"There is good science out there. It hasn’t entirely ceased."
No. Not entirely. But "it", (science in general), has fallen prey to the loathsome infection that is socialism. Some days we seem well on our way to re-implementing Lysenkoism. Think masks, or six foot spacing, or gene therapies promoted as "vaccines". Look at how governments have assumed virtually total control over research funding. ALL funding.
Then think about all of the research that DOES NOT happen!
I don't think you answered question 3 properly. Here's my take on what your answer would be (having read Uncertainty and followed your blog for some time now).
"Question 3: What type of evidence or studies could alter your opinion that significant changes are needed in the way statistics is practiced?"
There are two parts to the answer.
The first part is about the flaws Briggs has identified with common statistical practices. No evidence or studies could change his opinion on those flaws, because those flaws are logical - not scientific - in nature. Evidence and studies are therefore not the appropriate tool to invalidate his opinion. (Consider how absurd would it be to have a headline like "study shows 1 + 1 = 2"?) Instead, a valid logical argument needs to be made showing how his is opinion wrong.
The second part is about about what evidence one could use to know when statistics is being done right. This evidence would be that all/most models produced by the statistical practices have good predictive power, that is, enough to be appropriate for what they are used for. Which is a little handwavy, but the exact criteria varies a lot. Ultimately the logic doesn't matter if the predictive power is good, as logic is just the tool we use to improve our models. (Though if you reject logic then you are likely to have many headaches in the future when you need to understand or modify the model!)
"There is good science out there. It hasn’t entirely ceased."
No. Not entirely. But "it", (science in general), has fallen prey to the loathsome infection that is socialism. Some days we seem well on our way to re-implementing Lysenkoism. Think masks, or six foot spacing, or gene therapies promoted as "vaccines". Look at how governments have assumed virtually total control over research funding. ALL funding.
Then think about all of the research that DOES NOT happen!
I have more questions. None I can put into words, as statistics has always confounded me. Perhaps I shall read Uncertainty for some clarity.
Thank you for answering my questions. Great stuff as always.
I’ve always felt that using Bayesian theories grants a degree of freedom that doesn’t exist
I don't think you answered question 3 properly. Here's my take on what your answer would be (having read Uncertainty and followed your blog for some time now).
"Question 3: What type of evidence or studies could alter your opinion that significant changes are needed in the way statistics is practiced?"
There are two parts to the answer.
The first part is about the flaws Briggs has identified with common statistical practices. No evidence or studies could change his opinion on those flaws, because those flaws are logical - not scientific - in nature. Evidence and studies are therefore not the appropriate tool to invalidate his opinion. (Consider how absurd would it be to have a headline like "study shows 1 + 1 = 2"?) Instead, a valid logical argument needs to be made showing how his is opinion wrong.
The second part is about about what evidence one could use to know when statistics is being done right. This evidence would be that all/most models produced by the statistical practices have good predictive power, that is, enough to be appropriate for what they are used for. Which is a little handwavy, but the exact criteria varies a lot. Ultimately the logic doesn't matter if the predictive power is good, as logic is just the tool we use to improve our models. (Though if you reject logic then you are likely to have many headaches in the future when you need to understand or modify the model!)