The Student Room Group

S2 - Hypothesis Testing Quick Question

Hey guys.

I was just wondering, when doing a hypothesis test for the poisson distribution, which values of lambda should I quote for H0 and H1?

Say I had a question which said, a manufacturer produced 4 defective products out of every 100.

Then I was testing for a sample of 300, would I say H0: Lambda = 4, or H0: Lambda = 12.

Do I quote lambda with respect to the original observations or with respect to the sample?

The mark schemes seem to vary with the answers they give. :s

Thanks
Reply 1
I would say you'd go for Lambda=12 for this question. You do it in relation to the sample as you are doing a hypothesis test on the sample not the original observations. If you used 4 as your lambda in a sample of 300 the results would be incorrect.
If you are solving the problem with the context involving a sample of 300, then use λ=12.

Unless you are further required to perform calculations based on a sample of 100, then λ=4 . Do make sure you assign different labels to the random variables describing the two distinct scenarios.

Hope it helps. Peace.
Reply 3
Original post by Lizzie232
I would say you'd go for Lambda=12 for this question. You do it in relation to the sample as you are doing a hypothesis test on the sample not the original observations. If you used 4 as your lambda in a sample of 300 the results would be incorrect.



Original post by WhiteGroupMaths
If you are solving the problem with the context involving a sample of 300, then use λ=12.

Unless you are further required to perform calculations based on a sample of 100, then λ=4 . Do make sure you assign different labels to the random variables describing the two distinct scenarios.

Hope it helps. Peace.


Alright thanks guys, thats what I thought too. Mark schemes can be confusing though with what they want
Original post by Hamburglar
Alright thanks guys, thats what I thought too. Mark schemes can be confusing though with what they want


No problem. Peace.

PS: Your nick is real awesome.

Quick Reply

Latest