In Wikipedia:


it says that $X_1+\cdots+X_n$ is a sufficient statistic for the parameter of the Poisson distribution and its proof follows by using the factorization theorem. However, the expression they obtain is:

$$ e^{-n\lambda}\lambda^{(x_1+x_2+\cdots +x_n)}\cdot {1 \over x_1!x_2!\cdots x_n!}$$

which also depends on $n$. Am I correct by assuming that this proof is flawed?

I know another proof using another result, but I wonder about the correctness of this claim in wiki.

  • 4
    $\begingroup$ This is not clashing with the definition of sufficiency. $\endgroup$ – Xi'an Jan 25 '19 at 18:59
  • 2
    $\begingroup$ You misread Wikipedia: the formula you quote is the probability for the vector $(x_1,x_2,\ldots,x_n).$ Please read on from that statement through the end of their demonstration that the likelihood depends only on the mean of the $x_i$ (or, equivalently, their sum, because $n$ is known). $\endgroup$ – whuber Jan 25 '19 at 20:12

A sufficient statistic is sufficient for a particular family of probability distributions, and in this case that family is actually not the family of Poisson distributions, each member of which is supported on the set $\{0,1,2,3,\ldots\},$ but rather it is a family of distributions supported on the set $\{0,1,2,3,\ldots\}^n$ of all $n$-tuples of finite cardinal numbers. Thus the random variable involved is an $n$-tuple $(X_1,\ldots,X_n)$ in which the components are independent and each has a $\operatorname{Poisson}(\lambda)$ distribution.

Within the set of all such $n$-tuples one finds the observable data $(x_1,\ldots,x_n)$. If the observable data are altered within that set, the number $n$ does not change; only the $x\text{s}$ change. Thus the proof is not erroneous.

If the number of independent identically distribued Poisson observations were a random variable in its own right, then one would have a different family of distributions and the sum of the observations would probably not be a sufficient statistic.


The proof of the factorization theorem is based on the sample size (the conditional distribution of the sample given the test statistic, of size $n$, for each $n$), thus, it is specific to each $n$. For this reason, the sample size $n$ is considered fixed/known.

Most "answers"/"comments" here focus on claiming "yes"/"obvious", without scientific basis, but this provides a more formal answer.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.