Sample Size / Statistical Significance 1

[LynnB]

New here. My search didn't turn up anything, but if I missed it, please point me in the right direction.

I need know how to determine the number of units we need to test (MIL_STD 810F) in order for the testing to be statistically significant.

All the info I can find on the web talks about marketing surveys, and I'm having trouble translating it into something useful.

Is there a tutorial around, or should I just use 32?

Thanks in advance.

 

[IRstuff]

Have you looked at MIL-STD-105 and MIL-HDBK-108?

TTFN

FAQ731-376

 

[LynnB]

Not yet...

Thanks!

 

[zubaj]

If you're using a statistical package for your analysis, these usually have power and sample size calculators built in.

 

[Schavlan]

I don't have access to a statistical engine for my computer or the military specs mentioned, but I'm looking at the same issue.

In trying to prove out a new batch process, how many batches do I need to run in order to show that I am able to confidently make specifications of the product.

Is 10 enough ?

 

[GregLocock]

That depends on how many there are in a batch, your required confidence, and the relationship between the sample distribution and the acceptable limts. and probably a couple other things.

Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[IRstuff]

http://assist.daps.dla.mil/quicksearch/

TTFN

FAQ731-376

 

[magoo2]

I've found some pretty good info in a report from the USGS on Statistical Methods in Water Resources. Chapter 3 of that report is titled Describing Uncertainty and has a lot of info on confidence intervals which relate to your question about sample size and statistical significance.

Here's the link for that chapter:

http://pubs.usgs.gov/twri/twri4a3/pdf/chapter3.pdf

The entire report is in the public domain and has a wealth of knowledge about statistical methods.

You can read the intro and get to the PDFs from this page:

http://pubs.usgs.gov/twri/twri4a3/

I'm a power engineer so I shouldn't be recommending a reference dealing with water, but I think you'll find this to contain a wealth of good basic statistical info and I think it's quite well written.

 

[HughL]

A little late for you may be but; MS Excel has the Confidence Worksheet function which may be of interest.