Try it out now with the @EstimationStats web application.
https://www.estimationstats.com/#/
-
-
Prikaži ovu nit
-
By the way, I’m looking for a postdoc to help us study synaptic structure and function. https://www.nature.com/naturecareers/job/postdoctoral-fellow-dukenus-medical-school-dukenus-691497 …
Prikaži ovu nit -
Novi razgovor -
-
-
Hey
@adamcchang - awesome paper, I like the concept of visually showing effect size! (Effect size is underappreciated IMO - differences need to be biologically / behaviorally relevant, not just 'significant'). (1/2) -
@adamcchang Do you have any examples of what this might look like with a 2x3 independent samples experiment design (classical two-way ANOVA, say something like genotype x treatment w/ an interaction)? I'm having a hard time visualizing how that would be organized graphically - Još 8 drugih odgovora
Novi razgovor -
-
-
What is wrong with the graph on the left side? Looks perfectly fine to me.
-
Which is why we wrote the paper, because it's very not fine.

- Još 10 drugih odgovora
Novi razgovor -
-
-
Congrats with discovering the
plot.
Novi razgovor -
-
-
Agreed that it's time to set aside ideal distributions and models for real biology. But did you consider or address non-parametric statistics? How about prediction/classification? The real question in almost always: could you predict to which group new, blind measurements belong?
-
Bootstrap is a non-parametric method. Due to the central limit theorem, it produces bell-shaped effect-size curves nevertheless. DABEST also helps visualize differences of medians and Cliff's deltas.
Kraj razgovora
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.