Analysis of ELISA results with low n?

lb682

Hello,

I am seeking help analysing data from an ELISA I recently ran on 48-hour eluates of some biological products. One replicate was run due to sample availability. 

 

Unknowns were tested for concentration of protein X. There were seven samples total: three "sheet" products and four "liquid" products. Thus, I will have two analyses...one with an n of 3 and one with an n of 4. 

 

I've subtracted background, adjusted for dilution factor, normalized to the original amount of sample used for elution, and made bar graphs. I would like to generate some statistics about these samples, but am not sure I can with such low n's...

 

Please advise. 

 

Peter_Bartell

When you say you want to 'generate some statistics' about the characteristics of interest, usually we start with simple descriptive statistics like means or variances. The issue of sample size is most directly related to interval estimates you might want to calculate for the given 'statistics' you generated, not generating the means and variances directly. As sample size increase, the width of the intervals will decrease. So a 'low n' as you put it doesn't influence your ability to estimate 'statistics'...but has lots to do with any intervals wrt to those statistics.