Dealing with limited overlap in estimation of average treatment effects
Estimation of average treatment effects under unconfounded or ignorable treatment assignment is often hampered by lack of overlap in the covariate distributions between treatment groups. This lack of overlap can lead to imprecise estimates, and can make commonly used estimators sensitive to the choi...
Gespeichert in:
| Veröffentlicht in: | Biometrika 2009-03, Vol.96 (1), p.187-199 |
|---|---|
| Hauptverfasser: | , , , |
| Format: | Artikel |
| Sprache: | eng |
| Schlagworte: | |
| Online-Zugang: | Volltext |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| container_end_page | 199 |
|---|---|
| container_issue | 1 |
| container_start_page | 187 |
| container_title | Biometrika |
| container_volume | 96 |
| creator | CRUMP, RICHARD K. HOTZ, V. JOSEPH IMBENS, GUIDO W. MITNIK, OSCAR A. |
| description | Estimation of average treatment effects under unconfounded or ignorable treatment assignment is often hampered by lack of overlap in the covariate distributions between treatment groups. This lack of overlap can lead to imprecise estimates, and can make commonly used estimators sensitive to the choice of specification. In such cases researchers have often used ad hoc methods for trimming the sample. We develop a systematic approach to addressing lack of overlap. We characterize optimal subsamples for which the average treatment effect can be estimated most precisely. Under some conditions, the optimal selection rules depend solely on the propensity score. For a wide range of distributions, a good approximation to the optimal rule is provided by the simple rule of thumb to discard all units with estimated propensity scores outside the range [0.1,0.9]. |
| doi_str_mv | 10.1093/biomet/asn055 |
| format | Article |
| fullrecord | <record><control><sourceid>jstor_proqu</sourceid><recordid>TN_cdi_proquest_journals_201696479</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><jstor_id>27798811</jstor_id><oup_id>10.1093/biomet/asn055</oup_id><sourcerecordid>27798811</sourcerecordid><originalsourceid>FETCH-LOGICAL-c493t-1dc49d7a1cee8983a86974164a1880c42e09a50b493db4f846fb28149889a7d33</originalsourceid><addsrcrecordid>eNqFkM2LFDEQxYMoOK4ePQpBELy0m-qk83GUXd0PF1RQEC8h0129m7G705tk1t3_3gw9zB49VB6hfryqeoS8BvYBmOHHax9GzMcuTaxpnpAVCCkq3gB7SlaMMVlxIcRz8iKlze4rG7kiZ6foBj9d078-39DBjz5jR8MdxsHN1E8UU_ajyz5MNPTUlYa7RpojujzilCn2PbY5vSTPejckfLXXI_Lz86cfJ-fV1dezi5OPV1UrDM8VdEU75aBF1EZzp6VRAqRwoDVrRY3MuIatC9ytRa-F7Ne1BmG0Nk51nB-Rt4vvHMPttixnN2EbpzLS1gykkUKZAlUL1MaQUsTezrEcER8sMLuLyi5R2SWqwl8ufMQZ2wMctvOeu7PcGVmeh1I1Y6aILwWl5p1qZcEYe5PHYvZuv6FLrRv66KbWp4NpDaBrDrJw7xeujPnvfm8WdJNyiI9WSpVcAB7v9Snj_aHv4h8rFVeNPf_12-pv-hLYl-_2lP8DJgOqbg</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>201696479</pqid></control><display><type>article</type><title>Dealing with limited overlap in estimation of average treatment effects</title><source>RePEc</source><source>JSTOR Mathematics & Statistics</source><source>JSTOR Archive Collection A-Z Listing</source><source>Oxford University Press Journals All Titles (1996-Current)</source><source>Alma/SFX Local Collection</source><creator>CRUMP, RICHARD K. ; HOTZ, V. JOSEPH ; IMBENS, GUIDO W. ; MITNIK, OSCAR A.</creator><creatorcontrib>CRUMP, RICHARD K. ; HOTZ, V. JOSEPH ; IMBENS, GUIDO W. ; MITNIK, OSCAR A.</creatorcontrib><description>Estimation of average treatment effects under unconfounded or ignorable treatment assignment is often hampered by lack of overlap in the covariate distributions between treatment groups. This lack of overlap can lead to imprecise estimates, and can make commonly used estimators sensitive to the choice of specification. In such cases researchers have often used ad hoc methods for trimming the sample. We develop a systematic approach to addressing lack of overlap. We characterize optimal subsamples for which the average treatment effect can be estimated most precisely. Under some conditions, the optimal selection rules depend solely on the propensity score. For a wide range of distributions, a good approximation to the optimal rule is provided by the simple rule of thumb to discard all units with estimated propensity scores outside the range [0.1,0.9].</description><identifier>ISSN: 0006-3444</identifier><identifier>EISSN: 1464-3510</identifier><identifier>DOI: 10.1093/biomet/asn055</identifier><identifier>CODEN: BIOKAX</identifier><language>eng</language><publisher>Oxford: Oxford University Press</publisher><subject>Applications ; Arithmetic mean ; Average treatment effect ; Biology, psychology, social sciences ; Causality ; Clinical outcomes ; Distribution theory ; Estimation methods ; Estimators ; Exact sciences and technology ; General topics ; Heart catheterization ; Ignorable treatment assignment ; Infinity ; Intubation ; Mathematics ; Overlap ; Population ; Population estimates ; Probability and statistics ; Probability theory and stochastic processes ; Propensity score ; Research methods ; Sample size ; Sciences and techniques of general use ; Statistical discrepancies ; Statistics ; Studies ; Treatment effect heterogeneity ; Unconfoundedness ; Weighting functions</subject><ispartof>Biometrika, 2009-03, Vol.96 (1), p.187-199</ispartof><rights>2009 Biometrika Trust</rights><rights>Oxford University Press © 2009 Biometrika Trust 2009</rights><rights>2009 INIST-CNRS</rights><lds50>peer_reviewed</lds50><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c493t-1dc49d7a1cee8983a86974164a1880c42e09a50b493db4f846fb28149889a7d33</citedby><cites>FETCH-LOGICAL-c493t-1dc49d7a1cee8983a86974164a1880c42e09a50b493db4f846fb28149889a7d33</cites></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktopdf>$$Uhttps://www.jstor.org/stable/pdf/27798811$$EPDF$$P50$$Gjstor$$H</linktopdf><linktohtml>$$Uhttps://www.jstor.org/stable/27798811$$EHTML$$P50$$Gjstor$$H</linktohtml><link.rule.ids>315,781,785,804,833,1585,4009,27929,27930,58022,58026,58255,58259</link.rule.ids><backlink>$$Uhttp://pascal-francis.inist.fr/vibad/index.php?action=getRecordDetail&idt=21182316$$DView record in Pascal Francis$$Hfree_for_read</backlink><backlink>$$Uhttp://econpapers.repec.org/article/oupbiomet/v_3a96_3ay_3a2009_3ai_3a1_3ap_3a187-199.htm$$DView record in RePEc$$Hfree_for_read</backlink></links><search><creatorcontrib>CRUMP, RICHARD K.</creatorcontrib><creatorcontrib>HOTZ, V. JOSEPH</creatorcontrib><creatorcontrib>IMBENS, GUIDO W.</creatorcontrib><creatorcontrib>MITNIK, OSCAR A.</creatorcontrib><title>Dealing with limited overlap in estimation of average treatment effects</title><title>Biometrika</title><description>Estimation of average treatment effects under unconfounded or ignorable treatment assignment is often hampered by lack of overlap in the covariate distributions between treatment groups. This lack of overlap can lead to imprecise estimates, and can make commonly used estimators sensitive to the choice of specification. In such cases researchers have often used ad hoc methods for trimming the sample. We develop a systematic approach to addressing lack of overlap. We characterize optimal subsamples for which the average treatment effect can be estimated most precisely. Under some conditions, the optimal selection rules depend solely on the propensity score. For a wide range of distributions, a good approximation to the optimal rule is provided by the simple rule of thumb to discard all units with estimated propensity scores outside the range [0.1,0.9].</description><subject>Applications</subject><subject>Arithmetic mean</subject><subject>Average treatment effect</subject><subject>Biology, psychology, social sciences</subject><subject>Causality</subject><subject>Clinical outcomes</subject><subject>Distribution theory</subject><subject>Estimation methods</subject><subject>Estimators</subject><subject>Exact sciences and technology</subject><subject>General topics</subject><subject>Heart catheterization</subject><subject>Ignorable treatment assignment</subject><subject>Infinity</subject><subject>Intubation</subject><subject>Mathematics</subject><subject>Overlap</subject><subject>Population</subject><subject>Population estimates</subject><subject>Probability and statistics</subject><subject>Probability theory and stochastic processes</subject><subject>Propensity score</subject><subject>Research methods</subject><subject>Sample size</subject><subject>Sciences and techniques of general use</subject><subject>Statistical discrepancies</subject><subject>Statistics</subject><subject>Studies</subject><subject>Treatment effect heterogeneity</subject><subject>Unconfoundedness</subject><subject>Weighting functions</subject><issn>0006-3444</issn><issn>1464-3510</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2009</creationdate><recordtype>article</recordtype><sourceid>X2L</sourceid><recordid>eNqFkM2LFDEQxYMoOK4ePQpBELy0m-qk83GUXd0PF1RQEC8h0129m7G705tk1t3_3gw9zB49VB6hfryqeoS8BvYBmOHHax9GzMcuTaxpnpAVCCkq3gB7SlaMMVlxIcRz8iKlze4rG7kiZ6foBj9d078-39DBjz5jR8MdxsHN1E8UU_ajyz5MNPTUlYa7RpojujzilCn2PbY5vSTPejckfLXXI_Lz86cfJ-fV1dezi5OPV1UrDM8VdEU75aBF1EZzp6VRAqRwoDVrRY3MuIatC9ytRa-F7Ne1BmG0Nk51nB-Rt4vvHMPttixnN2EbpzLS1gykkUKZAlUL1MaQUsTezrEcER8sMLuLyi5R2SWqwl8ufMQZ2wMctvOeu7PcGVmeh1I1Y6aILwWl5p1qZcEYe5PHYvZuv6FLrRv66KbWp4NpDaBrDrJw7xeujPnvfm8WdJNyiI9WSpVcAB7v9Snj_aHv4h8rFVeNPf_12-pv-hLYl-_2lP8DJgOqbg</recordid><startdate>20090301</startdate><enddate>20090301</enddate><creator>CRUMP, RICHARD K.</creator><creator>HOTZ, V. JOSEPH</creator><creator>IMBENS, GUIDO W.</creator><creator>MITNIK, OSCAR A.</creator><general>Oxford University Press</general><general>Biometrika Trust, University College London</general><general>Oxford University Press for Biometrika Trust</general><general>Oxford Publishing Limited (England)</general><scope>BSCLL</scope><scope>IQODW</scope><scope>DKI</scope><scope>X2L</scope><scope>AAYXX</scope><scope>CITATION</scope><scope>7QO</scope><scope>8FD</scope><scope>FR3</scope><scope>P64</scope></search><sort><creationdate>20090301</creationdate><title>Dealing with limited overlap in estimation of average treatment effects</title><author>CRUMP, RICHARD K. ; HOTZ, V. JOSEPH ; IMBENS, GUIDO W. ; MITNIK, OSCAR A.</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c493t-1dc49d7a1cee8983a86974164a1880c42e09a50b493db4f846fb28149889a7d33</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2009</creationdate><topic>Applications</topic><topic>Arithmetic mean</topic><topic>Average treatment effect</topic><topic>Biology, psychology, social sciences</topic><topic>Causality</topic><topic>Clinical outcomes</topic><topic>Distribution theory</topic><topic>Estimation methods</topic><topic>Estimators</topic><topic>Exact sciences and technology</topic><topic>General topics</topic><topic>Heart catheterization</topic><topic>Ignorable treatment assignment</topic><topic>Infinity</topic><topic>Intubation</topic><topic>Mathematics</topic><topic>Overlap</topic><topic>Population</topic><topic>Population estimates</topic><topic>Probability and statistics</topic><topic>Probability theory and stochastic processes</topic><topic>Propensity score</topic><topic>Research methods</topic><topic>Sample size</topic><topic>Sciences and techniques of general use</topic><topic>Statistical discrepancies</topic><topic>Statistics</topic><topic>Studies</topic><topic>Treatment effect heterogeneity</topic><topic>Unconfoundedness</topic><topic>Weighting functions</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>CRUMP, RICHARD K.</creatorcontrib><creatorcontrib>HOTZ, V. JOSEPH</creatorcontrib><creatorcontrib>IMBENS, GUIDO W.</creatorcontrib><creatorcontrib>MITNIK, OSCAR A.</creatorcontrib><collection>Istex</collection><collection>Pascal-Francis</collection><collection>RePEc IDEAS</collection><collection>RePEc</collection><collection>CrossRef</collection><collection>Biotechnology Research Abstracts</collection><collection>Technology Research Database</collection><collection>Engineering Research Database</collection><collection>Biotechnology and BioEngineering Abstracts</collection><jtitle>Biometrika</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>CRUMP, RICHARD K.</au><au>HOTZ, V. JOSEPH</au><au>IMBENS, GUIDO W.</au><au>MITNIK, OSCAR A.</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Dealing with limited overlap in estimation of average treatment effects</atitle><jtitle>Biometrika</jtitle><date>2009-03-01</date><risdate>2009</risdate><volume>96</volume><issue>1</issue><spage>187</spage><epage>199</epage><pages>187-199</pages><issn>0006-3444</issn><eissn>1464-3510</eissn><coden>BIOKAX</coden><abstract>Estimation of average treatment effects under unconfounded or ignorable treatment assignment is often hampered by lack of overlap in the covariate distributions between treatment groups. This lack of overlap can lead to imprecise estimates, and can make commonly used estimators sensitive to the choice of specification. In such cases researchers have often used ad hoc methods for trimming the sample. We develop a systematic approach to addressing lack of overlap. We characterize optimal subsamples for which the average treatment effect can be estimated most precisely. Under some conditions, the optimal selection rules depend solely on the propensity score. For a wide range of distributions, a good approximation to the optimal rule is provided by the simple rule of thumb to discard all units with estimated propensity scores outside the range [0.1,0.9].</abstract><cop>Oxford</cop><pub>Oxford University Press</pub><doi>10.1093/biomet/asn055</doi><tpages>13</tpages><oa>free_for_read</oa></addata></record> |
| fulltext | fulltext |
| identifier | ISSN: 0006-3444 |
| ispartof | Biometrika, 2009-03, Vol.96 (1), p.187-199 |
| issn | 0006-3444 1464-3510 |
| language | eng |
| recordid | cdi_proquest_journals_201696479 |
| source | RePEc; JSTOR Mathematics & Statistics; JSTOR Archive Collection A-Z Listing; Oxford University Press Journals All Titles (1996-Current); Alma/SFX Local Collection |
| subjects | Applications Arithmetic mean Average treatment effect Biology, psychology, social sciences Causality Clinical outcomes Distribution theory Estimation methods Estimators Exact sciences and technology General topics Heart catheterization Ignorable treatment assignment Infinity Intubation Mathematics Overlap Population Population estimates Probability and statistics Probability theory and stochastic processes Propensity score Research methods Sample size Sciences and techniques of general use Statistical discrepancies Statistics Studies Treatment effect heterogeneity Unconfoundedness Weighting functions |
| title | Dealing with limited overlap in estimation of average treatment effects |
| url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-12T00%3A43%3A01IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-jstor_proqu&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Dealing%20with%20limited%20overlap%20in%20estimation%20of%20average%20treatment%20effects&rft.jtitle=Biometrika&rft.au=CRUMP,%20RICHARD%20K.&rft.date=2009-03-01&rft.volume=96&rft.issue=1&rft.spage=187&rft.epage=199&rft.pages=187-199&rft.issn=0006-3444&rft.eissn=1464-3510&rft.coden=BIOKAX&rft_id=info:doi/10.1093/biomet/asn055&rft_dat=%3Cjstor_proqu%3E27798811%3C/jstor_proqu%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=201696479&rft_id=info:pmid/&rft_jstor_id=27798811&rft_oup_id=10.1093/biomet/asn055&rfr_iscdi=true |