A Problem with Confidence Intervals
top of page
Stochastic Musings
Exploring statistics, data, science, and epistemology
Home: Welcome
Search
Confidence intervals may not adequately convey how certain a statistical model is
Jackson Curtis
- Jun 2, 2022
- 7 min
5 Problems with Using Linear Regression for MMM (and what to do instead)
An introduction to some of the complexities of marketing mix modeling
Jackson Curtis
- Feb 21, 2022
- 4 min
Why I'm Stoked to Join Recast
I'm so excited for my new career move
Jackson Curtis
- Dec 28, 2021
- 2 min
A Baby Name Picking App
A shiny app implementing a MaxDiff exercise for choosing baby names
Jackson Curtis
- Oct 15, 2021
- 2 min
This chart makes me grateful for computers
A chart I made for our family's baby guessing game
Jackson Curtis
- Sep 8, 2021
- 2 min
The Conditional Birthday Problem
A famous stats problem with a surprise personal twist at the end.
Jackson Curtis
- Sep 6, 2021
- 3 min
'Objective Ignorance' and the Limits of Predictability
How data scientists fool themselves into thinking prediction problems are easier than they are.
Jackson Curtis
- Aug 28, 2021
- 3 min
I got on FiveThirtyEight!
What I learned that got me a shoutout on fivethirtyeight.com
Jackson Curtis
- Jul 21, 2021
- 2 min
Can you interpret this simple linear regression?
Can you draw a valid conclusion from this regression and whether a player is paid appropriately?
Jackson Curtis
- Jul 9, 2021
- 4 min
Ranking NBA Refs, pt. 3 - Uncertainty in Rankings
I look at the confidence intervals from my Bayesian model to assess reffing accuracy
Jackson Curtis
- Jun 28, 2021
- 7 min
Ranking NBA Refs - pt. 2 Building a Bayesian Model
A walkthrough of my Bayesian model to rank NBA referees
Jackson Curtis
- Jun 22, 2021
- 3 min
Ranking NBA Refs, pt. 1 - Modeling Bad Calls
Modeling the bad call rate among NBA refs using Bayesian poisson regression.
Jackson Curtis
- Jun 9, 2021
- 3 min
XGBoost and Model Assumptions
My post this week is simply to emphasize the work Saupin Guillame found here. In his post, Saupin demonstrates that a major limitation of...
Jackson Curtis
- Jun 4, 2021
- 4 min
Doctor or Quack: which kind of data scientist are you?
Are data scientists adopting some of the worst traits of alternative medicine?
Jackson Curtis
- Jun 3, 2021
- 1 min
Questions about vaccine reporting
The primary endpoint in clinical trials for COVID-19 vaccines is "vaccine efficacy" which is essentially what percentage of COVID cases...
Jackson Curtis
- May 29, 2021
- 7 min
Predicting Regression toward the Mean Using Hierarchical Bayesian Models
Regression toward the mean is a fascinating topic to me. It's completely ubiquitous across the world of statistical measurement, but can...
Home: Blog2
Home: Subscribe
Home: Contact
bottom of page