Test of Comparison

[JohnK]

Is it alright to do a comparison on the reliability of two components if you have different amounts of data for the two component?

John

[Arai.M]

Comparison on reliability usually includes directly or indirectly the use of confidence bounds, now these inlcude a variance measure that should reflect the amount of data you have. As the number of data increases, your bounds on Reliability should get narrower. It should be fine then to compare components with different sample size as long as your test of comparison has an indicator of the spread of the data (e.g. a simple mean comparison where the standard deviation is assumed to be equal would not be representative)