Median of discrete and continuous random variables.

9k Views Asked by At

If case of continuous random variables, we define median 'm' as a value such that the $P(X\ge m)=0.5$ and $P(X\le m)=0.5$. This link also mentions the same

https://www.google.co.in/url?sa=t&source=web&rct=j&url=https://www.wiley.com/legacy/Australia/Landing_Pages/c12ContinuousProbabilityDistributions_web.pdf&ved=0ahUKEwiCvLm0va3YAhUBro8KHUaYA-sQFgg6MAI&usg=AOvVaw1qTWw5kE9ZsgWJ6RtMjngD

For discrete random variables, $P(X\ge m)\ge0.5$ and $P(X\le m)\ge0.5$

Why is the cumulative probability required to be greater than or equal to $0.5$ in case of discrete random variables and only equals to $0.5$ in case of continuous random variable?

2

There are 2 best solutions below

0
On BEST ANSWER

For strictly continuous random variables the value $m$ is unique and always in the support (the set of possible values) of $X$, thus there is no problem to define it as $P(X\ge m) = 0.5$. In discrete random variable whether with finite or infinite countable support, it is possible that no possible value of $X$ will satisfy $P(X\ge m )=0.5$. Let us consider a simple example. Let $X \sim \mathcal{B}in (3, 1/2)$, hence we have to find $m$ such that $ P(X \le m ) =0.5 $. Let us check possible value of $m$ $$ P(X\le 1) = \frac{1}{2^3} + \frac{3}{2^3} = 1/2. $$ Bingo! $1$ is the median. Now, take $p=1/4$ instead of $1/2$, $$ P(X\le 1) = \frac{3^3}{4^3} + \frac{3}{2^3} = 0.74, $$ that is way too high probability, however $$ P(X\le 0) = \frac{3^3}{4^3} = 0.42. $$ That is too low. I.e., there is no value that $X$ can get and it will satisfy $ P(X\le m) = 0.5, $ hence you have to chose $0$ or $1$ as the median. Any value in $(0,1)$ will not change a thing as $X$ cannot receive these values.

1
On

It is possible in the discrete case for the X=m not be possible. Simple example: toss one die, possible values of X 1,2,3,4,5,6. Median is 3.5.