Should I use binomial or Poisson distribution formula?

M

Mehmet Hascan

Hi All,

I'd be very grateful if you could help me with the following question. Many thanks in advance.

Question:
For a critical defect, the sampling plan is n=1200 and accept/reject = 0/1. If the process is running 0.1% defective, what's the probability of acceptance.

Should I use binomial or Poisson distribution formula?
 

Steve Prevette

Deming Disciple
Leader
Super Moderator
The exact answer would be HyperGeometric, but likely at a 1200 sample size the Binomial, Poisson, and HyperGeometric would all be pretty close.

By the way - is this for a college course or a real world application?
 
M

Mehmet Hascan

Hi Steve,

Thanks a million for your help. Yes, this is for a college course.
Shall I just calculate P(X = 0) to determine what the probability of acceptance is? or shall I calculate P(x = 1) and then P(A) = 1 - P(R)?

Many thanks in advance for your help.
 

Steve Prevette

Deming Disciple
Leader
Super Moderator
I sent you an email in response to the one you sent me. One thing - I would only be able to use the Hypergeometric if I knew the population size, so I would indeed default to the binomial. Again, all three likely will give close to the same answer (try it - I'll "leave the proof to the student").
 
A

aharte

hi, I'm currently stuck on this same question for a study course, could you share how to answer? thanks
 

Miner

Forum Moderator
Leader
Admin
You need to be more specific. In general terms, binomial is used for good/bad (i.e., percentage) scenarios while Poisson is used for count data.

What specifically does your scenario entail?
 

Steve Prevette

Deming Disciple
Leader
Super Moderator
hi, I'm currently stuck on this same question for a study course, could you share how to answer? thanks

I usually asked folks puzzled by this - do you have a situation like a coin flip? You flip a coin once and it either comes up heads or tails. Then its a binomial.

If you are counting things per some unit where you can't identify the individual coin flips (such as defects per 100 square yards of fabric), then you are likely dealing with poisson.

There are some more specific clarifiers, like each coin flip has to be independent from each other, and the probability of heads occurring does not change, but the above is usually sufficient.
 
Top Bottom