Control limits for varying subgroup sizes

I'm trying to create a control chart. I have 4 sample groups with different numbers of parts in each group. When trying to construct a control chart and picking the right constant I need "n" the problem is, is that n is different for each sub group. Can I not construct the control chart using this forumula if each subgroup has a different number of parts?



DaRe DaVil

Re: Reliability of a System - Estimation of R(t) having constant failure rate "Lambda

What are you trying to accomplish? Is this a test of the difference between the four groups (better done by test of hypothesis / ANOVA) or do you have a time sequence going on and you want to analyze the time trend in failures (SPC)?