> Using both Frequentist and Bayesian methods, we compared the groups and explored the associations of strengths knowledge and use with outcomes across both groups
This is such a strange statement to me. If you scroll down you see they mean they calculated statistics like cramers v or Bayes factors. I think it’s so strange to differentiate things into frequentist and Bayesian buckets. Like I get it, the authors want to do some kind of group wise comparison like an ANOVA or ANCOVA. But it’s all the same data. The only thing that matters is demonstrating the differences to test the hypothesis. I am sure this is field dependent. I have published a lot in physics education research. In that field there are some people who attack anyone who doesn’t use some hierarchical Bayesian modeling or mixed effects or random effects modeling because they think that’s best. I even wrote this paper:
Arguing that the type of model doesn’t matter what matters is your ability to recreate the underlying distribution. Once you have demonstrated that ability, then you can interrogate your model to see what it knows about your data.
I guess this is more of a comment than a question. And I’m trying to provide myself context as to why the authors would differentiate between frequentist and Bayesian methods to do group wise comparisons.
> Using both Frequentist and Bayesian methods, we compared the groups and explored the associations of strengths knowledge and use with outcomes across both groups
This is such a strange statement to me. If you scroll down you see they mean they calculated statistics like cramers v or Bayes factors. I think it’s so strange to differentiate things into frequentist and Bayesian buckets. Like I get it, the authors want to do some kind of group wise comparison like an ANOVA or ANCOVA. But it’s all the same data. The only thing that matters is demonstrating the differences to test the hypothesis. I am sure this is field dependent. I have published a lot in physics education research. In that field there are some people who attack anyone who doesn’t use some hierarchical Bayesian modeling or mixed effects or random effects modeling because they think that’s best. I even wrote this paper:
https://link.aps.org/doi/10.1103/PhysRevPhysEducRes.17.02010...
Arguing that the type of model doesn’t matter what matters is your ability to recreate the underlying distribution. Once you have demonstrated that ability, then you can interrogate your model to see what it knows about your data.
I guess this is more of a comment than a question. And I’m trying to provide myself context as to why the authors would differentiate between frequentist and Bayesian methods to do group wise comparisons.