Hypothesis Testing: Poisson Distribution

Рет қаралды 3,840

11 жыл бұрын

Watch more tutorials in my Edexcel S2 playlist: goo.gl/gt1up
This is the fourth in a sequence of tutorials about hypothesis testing. I explain how to carry out a hypothesis test about the parameter lambda of a Poisson distribution. The video includes examples where a Poisson distribution is used to approximate a binomial distribution, and where a normal distribution is used to approximate a Poisson distribution.
Tutorials on hypothesis testing
Introduction: kzfaq.info/get/bejne/p7iKrbeZlpjaiKs.html
Critical Regions: kzfaq.info/get/bejne/ich5rdhmnqycXXk.html
Binomial Distribution: kzfaq.info/get/bejne/euCTlNKf2MXSlqs.html
Poisson Distribution: kzfaq.info/get/bejne/p9qDgNyI2LvQep8.html
Past Paper Questions: kzfaq.info/get/bejne/hruVibyk1qnXd30.html
Watch more tutorials in my Edexcel S2 playlist: goo.gl/gt1up
Visit my channel for other maths videos: kzfaq.info
Subscribe to receive new videos in your feed: goo.gl/7yKgj

Пікірлер: 7

@MrNichollTV 10 жыл бұрын

Good question! It should be ok to use both approximations if the conditions for both are satisfied. On the other hand, I have come across at least one Edexcel past paper question (S2 June 2002 q4) where both approximations appeared ok but the mark scheme only allowed the Poisson approximation. My advice is to use the Poisson approximation if you have a choice. It is easier (because you don’t have to worry about making a continuity correction or standardising) and usually more accurate.

@mengshuliu 11 жыл бұрын

Excellent videos! Thanks!

@MrNichollTV 11 жыл бұрын

I am very well, thank you. With a two-tailed test you think about the expected outcome. Here the expected outcome is 7.5 and the actual outcome is 2. Since 2 is less than 7.5, the more extreme values are 0 and 1, so you find the probability of getting less than or equal to 2. If the actual outcome were 10, the more extreme values would be 11 and higher, in which case you would find the probability of getting 10 or more. Hope this helps. Good luck tomorrow!

@sammanirox 10 жыл бұрын

In the Wi Fi problem why use poisson approximation to the binomial?Can't we use norml approximation to the binomial as np>5 and n(1-p)>5

@Virus95j 11 жыл бұрын

Hey Mr Nicholl, how are you?. Just wanted to ask, when it comes to two-tailed tests, how would I know if X is less than or equal to x or more than equal to x ?For example at 11:45 you have made X be less than or equal to 2. In one tailed tests I find it simple to designate the sign to the X beacuse the question actually tells you, however in two-tailed you get no hint of which way round the sign should be.Thanks very much Mr Nicholl. just a bit confused :s