John Flournoy @flourneuro 2663 days ago

Why hello problematic distribution of test statistics in economics. Also, tools to help thinking about underlying models of published test-stat distributions c/o @lakens https://t.co/7dnTlWCV2K https://t.co/5L2uTKxrfL

Daniël Lakens @lakens 2664 days ago

@actpredictlab @rogierK @StuartBuck1 @f2harrell @revodavid @StatGarrett No, because https://t.co/HNyInPv1xY :)

Daniël Lakens @lakens 2664 days ago

@aserebrenik Lakens, D. (2015). On the challenges of drawing conclusions from p-values just below 0.05. PeerJ, 3, e1142. https://t.co/m7iSjbZAO5.

Daniël Lakens @lakens 2664 days ago

@rogierK @LucioMM1 @StuartBuck1 @f2harrell @revodavid @StatGarrett I discuss this briefly somewhere in https://t.co/HNyInPv1xY

Daniël Lakens @lakens 2664 days ago

@jonatanpallesen See https://t.co/HNyInPv1xY

Daniël Lakens @lakens 2849 days ago

@statsepi This is completely wrong, and the author is knowingly citing very selectively. See https://t.co/7uCWBPg2z4 and https://t.co/HNyInPv1xY. The comments in this section by the author are completely unfounded.

PsychBrief @PsyBrief 3015 days ago

More p-values just under 0.05 threshold doesn't necessarily mean more p-hacking; could b decreasing power https://t.co/qLLHG7uk1M by @lakens

Brian D. Earp @briandavidearp 3018 days ago

On the challenges of drawing conclusions from p-values just below 0.05 [PeerJ] @lakens https://t.co/HhSUg5RJHw

Brian D. Earp @briandavidearp 3018 days ago

RT @lakens: @briandavidearp I discuss this paper, quite critically, here: https://t.co/HNyInPv1xY

George Siemens @gsiemens 3018 days ago

RT @lakens: @briandavidearp I discuss this paper, quite critically, here: https://t.co/HNyInPv1xY

Daniël Lakens @lakens 3018 days ago

@hansijzerman See https://t.co/HNyInPv1xY

Daniël Lakens @lakens 3018 days ago

@briandavidearp I discuss this paper, quite critically, here: https://t.co/HNyInPv1xY

Daniël Lakens @lakens 3129 days ago

@timothycbates @uri_sohn For data on this, see https://t.co/HNyInPv1xY - marginally significant results more likely published than ns, but decreasing over years.