Multiplying binomial distributions

2.9k Views Asked by At

Suppose I have X ∼ Bin(1,p). Is the distribution of X^2 the same as X?

I know that X+X ∼ Bin(1+1,p), which would lead me to the (probably incorrect) assertion that X^2 ∼ Bin(1*1,p).

I also know that if for Bin(n,p), where n = 1, then only one trial happens. Is this the only situation in which X^2 has the same distribution as X?

1

There are 1 best solutions below

2
On BEST ANSWER

Yes, when $X$ is a Bernoulli random variable $X^2$ has the same distribution; it has the same support $\{0,1\}$ and realises each outcome with the same probability.   Moreover, as Robert comments, it is not just the same distribution, this makes it the same random variable.

However, $X^2$ does not generally have the same distribution as $X$ when $X$ is the count of successes in a series of iid Bernoulli trials.   This can be seen by examining the supports, as $X$ has the support of $\{0,1,2, ..., n\}$ while $X^2$ is supported on $\{0,1,4, ..., n^2\}$.   Only when $n=1$ are these identical.

Further $2X\nsim\mathcal{Bin}(2,p)$ for the same reason.

If you had an independent and identically distributed variable, $Y\sim\mathcal {Bin}(1,p)$, then $X+Y\sim\mathcal{Bin}(2,p)$.   But $X$ is clearly not independent from itself.