14/14 We suspect that similar studies remain unpublished or hidden in the "grey" literature - often in non-English languages. One reviewer cited a nice Swedish study as an example https://shorturl.at/DGPRU.

A negative result is a result they say. But good luck to publish that… #PublicationBias 📚🔍

@colatkinson

AHA - I found the thingy I was on about, of course it's #BenGoldacre :

https://www.ted.com/talks/ben_goldacre_what_doctors_don_t_know_about_the_drugs_they_prescribe?language=en

Over a decade old now, nothing's changed.

#BadPharma #PublicationBias

Evidence from psychology shows that #registeredreports have a lower positive result rate than the standard literature & likely reduces #publicationbias & QRPs

📰by annescheel@mastodon.online Mitchell Schijen & @lakens https://doi.org/10.1177/25152459211007467

Time for eco & evo to support Registered Reports!

Beyond #rationality?
https://daviddfriedman.substack.com/p/beyond-rationality

#RationalChoice #BoundedRationality
#science #research #publishing #publicationbias #preregistration

Nature introduces Registered Reports as an article type for cognitive neuroscience and the behavioural and social sciences. This approach assesses the research question and methods over the results, to help tackle the publication bias against negative results.

#PublicationBias #NeuroMastodon #AcademicMastodon #NegativeResults #SocialScience

@mjb @socialpsych They could probably use that skewed distribution in text books to illustrate publication bias! #PublicationBias #MetaAnalysis #SystematicReviews #Psychology

RT @sortecoevo@twitter.com

From #SORTEEvoices summary blog: https://www.sortee.org/blog/2022/12/02/2022_sortee_anniversary/

Which open research practice do you think deserves more attention?

SORTEE members answered:
- Code- and data-sharing to increase verifiability and reusability of research
- @RegReports@twitter.com to fight #publicationbias

🐦🔗: https://twitter.com/sortecoevo/status/1601124489374285824

From #SORTEEvoices summary blog: https://www.sortee.org/blog/2022/12/02/2022_sortee_anniversary/

Which open research practice do you think deserves more attention?

SORTEE members answered:
- Code- and data-sharing to increase verifiability and reusability of research
- Registered Reports to fight #publicationbias and threats to reliable science

Five years ago I predicted that
1 there will be mandatory #preregistration of #behavioraleconomics
#experiments
2 we will soon discover #publicationbias and less than satisfactory levels of #replication

https://twitter.com/renebekkers/status/938171836327743488

Findings from 2016 gave reason for #optimism https://www.science.org/doi/epdf/10.1126/science.aaf0918

but subsequent replication projects such as https://doi.org/10.1038/s41562-018-0399-z and https://doi.org/10.1016/j.joep.2019.102199
were less positive.

prediction #1 was clearly too optimistic.

https://renebekkers.wordpress.com/2022/12/03/ten-meta-science-insights-to-deal-with-the-credibility-crisis-in-the-social-sciences/

@rausch Right, one would need to assume a certain #publicationBias process to derive the quantity we are interested in. But this information is helpful to calibrate our #ABM model: After applying publication bias, similar numbers should show up in our model.

#Iguana #Zcurve #PublicationBias #SocialPsychology #Phacking #CFA #LatentVariables #ESM #Psychology #MetaScience #Personality #Statistics

Does anyone have recommendations on how to examine #PublicationBias in MASEM (#MetaAnalysis + #SEM)? I haven't seen respective applications. Currently, I am thinking about conducting a multivariate meta-regression of the correlation matrices on their respective standard errors (similar to PET/PEESE analyses) and then fitting the SEM to the "corrected" pooled correlations. But I am not sure about this idea...

@xrcologne @XRBerlin @xrgermany

Also das sagt #EricaChenowth:

"...these findings are qualified by data constraints"

"...not common enough for there to be a real pattern from which to draw inferences."

"...this effect is not statistically significant.."

"In sum, having an armed wing is risky, but not necessarily decisive."

" terrorist groups targeting civilians lose public support compared with groups that limit their targets to the military or police"
...und vieles mehr.

Hast du/ihr überhaupt irgendwas davon gelesen oder wird einfach stumpf #RogerHallam nachgeplappert.
[ich bin verärgert über deine/eure Resistenz sich mit dem Thema ernsthaft auseinander zusetzten]

Aber ja, nur weil es Kritik gibt heißt es nicht das eure Grundannahmen falsch sind, sondern weil sie widerlegt wurden.

zum Thema "Studien sind wissenschaftlich und deswegen müssen die stimmen", ist es empfehlenswert sich mit folgenden Themen auseinander zusetzten:

- #Publicationbias
https://en.wikipedia.org/wiki/Publication_bias
- #Reportingbias
https://en.wikipedia.org/wiki/Reporting_bias
- #Mediabias
https://en.wikipedia.org/wiki/Media_bias

Les ravages de la fétichisation des p-valeurs et du biais de publication, en une diapo.
(via http://audimath.math.cnrs.fr/paradoxes-aleatoires.html )
#pvalue #publicationBias