Thread: Poisson approximated by Normal for large numbers?

Just a quick question:

Am I correct in thinking that X~Poi(36) may be approximated using X~N(36,36)? What would be a sufficiently large number to do this?

Cheers,

Ivar

Originally Posted by Ivar_sb

Just a quick question:

Am I correct in thinking that X~Poi(36) may be approximated using X~N(36,36)? What would be a sufficiently large number to do this?

Cheers,

Ivar

It depends on what you are trying to do. For rough and ready calculations 10 will often do, 20 is better. Google for "error normal approximation poisson".

CB