It’s really funny, though not a comedy

Here’s a paper doing imputation. I read and earlier version, not the latest, and was frustrated that they didn’t compare their ML imputation method to other more traditional imputation approaches. arxiv.org/pdf/2305.096...

This is the closest I’ve found: arxiv.org/abs/2401.08572

How do you make a scientific case about a method’s usefulness without a proper lit review?

by synthetic samples, I mean using LLMs to generate survey responses "imagine you are a 54 year old African American man. Please answer the following survey question..."

Can anyone point me to a paper, WP, blog post that reviews the literature on synthetic samples? Thank you

Has anyone used DC’s school lottery as an instrument to evaluate effect of pre-k on subsequent educational achievement ?

IMO, the worst problem with data analysis a in Excel is how hard it is to trace back where a result came from. Yes, there are now trace dependence and trace precedence buttons, but it still is not nearly as clear as it would be in a script file

Took us a while, but we finally put another chapter of our book "Data Collection with #Wearables#Apps#Sensors#Sampling#Coveragebookdown.org/wasbook_feed...@stephnie.bsky.social

They really are : “the importance…in some vague sense” IIRC