I get confused on the proper notations of meanings, as well as the meanings of some notations relating to random variables and their distributions. Below, I will list things that I think are true, as well as things that I don't understand, and I would love input/corrections. I have labeled each point/question with a number for ease of reference. If it is not appropriate to list items in a single question like this, please let me know. I thought it would be ok since they are all short.

  1. A random variable is notated by a capital letter, e.g. $X$.

  2. What does an operation on a random variable mean? (e.g., how do you interpret $X^2$ in words?).

  3. A specific draw from a random variable is notated by either the lowercase letter (e.g. $x$) or the lowercase letter with a subscript (e.g. $x_1$) or an uppercase number with a number(e.g. $X_1$).

  4. The random variable that is the $kth$ order statistic of $n$ draws from a random variable $X$ is notated as $X_{kn}$.

  5. Is there a shorthand way to write "X is the random variable that is distributed by F(x) (or "cdf F(x)" or "B(a,b)" or any way to characterize a distribution)"?

  6. Can I write $\mathbb{E}F(x)$ to mean the expectation of the variable distributed according to $F(x)$?

  7. If I perform an operation on a variable X's cdf, for example, $F_{new}(x) = F_{old}(x)^2$ to get the cdf of the maximum of 2 draws from $X$, can I notate that in terms of $X$ somehow?

  8. Is the appropriate way to write $(F(x))^2$ succinctly $F^2(x)$ or $F(x)^2$?

  9. Is there any notational difference between a discrete and a continuous variable?

  • 2
    $\begingroup$ Karl already summarized everything perfectly, I just want to add that $EF(x)$ is understood as a expectation of random variable $y=F(x)$, where $x$ is the random variable. If $x\sim F$, then $F(x)$ is uniformly distributed in interval $[0,1]$, so $EF(x)=1/2$, for any $x\sim F$. Definitely not the kind of definition you would want to use :) $\endgroup$
    – mpiktas
    Oct 14, 2011 at 11:28

1 Answer 1

  1. I like to say: a random variable assigns a number to each possible outcome of a random "experiment", where a random experiment is some well-defined process with an uncertain outcome.

  2. $X^2$ is another random variable; whenever $X = x$, $X^2 = x^2$.

  3. I would generally use lower cases letters as realizations of random variables. I wouldn't use $X_1$ this way; it would be another random variable.

  4. I wouldn't talk about $n$ draws from a random variable. I would talk about $n$ draws from a distribution, which would give $n$ independent and identically distributed random variables, $X_1$, ..., $X_n$. I would generally write the $k$th order statistic not as $X_{kn}$ but as $X_{(k)}$, and note that it is a random variable.

  5. You generally write $X \sim F$ to say $X$ is a random variable with distribution $F$.

  6. I've never seen that notation for the mean of a distribution. I'd say $\mathbb{E} X$ where $X \sim F$.

  7. I would just write $Y = \max(X_1, X_2)$ where $X_i \sim \text{iid } F$.

  8. I guess either might be understood, but probably $[F(x)]^2$ is most clear, and while it's more cumbersome to type, it doesn't really take up much more space.

  9. There's not generally a notation difference between discrete and continuous variables, except that you generally wouldn't choose $N$ to be a continuous random variable.

  • $\begingroup$ Thanks so much, Karl! One question on #5: Is "F" a cdf, or a pdf, or the name of known (parameterized?)distribution, such as U(0,1) or B(a,b)? $\endgroup$
    – OctaviaQ
    Oct 11, 2011 at 19:22
  • 1
    $\begingroup$ @JandR - you could use any of those (generally upper-case for cdfs and lower-case for pdfs), since the cdf implies a particular pdf and vice versa. $\endgroup$
    – Karl
    Oct 11, 2011 at 19:40
  • $\begingroup$ So, I could also say $X \sim f$ and it would be inferred that $f$ is a pdf? Thanks! $\endgroup$
    – OctaviaQ
    Oct 11, 2011 at 19:41
  • 1
    $\begingroup$ @JandR - probably, but usually you'd be explaining $f$ further anyway, and it'd be better to be precise (though more verbose) in saying "$X$ is a random variable with pdf $f$." $\endgroup$
    – Karl
    Oct 11, 2011 at 19:45
  • 1
    $\begingroup$ @JandR - that was a bad typo; I wrote "not" but meant "note" $\endgroup$
    – Karl
    Oct 13, 2011 at 23:08

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.