Double robust, flexible adjustment methods for causal inference: An overview and an evaluation

{"Double":[0,109],"robust":[1,110,199],"methods":[2,20,33,38,56,70,75,111,135,139,179,200],"for":[3],"flexible":[4,60,77,161,170],"covariate":[5,61],"adjustment":[6,62],"in":[7,12,39,57,103,140,158],"causal":[8],"inference":[9],"have":[10,180],"proliferated":[11],"recent":[13],"years.":[14],"Despite":[15],"their":[16],"apparent":[17],"advantages,":[18],"these":[19,32,69,134],"are":[21,154,186],"rarely":[22],"used":[23],"by":[24,84,98],"social":[25],"scientists.":[26],"It":[27,47,81],"is":[28,48,94],"also":[29],"unclear":[30],"whether":[31],"actually":[34],"outperform":[35,137],"more":[36,72,188],"traditional":[37,73,138],"finite":[40],"samples.":[41],"This":[42],"paper":[43],"has":[44],"two":[45],"aims:":[46],"a":[49,141],"guide":[50],"to":[51,71],"some":[52,100,132],"of":[53,133,144,197,207,213],"the":[54,90,151,168,192,195,208,211],"latest":[55],"double":[58,198],"robust,":[59],"using":[63,85],"machine":[64,162],"learning,":[65],"and":[66,76,96,124,156,176,185],"it":[67],"compares":[68],"statistical":[74],"\\"single":[78],"robust\\"":[79],"methods.":[80,108],"does":[82],"this":[83],"both":[86],"simulated":[87],"data":[88],"where":[89],"treatment":[91],"effect":[92],"estimate":[93],"known,":[95],"then":[97],"replicating":[99],"prominent":[101],"articles":[102],"sociology":[104],"that":[105,131],"use":[106],"simpler":[107],"covered":[112],"include":[113],"Augmented":[114],"Inverse":[115],"Probability":[116],"Weighting":[117],"(AIPW),":[118],"Targeted":[119],"Maximum":[120],"Likelihood":[121],"Estimation":[122],"(TMLE),":[123],"Double/Debiased":[125],"Machine":[126],"Learning":[127],"(DML).":[128],"Results":[129],"suggest":[130],"do":[136],"wide":[142],"range":[143],"simulations,":[145],"but":[146,204],"only":[147,181],"slightly.":[148],"In":[149,191],"particular,":[150],"top":[152],"performers":[153],"TMLE":[155],"AIPW":[157],"conjunction":[159],"with":[160,167],"learning":[163],"estimators.":[164,218],"But":[165],"G-computation":[166],"same":[169],"estimators":[171],"obtains":[172],"almost":[173],"identical":[174],"results,":[175,209],"simple":[177],"regression":[178],"slightly":[182],"higher":[183],"bias":[184],"much":[187],"computationally":[189],"efficient.":[190],"article":[193],"replications,":[194],"application":[196],"substantively":[201],"changes":[202],"some,":[203],"not":[205],"all,":[206],"highlighting":[210],"importance":[212],"comparing":[214],"performance":[215],"from":[216],"multiple":[217]}