8

Examples online typically write "F(x)", but that seems confusing to readers.

gung - Reinstate Monica
  • 132,789
  • 81
  • 357
  • 650
David Kelley
  • 171
  • 2
  • 5
  • 1
    I'd usually just use F(x) or occasionally $F_n(x)$ or more rarely $F^*(x)$ if I wasn't confident that it was clear we were discussing an ECDF. Sometimes I have used ECDF(x). Sometimes perhaps something else, depending on context. An ECDF *is* a cdf, so I don't think it's a big problem, but where needed, the $n$ subscript usually suffices to clearly indicate a sample quantity. – Glen_b Nov 02 '14 at 03:58
  • Readers in which field? What can we assume about that field and its readers? The answer depends on the precise question. – Nick Cox Feb 15 '17 at 19:23

1 Answers1

8

For the y axis of an ecdf plot, you can use "Fraction of Data" which is easier to interpret than F(x)

Josh
  • 471
  • 4
  • 15
  • 3
    I think this is a reasonable answer. Others might be "Cumulative probability" and so forth. What's unclear about the question is that "publication" could mean anything from a mainstream statistics publication for which readers might be expected to recognise $F$ as an old friend to a publication in an applied field where these plots are not standard at all. (A horror example of what may need to be explained is that in fields like data journalism and business presentation a scatter plot is widely discussed as a plot that many people will find unfamiliar and difficult to understand.) – Nick Cox Feb 15 '17 at 19:21