Will a paired t-test work for my lab data?

Question

I have a set of 10 samples, each having an unknown concentration of analyte. Each sample is quantitated by method A and method B. I need to know if the quantitative results from method A differ significantly from method B. Can I use a paired t-test for this?

score 1 · Answer 1 · answered Dec 01 '16 at 16:22

1

Having a sample size of only 10, I would suggest using a non-parametric test such as the Wilcoxon signed-rank test.

answered Dec 01 '16 at 16:22

user140571

71
4

score 1 · Accepted Answer · edited Apr 13 '17 at 12:44

1

I think you might want to investigate other ways of doing this like the Bland and Altman plot which is discussed in a long and detailed answer by Stephan Kolassa in Creating and interpreting Bland-Altman plot This method has become very widely accepted in health although I do not know about other arenas.

edited Apr 13 '17 at 12:44

Community

1

answered Dec 01 '16 at 16:54

mdewey

16,541
22
30
57

Thanks, I like the look of these plots and will try it out. – 100000eV Dec 01 '16 at 18:09

Will a paired t-test work for my lab data?

2 Answers2