pontusps

Research at Karolinska Institutet and Copenhagen University Hospital, R, modeling, multi-modal imaging, image-analysis and Star Trek. 🤓

[pontusps @ppsigray]

@GeroScienceAGE @jaralaus will make a longer thread about this paper later, but a quick shout out to @MartinBJensen and his @impetusgrants initiative: without Round 1 funding we would not have written this.

[pontusps @ppsigray]

@JamieLarsH @ParkvilleGeek Thanks, very good work! This slipped our radar completely. Sorry about that – we’ll update the manuscript referring to this.

[pontusps @ppsigray]

@richarddmorey @EJWagenmakers @nicebread303 @JeffRouder @TurkuPETCentre @dsquintana But as we talked to people we realised we need to first anchor it something they knew of: which is only being concerned with “proving” H1 and a freakish obsession with 5% fpr. We hope they read on and start thinking about other options as well.

[pontusps @ppsigray]

1/ Me and @jaralaus made a tutorial on how to perform sequential bayes factor testing (using #rstats) to stop a study early in research fields were samples are small and comes with large monetary and ethical costs. https://t.co/QCbwhVdnZZ Below, a thread:

[medRxiv @medrxivpreprint]

Early stopping in clinical PET studies: how to reduce expense and exposure https://t.co/aWgUiNEzJG #medRxiv

[pontusps @ppsigray]

@richarddmorey @EJWagenmakers @nicebread303 @JeffRouder @TurkuPETCentre @dsquintana Yeah, we kind of figured that this would amount to only a transformation of a freq seq-design. The idea was to start there and then introduce what happens when also stopping for H0. Our prime purpose with this was to get the field to start thinking about seq-designs and H0.

[pontusps @ppsigray]

@lakens @jaralaus @EJWagenmakers @nicebread303 @JeffRouder @richarddmorey @TurkuPETCentre @dsquintana We clearly show that power suffers compared to classical NHST when using this approach. And I think that’s enough to turn people away in our field. It’s about showing trade-offs and opening up new inferential options, not proving “superiority”.

[pontusps @ppsigray]

@lakens @jaralaus @EJWagenmakers @nicebread303 @JeffRouder @richarddmorey @TurkuPETCentre @dsquintana Well, I’m not sure classifying the only way inference is made today in the field of positron emission tomography as “a straw man” is correct. We are trying hard to avoid the “stat war issue” that seem to cripple much of psych, and make it into an educational stepping stone.

[pontusps @ppsigray]

14/14 We would love to get some feedback, @EJWagenmakers, @nicebread303, @JeffRouder, @richarddmorey, given time and interest of-course. Maybe also of interest to @TurkuPETCentre, @dsquintana?

[pontusps @ppsigray]

13/ Perhaps weird to discuss “power/error control” when it comes to BF, and many would likely think this is a silly use of the metric. But we think it makes sense when planning to perform a study with seq BF, and it’s a nice gateway for a field that only heard about NHST.