to_read   32603

« earlier    

[1902.04114] Using Embeddings to Correct for Unobserved Confounding
"We consider causal inference in the presence of unobserved confounding. In particular, we study the case where a proxy is available for the confounder but the proxy has non-iid structure. As one example, the link structure of a social network carries information about its members. As another, the text of a document collection carries information about their meanings. In both these settings, we show how to effectively use the proxy to do causal inference. The main idea is to reduce the causal estimation problem to a semi-supervised prediction of both the treatments and outcomes. Networks and text both admit high-quality embedding models that can be used for this semi-supervised prediction. Our method yields valid inferences under suitable (weak) conditions on the quality of the predictive model. We validate the method with experiments on a semi-synthetic social network dataset. We demonstrate the method by estimating the causal effect of properties of computer science submissions on whether they are accepted at a conference."
to:NB  causal_inference  statistics  blei.david  re:homophily_and_confounding  to_read 
yesterday by cshalizi
The Bias Is Built In: How Administrative Records Mask Racially Biased Policing by Dean Knox, Will Lowe, Jonathan Mummolo :: SSRN
"Researchers often lack the necessary data to credibly estimate racial bias in policing. In particular, police administrative records lack information on civilians that police observe but do not investigate. In this paper, we show that if police racially discriminate when choosing whom to investigate, using administrative records to estimate racial bias in police behavior amounts to post-treatment conditioning, and renders many quantities of interest unidentified---even among investigated individuals---absent strong and untestable assumptions. In most cases, no set of controls can eliminate this statistical bias, the exact form of which we derive through principal stratification in a causal mediation framework. We develop a bias-correction procedure and nonparametric sharp bounds for race effects, replicate published findings, and show traditional estimation techniques can severely underestimate levels of racially biased policing or even mask discrimination entirely. We conclude by outlining a general and feasible design for future studies that is robust to this inferential snare."
to:NB  to_read  causal_inference  police  discrimination  statistics  to_teach:undergrad-ADA  via:henry_farrell 
2 days ago by cshalizi

« earlier    

related tags

aha!  angst  apolloxi  archives  astronomy  avengers_au  bigbang  blackpanther  blei.david  books!  books  business  c++  captainmarvel  causal_inference  censorship  coffee_shop  collective_cognition  culture  deltoro  depression  development  discrimination  doctorwho  escape-room  experimental_design  experimental_sociology  facebook  fantastic_literature  farming  fic  freelancing  h/c  hci  history  hope  iarichards  illness  information-theory  instagram  interesting!  internet  ios  learning_theory  libraries  machine-learning  management  math:dynamical_systems  math_and_stats  mathematics  medical  model_checking  movies  mushrooms  neil_gaiman  network_data_analysis  networked_life  neural_networks  news  numeric_methods  openaccess  philosophy  police  positive  practicalcriticism  prediction  professional_issues  professional_skills  programming  przed  pterry  q/007  queen  r  rakhlin.sasha  re:actually-dr-internet-is-the-name-of-the-monsters-creator  re:democratic_cognition  re:do_not_adjust_your_receiver  re:homophily_and_confounding  rendering  research  samanthahirr  science_fiction  scifi  singin_in_the_rain  six@six  social_influence  social_networks  socialmedia  sociology  software  space!  spatial_ecology  spatial_statistics  speaking  splines  star_wars  startrek  statistics  statistics:spatial  statistics:time_series  steverogers/buckybarnes  stochastic_methods  stochasticity  swift-lang  tattoos  technology  this!  thor  time  to:nb  to_be_shot_after_a_fair_trial  to_buy  to_do  to_teach  to_teach:baby-nets  to_teach:data_over_space_and_time  to_teach:undergrad-ada  to_try  travel  tutorial  useful  ux  web2.0  wordcount:20.000-29.999  wordcount:50.000-74.999  work  writing 

Copy this bookmark: