, if the allowable risk of incorrect acceptance is incredibly low, then our confidence level must be correspondingly high and therefore we need to have a larger sample.

FAR: May 1st, 2014 - 91
AUD: May 29th, 2014 - 97!
BEC: July 16th, 2014 - 91
REG: August 29th, 2014 - 88

Licensed December 2015

Feel free to add me on LinkedIn by clicking my username!

[JamesBJames]

That goes back to the acceptable level of risk (or confidence level, if you prefer that term). The acceptable level of risk encompasses audit risk, the risk of incorrect acceptance, and the risk of incorrect rejection.

Changing any of these parameters should affect the sample size, as far as I'm aware. For example, if the allowable risk of incorrect acceptance is incredibly low, then our confidence level must be correspondingly high and therefore we need to have a larger sample.

FAR: May 1st, 2014 - 91
AUD: May 29th, 2014 - 97!
BEC: July 16th, 2014 - 91
REG: August 29th, 2014 - 88

Licensed December 2015

Feel free to add me on LinkedIn by clicking my username!

[osoderquist]

It's actually the same thing. First decide what is the worst case; incorrect rejection or incorrect acceptance, choose that as your h0-hypothesis, or alpha (sorry couldn't find my greek keys). Then the risk has to be evaluated, 95, 98, or 99.5% certitude, 1, 2 or 3 sigma (still haven't found the keys:). Chose the acceptable risk and document it, as long as your documentation is there you have done your duty, reasonable assurance, or whatever the proper name is. If you're in doubt, increase the sample size.

[osoderquist]

It's actually the same thing. First decide what is the worst case; incorrect rejection or incorrect acceptance, choose that as your h0-hypothesis, or alpha (sorry couldn't find my greek keys). Then the risk has to be evaluated, 95, 98, or 99.5% certitude, 1, 2 or 3 sigma (still haven't found the keys:). Chose the acceptable risk and document it, as long as your documentation is there you have done your duty, reasonable assurance, or whatever the proper name is. If you're in doubt, increase the sample size.

[SamCPA]

Thanks that helps clarify things. I have another question on Attribute sampling now. Is the risk of assessing control risk too low…does it have a inverse or direct affect on sample size?

[SamCPA]

Thanks that helps clarify things. I have another question on Attribute sampling now. Is the risk of assessing control risk too low…does it have a inverse or direct affect on sample size?

[SamCPA]

I also read how the ASSESSED (NOT ACCEPTABLE) level of risk affects the variable sample size in a direct way. Is it because the higher (IR and CR) is the larger we would need our sample to be?

[SamCPA]

I also read how the ASSESSED (NOT ACCEPTABLE) level of risk affects the variable sample size in a direct way. Is it because the higher (IR and CR) is the larger we would need our sample to be?

[Author]

Replies