2
$\begingroup$

I usually use mean±SD notation for reporting values. I know that many people use mean (SD) notation and I understand why.

Right now in the document I am writing I am also reporting medians but I don't know which notation I should use for the 3 measures. Any suggestions?

$\endgroup$
4
$\begingroup$

Both conventions you mention are ambiguous - for example you can't tell whether the number after the $\pm$ is a standard deviation, a standard error, or the half-width of an interval. (a problem that has led to numerous questions on this site).

If you're publishing somewhere that offers precise guidelines follow those, but otherwise I suggest being explicit, for example: $``\text{median } 28.5,$ $\text{ mean } 31.2,$ $\text{ standard deviation } 3.6\!"$ -- then there's no possibility of misinterpretation.

$\endgroup$
  • 2
    $\begingroup$ +1. I usually write $``31.2 \pm 3.6\text{ (mean}\pm\text{SD)}\!"$ and hope that this is clear enough... $\endgroup$ – amoeba Mar 16 '15 at 14:38
  • 1
    $\begingroup$ @amoeba That's explicit, so it looks okay to me. I wasn't trying to hold people to any one format, just avoiding ambiguity; yours does that. $\endgroup$ – Glen_b Mar 16 '15 at 14:47
  • $\begingroup$ Well, I end up using 5.0 (5) ± 0.5 (2.5) [0 - 10] for the hypothesis variable (the numbers I show here are not the true values) and inserted a footnote explaining the format Mean (Median) ± SD (Q1-Q3) [Min - Max]. For all other variables, since they are presented in tables, it was inserted a table footnote explaining the format: Mean ± SD [Min - Max] or Median (Q1-Q3) $\endgroup$ – yohanna Mar 23 '15 at 14:55
1
$\begingroup$

I personally usually use Mean (Median) $\pm$ SD (IQR) [Min - Max]. That's quite maximalist, so may be a bit too much for certain purposes/readers, but is a quite logical grouping of indicators (central tendency $\pm$ dispersion, with traditional indicator first, robust alternative in parenthesis), I believe. (Minimum and maximum is of course added not because they're statistically especially important or good indicators, but because on the biomedical field it is often required for ''clinical'' purposes.)

$\endgroup$
  • 1
    $\begingroup$ I actually like this idea, but only if the order is explicitly specified in the paper so the reader knows what they're looking at $\endgroup$ – shadowtalker Apr 30 '15 at 18:48
  • $\begingroup$ @ssdecontrol: Of course! I usually specify it in the - say - table caption. $\endgroup$ – Tamas Ferenci May 1 '15 at 18:55

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.