Wash your hands for as long as it takes to explain to your collaborator why the Wilcoxon is not necessarily testing the medians and then your hands will be so raw you'll be more worried about a skin infection than a respiratory infection boom crisis averted.
NOVELISTS WRITING ABOUT BIOSTATISTICS: The nerd took out a calculator, said some words, and pissed everyone off.
MY ACTUAL EXPERIENCE AS A BIOSTATISTICIAN: Armed with the sword of simulations, the shield of theory, and my duty as a Guardian of Science, I plunged into the fray.
This Post is from an account that no longer exists. Learn more
4
THE 48 LAWS OF POWER CALCULATIONS
1. NO
2. I
3. CAN'T
4. JUST
5. "FILL
6. OUT
7. THE
8. STATS"
9. BY
10. TOMORROW
11. ARE
12. YOU
13. SERIOUS
14. DID
15. WE
16. NOT
17. GO
18. OVER
19. THIS
20. LAST
21. TIME
22. U
23. G
24. H
25. H
26. H
27. H
28. H
.
.
.
46. H
47. OK
48. BOOMER
Pediatrician: *hands medicine to my kid* take one a day and you'll get better, ok? Have a nice day!
Me: Ok kiddo be polite now - what do you say?
Kid: Are you sure you're not confusing population-level effects with individual-level effects?
Me: *proudly sheds one tear of joy*
Literally all I do as a biostatistician:
Oh.
Oh.
Seriously?
Oh.
Well in that case what you can do is-
Oh...
Wait so your huge cluster randomized trial only has two clusters?
Oh.
And you're submitting tomorrow.
Oh.
Oh?
OMG I just realized I have to uhhh feed my cat gotta go bye
Literally all I do as a statistician:
No.
No.
That's not the definition of a p-value.
No.
Trending towards significance is not a thing.
No.
No pie charts!
That only works if data is normal.
No.
That's logistic regression not AI.
No.
Your "novel" method was invented in 1918.
No.
6
Carbon footprint of
t-test: *
GLM: **
Random Forest: ***
Gradient Boosting Machine: ****
Recurrent Neural Network: *****
Stress eating hamburgers because this EHR data is just an endless hell of bias and missingness: *********************************************************
tbh kind of disappointed that i haven't seen a data science venn diagram made with overlapping hex stickers yet, you millennials are so lazy
1
Parent: please tell my kid that vaping's bad
Doctor: we know little about its harmful effects but it's still addictive
Teen: ha! no harmful effects
Parent: still addictive!
Biostatistician: *bursts thru wall* BUT MOST IMPORTANTLY ABSENCE OF EVIDENCE IS NOT EVIDENCE OF ABSENCE
ggblotbot is a chrome plug-in that uses neural networks to ID any plots made with base R graphics and covers them with angry blots of ink
If you stare long enough at the eclipse it starts to look like a pie chart and then not only do your eyes start burning but your soul too
Suppose P=NP. Then N=1. However big data is where it's at and N=1 is small data and thus not where it's at. Thus N!=1. Therefore P!=NP. QED.
3
Oh wow big news pretty sure I just proved the P=NP problem using a novel BIG DATA approach, will tweet proof soon plz stand by..............
1
Riding on my antelope, watching the cows lined up to return to their barn brought me a sudden joy. A back of the antelope cow queue elation.
1
Planning to tweet the primary results as soon as they're done to preempt peer review and formal publication is an intent-to-tweet analysis
Did you know that if you have unmeasured confounding AND informative missingness they cancel out and you get a blinded RCT and also tenure
5
Let the below assumptions hold:
1 Cauchy data
2 Super informative missingness
3 Tons of unmeasured confounding
4 We're all gonna die anyway
3

