confounding   68

« earlier    

[1902.10286] On Multi-Cause Causal Inference with Unobserved Confounding: Counterexamples, Impossibility, and Alternatives
Unobserved confounding is a central barrier to drawing causal inferences from observational data. Several authors have recently proposed that this barrier can be overcome in the case where one attempts to infer the effects of several variables simultaneously. In this paper, we present two simple, analytical counterexamples that challenge the general claims that are central to these approaches. In addition, we show that nonparametric identification is impossible in this setting. We discuss practical implications, and suggest alternatives to the methods that have been proposed so far in this line of work: using proxy variables and shifting focus to sensitivity analysis.
causality  confounding  via:arthegall 
11 days ago by arsyed
[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."
causality  confounding  via:arthegall 
4 weeks ago by arsyed
[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."
victor-veitch  arxiv  research-article  machinelearning  confounding  causal-learning  causality 
4 weeks ago by arthegall
[1803.07276] Fair Deep Learning Prediction for Healthcare Applications with Confounder Filtering
The rapid development of deep learning methods has permitted the fast and accurate medical decision making from complex structured data, like CT images or MRI. However, some problems still exist in such applications that may lead to imperfect predictions. Previous observations have shown that, confounding factors, if handled inappropriately, will lead to biased prediction results towards some major properties of the data distribution. In other words, naively applying deep learning methods in these applications will lead to unfair prediction results for the minority group defined by the characteristics including age, gender, or even the hospital that collects the data, etc. In this paper, extending previous successes in correcting confounders, we propose a more stable method, namely Confounder Filtering, that can effectively reduce the influence of confounding factors, leading to better generalizability of trained discriminative deep neural networks, therefore, fairer prediction results. Our experimental results indicate that the Confounder Filtering method is able to improve the performance for different neural networks including CNN, LSTM, and other arbitrary architecture, different data types including CT-scan, MRI, and EEG brain wave data, as well as different confounding factors including age, gender, and physical factors of medical devices etc
deep-learning  healthcare  confounding 
august 2018 by arsyed
Randomization never ensures zero bias. It provides probabilistic bounds on confounding.

Therefore, by…
confounding  from twitter_favs
july 2018 by csabatino
What are radiological deep learning models actually learning?
CNNs appear to exploit information beyond specific disease-related imaging findings on x-rays to calibrate their disease predictions
deep-learning  machine-learning  causality  confounding 
july 2018 by tarakc02
Education and Political Participation: Exploring the Causal Link | SpringerLink
To test for a causal effect of education, we exploit the rise in education levels among males induced by the Vietnam draft. We find little reliable evidence that education induced by the draft significantly increases participation rates.
study  politics  polisci  education  human-capital  correlation  confounding  endogenous-exogenous  natural-experiment  null-result  military  war  history  mostly-modern  cold-war  asia  developing-world  sociology  phalanges  stylized-facts 
december 2017 by nhaliday
Red State Families: Better Than We Knew | Institute for Family Studies
Second, adjusting for education and race/ethnicity transforms the relationship between the Red State Index and marital upbringing from a curvilinear to a linear one. The redder the state, the more likely is a teen to grow up with his or her married birth parents. The relationship is modest but statistically significant. For every ten-point increase in the Red State Index, the proportion of teens living with both parents rises by one percentage point. This suggests that red state family culture is associated with increased odds of being raised in an intact, married family.
news  org:ngo  douthatish  sociology  data  analysis  usa  within-group  politics  ideology  coalitions  right-wing  correlation  confounding  race  demographics  education  class  life-history  dignity  wonkish  multi  gnon  general-survey  pro-rata  org:mag  intricacy  religion 
november 2017 by nhaliday
Do High School Sports Build or Reveal Character?
We examine the extent to which participation in high school athletics has beneficial effects on future education, labor market, and health outcomes. Due to the absence of plausible instruments in observational data, we use recently developed methods that relate selection on observables with selection on unobservables to estimate bounds on the causal effect of athletics participation. We analyze these effects in the US separately for men and women using three different nationally representative longitudinal data sets that each link high school athletics participation with later-life outcomes. We do not find consistent evidence of individual benefits reported in many previous studies – once we have accounted for selection, high school athletes are no more likely to attend college, earn higher wages, or participate in the labor force. However, we do find that men (but not women) who participated in high school athletics are more likely to exercise regularly as adults. Nevertheless, athletes are no less likely to be obese.
pdf  study  broad-econ  economics  econometrics  microfoundations  human-capital  sports  intervention  null-result  wonkish  endo-exo  selection  confounding  labor  education  health  fitness  fitsci  org:ngo  white-paper  methodology  compensation  cost-benefit  input-output  endogenous-exogenous  branches 
november 2017 by nhaliday
Places and Preferences: A Longitudinal Analysis of Self-Selection and Contextual Effects | British Journal of Political Science | Cambridge Core
preferences -> place, not place -> preferences (mostly)
Cosmopolitan immigration attitudes in Europe's large cities: Adaptation or selection:
The myth of London exceptionalism:
London is not as invulnerable to the appeal of UKIP as commonly reported, finds new research from Eric Kaufmann.

Are White British Londoners more accepting of immigration than White British elsewhere? The British Election Study (BES)’s 2015 panel survey asks whether immigration enriches or undermines cultural life. 34.7 per cent of White British outside London say immigration strongly undermines cultural life. But so do 34.4 per cent of White British Londoners. Not much difference there. 44 per cent of White Brits outside London want to leave the EU, but so do 42.3 per cent of White British Londoners. Again, not much in it.


Finally, when we control for a fuller range of demographic and attitudinal characteristics, as in figure 4, London and the South East emerge as significantly more likely than the rest of England and Wales to have voted UKIP in 2014, according to the BES.

Cosmopolitan cities and their country cousins – UK in a changing Europe:
study  sociology  politics  polisci  ideology  elections  data  correlation  null-result  confounding  endo-exo  migration  causation  stylized-facts  selection  britain  urban  polarization  longitudinal  context  intervention  roots  piracy  pdf  class-warfare  multi  eric-kaufmann  brexit  trust  social-capital  race  demographics  org:ngo  summary  values  poll  dropbox  phalanges  geography  wonkish  analysis  variance-components  chart  populism  nationalism-globalism  org:edu  org:anglo  psychology  social-psych  org:mag  news  endogenous-exogenous  preprint  urban-rural  hari-seldon  sentiment 
may 2017 by nhaliday

« earlier    

related tags

2016-election  academia  acm  acmtariat  age-generation  aging  akrasia  albion  altruism  analysis  anglo  anthropology  applications  article  arxiv  asia  attaq  autism  bacanu.silviu-alin  backup  bad_data_analysis  badscience  bayesian  behavioral-gen  best-practices  bias  biases  bigdata  bio  biodet  bioinformatics  biophysical-econ  blogged  blowhards  books  branches  brexit  britain  broad-econ  c:**  causal-learning  causal_inference  causality  causation  chart  cjones-like  class-warfare  class  cliometrics  cluster  cmu  coalitions  cog-psych  cold-war  coming-apart  commentary  comparison  compensation  concept  contagion  context  contrarianism  control  cooperate-defect  corporation  correlation  corruption  cost-benefit  counter-revolution  crime  criminology  critique  cultural-dynamics  current-events  data  deep-learning  deep-materialism  definition  demographic-transition  demographics  density  dependence-independence  descriptive  developing-world  developmental  devlin.bernie  dignity  direct-indirect  direction  dirty-hands  discipline  discussion  disease  distribution  diversity  douthatish  draft  dropbox  dysgenics  early-modern  econometrics  economics  econotariat  education  effect-size  effects  egalitarianism-hierarchy  ego-depletion  egt  elections  elite  embodied-cognition  embodied  encyclopedic  endo-exo  endogenous-exogenous  environmental-effects  envy  epidemiology  epistemic  eric-kaufmann  error  essay  ethics  eu  europe  events  evidence-based  evisceration  evolution  evopsych  experimental_psychology  explained  explanation  expression-survival  faq  fertility  field-study  fitness  fitsci  flynn  food  futures  galton  garett-jones  gelman  gender-diff  gender  general-survey  generalization  generative  genetic-load  genetics  genomic_control  genomics  geography  germanic  giants  gilens-page  gnon  gnosis-logos  gotchas  government  graphical-models  graphical_models  graphs  greenland.sander  gregory-clark  growth-econ  guide  gwas  gwern  gxe  haidt  hanson  hari-seldon  harvard  have_read  health  healthcare  heritability  heterodox  higher-ed  history  hmm  homophily  honor  housing  human-capital  human_genetics  hypothesis-testing  identifiability  ideology  impetus  in_nb  incentives  industrial-revolution  inequality  info-dynamics  input-output  inspiration  institutions  integrity  intelligence  interdisciplinary  interests  intervention  intricacy  iq  islam  jargon  journos-pundits  justice  kaggle  kanye  kith_and_kin  labor  language  latent-variables  latent_variables  latin-america  law  left-wing  legacy  lens  life-history  life  links  list  long-term  longevity  longitudinal  machine-learning  machinelearning  madisonian  malaise  maps  marginal  matter  measurement  media  medicine  mena4  mendel-randomization  mental_testing  meta-analysis  meta:prediction  meta:science  metameta  methodology  methods  micro  microfoundations  migration  military  modeling  models  money  morality  mostly-modern  multi  multilevel  nationalism-globalism  natural-experiment  network-structure  network_data_analysis  neuro-nitgrit  neuro  new-religion  news  nibble  noblesse-oblige  nonlinearity  nordic  null-result  obesity  okcupid  old-anglo  org:anglo  org:biz  org:bleg  org:data  org:econlib  org:edu  org:junk  org:lite  org:mag  org:mat  org:nat  org:ngo  org:rec  organizing  outliers  papers  parenting  paternal-age  patience  pdf  pearl.judea  personality  phalanges  philosophy  pic  piracy  planning  plots  poast  polarization  policy  polisci  political-econ  political  politics  poll  pop-structure  population-genetics  population  populism  postmortem  power  pre-ww2  prediction  preprint  pro-rata  productivity  propaganda  psych-architecture  psychiatry  psychology  psychometrics  public-health  putnam-like  qtl  quality  r  race  ratty  re:critique_of_diffusion  re:g_paper  re:homophily_and_confounding  realness  reddit  redistribution  reference  regression  regularizer  religion  rent-seeking  replication  research-article  research  rhetoric  right-wing  robins.james  roeder.kathryn  roots  rot  s-factor  sampling-bias  sapiens  science-anxiety  science  scitariat  selection  self-centered  self-report  selfish-gene  sentiment  sex  sexuality  shift  sib-study  simulation  social-capital  social-choice  social-norms  social-psych  social-science  social-structure  social  social_influence  social_networks  society  sociology  solid-study  spatial  spearhead  speculation  sports  ssc  stamina  state-of-art  statistics  stats  stereotypes  stream  stress  study  stylized-facts  summary  supply-demand  survey  techtariat  terrorism  the-bones  the-world-is-just-atoms  things  thinking  time-preference  time-series  to:blog  to_teach:complexity-and-inference  track_down_references  tradeoffs  trends  trump  trust  tutorial  twin-study  twitter  unaffiliated  unit  urban-rural  urban  usa  validity  values  vampire-squid  variables  variance-components  victor-veitch  video  visualization  volo-avolo  voter_model  war  wasserman.larry  wealth  welfare-state  wermuth.nanny  west-hunter  westminster  west’s  white-paper  why  wiki  winner-take-all  within-group  within-without  wonkish  world-war  world  yoga  yvain  zeitgeist  zero-positive-sum  🌞  🎩  🐸  👽  🔬 

Copy this bookmark: