1
$\begingroup$

If I plot a symmetrical errorbar (let's say SEM = 0.5), should it show 0.5 on each side, or 0.25 (half) on each side?

$\endgroup$
2
  • $\begingroup$ If you work under the assumption of approximate normality (usually you do), you should multiply by 1.96 which is the rule of thumb of multiplying by 2, as this corresponds to a symmetric two-sided 95% confidence region. $\endgroup$
    – Alex2006
    Commented Jun 26, 2018 at 17:02
  • 1
    $\begingroup$ The error bars should be the mean plus or minus the standard error of mean, so you would not halve the values. $\endgroup$ Commented Jun 26, 2018 at 17:34

1 Answer 1

1
$\begingroup$

It should ideally show twice on each side, representing an approximate 95% confidence interval for the true value of the parameter. This makes it easy to see for which possible values you have enough evidence to claim are not candidates for the true parameter value. You 100% should not display halved error bars, as this is never done and is misleading. Display the full error bar on each side is fine, but doesn't really convey much information, and many readers will simply double its length in their head to approximate a 95% confidence interval.

$\endgroup$
3
  • $\begingroup$ So, if I understood you correctly, I should actually double it on each side? I.e., show 1.0 on each side? $\endgroup$
    – RV1994
    Commented Jun 26, 2018 at 16:46
  • 1
    $\begingroup$ If you work under the assumption of approximate normality (usually you do), you should multiply by 1.96 which is the rule of thumb of multiplying by 2, as this corresponds to a symmetric two-sided 95% confidence region. $\endgroup$
    – Alex2006
    Commented Jun 26, 2018 at 17:03
  • 2
    $\begingroup$ I'm not as convinced as @Noah is that the confidence interval is necessarily the most usual, best, or natural value to use for error bars. It's pretty common to report standard error of the mean or other measures of dispersion. In all cases, this is why it's important to say what the error bar represents in the figure caption. That being said, I probably do agree that the confidence interval is the most useful. $\endgroup$ Commented Jun 26, 2018 at 18:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.