I know that you can to quite a degree of accuracy use an approximating normal distribution to estimate a binomial distribution with large value of n and a value of p which is close to 0.5
I was wondering however if the normal distribution become a perfectly accurate representation of the binomial given that p=0.5 for any value of n.
Any /stat/ pros willing to help me out here?
>>8971623
no it doesn't. it just mean you need a much smaller n to get a good approximation
>>8971631
Alright thanks
binomial distribution is discrete, normal distribution is continuous, so even as n approaches infinity, the normal approximation won't be equivalent to the binomial.
>>8971639
Even if you are only approximating integer values of n? Of course with continuity correction
>perfectly accurate
no. in fact, it's strictly impossible. binomial is a discrete distribution, normal is continuous.
you will fail s2