Erik J. Ringen @err_ring profile

Erik J. Ringen

@err_ring

Followers

322

Following

2K

Statuses

215

Data Scientist (Applied Statistician) at the University of Zurich PhD in Bio-Anth. Behavioral ecology, cultural evolution, statistics & causal inference.

Zurich, Switzerland

Joined January 2018

Don't wanna be here? Send us removal request.

Erik J. Ringen

@err_ring

3 months

RT @mcmc_stan:

0

6

0

Erik J. Ringen

@err_ring

3 months

New blog: If you use observation-level random effects, your predictive checks are probably wrong. Here's how to do it the right way in both #brms and #PyMC tl;dr: you need to sample new levels of the OLRE to avoid a false goodness-of-fit

0

1

2

Erik J. Ringen

@err_ring

3 months

RT @PinkRangerLB: Twin Peaks posting again

0

2K

0

Erik J. Ringen

@err_ring

3 months

RT @ChelseaParlett: It turns out you can’t teach a decades worth of stats experience/learning in a single semester

0

14

0

Erik J. Ringen

@err_ring

3 months

It feels like whenever OLS comes up on this site, one camp is talking about model checking (residuals, predictive plots) and the other about standard errors. Both talk past each other. Onlooks become more confused.

Dan

@dan_p_simpson

3 months

Yeah no. Under Gaussian errors the distribution of the residual (with fixed covariates, or more generally conditioned on the covariates) is Gaussian. It’s a linear projection. People mathsplaining regression always get it wrong 🙄

0

1

Erik J. Ringen

@err_ring

4 months

@ed_kroc poisson with observation-level random effects that follow t-dist rather than normal

0

1

Erik J. Ringen

@err_ring

4 months

@cameron_pfiffer bottomless_pit

0

1

Erik J. Ringen

@err_ring

5 months

Has anyone had their @github account flagged before (all activity suddenly hidden from public)? I've had an open ticket with them for reinstatement for over a month with no response...

0

1

Erik J. Ringen

@err_ring

5 months

@camjpatrick yes, priors are also research misconduct. informative prior = simulation of work that has no been conducted. generic regularizing prior = hiding work that has been conducted.

0

5

Erik J. Ringen

@err_ring

6 months

RT @JessieAdriaense: Happy to announce the arrival of my 1st postdoc project 🙌, now out in preprint: we looked at marmoset play and use of…

0

10

0

Erik J. Ringen

@err_ring

6 months

@PhDemetri @camjpatrick R + targets -> quarto is the way

1

0

1

Erik J. Ringen

@err_ring

6 months

RT @guichitarivera: Mi gente! I'm thrilled to share the first preprint from the Period Biology Study I've started at Dartmouth! Led by amaz…

0

3

0

Erik J. Ringen

@err_ring

7 months

PSA: this no longer works as of cmdstan 2.35.0. A new solution, courtesy of Andrew Johnson:

Erik J. Ringen

@err_ring

3 years

Nice solution from @babeheim: extract_samples <- function(cmdstanrfit) { cmdstanrfit$output_files() |> rstan::read_stan_csv() |> rethinking::extract.samples() }

0

1

Erik J. Ringen

@err_ring

8 months

@DStibbardHawkes it me

0

1

Erik J. Ringen

@err_ring

9 months

RT @sheinalew: I am advertising for a postdoc! You get to do cross-cultural research from the comfort of an office, write a lot of papers,…

0

72

0

Erik J. Ringen

@err_ring

1 year

@SolomonKurz It is okay. A researcher's ability to intervene is contextual. Ethics, resources, and technology determine what one can intervene on in a study. The limitations of researchers has nothing to do with the existence of causal effects.

0

Erik J. Ringen

@err_ring

1 year

RT @TanjaStadler_CH: We are hiring! Come join us for a PhD in Computational Language Evolution @ETH in Basel @ETH_BSSE and the National Res…

0

34

0

Erik J. Ringen

@err_ring

1 year

@PhDemetri I only teach stats for MA/PhD+ researchers, maybe some replies misunderstood your audience (undergrads?)

0

Erik J. Ringen

@err_ring

1 year

More evidence of the prior recession. In the old days a high school graduate could afford a fat-tailed, Cauchy(0, 5) prior on a fixed effect. Now they struggle to afford rent on a N(0, 0.5). @rlmcelreath

0

3