11
$\begingroup$

I've run a Kruskal Wallis test, and for some of the questions the p value is not significant. Would I report this in the same way as if it was significant, stating the df, test statistic and p-value? So it would be something like this a Kruskal Wallis test was conducted but the results were found not to be significant H(3) = 2.119, p>0.05 (or would I state the exact p value here (.548))

Improve this question
$\endgroup$
4
  • 3
    $\begingroup$ Do not publish non-significant results. Otherwise, you'll bias publication bias. $\endgroup$
    – Mark L. Stone
    Apr 1, 2017 at 13:49
  • $\begingroup$ But publication bias isn't a good thing in itself? Anyway, I'm not 'publishing' the results so to speak, it's for my dissertation but thanks for your comment $\endgroup$
    – Dragonfly
    Apr 1, 2017 at 13:59
  • 2
    $\begingroup$ Dragonfly, The comment by @Mark is tongue-in-cheek. Remember what day it is, too. BTW, your dissertation is worthless unless it's publishable. $\endgroup$
    – whuber ♦
    Apr 1, 2017 at 19:11
  • 1
    $\begingroup$ Do report it in our thesis. And please do read Andrew Gelman's blog regulary: andrewgelman.com/?s=p+value . I disagree with @whuber - you may have good stuff in your thesis even if it's not publishable. If it's accepted you have your degree. $\endgroup$
    – Ethan Bolker
    Apr 1, 2017 at 23:57

2 Answers 2

Reset to default
18
$\begingroup$

Yes, non-significant results are just as important as significant ones. If you are reporting any result, always include the df, test statistic, and p value. And in that case, you should state the exact p-value, rather than generalising to >0.05

Improve this answer
$\endgroup$
1
  • $\begingroup$ Our answers agree and Conor is correct to point out that you should report the exact p-value if it is available. $\endgroup$
    – Michael R. Chernick
    Apr 1, 2017 at 13:34
7
$\begingroup$

If you are publishing a paper in the open literature, you should definitely report statistically insignificant results the same way you report statistical significant results. Otherwise you contribute to underreporting bias.

Improve this answer
$\endgroup$
1
  • $\begingroup$ But why? This implies insignificant values are just as important as significant ones. Is this because there is still a chance the “insignificant” values are actually significant. Statistical significance is not a definitive claim that something is meaningless. Yes? $\endgroup$
    – user3138766
    Jun 29, 2021 at 17:41

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.