Biasing the input: A yoked-scientist demonstration of the distorting effects of optional stopping on Bayesian inference

2021 ◽  
Author(s):  
Richard B. Anderson ◽  
Jennifer C. Crawford ◽  
Michael H. Bailey
Keyword(s):  
Bayesian Inference ◽  
Optional Stopping

Bayesianism vs. Frequentism in Statistical Inference

2017 ◽  
Author(s):  
Jan Sprenger
Keyword(s):  
Bayesian Inference ◽  
Statistical Inference ◽  
Confidence Intervals ◽  
Mathematical Methods ◽  
Inference Models ◽  
P Values ◽  
Scientific Inference ◽  
Optional Stopping ◽  
Degrees Of Severity

Bayesianism and frequentism are the two grand schools of statistical inference, divided by fundamentally different philosophical assumptions and mathematical methods. Bayesian inference models the subjective credibility of a hypothesis given a body of evidence, whereas frequentists focus on the reliability of inferential procedures. This chapter gives an overview of the principles, varieties and criticisms of Bayesianism and frequentism, compares both schools, taking in an examination of Deborah Mayo’s account of frequentism, an innovative proposal in which she presented as crucial the concept of degrees of severity; and applies them to salient topics in scientific inference, such as p-values, confidence intervals and optional stopping. author OK

How do academics assess the results of multiple experiments?

2018 ◽  
Author(s):  
Olmo Van den Akker ◽  
Linda Dominguez Alvarez ◽  
Marjan Bakker ◽  
Jelte M. Wicherts ◽  
Marcel A. L. M. van Assen
Keyword(s):  
Bayesian Inference ◽  
Secondary Analysis ◽  
Individual Participant Data ◽  
Additional Experiment ◽  
Underlying Theory ◽  
Individual Participant ◽  
Vote Counting

We studied how academics assess the results of a set of four experiments that all test a given theory. We found that participants’ belief in the theory increases with the number of significant results, and that direct replications were considered to be more important than conceptual replications. We found no difference between authors and reviewers in their propensity to submit or recommend to publish sets of results, but we did find that authors are generally more likely to desire an additional experiment. In a preregistered secondary analysis of individual participant data, we examined the heuristics academics use to assess the results of four experiments. Only 6 out of 312 (1.9%) participants we analyzed used the normative method of Bayesian inference, whereas the majority of participants used vote counting approaches that tend to undervalue the evidence for the underlying theory if two or more results are statistically significant.

Sign in / Sign up

Export Citation Format

Share Document