Mayo's work on severity has been a guiding principle in my philosophy on statistics, and consequently, in my work on how to design better studies (such as when and why to preregister https://t.co/21k6F1tKA7) - go and listen! https://t.co/wPc9oeUQib
One of my articles I like best is my conceptual analysis of the value of preregistration. It is common to be conceptually vague about the function preregistration has, which leads to endless discussions because it is easy to attack an incoherent argument. https://t.co/XKGmbntfyD
@JessicaHullman I think the most coherent view on this comes from Mayo's work on severe testing and an error-statistical philosophy of inferences. I would recommend her 2018 book. I translate this idea to preregistration here (see screenshot) https://t.co/DbU4g0hrPj https://t.co/9GvtsXXp2n
@RafMBatista @OlegUrminsky @uri_sohn @jpsimmon @m_sendhil @betsylevyp @alexoimas See the answer in my follow up tweets (and to repeat:) my paper here https://t.co/bkuQjA5MSC and especially the section in this screenshot https://t.co/kfqx6pxWEX
@RafMBatista If you want to read more about this, my best explanation is in: https://t.co/bkuQjA5MSC The screenshot is exactly the scenario you described in these tweets, so hopefully that discussion helps: https://t.co/w9skSnhPKT
You always need to evaluate the severity of tests, and researchers should discuss test severity more explicitly. But it gets easier as you learn more about your topic of interest, and pre-reg makes the evaluation process easier throughout. For more, see https://t.co/DbU4g0hrPj
@rfarmer27 @eba_psych @OSFramework If you want to be state-of-the-art, you can try to write a machine readable hypothesis test: https://t.co/jdN5esl7mn. And you might want to read what preregistration is actually for - what it achieves, and if that is even what you want to achieve: https://t.co/21k6F1tKA7
@WynekenHenry @tunc_necip @uygun_tunc The answer is, as always, it seems: It depends. I discuss a related hypothetical situation in my paper on preregistration and test severity: https://t.co/2nIvp3nYnW https://t.co/5pSAwSc6dg
Most papers on preregistration are still very vague on what it is and does. In my conceptual analysis, what preregistration does is very limited, but very useful for certain philosophies of science. Accurately defining things prevents a lot of confusion.
https://t.co/21k6F1tKA7 https://t.co/z4eysM3bt9
The one year embargo of my paper "The value of preregistration for psychological science: A conceptual analysis" is lifted, and from now on it is available Open Access from Japanese Psychological Review: https://t.co/21k6F1tKA7 #openaccess
Special issue titled "The new style of psychological research: CHANGE we can believe in" edited by @asarin who did so much to introduce ways to improve research in Japan. Some articles in English, such as this on how to cheat with preregistration: https://t.co/L6hEFUwxjM https://t.co/dIrAPjhp3C
This paper, published in a Japanese local journal (but written in English!), attempted to develop a scale with Madagascar farmers. It nicely demonstrates weiredness of Likert scale for non-WEIRD ppl. Likert assume lots of things for responders.
https://t.co/LQgFKRpC7h
@IoanaA_Cristea This latest from @lakens is relevant here: https://t.co/hN4wGg55ka
HT @IMLahart
There’s the general issue of preregistration for pilot/exploratory/audit/descriptive studies vs. studies testing hypotheses/making inferences. https://t.co/VavPecVZkS