Problems in using p-curve analysis and text-mining to detect rate of p-hacking and evidential value

View article
1752 days ago
@richarddmorey @dstephenlindsay This is one where we had one hostile and 2 positive reviewers - may be useful to see the discussion - paper eventually accepted https://t.co/EktmhSZej2
reply repost on X like
1872 days ago
@dnunan79 @enemiesnet Hmm, not quite. I'm looking for a really easy explainer, preferably with epidemiological examples. (And I wrote a paper with @stat_PT where we took issue with some of the conclusions from Head et al! https://t.co/EktmhSZej2.)
reply repost on X like
1951 days ago
RT @peter1328: @bmwiernik @BoulesteixLaure sure here are some showing up so many issues I can't even count that far (there are quite some m…
reply repost on X like
@bmwiernik @BoulesteixLaure sure here are some showing up so many issues I can't even count that far (there are quite some more but I don't have a good collection): https://t.co/uaPKQbBMOG https://t.co/KOp3YeR3wm https://t.co/hQSAjEm3RN https://t.co/gs6hzWZTQT https://t.co/IKtUnXujff
reply repost on X like
RT @westwoodsam1: #124 food for thought from @deevybee & @stat_PT on the misleading effects of ghost and correlated variables on p-curve an…
reply repost on X like
#124 food for thought from @deevybee & @stat_PT on the misleading effects of ghost and correlated variables on p-curve analyses, and on the limits of text-mined p-values https://t.co/aVNSViBO2g @peanutbuttner you might find this interesting https://t.co/b4dtODpa4J
reply repost on X like
2571 days ago
@wgervais @hardsci I just know from the simulations we did on p-hacking that things get more complex if you have correlated variables. https://t.co/EktmhSZej2
reply repost on X like
2686 days ago
@EJWagenmakers @siminevazire @farid_anvari @mo279_mo I don’t think that’s right if there is really no effect: see our simulations - but you may not get ‘diagnostic ‘ L skew: https://t.co/EktmhSZej2. but big problem in automated analysis if no check on which pvalues included
reply repost on X like
RT @deevybee: Data-mining problematic for this kind of analysis - see Bishop & Thomspon https://t.co/EktmhSZej2 https://t.co/JPsrtXr5q2
reply repost on X like
2968 days ago
RT @deevybee: Data-mining problematic for this kind of analysis - see Bishop & Thomspon https://t.co/EktmhSZej2 https://t.co/JPsrtXr5q2
reply repost on X like
RT @deevybee: Data-mining problematic for this kind of analysis - see Bishop & Thomspon https://t.co/EktmhSZej2 https://t.co/JPsrtXr5q2
reply repost on X like
RT @deevybee: Data-mining problematic for this kind of analysis - see Bishop & Thomspon https://t.co/EktmhSZej2 https://t.co/JPsrtXr5q2
reply repost on X like
RT @deevybee: Data-mining problematic for this kind of analysis - see Bishop & Thomspon https://t.co/EktmhSZej2 https://t.co/JPsrtXr5q2
reply repost on X like
RT @deevybee: Data-mining problematic for this kind of analysis - see Bishop & Thomspon https://t.co/EktmhSZej2 https://t.co/JPsrtXr5q2
reply repost on X like
RT @deevybee: Data-mining problematic for this kind of analysis - see Bishop & Thomspon https://t.co/EktmhSZej2 https://t.co/JPsrtXr5q2
reply repost on X like
2968 days ago
Data-mining problematic for this kind of analysis - see Bishop & Thomspon https://t.co/EktmhSZej2 https://t.co/JPsrtXr5q2
reply repost on X like
RT @PsyBrief: P-curves are useful BUT you have to know where p-values came from to show evidence of publication bias https://t.co/Oy3618uxZ…
reply repost on X like
RT @PsyBrief: P-curves are useful BUT you have to know where p-values came from to show evidence of publication bias https://t.co/Oy3618uxZ…
reply repost on X like
RT @deevybee: @hildabast @BMJRtoP Disagree. This is a silly analysis. You have to know what analysis a p-value came from https://t.co/Ektmh…
reply repost on X like
RT @PsyBrief: P-curves are useful BUT you have to know where p-values came from to show evidence of publication bias https://t.co/Oy3618uxZ…
reply repost on X like
RT @PsyBrief: P-curves are useful BUT you have to know where p-values came from to show evidence of publication bias https://t.co/Oy3618uxZ…
reply repost on X like
3097 days ago
RT @deevybee: @hildabast @BMJRtoP Disagree. This is a silly analysis. You have to know what analysis a p-value came from https://t.co/Ektmh…
reply repost on X like
3097 days ago
RT @deevybee: @hildabast @BMJRtoP Disagree. This is a silly analysis. You have to know what analysis a p-value came from https://t.co/Ektmh…
reply repost on X like
RT @PsyBrief: P-curves are useful BUT you have to know where p-values came from to show evidence of publication bias https://t.co/Oy3618uxZ…
reply repost on X like
3097 days ago
P-curves are useful BUT you have to know where p-values came from to show evidence of publication bias https://t.co/Oy3618uxZR By @deevybee
reply repost on X like