28 replies to this topic

[Pizza&Sandwich]

I have taken some product samples: 10 using gloves, 10 without using gloves

 

How can I determine if there's a significant statistical difference?

 

The data is attached in Excel

 

 

Attached Files

•  Gloves.xlsx   9.2KB   32 downloads

Edited by Lacey, 04 December 2013 - 04:49 PM.

[Mr. Incognito]

Lacey,

 

What are you testing, what are you testing it with?  10 data points isn't a very good sampling pool so it's hard to be able to interpret your data without missing information.

 

I'm guessing it's a charm swab or a total plate count test.

 

Also when was it taken?  What are the conditions of the test? 

Edited by MerleW, 04 December 2013 - 05:46 PM.

[Charles.C]

Dear Lacey,

 

As per previous post, the actual experiment is so far unknown other than gloves/no gloves were used.

One typical statistical approach for a controlled  comparison of this type is a null hypothesis that using gloves makes no difference to the result (eg the mean).

Offhand, ie by eyeing the data, I (courageously) predict this is not the case.

Perhaps a t-test with appropriate significance levels, or a distribution-free alternative.

But it may depend on the experiment.

 

Rgds / Charles.C

[FurFarmandFork]

Thought I'd add this to the thread. Good article on some tips and tricks for compliance (note that one of them is the multi-glove use idea, which typically isn't accepeted by GFSI or regulators). 

 

http://magazine.qual...compliance.aspx