That would be my starting point. Then I’d look at the risk of bias and how that might be impacting on the observed association. Study design here already sets my prior to “interesting, but....”
i’ll believe it if it’s a preregistered studyhttps://twitter.com/ShaiCarmi/status/1222593017317601283 …
-
-
-
^ this. Replication and (to a slightly lesser extent) pre-registration do nothing when the flaws are inherent to the study design.
Kraj razgovora
Novi razgovor -
-
-
*laughcry*
-
N = ~350, no main effect of genotype, GxE p-value of *just* under 0.05... I think we've been here before.
- Još 1 odgovor
Novi razgovor -
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.