Aaron Fisher

Statistician at @FoundationATCG who likes to cook, run and draw.

[Aaron Fisher @ajfishr]

Hi friends! Are any of y'all going to #ICML24 next week and free to meet up for dinner, coffee, or a very low key run?

[Ben Ackerman, PhD @backerman150]

I *finally* got to experience the famous 6am #ENAR2024 running club (as well as the 12:30pm rendition 😅) 🏃

[Aaron Fisher @ajfishr]

I should maybe say, graph discovery isn't my area, but as a non expert I found this very approachable!

[Aaron Fisher @ajfishr]

There's a great 5 minute explanation of causal graph discovery here! (min 20:24)

Any DAG can fit the data equally well, so we search for the DAG that leads to fitted models that are (for example) sparse.

https://t.co/BhRAW396Lj

[Valeriy M., PhD, MBA, CQF @predict_addict]

Prof Larry Wasserman’s (CMU) talk “problems with Bayesian causal inference”

https://t.co/ZRQ0KnusjE

#bayesian #causality

[Valeriy M., PhD, MBA, CQF @predict_addict]

Prof Larry Wasserman’s (CMU) talk “problems with Bayesian causal inference”

https://t.co/ZRQ0KnusjE

#bayesian #causality

[Daniel Scharfstein @dscharf3]

Excited to share my latest paper with @amoafo_linda and Elizabeth Platz; it develops a sensitivity analysis tool for evaluating the influence of unmeasured confounding in observational studies with time-to-event outcomes https://t.co/zJ4lB7qtvE

[Aaron Fisher @ajfishr]

Full paper at https://t.co/e3pYhplkkF

[Aaron Fisher @ajfishr]

Super excited to present this poster at #ENAR2024!

Inverse propensity score weights can lead to unstable conditional treatment effect estimates 😢

but inverse *variance* weights re-stabilize the estimate without adding bias 🎉

It turns out R-Learning does this implicitly!

[Aaron Fisher @ajfishr]

Full paper at https://t.co/e3pYhplkkF

[Aaron Fisher @ajfishr]

@maria__cuellar I'll be there!🙋‍♂️It'd be great to meet up if you have time!

(Wait, do you mean ENAR? I see you're presenting on Monday?)

[shira mitchell @shiraamitchell]

We (Blue Rose Research) are hiring again !

Come use statistics to advise progressive organizations as we gear up for the 2024 election.
https://t.co/1f2TIcUkSt

[Aaron Fisher @ajfishr]

If you're interested in working at the intersection of quantitative and qualitative medical research, I have heard great things about Modus!
https://t.co/6rFQMTEn0h

[Aaron Fisher @ajfishr]

A few years ago I posted about why we divide by n-1 for the sample variance

In case it's of interest, I just made some minor updates to that post, simplifying the steps involved!

[Aaron Fisher @ajfishr]

Confused about why we divide by n-1 for the sample variance?

The variance tells us how much people tend to be different from each other. And, for each person in our sample, there are only n-1 people we can compare them to.

More in my blog post: https://t.co/I9FrRhusmK

[🇲🇽 Leonardo Collado-Torres @lcolladotor]

Kristen @kr_maynard & I are #hiring a Research Associate to work on 📊 funded by R01DA055823 involving 🤩 technologies Multiomics & Spatial Transcriptomics (Visium) by @10xGenomics https://t.co/TRW1xfOfCM

Come join us! Requires undergrad + yrs or Masters

https://t.co/apQIjLzpEn

[Aaron Fisher @ajfishr]

A thought that's kinda ruining a lot of sci fi for me recently:

If sci fi writers abandoned the epic idea that terraforming new planets would somehow be more promising than fighting climate change, would billionaires spend less energy building space ships?

[C. Andrew Frank @CAndrewFrank]

How to read a paper in your field (thread):

1. Run across the title. Small panic. Isn’t that exactly the same idea you were testing??

2. Read Abstract. Relief. It’s clearly not the same. Or is it? 😱

[Aaron Fisher @ajfishr]

@timtriche @paperpile It's a somewhat tedious and stressful process, but, of course, it's very important too.

[Aaron Fisher @ajfishr]

Does anyone have a strong system they like for managing citations for a new project?

Excel? Hand-drawn citation tree diagrams? "Murder walls"?

[Aaron Fisher @ajfishr]

@timtriche @paperpile Ah, so say there's a paper called "paper X" that's highly relevant to my topic.

By upward & downward, I mean looking through everything paper X cites, and also everything that cites paper X.

I'd say that's the idealized level of thoroughness. Don't always achieve it, but I try

[Aaron Fisher @ajfishr]

I should clarify: by "managing citations" I really mean managing literature searches, and trying to do them in an exhaustive and organized way